The best and worst part of the web is the ability to freely share your opinions regardless of whether they’re technically accurate.
But imagine a scenario where you are held accountable not only for what you say online, but also for whether it is human- or AI-written?
A recent report came out that a German court is holding Google accountable for just this. The court treated AI Overviews as Google’s own content and rejected Google’s argument that users were responsible for fact-checking the results themselves.
Let’s talk more about what this could look like for businesses and individuals.
The ‘disclaimer’ defense is cracking
For the last few years, nearly every AI platform has relied on some version of the same disclaimer: “AI can make mistakes. Please verify important information.”
Most users recognized this as the cost of using these tools.
But the German court essentially said that merely warning users about potential errors doesn’t automatically remove responsibility when those errors cause harm. If your system creates new claims that never existed in the source material, those aren’t someone else’s words anymore. They’re yours.
Why? Because it moves the conversation away from whether AI is useful (AKA *) and toward who owns the consequences when it gets things wrong.
What this means for businesses
Many companies are blindly trusting rapidly adopting AI across content creation, customer service, product descriptions, reporting, legal reviews, hiring, and internal communications.
In many cases, the discussion centers around efficiency.
Can we create content faster?
Can we answer support tickets more cheaply?
Can we automate this process?
Those are reasonable questions.
But the German ruling introduces another question: Who is responsible when the output is wrong?
What if:
AI-generated support response gives inaccurate guidance.
An AI-written article damages a competitor’s reputation.
An AI-generated report contains fabricated information that influences a business decision.
The more people position AI as a trusted source of information, the harder it becomes to argue they shouldn’t be accountable for what it says.
The situation is kinda funny…
The irony in this is that most AI vendors already know this.
That’s why nearly every platform includes warnings, disclaimers, and usage policies.
Yet those same companies market AI as smarter, faster, more capable, and increasingly reliable.
You can’t, at the same time, tell users to trust the answer while arguing nobody should trust the answer.
At some point, those positions collide. We’re starting to see Google’s “solution”: an option to opt out of AI.
Germany may simply be one of the first courts willing to force Google (or any other LLM business) to take certain actions.
What SEOs should be paying attention to
Ironically, this ruling could end up benefiting everyone.
Today, the conversation is focused on whether AI companies should be responsible for the content their systems generate. But accountability could expand, well beyond AI.
The internet has spent decades creating distance between actions and consequences. Anonymous accounts, fake profiles, throwaway emails, and now AI-generated content all make it easier to say things without owning them.
That’s why I find this ruling so interesting.
Not because of Google. Because it introduces the idea that “I didn’t write it” may not be enough.
The image below shows a real email that Russell and Nina Westbrook received. A real person sat behind a keyboard and sent a message, hoping they would die in a car crash.
That isn’t free speech. It’s hate speech.
The internet, combined with AI, needs more confidence that the content is accurate and that people who create it are held accountable.
You don’t get to claim the productivity gains when AI is right and blame the algorithm when it’s wrong.
This post first appeared on the author’s website and is republished here with permission.
Competitive brand bidding has become a staple PPC strategy. Many brands assume there’s little they can do about it beyond running their own branded campaigns.
Industry case studies show that advertisers who gain visibility into competitor, affiliate, and trademark bidding activity often uncover substantial CPC inflation. In documented cases, brands reduced branded CPCs by 25% to 75% after identifying infringing advertisers and enforcing their policies.
In this guide, Bluepear’s brand protection experts show you how to monitor branded keywords, identify who’s advertising on them, and determine what actions may be available based on what you find.
Choosing keywords to avoid gaps in your monitoring setup
To identify who’s using your brand in search ads, start by deciding which keywords to monitor.
One of the most common mistakes is monitoring only your exact brand name. While it’s a valid starting point, it rarely shows the full picture. Some advertisers deliberately target brand-related coupon, discount, review, or alternative queries because they attract high-intent users while drawing less scrutiny.
For example, someone searching for “Brand coupon” or “Brand discount code” may be much closer to making a purchase than someone searching for the brand alone. These queries also tend to attract coupon affiliates, loyalty sites, and unauthorized advertisers seeking to intercept branded traffic.
Similarly, searches containing terms like “reviews” or “alternatives” often attract competitors and comparison sites that position themselves alongside your brand.
Misspellings deserve special attention. Some advertisers deliberately target spelling variations because they’re less likely to be monitored and often face less competition.
Your brand monitoring setup should be based on these types of keywords:
Bluepear’s built-in AI assistant generates keyword suggestions using this list as a starting point.
The number of terms you need to monitor depends on the size of your brand portfolio, including trademarks, local branches, and product names. Most small to medium-sized brands start with about 20 keywords and expand coverage over time as new risks and opportunities emerge.
Choosing locations and monitoring frequency
A search from your office, on your device, at a single point in time rarely reflects what real customers see. Search results are highly dynamic, so two users searching for the same keyword may see entirely different ads and organic listings.
Another factor to consider: Some advertisers deliberately try to hide their activity.
For example, a fraudster or affiliate violating your PPC policy may run campaigns outside normal business hours to reduce the chance of detection. If you only check search results manually during the workday, there’s a good chance you’ll never see these ads.
When monitoring branded search results, it’s important to consider:
Countries and markets where your brand operates.
Regional differences within those markets.
Mobile and desktop search results.
Different times of day.
Weekday versus weekend activity.
Monitoring frequency matters as well. Some violations appear only briefly before disappearing. Running checks multiple times throughout the day helps capture activity that might otherwise go unnoticed.
Keeping track of all these variables can be difficult when relying on manual checks, especially for brands operating across multiple markets.
Bluepear accounts for everything: locations, devices, time zones, and redirects that obscure the true destination of traffic. Set your parameters once and gain continuous visibility without disrupting your workflow.
Otherwise, you could end up spending hours each week on tedious manual work.
Reviewing search results and recording evidence
Not every advertiser bidding on your branded keywords is violating a policy. Competitors may legally bid on branded keywords as long as they don’t use your trademark in the ad copy. Affiliates may also be authorized to promote your brand under specific conditions.
However, you still need to determine whether an advertiser’s behavior crosses the line from legitimate brand bidding into trademark misuse, policy violations, or customer deception.
Some common signals to investigate include:
Trademark use in ad copy: Does the ad mention your brand name in the headline or description? For brands with trademark restrictions or affiliate PPC policies, this may indicate a compliance issue.
Misleading claims: Look for language that could confuse users about the advertiser’s relationship with your company. Claims of being an “official” partner, references to exclusive offers, or statements implying authorization when none exists warrant investigation.
Unauthorized coupon and discount promotions: Verify that the advertised discount, promo code, or offer is legitimate. Some affiliates use expired, misleading, or entirely fabricated offers to attract clicks.
Impersonation signals: Some ads and landing pages are designed to resemble a brand’s official website. Even without explicitly claiming to be your company, these tactics can create confusion and divert branded traffic.
Because advertisers can change ad copy, pause campaigns, or remove landing pages at any time, violations may disappear before they can be documented. Take screenshots and record key details, including:
Ad copy
SERP position
The keyword that triggered the ad
Location
URLs and redirects
Landing page content
Timestamps
Alternatively, Bluepear can handle this automatically by compiling a comprehensive report containing all relevant information.
Identifying who’s behind the activity
Sometimes you won’t immediately know whether an advertiser is a competitor, an affiliate, or something else entirely.
Branded search results often include a mix of participants with different motivations and risk levels. Understanding who you’re dealing with is essential before deciding whether any action is necessary.
The following signals can help guide your investigation:
Signal
Likely Competitor
Likely Affiliate
Likely Coupon / Loyalty Site
Possible Fraudster
Direct competitor domain
+
–
–
–
Coupon or cashback page
–
Sometimes
+
–
Affiliate network tracking links
–
+
+
Sometimes
Product comparison page
+
Sometimes
–
–
Uses your trademark in ad copy
Sometimes
Sometimes
Sometimes
+
Claims to be “official”
–
–
Sometimes
+
Multiple redirects before reaching the destination page
–
+
+
+
Promotes coupon codes that cannot be verified
–
Sometimes
+
+
Landing page closely imitates your brand’s design or messaging
–
Sometimes
–
+
No single signal provides a definitive answer. The goal is to evaluate multiple pieces of evidence before drawing conclusions.
Once you know who’s advertising on your brand terms, you can move beyond detection and assess whether their activity aligns with your policies and business objectives.
What to do next
Once you’ve identified who’s advertising on your brand terms and reviewed their ads, the next step is deciding how to respond.
Competitor brand bidding
Not every competitor bidding on your branded keywords requires immediate intervention.
Before acting, you need the answers to these questions:
How frequently does the competitor appear in your brand search?
Which keywords are being targeted?
Are they using trademarked terms in ad copy?
Are they directing users to comparison content or direct offers?
In many cases, the best approach is to monitor the activity and evaluate its business impact over time. Document patterns to establish a baseline that can support future compliance reviews or legal discussions if escalation becomes necessary.
Affiliate violations
If an affiliate is bidding on restricted branded keywords or otherwise violating program rules, the next step is usually to gather evidence and contact the affiliate or network.
The common workflow is:
Document the violation.
Verify the affiliate’s ID.
Show evidence to them.
Request removal or corrective action.
Apply program enforcement measures if needed.
Having screenshots, timestamps, and redirect data makes these conversations significantly easier.
Ads that create confusion about brand affiliation.
Impersonation attempts.
Misleading claims of being an official brand representative, partner, or reseller.
The appropriate response depends on the circumstances, your internal policies, and applicable laws.
In many jurisdictions, competitors are generally allowed to bid on trademarked keywords. However, ads that create confusion about the advertiser’s relationship with your brand may raise trademark or unfair competition concerns, depending on the specific facts and local laws.
The advertising platform’s own policies also matter. Google allows advertisers to bid on trademarked keywords, but may restrict trademark use in ad text when a valid trademark complaint is submitted. Google also prohibits ads that use trademarks in a confusing, deceptive, or misleading way.
Before taking action, collect as much evidence as possible, including screenshots, detection timestamps, URLs, and landing page content.
Once the facts have been documented, consider the following actions:
Contact the advertiser directly and request removal of the ad.
Submit a trademark complaint to the advertising platform.
Send a cease and desist letter requesting to stop the activity.
Escalate through legal channels if necessary.
Why you need to keep monitoring your brand search
The important thing to remember is that this isn’t a one-time exercise. Affiliates may activate and pause campaigns throughout the month. Some violations appear only on weekends, outside business hours, or in specific markets. An advertiser that disappears today may return next week with different ad copy, a new domain, or a different affiliate account.
That’s why effective brand protection isn’t built on occasional searches or periodic audits. It requires consistent monitoring and a repeatable investigation process that reveals who’s appearing on your brand terms, how they operate, and whether action is warranted.
Want effortless visibility into your branded search landscape?
Try Bluepear. The platform helps marketing teams identify issues earlier, respond faster, and make more informed decisions about protecting their traffic and advertising investments.
What Google said. Google wrote, “Released the June 2026 spam update, which applies globally and to all languages. The rollout may take a few days to complete.”
Timing. This update may only take just a few days to roll out.
Why we care. Google released search ranking updates several times per year. Spam updates should impact sites using manipulative techniques to abuse the search ranking algorithm. If your site is not doing any of that, then you should be okay.
“While Google’s automated systems to detect search spam are constantly operating, we occasionally make notable improvements to how they work. When we do, we refer to this as a spam update and share when they happen on our list of Google Search ranking updates.
For example, SpamBrain is our AI-based spam-prevention system. From time-to-time, we improve that system to make it better at spotting spam and to help ensure it catches new types of spam.
Sites that see a change after a spam update should review our spam policies to ensure they are complying with those. Sites that violate our policies may rank lower in results or not appear in results at all. Making changes may help a site improve if our automated systems learn over a period of months that the site complies with our spam policies.
In the case of a link spam update (an update that specifically deals with link spam), making changes might not generate an improvement. This is because when our systems remove the effects spammy links may have, any ranking benefit the links may have previously generated for your site is lost. Any potential ranking benefits generated by those links cannot be regained.”
Google is effectively closing the door on creating new Smart Campaigns through its API, suggesting a continued shift toward newer automated campaign types such as Performance Max and away from legacy automation products.
What’s happening. Google announced that developers will no longer be able to create new Smart Campaigns through the Google Ads API beginning August 3rd.
Existing Smart Campaigns will not be affected and will continue serving ads. Advertisers and developers will also retain the ability to update and manage those campaigns through the API.
The change applies specifically to the creation of new Smart Campaigns.
Why we care. Google is further steering advertisers toward newer AI-driven campaign types like Performance Max. While existing Smart Campaigns will continue running, businesses and platform partners that rely on API-based campaign creation will need to update their workflows before August 2026 or risk disruptions when creating new campaigns.
What Google recommends instead. As Smart Campaign creation is phased out, Google is encouraging advertisers to consider alternative campaign types.
The company specifically pointed to Performance Max, its AI-powered campaign type that runs across Google’s advertising inventory, as the primary alternative. Other options include Search campaigns and Demand Gen campaigns, depending on an advertiser’s objectives.
A broader automation shift. The move reflects Google’s ongoing consolidation around newer automated campaign formats.
Over the past several years, Google has increasingly prioritized AI-driven campaign types that automate bidding, targeting and creative optimization across multiple channels.
Limiting the creation of new Smart Campaigns suggests the company sees Performance Max and other automation products as the future of campaign management.
What developers need to know. Applications that currently create Smart Campaigns through the API should be reviewed before the August deadline.
Developers can identify affected requests by looking for campaign creation operations where:
advertising_channel_type is set to SMART
advertising_channel_sub_type is set to SMART_CAMPAIGN
After August 3, attempts to create new Smart Campaigns will fail.
In version 24 of the Google Ads API, developers will receive a SmartCampaignError.CREATION_FAILED error.
In version 23 and earlier, requests will return an OperationAccessDeniedError.CREATE_OPERATION_NOT_PERMITTED error.
What to watch. While Google is not shutting down existing Smart Campaigns, removing API support for new campaign creation is another sign that advertisers should prepare for a future centered on Performance Max and other AI-driven campaign types.
For agencies, software providers and advertisers with automated campaign creation workflows, now is the time to evaluate migration plans before the August 2026 deadline.
Users who saw a brand recommended by ChatGPT were much more likely to visit that brand’s website within a week, according to Similarweb.
What happened. Users were 2.5 times more likely, on average, to visit an AI-recommended brand than a direct competitor, based on Similarweb’s study of U.S. desktop behavior across finance, travel, and beauty.
Similarweb tracked users who asked ChatGPT industry-relevant questions, received a specific brand recommendation, and then visited either that brand’s website or a competitor’s site within seven days.
The study excluded users who had visited the brand’s site in the prior four weeks or named the brand in their prompt.
Recommendations shifted traffic. The pattern appeared across all three industries:
Finance: After an American Express recommendation, 7.2% of users visited American Express, compared with 3.1% who visited Capital One. After a Capital One recommendation, 14.2% visited Capital One, compared with 3.8% who visited American Express.
Travel: After a Skyscanner recommendation, 9.5% visited Skyscanner, compared with 7.6% who visited Kayak. After a Kayak recommendation, 12% visited Kayak, compared with 3.4% who visited Skyscanner.
Beauty: After a Sephora recommendation, 7.9% visited Sephora, compared with 3.3% who visited Ulta. After an Ulta recommendation, 7.6% visited Ulta, compared with 4.6% who visited Sephora.
AI demand showed up in search. Most AI-influenced visits didn’t appear as AI referral traffic. ChatGPT may influence brand choice, but the later visit often appears in analytics as search traffic.
55.9% of AI-influenced visits came through search, compared with 40.4% of non-AI-influenced visits, Similarweb found.
Direct traffic accounted for 19.9% of AI-influenced visits, compared with 38.8% of standard visits.
Recommended users stayed longer. AI-influenced visitors also engaged more deeply once they reached a brand’s site. They viewed 12 pages and spent 11.8 minutes on site, on average, compared with 6.5 pages and 5.6 minutes for non-AI-influenced visitors. Those users likely arrived after narrowing their options during the AI conversation, Similarweb said.
Why we care. AI visibility can drive visits even when referral reports miss the source of influence. You need to know whether ChatGPT is creating demand for your brand or sending it to a competitor.
About the data. Similarweb used its opted-in U.S. desktop web panel to track user journeys from July through December 2025. The report focused on finance, travel, and beauty brand pairs with competitive overlap.
OpenAI is signaling early momentum in its advertising business, with executives reporting that users are increasingly engaging with ads inside ChatGPT rather than dismissing them — a key metric as the company builds a new revenue stream beyond subscriptions and enterprise AI.
What’s happening. OpenAI says the rate at which users dismiss ads within ChatGPT has fallen by 50% since the company launched its advertising business in February. The company views ad dismissals as a proxy for relevance, making the decline an indication that users are finding the ads they encounter more useful and aligned with their interests.
The update was shared by OpenAI Chief Revenue Officer Denise Dresser.
Why we care. With data suggesting users becoming more receptive to ads within ChatGPT, conversational AI could improve it’s viability of being a new advertising channel. A 50% reduction in ad dismissals indicates improving relevance and engagement, potentially giving brands an opportunity to reach consumers in high-intent, task-oriented moments rather than through traditional interruptive ad formats.
Why relevance matters. Unlike traditional display advertising, ads within AI experiences face a higher bar for usefulness. Users typically engage with ChatGPT to accomplish a task, answer a question or solve a problem. Ads that feel disconnected from that experience risk creating friction and undermining trust.
According to Dresser, improving relevance has been a central focus.
“This form factor is about usefulness,” she said. “That’s great for the consumer, great for the user.”
The bigger picture. The results offer an early glimpse into how advertising may evolve inside generative AI platforms. Rather than interrupting content consumption, AI-powered advertising is increasingly focused on integrating with user intent and ongoing conversations. Success may depend less on grabbing attention and more on delivering recommendations that users find genuinely helpful.
The lower dismissal rates suggest OpenAI is making progress toward that goal.
Competition extends beyond advertising. The advertising update comes as OpenAI continues expanding its business on multiple fronts. While building an ads business, the company is also competing aggressively for enterprise AI spending, where rivals such as Anthropic are fighting for market share among corporate customers.
The cardinal sin of positioning in the AI era is replacement. I call it “substitution positioning.” Short term, it’s yummy. Long term, it plants cavities in your perception.
In January, Anthropic CEO Dario Amodei predicted software engineering jobs to go away in the following 6-12 months. Demand for software engineers is reaching new heights today.
“I think we might be 6 to 12 months away from when the model is doing most, maybe all of what SWEs do end to end.”
In September 2025, OpenAI CEO Sam Altman predicted customer support jobs will go away. Almost immediately after he said that, customer service hiring started outpacing the broader job market.
“I’m confident that a lot of current customer support that happens over a phone or computer, those people will lose their jobs, and that’ll be better done by an AI.”
AI companies use fear as a marketing tool. Specifically, the fear of being replaced. They got me, too. I was deeply worried about my future when Opus 4.5 got traction.
However, as long as I see that even Anthropic is hiring copywriters and SEOs, I sleep much better.
Times were more uncertain before AI came on the scene (I’ll spare you the TED talk about unprecedented times). Fear sells. Tapping into our primal fight-or-flight response is a sure way to get attention. And layoffs act as a double agent in this environment: They make companies that underperform or overhired look good (“Look, we don’t need as many people anymore because we’re so innovative!”) and feed the substitution positioning narrative.
But the facts show that recent layoffs are closer to AI washing than AI spring cleaning:
New York state added an option for companies to indicate when they lay staff off due to AI (technological innovation or automation). In March of this year, over 160 companies filed mass layoffs of ~28,300 workers. Not even one chose AI as the reason. That list includes Amazon and Goldman Sachs.
Researchers at Yale looked at the CPS, a government survey that tracks employment across the US, over the last 33 months and found no evidence of job displacement from AI. The way AI impact works looks very similar to how computers and the internet changed work.
Premium is $150/year. That’s less than one walked-back AI announcement. The Growth Memo resource library includes the positioning reframe asset in this memo… and the rest of the audits, workflows, and decks built for the AI visibility era.
“Stop trying to make ‘replacement’ happen. It’s not going to happen.”
— Regina George from Mean Girls in an alternate multiverse
AI is just not there… yet.
Synthetic intelligence in its current form is heavily paradoxical: AI can do some things better than humans and others not even at the level of a chimpanzee.
This phenomenon is called the Jagged Frontier. It’s based on a paper where BCG and Harvard researchers conducted a randomized controlled trial of 758 knowledge workers and found that you get the most (12.2% more tasks, 25.1% faster) out of AI if you know what it’s good and bad at.
That’s why some people are lukewarm on AI and others see it as new salvation.
“Currently, measures of exposure, automation, and augmentation show no sign of being related to changes in employment or unemployment.”
Microsoft came to the same conclusion. In its 2026 Work Trend Index Annual Report, the Windows mother found that ~16% of the 20,000 AI users surveyed are “Frontier Professionals” or the most advanced AI users in the research set who use agents for multi-step workflow systems.
80% of those Frontier Professionals said they produce work they couldn’t have a year ago, compared to 58% across the board. They are not just using AI more, but also know which mode a task calls for. They have the judgment to navigate the Jagged Frontier.
Caption: Microsoft’s 4 modes of using AI at work
Despite some of the gains you can see from AI, we’re far away from replacing people.
Do you truly trust any AI workflow enough to leave it on full automation without any maintenance or quality assurance?
Look at a company that tried: Klarna publicly bragged its AI did the work of 700 agents and cut headcount ~40%, Then they reversed and rehired humans after the CEO admitted cost-cutting produced “lower quality” and that customers want a human option.
As a result, choosing substitution positioning means you gain short-term attention but sacrifice long-term credibility. Anthropic calls for a public pause on AI development on June 4 and launches Fable 5 on June 9. OpenAI did the same: Altman told the U.S. Senate in May 2023 that AI needs licensing and a global watchdog, then opposed binding rules and lobbied against California’s SB-1047 and mandatory federal pre-launch approval. The words don’t match the actions.
Substitute positioning could work if customers (1) actually wanted the replacement and (2) if the technology was there. Neither are true.
Cost reduction is a much stronger AI-use argument than productivity growth because it hits the P&L immediately. Productivity gains show up later. They build slowly inside a company and even slower across the economy, the same lag we saw with computers and the internet.
However, because replacement positioning goes way beyond just cost-cutting (it’s people-cutting), you also directly antagonize the people you sell to:
Duolingo’s AI-first memo drew a “torrent of pushback” on social media/TikTok. CEO von Ahn said he “did not expect the amount of blowback” and walked it back, reframing AI as a “tool to accelerate,” not replace contractors.
Surveys show that leadership is commonly more excited about AI than the employees who use it.
In April, Uber execs bragged about burning their annual token budget in just four months. Today, that’s a massive liability. Token cost is growing despite being heavily subsidized. But soon, it will reach a price point that makes junior employees start to look interesting again.
Entry-level jobs might soon be cheaper than tokens accomplishing the same output, and young people don’t have the same baggage and reflexes that experienced experts have.
What is the replacement for ‘replacement’?
Instead of trying to replace people, the solution is to position your AI use as enhancement. The opposite of fear-based marketing is aspiration and empowerment.
AI goes both ways: You can reduce the number of people or grow output with the same number of people. And the data shows that productivity gains have a higher ROI than substitution.
In a paper by the National Bureau of Economic Research, Researchers asked 750 executives about the impact of AI on productivity and labor markets. While larger firms have a higher interest in replacing labor cost (people), the highest ROI comes from growing productivity.
The moral of the story? Doing more with what you have works better than firing copywriter Alex because only Alex can judge good copy.
Premium: Your positioning matters more than ever + Claude skill
Building products has gotten easier, but distribution has not.
When supply explodes, the scarce thing is being the product that actually gets chosen, and positioning (right alongside product quality) is exactly how you get chosen.
Premium subscribers get the Claude skill that performs a quality check on your content’s positioning for any AI-based tooling within your product, offering quality reframes for your team that avoid substitution positioning.
This post first appeared on the author’s website and is republished here with permission.
Generating demand is one of the biggest challenges in digital marketing. Measuring where that demand came from is even harder.
For nearly two decades, marketers have evaluated paid search and paid social as separate channels. Search campaigns are typically measured on clicks, conversions, and ROAS, while social is often judged by platform-reported metrics and attributed conversions.
The problem is that consumers don’t experience marketing channels in isolation as they move through the conversion process.
A prospective customer may discover your brand through an ad on Meta, ignore it initially, see another ad days later, and eventually search for your brand or products on Google before adding an item to their cart and converting. In most reporting platforms, paid search gets the credit because it captured the last click. But is that fair if it didn’t create the demand?
As privacy regulations, technology, and attribution limitations continue to evolve, marketers need new ways to understand how paid social influences search behavior. Here are several practical methods for measuring the relationship between the two.
Signs paid social is influencing search performance
Paid social’s influence on search isn’t always visible in attribution reports, but it often appears in performance data. These indicators can help you identify whether your social campaigns are creating awareness that later translates into search activity and conversions.
Branded search volume increases
One of the clearest signs that paid social is helping drive downstream search activity is an increase in branded search queries.
When people encounter a compelling, relevant social ad on Meta, TikTok, or other platforms, many don’t immediately click the ad. Instead, they may later search for the brand name, product name, founder, or other branded terms.
For example, when you launch a new Meta Ads campaign, you might notice increases in searches for:
Brand name.
Brand + product category.
Brand + reviews.
Brand + pricing.
Brand + competitor comparisons.
Monitor the volume of these and other branded searches over time. This can help identify whether your paid social efforts are generating awareness that later translates into search behavior.
Review data from your Google Ads and Microsoft Advertising search campaigns, Google Analytics, Google Search Console, Google Trends, and any third-party SEO platforms you use.
Note trends before, during, and after major paid social campaign launches and adjustments. If branded search volume continues to rise as you increase investment in paid social, there’s a strong possibility that those efforts are contributing to demand generation.
Not every increase in branded search volume is caused by paid social. The goal isn’t to prove perfect causation. It’s to identify a meaningful directional relationship.
Other factors can also contribute to growth in branded searches, including:
Another way to identify a relationship between paid search and paid social is by looking at click-through rates (CTRs).
Consumers are more likely to click ads from brands they recognize or have previously engaged with. If your paid social campaigns are increasing brand familiarity, people may become more inclined to click your search ads over a competitor’s.
For example, someone may see your Instagram video ads over a two-week period and later search for a related topic on Google. When presented with several options, they may be more likely to click a brand they’ve already seen elsewhere.
The same principle appears in the brand recognition surveys Meta and LinkedIn sometimes display in users’ feeds. I often find myself answering yes, even for brands I’ve never purchased from, simply because I’ve seen their ads on social media.
That familiarity, even at a basic level, can help improve CTR on branded search campaigns, increase CTR on non-branded campaigns, and lower CPCs over time.
The next time you launch a new paid social campaign or make a significant adjustment, review your paid search performance and compare CTR trends before and after.
Source: LinkedIn
Search conversion rate improvements
Brand familiarity can also influence conversions, as consumers who’ve previously engaged with your brand often arrive on your website with greater confidence and trust than first-time visitors. As a result, search traffic may become more likely to convert after periods of strong paid social activity.
Look for signs of higher search conversion rates, better lead quality, lower search CPAs, and higher revenue per visitor. If you sell products or services with longer consideration cycles and multiple brand touchpoints before conversion, you may notice this effect even more.
Conversion efficiency can be a valuable indicator of paid social’s influence on search behavior.
How to validate social’s impact on search
The indicators above provide directional insight. To build stronger evidence, use these methods to measure whether paid social activity is influencing search performance.
Pre- and post-campaign analysis
One of the easiest ways to evaluate the impact of a paid social campaign on search performance is through a pre- and post-campaign analysis.
Measure the following before your paid social campaign launches and compare it with performance after launch:
Branded search impressions.
Branded search clicks.
Search CTR.
Search CVR.
CPA.
Total search conversions.
While the data alone won’t fully prove causation, it can provide evidence that increased social activity may be influencing search performance. As you conduct your analysis, account for seasonality where relevant, compare similar time periods, and monitor changes in competitor activity.
For stronger evidence, consider running a geotargeted holdout test. Run a paid social campaign in a specific geographic market while intentionally withholding it from other areas. Then compare search performance across both groups.
For example, instead of running paid social campaigns across all markets, a nationwide company could divide its audience into two groups:
Test market(s): Paid social campaigns active.
Control market(s): No paid social campaigns or exposure.
Run the test for several weeks and monitor the following metrics for both groups:
Branded search volume.
Search CTR.
Search CVR.
Leads.
Revenue.
If the test market sees significantly stronger search performance than the control market, you’ll have a better opportunity to isolate the impact of your paid social campaigns.
Geotargeted tests work particularly well because they help reduce attribution bias. They allow you to evaluate business outcomes across similar populations rather than relying solely on platform-reported conversions while navigating privacy-related blind spots.
If you run a holdout test, choose comparable markets, allocate sufficient budget, and allow enough time to achieve statistically significant results. This approach typically works best for larger advertisers running regional or national campaigns. If you’re a smaller brand, start with the pre- and post-campaign analysis.
The relationship between paid search and paid social is often stronger than reporting platforms suggest. Rather than evaluating channels in isolation, look at how they complement one another. Search captures demand, and paid social helps create it.
Dive into your data to identify opportunities to invest more effectively and generate future demand and conversions across platforms. By monitoring these signals and applying the methods above, you’ll gain a stronger understanding of how paid social contributes to business growth.
Attribution isn’t perfect, but learning how to measure influence across channels can help you make better investment decisions and drive stronger results.
Cornell Tech researchers found that deep-research AI agents can be manipulated by short edits to public user-generated pages, allowing a single injected Reddit-style comment to become a cited recommendation for fake products, services, or entities.
The paper called those altered pages “poisoned” because the added text was designed to steer what the AI system cited and repeated. It identified the weakness in systems that search the web, gather sources, and write cited reports. The researchers called the attack WARP, short for Web Agent Retrieval Poisoning.
How injected text reaches reports. The attack doesn’t require access to the model, prompts, search engine or retrieval system. Instead, an attacker edits or appends text to a page the agent already tends to retrieve, such as a Reddit thread, Wikipedia page, or forum post.
When the agent later searches related topics, it may pull in that page, cite it, and repeat the attacker’s chosen message.
Deep-research tools often run many related searches for one user request, and the paper found the same user-generated pages surfaced across related queries.
Reddit was the biggest opening. Across STORM, Co-STORM, and OmniThink, 17% to 23% of retrieved URLs came from user-generated platforms, including Reddit, YouTube, Facebook, and Wikipedia.
Reddit made up the largest share of those pages. It accounted for 54% to 71% of user-generated URLs retrieved by the three open-source systems.
The researchers didn’t alter live websites. They used a simulation framework called GeoStorm to insert manipulated text into retrieved content during testing.
A few words worked. The researchers found the attack worked with snippets as short as about 13 words:
In one test, a 15-word sentence pushed a fake cryptocurrency, BananaCoin, into a Co-STORM report as an “emerging” long-term investment option. The report cited the altered source alongside legitimate crypto sources.
When the manipulated page was retrieved, the fake entity appeared in 38% to 51% of reports across systems. Targeting multiple pages raised that range to 42% to 62%.
The attack still worked when systems retrieved full Reddit threads, though mention rates were lower. When injected text was added to complete Reddit threads and made up less than 4% of the retrieved content, the fake entity still appeared in 30% to 53% of reports when the page was retrieved.
Defenses struggled. Blocking user-generated domains stopped this attack path, but it also removed sources such as firsthand product experiences and local recommendations.
The tested text filters failed to reliably separate injected passages from normal user content. The manipulated passages were fluent because they were written by an AI model, so perplexity-based filters were more likely to flag normal user content than the injected text.
Report-level checks also missed the manipulation. Altered reports looked similar to clean reports because the agent itself folded the fake recommendation into an otherwise normal answer.
Why we care. A small edit to a public page can become part of a cited AI answer, even when the underlying source is user-generated. Misinformation planted on sites like Reddit or in forums can move from discussion threads to cited recommendations in AI answers that look credible to users.
About the research. The paper, Deep-Research Agents Can Be Poisoned via User-Generated Content, was written by Tingwei Zhang, Harold Triedman, and Vitaly Shmatikov of Cornell Tech and posted to arXiv on May 22. The researchers tested the full attack on three open-source systems: STORM, Co-STORM, and OmniThink. They analyzed OpenAI Deep Research and Gemini Deep Research for user-generated citations, but didn’t run live manipulation tests because that would require publishing altered content to the open web.
Most B2B brands rank for thousands of keywords in Google, but appear in only about 3% of AI-generated answers, according to Walker Sands’ B2B AI Search Visibility Benchmark of 828 enterprise companies. (Disclosure: I’m the director of SEO and GEO at Walker Sands.)
The benchmark analyzed more than 45 million search queries in March across 828 enterprise B2B companies spanning 14 industries. It evaluated domains across four metrics:
Keyword coverage: The number of keywords for which a company ranks in Google.
Keywords with AI Overviews: The number of ranking keywords that trigger AI Overviews.
AI Overview incidence: The percentage of ranking keywords that display AI-generated responses.
Citation inclusion rate: How often a company’s domain is cited within those AI-generated answers.
Together, these metrics establish a baseline for how often AI Overviews appear and how often B2B brands are cited within them.
A baseline for B2B AI search visibility
The benchmark reveals a significant gap between ranking visibility and AI citation visibility.
AI Overviews appear in 50% of search results where enterprise B2B brands rank.
The median enterprise B2B brand is cited in just 3% of relevant AI Overviews.
4.6% of enterprise B2B companies aren’t cited in AI Overviews for any of their relevant keywords.
The typical enterprise B2B company ranks organically for about 9,700 search queries, and AI Overviews appear in nearly half of them. Yet across those opportunities, the median brand is cited in just 3% of AI Overviews.
In other words, B2B brands are present in the search results AI Overviews summarize, but they’re largely invisible within the summaries themselves.
While 4.6% represents a small portion of the market, it highlights a critical visibility gap. These companies may still rank well in traditional search results, but some brands are effectively absent from AI Overviews, which have become an influential part of the buyer journey.
In many cases, a lack of citations signals larger issues, such as:
Limited topical authority.
Unstructured or inaccessible content.
A lack of content that directly answers buyers’ questions.
Addressing these gaps is crucial for visibility in AI-driven search experiences.
The benchmark frames AI search performance as four sequential layers, and the value lost at each step tells the real story.
It starts with keyword coverage, or the number of keywords for which a brand ranks in Google’s top 100 organic results. Here, the leaders look healthy. The median company ranks for about 9,700 keywords, while top-quartile brands rank for more than 37,000.
The next layer, keywords with AI Overviews, narrows the field. These are ranking keywords that trigger an AI Overview. The median company has roughly 4,500 of them, already less than half of its ranking footprint.
The third layer, AI Overview incidence, measures how often AI-generated answers appear across a brand’s relevant searches. The median is 48.8%, meaning AI now intercepts roughly half the queries where these companies compete. Top-quartile brands operate in even more AI-saturated environments, at 61.7%.
The final layer is the one that matters most and where almost everyone falls off a cliff: citation inclusion rate, or how often a brand is cited as a source within an AI Overview. The median is 3.0%. Even the top quartile reaches only 4.5%, while the bottom quartile sits at 1.7%.
Viewed from top to bottom, the funnel is brutal. Tens of thousands of ranking keywords compress into a single-digit share of AI citations. Most of the visibility B2B brands have built through organic search doesn’t carry through to the layer of search that increasingly shapes buyers’ first impressions of a category.
Ranking breadth doesn’t buy you AI citations
The most important takeaway is also the most counterintuitive: Ranking breadth alone doesn’t predict AI citation rates.
The study found that some companies rank for thousands of keywords yet rarely surface in AI-generated answers. The strengths that won traditional SERP real estate — sheer page volume, broad keyword targeting, and domain authority accumulated over the years — don’t automatically translate into becoming the source an AI system chooses to cite.
This is a meaningful break from how most B2B SEO teams are still resourced and measured. If your dashboard tracks ranking keywords and estimated organic traffic, it may be telling a flattering story about a layer of search that’s shrinking in influence while saying nothing about the layer that’s growing.
The benchmark argues that brands consistently cited in AI-generated answers share three characteristics: deep topical authority across related content areas, clear and structured explanations that directly answer buyer questions, and consistent coverage across multiple relevant pages.
The common thread is specificity. Generative systems appear to reward content that resolves a buyer’s question clearly and demonstrates sustained expertise on a topic, rather than content that simply ranks for a query.
That reframes the work. Optimizing for AI citations looks less like chasing keyword volume and more like building genuine, well-structured subject-matter depth.
Some industries are far more exposed than others
AI search visibility isn’t distributed evenly across B2B technology. The industry breakdown reveals sharply different competitive dynamics depending on the category.
Cybersecurity leads on both fronts. AI Overviews appear in a median of 59.9% of cybersecurity-related searches, and cybersecurity brands earn the highest median citation rate in the study at 4.2%. Enterprise software (55.3% AI Overview incidence) and martech (56.3%) also see AI-generated answers in well over half of relevant queries.
At the other end, professional services and distribution and logistics trail in citations, both with a median rate of just 2.1%. Distribution and logistics also have the lowest AI Overview incidence, at 29.6%, meaning buyers in that category encounter AI-generated summaries far less often than in cybersecurity.
These differences create both risks and opportunities. In categories where AI-generated answers are already pervasive, such as cybersecurity, the cost of being invisible is immediate. Buyers are forming impressions inside AI summaries right now.
In categories where citation rates are low and few brands have figured out the new mechanics, there’s a genuine first-mover opportunity. Brands that learn to earn citations before their competitors can shape how an entire category is framed in AI-generated answers, much as early SEO adopters captured outsized shares of organic visibility.
The most striking number in the report is that 4.6% of enterprise B2B companies aren’t cited at all in AI-generated answers for their relevant keywords.
These aren’t fly-by-night operations. They’re companies with $100 million or more in revenue that, in many cases, still rank well in traditional search. They’ve simply become invisible in the AI layer — present in the index, absent from the answer.
Near-zero citation rates can be attributed to deeper structural issues: thin topical authority, content that’s unstructured or difficult for systems to parse, and a lack of material that directly answers the questions buyers are asking.
For a small but meaningful slice of the market, AI search isn’t a place where they’re losing share. It’s a place where they don’t exist.
What this means for B2B search teams
The benchmark provides a baseline, but the strategic implications for SEO, GEO, and marketing teams are already clear.
Measurement must evolve. Citation inclusion rate is now a distinct KPI from ranking. Teams that can’t see whether their content is being cited in AI-generated answers are flying blind on one of the fastest-growing parts of the funnel. Knowing your own citation rate — and benchmarking it against the 3% median and 4.5% top-quartile benchmarks — is a sensible first step.
The content mandate shifts from breadth to depth. The drivers all point toward consolidating authority around the topics that matter to buyers, structuring content so machines can interpret it, and answering real questions directly rather than producing volume for volume’s sake.
The window is open but closing. Generative AI will influence more than 75% of B2B search queries within the next one to two years. If that projection holds, even approximately, the median 3% citation rate isn’t a stable equilibrium. It’s a snapshot of an early, contested market that rewards brands moving now.
The uncomfortable truth is that much of the SEO equity B2B brands have built is being summarized by AI systems that don’t cite the companies that created it. For most enterprise brands, the question is no longer whether they rank. It’s whether they’re in the answer at all.
Google is experimenting with new labels that highlight the ads it believes are most relevant to a user’s search query, a move that could influence how users engage with paid results and how advertisers think about ad relevance.
What’s happening. Google began testing a new Search ads feature that adds “Strongest match” or “Strong match” labels to certain ads appearing in search results. The experiment was revealed by Google Ads Liaison Ginny Marvin, who said the labels are designed to help users quickly identify ads that are highly relevant to their search intent.
The test is currently limited to a small percentage of users in the U.S.
Why we care. Google is testing a new visual signal that could influence which ads users perceive as the most relevant and trustworthy. If the experiment expands, advertisers with strong ad relevance and quality signals may gain greater visibility and potentially higher click-through rates, while less relevant ads could become easier for users to overlook.
How it works. According to Google, the labels rely on existing ad quality and relevance signals already used within its advertising systems. Rather than introducing a new ranking factor, the experiment surfaces Google’s assessment of an ad’s relevance directly within the user interface.
The goal is to make it easier for users to identify the information most closely aligned with their search query.
Why Google is testing it. Google says the experiment is intended to improve the Search ads experience for both consumers and advertisers.
For users, the labels could provide an additional signal about which ads are most likely to address their needs.
For advertisers, the feature could help connect highly relevant ads with high-intent audiences, potentially increasing engagement and click-through rates.
Reading between the lines. The test reflects Google’s ongoing efforts to make ad relevance more visible.
Historically, relevance signals have largely operated behind the scenes through auction and quality systems. By displaying those signals more prominently, Google could be seeking to build greater user trust in sponsored results while rewarding advertisers that closely align ads with search intent.
The experiment also arrives as search platforms face increasing pressure to demonstrate the usefulness and quality of advertising experiences.
What to watch. Google emphasized that this is an early-stage experiment and has not indicated whether the labels will become a permanent feature. For now, advertisers may want to keep an eye on the test as another signal that ad relevance and quality remain central to Google’s vision for Search advertising.
“Tailor Your Feed” is the first time a user can shape their Discover feed by typing, in natural language, what they want to see. We have tracked it from its first appearance in Search Labs^search-labs to the pipeline[^pipeline] that powers it. Ten key points:
An explicit-control layer. Your prompt is turned into SEE_MORE / SEE_LESS actions, applied after a feed refresh.
Seemingly an LLM[^llm] under the hood. A persistent chat thread, and your prompt turned into instructions applied to your feed (in real time and over time).
The rebrand. “Tailor Your Feed” became “Add topics to your feed” in spring 2026, with a chat-style entry point.
The back-end pipeline.historicalnaturallanguagetuningcontent.f[^pipeline-id], the “historical” twin of naturallanguagetuningcontent.f.
Two ways content is chosen. Entity[^entity] / interest expansion (the majority) vs a query-intent[^query-intent] fan-out[^fan-out] (the minority), the latter being the GEO[^geo] mechanism inside Discover.
Visible attribution. The “You asked to see” label, the “resulting from natural language tuning” tag, and a prompt history in My Activity.
Niche sites and small creators surfaced. Vegan recipe creators, Mississippi Today, a LinkedIn post, niche Japanese-property blogs and, as an illustration of the retrieval’s[^retrieval] behaviour, publishers outside the usual mainstream (VentureBeat surfaced on a “niche sites” prompt, though not itself a small site).
A popularity bypass. This pipeline mostly carries content that had barely circulated in Discover before, the opposite of the classic pipelines that re-serve already-popular articles.
What it changes for publishers. Selection power shifts to the user, opening a third path to visibility for small, niche sites.
Still EN-only, still nascent. Search Labs US only (FR ≈ 0%), adoption still early. What’s next.
Methodology
This article combines two observation streams:
Field tracking of the feature in the Google app since December 2025: UI states, server responses, attribution tags, and feed behaviour after each “Refresh / Update your feed”. Captured on our test devices, US (English) Search Labs accounts.
A close reading of the feed itself: each card can be traced back to the pipeline that selected it. By isolating the cards served by historicalnaturallanguagetuningcontent.f, we describe how this pipeline behaves relative to the rest of the feed, drawing on 1492.vision tracking data.
Three deliberate notes on how we phrase things:
We describe distribution outcomes[^distribution], whether an article had ever circulated in Discover before, not raw audience numbers. When we say a card has “no prior Discover distribution”, we mean we find no trace of earlier serving in our Discover tracking dataset.
No account identifier appears in this article. Examples are shown as prompt → result, anonymised.
The internal mechanisms below are our interpretation of observed data and public research. Where a date is inferred rather than anchored, we say so.
1. What “Tailor Your Feed” is: an explicit-control layer
For years, Discover personalization was implicit: Google inferred your interests from clicks, dwell time, follows. “Tailor Your Feed” adds the opposite, an explicit layer where you simply type what you want.
The current entry point at the top of the feed: “What do you want to see?” with an “Add topics to your feed” field.
The feature opens a chat-style panel. You can pick a suggested template or write freely.
The original “Tailor your feed” intro card: “Say what you want in your own words”, with a “Try now” button.
Suggested prompts: “Start showing me women’s basketball”, “Keep me updated on country music”, “Show me more of Cara Nicole’s videos”, with a free-text box “Ask for the kind of content you want.”
These suggestions correspond to four intents the server responses expose: SEE_MORE, KEEP_UPDATED, CREATOR_MORE and SEE_LESS. Whatever you type is interpreted into one (or several) of these actions, then applied to the feed once confirmed with “Refresh / Update your feed.”
The feature shipped through Search Labs, US only.
The Search Labs entry: “Make your Discover feed truly yours by saying what you want to see.” Beta, US only.
Later iterations turned the free-text box into explicit starter templates: “Show me content from…”, “I want videos about…”, “Keep me updated…”. The same verbs, now surfaced as chips.
The “What do you want to see?” panel with its starter templates: “Show me content from…”, “I want videos about…”, “Keep me updated…”.
2. Under the hood: seemingly an LLM that turns prompts into actions
The flow is simple: prompt → interpretation → readable answer + an actionable result. You type a prompt, the assistant replies in plain language and proposes concrete changes, and a tap on “Update your feed” commits them.
A representative response, observed in the data exchanges for the prompt “show me more content on seroundtable.com”, looks like this:
{
"feature": "Discover • Tailor your feed",
"locale": "en-US",
"thread_key": "chat_thread_key_082fa565-234a-451c-9318-1e9af8b9d734",
"user_prompt": "show me more content on seroundtable.com",
"assistant_text": "I can show you more content from Search Engine Roundtable (SERoundtable.com) related to your interests. Refresh your feed to apply these changes.",
"result": {
"status": "UNDERSTOOD_AND_ACTIONABLE",
"actions": ["SEE_MORE"],
"show_call_to_action": true,
"count": 1
},
"ui_state_code": 2
}
The in-app prompt that produced the response above: “show me more content on seroundtable.com”, and the assistant’s reply ending on “Refresh your feed.”
Three things stand out:
A persistent thread. The thread_key is stable across exchanges; your tuning is a conversation, not a one-shot. The same thread is referenced again when, later, a card is attributed to one of your past prompts.
Actions, not topics. The response returns actions: ["SEE_MORE"]. Ask to remove a topic and you get ["SEE_LESS"]; a nuanced prompt can return both, e.g. “new country music releases… but no celebrity gossip” yields ["SEE_MORE", "SEE_LESS"].
Local context injection. Responses are interpreted with your context (locale, language, location). A generic “keep me updated about NBA” came back with “Updates on the Brooklyn Nets“, a local team injected from context.
“keep me updated about nba” → the assistant proposes scores, team and player news, and “Updates on the Brooklyn Nets”, a locally-injected entity.
On Google’s side, your sentence is turned into a set of instructions that feed the retrieval stage, with an “offline” path (applied over time) and a real-time one. This is, as we read it, the architectural shift the feature embodies: from inferred interest vectors to a natural-language profile you write yourself.
3. A six-month timeline (December 2025 → June 2026)
We documented the rollout as it happened. Dates anchored to in-app captures and the feature’s own changelogs; a few are approximate (marked with “~”).
December 2025: first sighting (Search Labs, US). The feature appears with everything described above: the chat panel, the JSON response, the persistent thread, the four intents, local context injection. First impression: the effect is subtle; after several refreshes you see a few on-topic cards, but nothing dramatic. (field note, example)
“I need a break from negative news. Show me more feel-good stories, but keep the local and breaking news.” The assistant lists the kinds of uplifting content to expect, then offers “Refresh your feed.”
Another early example: “can you show me https://dev.to/ on my feed” → the assistant offers programming articles, web-dev tutorials and software news.
~January 13, 2026: the attribution tag. Google starts marking cards “resulting from natural language tuning” after a refresh and the SEE_MORE/SEE_LESS arbitration, making it possible, for the first time, to tell which cards a prompt actually changed. A prompt history also appears in My Activity. (field note)
“‘Tailor your feed’ preferences in Discover. View and manage your prompts.” A dedicated My Activity surface (here, empty).
~February 2026: “historical” tuning, and what SEE_LESS really does. A second tag appears, “historical natural language tuning”, for cards influenced by a past prompt. Testing “fewer X posts”, we saw Google replace X (Twitter) cards with YouTube videos, and, notably, asking to remove a topic does not truly remove it: you get SEE_LESS, but the topic isn’t deleted from the feed. (field note)
~February 2026: the “niche” test. Asked for “more niche / small sites”, the feed came back, on the first refresh, with 2 of 10 cards modified by the prompt (a one-off snapshot, not an average), surfacing VentureBeat and Mississippi Today, with the very first result driven by the request. (field note)
~February 2026: the “entity” test. Asked for “more articles from a specific creator”, Google understood the topics related to that creator (entities), refreshed, and surfaced a LinkedIn post from them, tagged “natural language tuning content”. (field note)
~April 2026: the rebrand + the “You asked to see” label. “Tailor your feed” becomes “Add topics to your feed” with a chat UI, and a visible “You asked to see” label now marks the cards served by the pipeline. (rebrand, label)
A card labeled “You asked to see”: a historical-figures listicle from AOL.
“You asked to see” on a Guardian politics story.
“You asked to see” on Search Engine Roundtable cards; the publisher requested earlier now surfaces explicitly.
May 22, 2026: the query intents (the fan-out). We confirm that, beyond entity expansion, some pipeline cards carry a stored query intent, the prompt decomposed into specific retrieval queries, on the same pattern as the GEO “fan-out”. (More in section 5.) (field note)
“You asked to see” also applies to video: a Crunchyroll/One Piece YouTube card surfaced by a prompt.
June 2026: current state. The entry point reads “Add topics to your feed”, and the feature now surfaces small creators well outside the major publishers (section 6).
4. The pipeline behind it: historicalnaturallanguagetuningcontent.f
Every Discover card can be traced back to the pipeline that selected it. “Tailor Your Feed” maps to a dedicated pair:
naturallanguagetuningcontent.f, content based on your current natural-language preferences.
historicalnaturallanguagetuningcontent.f, content based on past prompts that keep influencing the feed (the “historical” tag from the timeline).
The pipeline retrieves content in two distinct ways:
Mode A: entity / interest expansion (the majority). Based on the observed behaviour, the prompt is mapped to entities and topics, and the feed expands around them. This is why asking for one publisher surfaces related sources and topics, not just that publisher, the same logic as the Follow button. Most cards work this way, expanding around your topics rather than echoing the exact words you typed.
Mode B: query-intent fan-out (the minority). For a fraction of cards, the prompt is decomposed into explicit query intents, natural-language retrieval queries that fetch the article. This is the GEO “fan-out” mechanism, and it is the subject of section 5.
One behaviour worth flagging: in our tracking data, Google appears to promote these cards cautiously, on average less than other pipelines, and pulls them back more often than any other, consistent with a retrieval that sometimes matches loosely (we’ll see a concrete false positive in section 5). It is, by design, a targeted pipeline, not a mass-distribution one: its cards show essentially no growth over time, the lowest of any pipeline. It serves what was asked for, to the user who asked. It doesn’t snowball.
5. Query intent: the GEO “fan-out” inside Discover
This is the most interesting mechanism. For a slice of cards, the prompt is broken down into a specific query intent that matched the article, the prompt turned into precise, natural-language retrieval queries. It is the functional analogue of the fan-out described for Generative Engine Optimization: a single prompt is decomposed into sub-queries that retrieve content by semantic relevance, with no popularity prerequisite.
The decomposition is visible. A prompt about SEO becomes a set of informational queries:
User prompt (approx.)
Decomposed query intents
“Show me content from SEO”
“SEO strategies algorithm changes” · “Google ranking system updates” · “tips for getting content into google discover”
And those query intents then retrieve real articles. Here are anonymised query intent → URL pairs we observed (the formulations are the exact internal query intents):
Our analyzer groups these query intents into clusters:
Clusters of query intents observed in the pipeline: “learning Google algorithms”, “AI content machine learning SEO”, “buying Japanese property guide”, “healthy cooking techniques”, “anime recommendations 2026″…
The same clusters, with the article count per intent.
Drilling into one cluster shows both the strength and the limit of the fan-out:
The “buying Japanese property guide” cluster: japantoday.com (how to buy a home in Japan) and maigomika.com (rural Japan) are spot-on niche matches, but rockpapershotgun.com (Forza Horizon 6 in-game home locations) is a false positive, a video-game article pulled in by surface word overlap. Loose matches like this are why Google ranks this pipeline so cautiously (section 4).
Why this matters for SEO: the query intent reveals the exact vocabulary Google uses to map a prompt to your content. These are natural-language informational queries, not raw keywords. Aligning titles, H1s and intros with these formulations is the Discover-side equivalent of optimizing for the AI fan-out.
6. Niche sites and small creators: the popularity bypass
Classic Discover pipelines mostly re-serve content that is already popular, articles that have already circulated widely and built engagement. “Tailor Your Feed” works differently: the cards we observe show a retrieval that reaches for semantically relevant content regardless of whether it ever circulated in Discover before.
1492.vision tracking data backs this up. On historicalnaturallanguagetuningcontent.f, a majority of the cards point to articles with no detectable prior Discover distribution in our dataset, content that had never (or barely) been served in the feed before. This is, by a wide margin, the highest share of any pipeline: the classic news pipelines show the opposite profile, where almost every card has a long distribution history. A minority of the pipeline’s cards are the exception, mainstream articles (a Reuters story, for instance) already widely distributed and simply re-surfaced here.
The clearest illustration is a recipe prompt. Asking for vegan recipes surfaced, not the big food publishers, but small independent creators:
“Show me content from recipes vegan” → the assistant proposes plant-based weeknight dinners, vegan stews, high-protein tofu, vegan baking… then “Update your feed.”
The result: a “Sweet Potato Tacos” recipe from an independent creator, and “72 Vegan BBQ Recipes” from a small vegan blog, each labelled “You asked to see”, with a rating widget (“How would you rate this suggestion?”).
Across our tracking, the same pattern recurs: a niche gardening blog (buzzyseeds.com) for a seed-starting prompt; Mississippi Today for a “niche sites” prompt (with VentureBeat surfaced on the same prompt, as an illustration of the retrieval’s behaviour); a LinkedIn post for a creator prompt; niche Japanese-property blogs (japantoday.com, maigomika.com) for a property prompt. Most are not the usual high-volume Discover winners.
The takeaway, carefully put: the feature surfaces articles that had barely circulated in Discover before. The retrieval appears to be driven by relevance to the prompt, not by prior popularity: what works like a popularity filter in the classic pipelines is, here, bypassed.
7. What this changes for publishers
This is the part that matters most looking forward, and it’s a genuine shift in how Discover visibility can be earned.
Selection power moves to the user. In the classic feed, Google decides what you see from inferred signals. Here, the user writes the prompt (“show me more of X”, “less of Y”), and Google turns it into entities, interests and query intents that drive retrieval. Demand becomes explicit.
Consequence: niche sites can surface without a Discover track record. Because the retrieval appears to reach for relevance rather than prior popularity, a small site can be served the moment a user asks for its topic, even if it had never really circulated in Discover before (section 6). That’s new.
It’s a third path to visibility. Until now, a niche site broke into Discover only two ways: through strong implicit affinity (Google infers, from repeated engagement, that you love a topic, and a re-surface pipeline keeps feeding you that niche site), or through an explicit follow. “Tailor Your Feed” adds a third, user-initiated path that depends on neither.
The concrete levers for publishers:
Optimize for entities/topics (the dominant Mode A). Be unambiguously about what users will name. A clear topical focus → cleaner entity association → you’re in the expansion set when someone asks for your subject.
Optimize for query-intent vocabulary (the Mode B fan-out). Phrase titles, H1s and intros to match the natural-language informational queries a prompt decomposes into (the Discover-side of GEO). Section 5 shows the exact formulations Google uses.
What it is not. Publishers should be clear-eyed:
Not a mass channel. The pipeline shows essentially no growth, and Google promotes its cards cautiously. It serves the user who asked; it doesn’t broadcast.
Not publisher-triggerable. Only the user can fire it. You can be retrieval-ready, but you can’t activate it for your own site.
Geographically and adoption-limited. It is EN-only (Search Labs, US; ≈ 0% in French feeds), and adoption is still early; the My Activity surface was empty in our tests. The future impact depends on (a) a general rollout and (b) whether users actually adopt prompt-based tuning at scale.
The strategic read: if “Add topics to your feed” graduates from Search Labs and users embrace it, the demand-driven, popularity-agnostic retrieval it relies on is structurally favourable to small, focused, well-described sites, the kind that classic, popularity-dominated pipelines rarely reward.
What’s next?
A snapshot, not an endpoint. What we’re watching:
The French (and EU) rollout. The feature is EN-only. Depending on how independent it is from AIO[^aio] and AI Mode[^ai-mode] features, it could reach France sooner or later.
Adoption. A powerful mechanism with no users changes nothing. The empty My Activity surface suggests prompt-based tuning is still a niche behaviour. Mass adoption is the variable that decides whether this matters for publishers.
The current / historical pair.naturallanguagetuningcontent.f (live) and historicalnaturallanguagetuningcontent.f (persistent) suggest tuning is meant to last over time, a standing instruction, not a one-off.
Convergence with generative retrieval. A nascent generativeretrieval.f pipeline, spotted recently in our tracking data, suggests LLM-driven retrieval may reach beyond this one feature (to be confirmed).
The bigger picture: Discover is moving from observed personalization (Google infers) toward declared personalization (you tell it), and the retrieval that serves declared intent doesn’t lock onto popularity. That’s the structural opening for niche publishers, if and only if the feature ships broadly and users adopt it.
Notes
[^pipeline]: In Discover, a pipeline is a selection circuit that chooses and serves the cards. Each card can be traced to the identifier of the pipeline that produced it; that is what we exploit in our 1492.vision tracking data.
[^pipeline-id]: The .f suffix in identifiers such as historicalnaturallanguagetuningcontent.f is an internal marker we observe in the metadata of Discover cards; it distinguishes the selection circuits.
[^search-labs]: Google’s beta program that lets users test experimental features in Search and Discover before any wider rollout.
[^llm]: Large Language Model. We assume one here, without direct proof of the model used, from the conversational behaviour and the structured responses observed.
[^entity]: In the sense of Google’s Knowledge Graph: a named entity; a person, topic, organisation or concept that is identified and linked to others, distinct from the everyday sense of the word “entity” (a company, a legal structure).
[^fan-out]: A mechanism by which a single prompt is broken into several retrieval sub-queries, each targeting a different angle of the topic.
[^query-intent]: A natural-language informational query, derived from the user’s prompt, that was used to retrieve a specific article. Observable in our data as metadata attached to certain pipeline cards.
[^geo]: Generative Engine Optimization: the practice of optimizing content to be visible and cited in the answers of generative engines (AI Overviews, ChatGPT, etc.), notably via query fan-out.
[^retrieval]: The stage where Google fetches and selects the content to display in the feed, from interest signals, entities, or queries.
[^distribution]: In this article, the fact that an article had already circulated in Discover, as observed in our 1492.vision tracking dataset, not an audience figure nor a Search Console metric.
[^aio]: AI Overviews: generative answers shown at the top of certain Google Search results.
[^ai-mode]: AI Mode: Google Search’s conversational interface, distinct from AI Overviews.
Field tracking: Google app, Search Labs (US/English) accounts, December 2025 – June 2026. Pipeline behaviour derived from close observation of the Discover feed via 1492.vision tracking data. “Distribution” here means whether an article had already circulated in Discover, as observed in our tracking dataset, not private audience figures. The internal mechanisms are our interpretation of observed data and public research; inferred dates are flagged as approximate.
Google updated its AI Max for Search reporting documentation with new guidance on performance reporting, optimization best practices and a key timeline for Dynamic Search Ads advertisers.
What’s happening. Google refreshed its help documentation for AI Max for Search campaigns, expanding reporting guidance and providing more detail on how advertisers should evaluate campaign performance.
The update doesn’t introduce new product features, but it offers insight into how Google wants advertisers to manage and interpret AI Max campaigns going forward.
Why we care. The document update provides an indication yet of Google’s long-term plans for AI Max and the eventual phaseout of Dynamic Search Ads. With automatic DSA upgrades scheduled for February 2027, advertisers have more visibility into how their Search strategies may need to evolve over the next year.
The headline change. Google formally documented the upcoming transition from Dynamic Search Ads to AI Max.
According to the updated help page, campaigns using DSA will be automatically upgraded to AI Max starting in February 2027 as Google works to expand adoption of its AI-powered Search campaign format.
What’s new in reporting: Google added new reporting views that allow advertisers to analyze performance across:
Search terms.
Search terms and landing pages from AI Max.
Search terms from Dynamic Search Ads.
Search terms and landing pages from Dynamic Search Ads.
The company also clarified that search term reports show where users are directed after clicking an ad and highlighted new options for excluding underperforming search terms or landing pages through negative keywords and negative URLs.
Travel advertisers get new guidance. Google also introduced a new section dedicated to Search Campaigns for Travel.
The documentation explains how reporting consolidates performance data across multiple campaign components into a unified view, helping advertisers evaluate search terms, inventory performance and conversion outcomes.
Travel advertisers can also segment reports by ad format to compare performance across Travel Promotion Ads, Booking Links and Travel Feed-based ads.
A shift in optimization philosophy. The updated best practices place greater emphasis on intent-based targeting rather than strict keyword matching.
Google now advises advertisers to:
Focus on conversion goals over keyword relevance.
Review search term and item group performance every one to two weeks.
Use negative keywords sparingly.
Avoid over-filtering traffic that could benefit from AI-driven intent matching.
Bottom line. Google’s documentation update provides more than reporting guidance—it offers a roadmap for how advertisers should prepare for an AI Max-first future as Dynamic Search Ads move closer to retirement.
Cloudflare and beehiiv added AI crawl controls to beehiiv’s platform. This gives newsletter publishers a way to see, allow, or block AI bots from their dashboard as AI search becomes a new discovery path for web content.
The integration, announced Tuesday, embeds Cloudflare’s Crawl Control technology into beehiiv. It lets publishers manage how AI search engines and agents access their content, either by allowing crawlers for broader discovery or blocking scraping to protect archives for future licensing and monetization.
AI bot data comes to the dashboard. beehiiv publishers will get an on-platform dashboard showing which AI crawlers tried to access their content, which were blocked, and how much referral traffic those crawlers sent back. The dashboard gives publishers a side-by-side view of crawler activity, blocking decisions and referral traffic from AI services.
Publishers get simpler controls. The companies said publishers will be able to allow or block specific AI models with one-click permissions. Cloudflare will also update the system as new AI crawlers appear, reducing the need for publishers to manage robots.txt files, firewalls, or code changes themselves.
What they’re saying. Cloudflare CEO Matthew Prince said the partnership gives newsletter operators “transparency and control” as the internet changes; beehiiv CEO Tyler Denk said publishers need “real leverage” as AI changes how people find and consume content. From Cloudflare’s announcement:
“As AI models evolve to offer new forms of search and discovery, independent creators are looking for flexible ways to understand and manage how their content is accessed. This integration simplifies the process by letting beehiiv users manage their digital footprint through two clear choices: publishers can either opt-in to maximum discovery to allow AI search engines and agents to crawl their work freely for broader distribution, or choose content protection, blocking AI scraping to preserve their archives for future monetization and licensing opportunities.”
Why we care. The key question is whether publishers will actually use these controls once they are available. AI crawling has outpaced many creators’ ability to manage it, and adoption will show whether simple dashboard controls are enough to change publisher behavior.
Rollout starts now. The new controls are rolling out through beehiiv’s standard dashboard settings. All beehiiv users will get beta access to AI Crawl Control for visibility into AI crawler activity and traffic. beehiiv Max customers will also be able to block AI crawlers.
Shopify is introducing Campaign Autopilot, a new AI-powered marketing tool that automatically creates, manages and optimizes campaigns across multiple channels, reducing the need for merchants to manually manage advertising and email marketing.
The feature is launching in early access and is available directly within the Shopify admin.
What’s happening. Campaign Autopilot uses AI to plan and run marketing campaigns on behalf of merchants across channels including Meta, Shop Campaigns and email.
Additional channels are already on the roadmap, including ChatGPT Ads, Microsoft Advertising and Snapchat.
Rather than requiring merchants to build campaigns manually, the system handles campaign creation, budget allocation and ongoing optimization automatically.
Why we care. Campaign Autopilot lowers the barrier to running multi-channel marketing campaigns by automating much of the work traditionally handled by agencies or in-house specialists. Instead of managing separate campaigns across Meta, email and other channels, merchants can set a budget and goals while Shopify handles campaign creation, optimization and budget allocation.
How it works. Merchants set a monthly budget, choose which channels to connect and define approval rules and guardrails.
From there, Campaign Autopilot:
Creates and launches campaigns.
Allocates budget across channels.
Adjusts spending based on performance.
Recommends email automations.
Monitors results and makes ongoing optimizations.
Merchants can approve campaigns before launch, modify budgets or pause activity at any time.
What’s different. Shopify is positioning Campaign Autopilot as an alternative to traditional campaign management tools and agency-led marketing.
The company says the system leverages performance data and patterns across millions of Shopify stores to inform recommendations and budget decisions.
Campaign Autopilot also operates separately from existing campaigns, meaning merchants already running Meta or Shop ads won’t see those campaigns altered.
The bigger picture. Shopify is increasingly embedding AI into merchant workflows, moving beyond ecommerce infrastructure and into growth and customer acquisition.
The launch reflects a broader industry trend toward autonomous marketing systems that can execute campaigns with limited human involvement while continuously optimizing performance.
What to watch. Shopify plans to expand channel support in the coming months, including integrations with ChatGPT Ads, Microsoft Advertising and Snapchat.
The company also says merchants can use its AI assistant, Sidekick, to review recommendations, trigger actions and monitor campaign performance.
First spotted. The update was spotted by Digital Marketing Consultant Susan Richards-Benson who suggested this feature for smaller ecommerce brands on Linkedin.
Since its inception in 2015, the Search Engine Land Awards have recognized exceptional marketers on an annual basis — showcasing outstanding work, providing well-earned exposure in coverage and interviews, and bestowing upon them the highest honor in search.
While there’s no single formula for creating a winning entry, our judges have seen enough submissions over the years to know what separates the truly exceptional from the merely good. The strongest applications don’t just share results, they tell a story. They provide context, demonstrate strategic thinking, and clearly communicate why the work mattered.
And because great advice shouldn’t be gatekept, we thought we’d bring some of those insights directly to you.
We asked several of this year’s Search Engine Land Awards judges to share their best advice for prospective entrants. From common mistakes to avoid to the elements that consistently stand out, their insights offer a valuable look inside the judging process and can help you build a stronger, more compelling submission.
Keep reading for a roundup of fresh insights from some of our judges. (And see the complete list of 2026 judges here!)
“A great entry is a story with a goal, an action, and a measurable outcome that ties back to it. Tell that story as well as you can, and include a deck that makes it easy to see exactly what you accomplished.”
– Amy Hebdon, Founder, Paid Search Magic
“Explain your tactics. Many entries just say “we used best practices”. Everyone’s best practices and tactics differ. Explaining the process that lead to your results will highlight your creative thinking, problem solving, and uniqueness. Showing your insights and thought processes helps your entry standout and showcase your company’s competitive edge.”
– Brad Geddes, Co-Founder, Adalysis
“I look for SAY which stands for: Situation, Action and Yield. Applicants should write a clear example of the situation, what they did and the result achieved over the time period,”
– Jo Juliana Turnbull, Growth Marketing Senior Manager, Holafly
“Show me the humans behind the metrics. We’re in a time where AI is reshaping search at a pace none of us have seen, and that shift matters…but the applications that rise to the top will lead with empathy, not just analytics. I want stories where I can see how your work built genuine trust with real people, not simply visibility in search and AI engines. I’m especially drawn to entries that embrace a wellness-based approach to their craft, and to teams who pair their quantitative wins with qualitative insight: the quote, the aha moment, the change in how someone felt about a brand or experience. Tell me how you held the human at the center – as strategy. If your project made people feel seen, understood, or genuinely helped, lean into that. Those are the stories I’ll be looking for.”
“Clearly state the challenge you solved, and back it up with data. Explain the strategy behind the tactics you used and the results they drove. Tell me not just what happened, but what impact did it have on your campaigns? What did you do differently as a result?”
– Melissa Mackey, Director of Paid Search, Compound Growth Marketing
“Evidence: charts, analytics, screenshots. Be detailed, specific, and share data.”
– Barry Schwartz, Editor, Search Engine Land
“Tell a story. Numbers get you in the room, but the story is what stays with the judges. I want to know what the problem was, why it was hard, what you tried, and what finally worked. That arc, the messiness of real work, is what separates a memorable entry from a forgettable one. The submissions that stick with me are the ones where I can feel the thinking behind the decisions, not just the outcome. You did great work this year; now, make the judges feel the weight of what you solved before you show them the numbers.”
– Ameet Khabra, Founder, Hop Skip Media
“The two main things all award-winning entries share are that they explain the whys behind the hows, and they bring receipts (data to back up claims). If you can’t share the data behind your entry (budgets, revenue, etc.), you are putting yourself at a distinct disadvantage and may end up wasting the entry fee. A lot of people submit the same practices – if you can distinguish yourself by showing innovative thinking, you’ll do well!”
– Navah Hopkins, Product Liaison, Microsoft
“Give me all the data you can. Show the numbers and the real impact of whatever you did; conversions, ROI, and whatever monetary increases you were able to cause.”
– Celeste Gonzalez, Content Implementation and Product Specialist, Lastmile Retail
“Show me something I haven’t seen before, then prove it worked. The applications that land are the ones with a genuinely unexpected approach backed by numbers that make the result undeniable.”
– Adam Tanguay, Head of Growth, Jordan Digital Marketing
“I am looking for an approach or strategy that challenges the norm of SEM. A unique approach that focuses on achieving the business goals by way of campaign structure across Brand, Non-Brand, Performance Max, Conquesting, and general upper-funnel tactics. An advanced way of thinking about the target audience, messaging, conversion goals, etc. that helps show a sophisticated way of managing the campaigns & overall strategy to exceed business goals.”
– Matt Devinney, Director, Client Partner, Tinuiti
“I am looking for projects that break new ground with innovative takes on SEO, and are backed up by data and numbers-driven insights every step of the way.”
– Olya Ianovskaia, Founder and Lead Consultant, MycoMinds SEO
“Make your entry easily readable. We are going to need to go through several entries – I know the entries could be quite technical (and the quality of that will take precedent), but I am more likely to vote for you if I enjoyed reading your entry.”
– Anu Adegbola, Paid Media Editor, Search Engine Land
“My #1 piece of advice is to showcase strategy that truly breaks new ground. Award-winning applications demonstrate innovation that anticipates where SEM is heading, whether that’s leveraging AI in novel ways, pioneering audience-targeting approaches, or developing unique cross-channel integration. But innovation alone isn’t enough. The most compelling entries connect these forward-thinking strategies directly to measurable business outcomes, providing clear evidence of how your work translated to client growth metrics that matter. We’re looking for that perfect balance: creative execution that pushes boundaries while delivering documented ROI that proves your approach wasn’t just innovative—it was transformative.”
YouTube is expanding its suite of creator marketing and campaign intelligence tools with new Gemini-powered features designed to help brands identify trends, understand creator audiences and improve campaign performance.
What’s happening. Google is introducing several new insights and optimization tools across YouTube and Google Ads that give marketers more visibility into trends, creator performance and audience behavior.
The company says the new capabilities are intended to help advertisers make better creative and media planning decisions in an increasingly AI-driven marketing landscape.
Why we care. These updates provide deeper visibility into what’s trending on YouTube, which creators are resonating with audiences, and how their brand is performing across both paid and organic content. That can help marketers make smarter decisions about creator partnerships, campaign planning and creative strategy.
What’s new:
More detailed trend insights.
Google Ads’ Insights Finder is gaining expanded trending insights in the U.S., providing advertisers with a more granular view of what’s gaining traction on YouTube.
Brand Pulse data comes to Insights Finder.
Select Brand Pulse metrics are now being integrated into Insights Finder, allowing brands to evaluate both their paid and organic presence in a single location.
New creator insights API.
The new Content & Creator Insights API gives agencies and partners deeper information about YouTube creators and their audiences, helping improve media planning and creator selection.
Gemini-powered creative recommendations.
Google says Gemini will soon provide creative optimization tips for Demand Gen campaigns, including recommendations on visuals and creative elements that may improve performance.
The bigger picture. As creator-led content becomes increasingly influential in purchase decisions and brand discovery, advertisers are looking for better ways to identify emerging trends and measure creator impact.
Google is betting that AI can help marketers surface those insights faster and make campaign planning more efficient.
There’s a trap door waiting for DTC brands that invest in Google Ads that makes your dashboards look amazing, but absolutely wrecks your P&L.
It’s the danger of recycling traffic from Meta.
Thanks to the overlap between paid search and paid social traffic, running Google as a standalone channel is incredibly difficult if you don’t know how to set it up. Ad platforms refuse to share data with one another, and they love to claim credit for the same conversion — even if those sales would’ve happened without the influence of ads.
The DTC brands I speak to are often proud to show off their new customer numbers: month-over-month growth, a steady upward trend, and a fantastic dashboard. But when we go deeper into the data, we often find that a big chunk of those “new” customers are:
Conversions that would’ve happened because of brand or content efforts.
Customers who aren’t truly incremental because they consumed ads on multiple platforms.
The same people signing up with multiple email addresses.
You could argue that these overlapping sales still count as revenue, and they do. But when you look at the contribution margin from those sales, they cost far more than they should and erode actual profit.
In other words, you lose money when you run ads on both platforms without guardrails.
But that doesn’t mean you need to stop or limit yourself to one channel. Instead, you need a better system for measuring actual customer acquisition.
PSA -> removing brand searches from a PMAX campaign doesn’t change the audience you’re targeting
it’s still warm traffic (either existing or already multiple touchpoints)
You’re thinking about it wrong
Very common
— Collin Schmelebeck (@SchmelebeckPPC) July 24, 2025
Why the new exclusions matter
If you’re spending five figures or more on Meta, TikTok, AppLovin, or any other top-of-funnel channel, you’ll want to minimize overlap with other channels to drive actual new customer acquisition.
Here’s what that looks like:
Someone sees your ad on Facebook or Instagram.
They visit your site, browse, and leave without buying.
A while later, they search for your brand on Google or get retargeted on YouTube.
Performance Max swoops in, grabs the conversion, and reports strong ROAS.
You may have won that order anyway, but now Google and Meta both want credit for it.
Now you’re paying two or more platforms to recycle a conversion that you might have earned with just one.
1/6 Stop guessing what drives your Performance Max results and start taking control with our latest updates.
Ever since Performance Max launched, there wasn’t much you could do about this. It’s been a bit of a black box that automatically goes after the warmest traffic it can find: branded search, site visits, email subscriptions, and existing customers.
It lets you bid more for new customers, but you can’t really stop the campaign from defaulting to easy mode.
A while ago, Google began letting you exclude people searching for your brand on Search and Shopping. Performance Max still targeted warm audiences through YouTube, Gmail, and the Display Network.
The latest round of updates from Google has finally addressed this problem. You can now force Performance Max to focus on net new customer acquisition through a combination of brand exclusions, audience exclusions, and Customer Match data.
First-party audience exclusions, announced in March, are the final piece that makes this possible (though not foolproof – customer list matching is never perfect).
A four-step framework for net new customer acquisition
Here’s a four-step framework we’re using at my agency to help clients maximize incrementality.
Google PMax is probably taking credit for conversions your Meta Ads created.
Here's the 4-step framework to force it into real new customer acquisition: https://t.co/BmgDYpLzoL— Menachem (Google Ads) (@MenachemAni) April 28, 2026
Step 1: Exclude your brand
This one has been around for a while, but it’s the foundation, so we have to start here.
For smaller brands, brand exclusions usually aren’t necessary. But once you’re spending real money and seeing more than 15% to 20% of your cost or revenue coming from brand searches, it’s time to take action.
There are two parts to this.
Go into your campaign settings and add a brand exclusion. If your brand isn’t already on the list, click New brand list, create one, and add your brand. Google will do its best to block branded queries from this list.
Because brand exclusions aren’t foolproof, go to the Keywords tab inside the campaign and add your brand name as a phrase match negative keyword. Add a few common variations, too. This catches anything the brand list misses.
If you’re excluding brand terms from Performance Max, you need a dedicated brand Search campaign and a brand Shopping campaign to capture those searches. Otherwise, you’re just leaving money on the table for competitors.
Step 2: Exclude website visitors and email subscribers
Even if you blocked brand searches, Performance Max would still retarget people who visited your website, opened your emails, or interacted with your brand on YouTube, Gmail, Discover, and Display. So even with brand exclusions in place, a big chunk of your spend still went to warm traffic.
Now you can change that. Go to your campaign settings and find the new audience exclusions option. Then build a few remarketing lists:
All website visitors: Set this up through the Google Ads pixel or Google Analytics. It captures anyone who has visited your site.
Email subscribers: Connect Klaviyo (or whatever ESP you’re using) directly to Google Ads. The benefit of the Klaviyo integration is that the audience updates in real time, so new subscribers are added automatically.
Once you exclude these audiences, Performance Max can only go after people who haven’t interacted with your brand in any meaningful way. What we typically do, and what I recommend, is to come up with an engagement metric that fits each account’s business goal, such as cart adds rather than visitors from the past seven days.
What a change from how this campaign type used to work.
Same idea as Step 2, but specifically for people who have already bought from you. You can do this two ways.
Through a pixel-based audience that captures anyone who has triggered the purchase event.
By uploading your customer list directly. Shopify now lets you set up Customer Match lists right inside the Google Shopping app, and Klaviyo can do this, too.
Add these audiences to the exclusions section of your campaign, and you’re done.
A small caveat to keep in mind: audience matching is never 100%. If you upload a customer list of 1,000 people, Google might only match 900 of them. So you’ll still see some level of bleed. But going from “the campaign is targeting all my existing customers” to “the campaign is targeting maybe 10% of them” is still a huge win.
Step 4: Use ‘New Customer Bidding’ in campaign settings
The last piece is to tell the campaign explicitly that you want new customers.
In your campaign settings under customer acquisition, you’ll see two options: bid only for new customers, or bid higher for new customers. Both require you to connect a customer list (which you’ve probably already done by Step 3).
The “only new customers” option is the most aggressive setting. The campaign simply won’t bid on existing customers. Combined with the audience exclusions from Steps 2 and 3, this gets you as close to pure new customer acquisition as Performance Max will allow.
The “bid higher for new customers” option is more flexible. You set a dollar value that represents the additional value of a new customer, and the system bids more aggressively when it thinks an auction will result in one.
Here’s where you need to be careful. If you tell Google a new customer is worth an extra $100, and you get a $200 sale from a new customer, Google will report it as $300 in revenue. That extra $100 is a fictional reporting value, not real revenue. It will inflate your ROAS numbers and distort your target ROAS bidding.
Our recommendation is to use a small placeholder value, such as a penny or a dollar, when you want to nudge the system toward new customers without distorting your reporting. Or use a number that genuinely reflects the lifetime value premium of a new customer to your business.
What to expect from this approach
It’s still early, so we can’t draw firm conclusions yet. But based on my experience managing PPC for ecommerce brands, here’s what I expect to happen.
Many advertisers who walked away from Performance Max did so because it was simply recycling Meta traffic. By splitting it out, you force it to go after net new traffic.
This will likely benefit brands that don’t have a ton of video creative for YouTube, which is another platform where brands try to drive net new acquisition at the awareness stage.
One of the big differences between Performance Max and Demand Gen is that the former is much more conversion-focused. Any brand considering excluding branded Search and Shopping from Performance Max should also consider this tactic, as it tends to over-index on hot traffic.
In terms of outcomes, I expect the reported ROAS attributed to Performance Max to be lower than what you may have seen in the past.
But when you look at the breakdown of new versus returning customers, it should align much more closely with new customer acquisition. Without advanced configuration, it might be a 60/40 split, even in the best situations.
Limitations and realistic expectations
Nothing about this is foolproof. Audience exclusions don’t match perfectly. Brand exclusions don’t catch every variation. Customer Match has its gaps. So even with all four steps in place, some percentage of your spend will still hit warm audiences.
But for the first time, you actually have the levers to push Performance Max into upper-funnel territory. You can make it work like a real prospecting channel instead of a retargeting channel that takes credit for demand created elsewhere.
This matters most for brands spending heavily on Meta, TikTok, or other channels and wanting Google to actually grow the customer base rather than recycle the traffic those channels generate. If you’re seeing strong ROAS in Performance Max but flat new customer numbers month over month, this framework is for you.
If you’re a smaller brand still trying to find product-market fit or build initial momentum, this is probably overkill. Let Performance Max do its thing and pick up conversions without too many restrictions.
But once you’re scaling and the question is no longer “Can we be profitable?” but “Can we be profitable while growing the customer base?” these settings become some of the most important levers you have.
Google’s giving you more control over PMax. Use it.
The conversation around brand versus non-brand is everywhere. You can’t throw a dart at a paid media conference without hitting someone with a strong opinion on it. But for some reason, almost no one seems to be testing this new option.
I just finished auditing an account spending $100,000 a month on Search with no Performance Max or Shopping, so they get purely new customer acquisition. We looked at their numbers and said maybe now’s the time to try this, exclude all these segments, and let it rip.
So here’s when I recommend implementing this test: if your ad spend is high enough (it doesn’t need to be $100,000 a month or anywhere near it), or you’re revisiting Performance Max. Your hypothesis should be that this approach increases the proportion of actual new customer conversions.
I think you’ll find that the needle moves further than you think.
AI has made SEO teams ambitious about what they can automate. Tasks that previously required engineering support can now be solved with the help of Claude or ChatGPT.
That’s exciting, but it also creates a new problem: thinking you can automate everything. In modern language, that often comes down to one question: Should we build or buy this new tool?
This build-versus-buy dilemma has never been simple, and AI has made it even more complicated. The challenge goes beyond cost. It involves security, maintenance, data access, internal capabilities, workflow fit, and whether a custom solution will remain maintainable, reliable, and useful six months from now.
How AI lowers the barrier to building
AI has lowered the barrier to experimentation. Even without technical knowledge, you can now create a custom GPT, build a workflow, connect data sources, or create an internal AI assistant.
But that doesn’t mean the same person can build and maintain a tool that will remain reliable over the next few years.
In most cases, AI can help SEO teams analyze data, identify patterns, summarize information, and recommend actions. It can save a lot of time, and teams that ignore AI are clearly falling behind.
But, at least for now, AI isn’t doing truly creative work in the same way humans do. It works from existing patterns and predicts likely outputs. That may change in the future.
AI also comes with hidden costs. Internally built tools are often treated as free because the invoice usually doesn’t sit with the SEO team. But that doesn’t mean token usage, API calls, infrastructure, engineering time, security reviews, and maintenance don’t cost money.
We are already seeing this effect. Reuters has described it as “corporate AI sticker shock,” with companies struggling to forecast usage-based AI costs. TechCrunch also reported that Uber introduced AI spending caps after blowing through its annual AI budget in four months.
Today, marketing teams aren’t the heaviest AI users, especially compared with engineering teams. But that can change quickly.
And when usage grows, the bills will grow too. That will naturally make companies ask which AI tools and AI-powered workflows create value and which ones only consume budget.
Before deciding whether to build or buy, SEO teams need to define what they really need.
Different ways to use AI and automation
Many teams group these solutions together, but they vary significantly in cost, complexity, and maintenance requirements.
A custom tool: A more complex internal system that usually needs engineering support. It is often more about automation, but it can have an artificial intelligence aspect.
A custom workflow: A repeatable process built with different tools, such as a custom GPT, Claude project, spreadsheet, reporting template, and so on. It often includes automation, for example, a scheduled task in an AI tool, and it usually has an artificial intelligence layer.
A custom layer on top of SaaS: Using data from existing tools and shaping it into your own reporting, prioritization, or recommendation workflow.
A true AI agent: A system that can take more autonomous actions. For example, it can scan your Slack and follow up with people you are still waiting on.
These aren’t the same, but people often label them incorrectly. Calling everything an “AI agent” creates confusion and can lead to wrong estimates about cost and complexity.
Look for repetitive, context-rich tasks
We’re still experimenting. Most of what our team has built focuses on daily tasks that require a lot of manual work.
For example, we’ve created a custom GPT that evaluates whether our content matches our personas and their pain points. The goal is not to replace the human copywriter or reviewer. It is to determine whether a piece remains generic and whether a few additions can make it more relevant.
We are also using AI for translations, monthly reporting, and a weekly summary that combines meeting notes, Slack, and Jira, and helps me see whether I have missed adding a task to Jira or where I still need to follow up.
One of our latest workflows transforms recorded internal meetings into organized landing page briefs.
These types of tasks are good candidates for AI-powered custom workflows because they rely on internal context, repeatable processes, and company-specific knowledge.
One example from our team was a prompt tracking tool that my colleague vibe-coded. It worked well as a starting point. But the data presentation was not perfect, and it was hard to create a trend graph without additional manual steps.
Soon, it became a maintenance burden because every external change in any of the LLM tools required fixes, for which we needed engineering help.
The real issue was reliability. For AI visibility and prompt tracking, we needed consistent data in one place, presented in a way we could analyze over time. That is why we moved to a specialized platform like Peec AI instead of continuing to maintain our own version.
That experiment was still valuable. It helped us understand the problem, the complexity, and the features we actually needed from an external vendor.
And this is one of my pieces of advice: whether you want to build a tool internally or buy one, always test what is already available on the market. Only then will you really understand what you actually need. You may think you need 10 features, only to realize you use only three.
For business-critical tools such as rank and AI visibility tracking, and website crawling, small SEO teams without dedicated technical support should usually be careful about building from scratch. If the data is fundamental to decision-making, reliability should be your main decision factor.
Use AI where your data already lives
Buy the crawler, rank tracker, or AI visibility platform. Then focus your internal efforts on connecting data from these tools to custom information, such as your GA and GSC accounts or even CRM data. Once connected, create reports that combine all these sources and enable you to analyze everything in one place.
MCP connections are also worth considering. The Model Context Protocol is an open standard for connecting AI applications to external systems, data sources, tools, and workflows. With MCP servers, you can analyze data from your primary tools directly using AI, taking your current workflows to the next level.
This doesn’t mean you’re required to learn how to code. But they need to know enough to ask the right questions.
If a tool connects to an internal knowledge base, customer data, or proprietary research, you should be aware that this could pose a security risk. And it might turn out that it is better for the company to dedicate an engineer to support you rather than risk exposing sensitive information.
You should also understand what the final cost will be for your company when you decide to go with a custom tool. Custom tools aren’t free just because the invoice doesn’t sit with SEO. Engineering time, security reviews, AI tokens, and API usage are all part of the cost.
Before asking leadership for a tool, SEO teams should be able to explain the workflow problem, the expected value, the cost of buying compared with the estimated cost of building, and what might happen if nothing is done.
The best requests don’t start with: “We need this tool.”
They start with: “Here is the problem, here is why it matters, here is what we’ve tested, and here is the best way we think we can solve it.”
How to prioritize what to build first
There’s no single prioritization matrix that will work for every situation.
A website crawler, a content evaluation tool, a report builder, or a competitive intelligence system can’t be judged by the same criteria.
If you are in a situation where you think you need more than one tool, start by mapping your current workflow and what your ideal situation looks like.
Once you do that, the patterns will be clear. Often, your strongest priorities will fall into two groups.
The first are tools that can support revenue creation. SEO teams are usually part of the marketing organization, and marketing is expected to bring visibility or leads. If a tool can help identify content opportunities, improve conversion rates, increase AI visibility, or surface gaps versus competitors, it can be seen as a priority.
The second group is workflows and tools that can help you minimize repetitive manual work. This category may not create revenue, but it will give your team time back to focus on more strategic work.
Don’t forget that quick wins also matter. Stakeholders don’t want to wait three months before seeing results. A smaller project that can bring value in three weeks will help you build trust and make it easier to get support for bigger initiatives.
Cross-team value should also be part of your decision.
SEO problems are often not just problems for your team. Competitive intelligence, for example, matters to PPC, ABM, content, product marketing, and sales, too. If several teams share the same pain, the business case becomes stronger.
So don’t be afraid to act as a cross-team synchronization layer when needed. Talk to the same teams you have already worked with, and try to understand their workflows and pain points, and where your needs overlap.
And remember, the best tool is not always the most ambitious one. Starting with something small is often the smartest move.
Google has confirmed that it has expanded access to the new Google Search Console AI performance reports to more users. Google’s John Mueller wrote on Bluesky, “We’re just rolling these out incrementally to sites, and reviewing the feedback along the way. I know everyone wants the new shiny thing immediately… but first, patience.”
AI performance report. The report shows you how well your content and websites are performing in AI responses, AI Mode, and AI Overviews in Google Search. The reporting includes impressions, pages, countries, devices, and dates, but does not include click data.
Expanding access. This morning, I spotted a number of SEOs posting that they are now seeing the report, and that it is not restricted to sites in the United Kingdom. Some are seeing the report for sites in the United States, India, Switzerland and so forth.
And as I quoted above, John Mueller from Google confirmed the search company is “rolling these out incrementally to sites.”
What it looks like. Here is a screenshot of this report:
Why we care. Site owners and publishers have been asking for controls over whether and how their content is shown in Google’s AI features since Google launched these a couple of years ago. Well, Google is rolling out this feature to more of its users today. It is unclear how soon everyone will gain access to these controls, but I am surprised that Google expanded access this quickly. Specifically within 20 days from its first release.
Many budget allocation strategies assume that every channel follows the same pattern: the first dollar is the most productive, and each additional dollar yields a slightly lower return.
The charts below show what that pattern looks like.
The log shape means that the first dollar is the most productive, and each subsequent dollar is worth a little less. When every channel looks like that, the game plan is to spread the budget to as many channels as possible and equalize the marginal CPAs to maximize profit.
But not every channel looks like that. Some have a warm-up region where the early spend is the least efficient, not the most. On those channels, the logic above breaks, and so does the “test small, scale the winners” playbook that most of the industry runs on autopilot.
The difference comes down to one question about the channel: Is the response curve C-shaped or S-shaped?
The answer can change how you approach channel testing and channel measurement, including any MMM analysis. Moreover, Google has been incorporating more S-shaped campaign types, and after its Google Marketing Live announcements, this trend seems set to continue.
The two shapes — and the only part that matters
The response curve plots output (conversions, revenue) against input (spend). This generally results in two types of curves in marketing.
C-shaped (concave): Diminishing returns from the very first dollar. A log or power curve. Picture the top-left quarter of a circle: steep at the start, flattening as you go.
S-shaped (sigmoid): A slow, inefficient start, then an inflection point where it gets steep, followed by a flattening into saturation. A logistic curve.
The response curve itself isn’t what you allocate against. You allocate against the marginal curve, the derivative, which answers the question: “What did the next dollar buy me?” That’s where the shapes diverge in a way that matters.
For a C-curve, marginal return is highest at the first dollar and falls in only one direction. Marginal CPA rises from the first dollar onward. If conversions are a*ln(s), marginal conversions per dollar are a/s, so marginal CPA is s/a, climbing in a straight line as you scale. There’s no warm-up. The cheapest conversion you’ll ever buy is the first one.
For an S-curve, marginal return starts low, rises to a peak at the inflection point, then falls. Marginal CPA is U-shaped. It’s expensive at the start, bottoms out around the inflection point, then climbs into saturation.
That region of increasing marginal returns is the whole story. It’s the difference between a channel where small budgets are productive and one where they are wasted.
Say your CPA goal is $50. Here is an S-shaped channel, modeled as Conversions = 1000 / (1 + e^(-0.25(s – 20))), with spend in the thousands and the inflection at $20,000/month:
Run the $10,000 test that a sane person runs before committing real budget. Average CPA comes back at $132, marginal around $94. If those two metrics are all you look at, you conclude that this channel can’t hit $50, so let’s kill it.
That verdict is wrong. At $20,000 to $25,000, the channel is running at an average of $32 to $40, and the marginal dollar in the $15,000 to $25,000 band costs $18. That’s not “barely viable.” In that band, it’s the best marginal buy you have. The small test fell within the warm-up and reversed the conclusion.
In a C-shaped channel, the small test would have shown you the best the channel can do. On an S-shaped channel, it shows you the worst.
This is the trap. The standard playbook is “test small, scale what works.” On S-curves, small tests systematically condemn channels that would’ve worked at scale because the test is structurally stuck in the inefficient region.
The optimization is convex. There’s one global optimum, the equimarginal rule from the marginal-CPA post applies cleanly, and the solution is usually interior, meaning lots of channels get funded.
Even a small allocation is productive because the first dollar is the best dollar. Run many channels lean, reallocate continuously at the margin, and pull back the instant marginal CPA crosses your goal.
S-shaped channels, go deep or skip
The optimization is non-convex. A small allocation can be strictly worse than zero because below the inflection your marginal return sits under your target, and you’ve sunk money to get nowhere.
The decision isn’t “how much.” It is binary: commit past the threshold, or don’t fund it at all. There’s a real minimum viable budget, and it’s often above normal test budgets. You can’t sprinkle an S-curve and expect efficiency, and you can’t evaluate one on an underfunded test.
Those two rules can look like they fight each other, but that’s only true to a certain point. Past the inflection, an S-curve is concave, so the equimarginal rule governs it exactly as it governs a true C. The S-specific instruction — commit a block instead of sprinkling — is only about the trip from zero to past the inflection.
Shape is therefore mostly a launch-and-evaluation problem. Getting a new prospecting channel into its efficient range requires a committed block and patience with ugly early numbers. Once it clears the inflection, you manage it at the margin like everything else, right up until you consider cutting it hard, where shape matters again because the downside is a cliff, not a ramp.
This is the part that’s genuinely counterintuitive, and it echoes the original marginal-return point: The right move isn’t always the one that looks most efficient at a small scale.
Which channels are which?
The historical default was concave. Simon and Arndt reviewed more than 100 studies and concluded that advertising follows the law of diminishing returns, a concave response.
The dissent came later: Vakratsas, Feinberg, Bass, and Kalyanaram found that threshold effects do exist and that response is not necessarily globally concave. Their explanation for why thresholds were so hard to find is the useful part. Mature accounts already operate inside the effective range, so the warm-up never shows up in the data, and most studies fit a concave model (the double-log) that can’t reject an S-curve even when one is present.
The platform shift has made the threshold visible again. Here is a fuller map, ordered roughly from C to S. The shape column is an inference from how each system targets and learns, not a measured constant, and the right shape for your account still has to be measured.
Two rows do most of the work.
AI Max is the live example of a channel migrating from C toward S. Swapping explicit keywords for broad and keywordless matching means it needs conversion volume to learn which queries convert, so below a data threshold, it explores badly.
The mixed independent results fit that: Google reports about 14% more conversions on average and up to 27% for exact-match-heavy campaigns, while independent testing reports 84% of advertisers seeing neutral or negative results. Much of that spread is accounts that turned it on without the conversion volume to clear the learning region.
Performance Max is the trap, because its curve is a composite. It blends a harvesting layer (branded, retargeting, Shopping against existing intent) with a prospecting layer (keywordless expansion across surfaces). The harvesting layer is a cheap C that pays off on the first dollar. The prospecting layer is the S underneath.
Blended, the early efficiency looks great, because you are mostly skimming demand you already had, and the average hides the prospecting warm-up entirely. That is also why the platform is glad to optimize it for you: the blend flatters the headline number. You can’t read PMax or run the shape analysis on it until you split the harvesting from the prospecting.
The throughline runs in two layers. Rules-based auctions capture the best inventory first, which yields concavity; machine-learning systems must be fed before they are efficient, which introduces a threshold. Underneath both, harvesting existing demand is concave and mostly non-incremental, while creating new demand is the S-shaped part where the real growth and the real warm-up cost both sit.
Average versus marginal: total over spend, or the slope where you stand.
What you allocate against is marginal incremental return, the slope of the incremental curve at your operating point. A holdout fixes the first axis only. Time-sliced marginal CPA on attributed data fixes the second only. A multi-cell scaling test gets both, at a cost.
MMM (method 1) estimates the whole curve from aggregate data and sidesteps click attribution entirely, but pays in identifiability and modeling assumptions instead. Most arguments about ‘what is working’ are two people standing on different axes.
There are two major cautions, and I would flag both as genuinely unsettled rather than settled facts.
Separating a true S-curve from “concave with a high half-saturation point” is hard, because a concave model will fit S-shaped data well enough to hide the inflection (this is the Vakratsas point, and it applies to your own dashboards as much as to academic studies).
The learning phase may be a one-time fixed cost to train the model rather than a permanent feature of the steady-state curve. If it is transient, the channel may behave concavely at the margin once it is trained, and the S you measured was a startup artifact. The truth is probably a mix: a one-time training cost, plus an ongoing minimum-volume requirement to stay efficient. Treat every shape call as provisional and re-check it.
One more failure mode, and this one is not unsettled science but a matter of where you are standing on the curve. An S only looks like an S if your data spans the inflection.
Above the inflection, an S is concave, mathematically identical to a C. Look at only the $20,000-and-up rows of the table above: marginal CPA rises monotonically from $18, a textbook C-curve, and the convex warm-up is invisible because you are no longer operating in it.
Established accounts usually sit past the inflection, which is exactly why Vakratsas found thresholds so hard to detect, and why you can run an S-shaped channel for years, correctly, while believing it is concave. The tell arrives the day you cut hard and fall off the inflection instead of easing down a slope.
When to go wide and when to go deep
The marginal-return post told you to equalize marginal CPAs across the program. That rule is still correct, but the shape of the curve tells you how you’re allowed to get there.
On C-shaped channels, you can get there by sprinkling, because every dollar is productive and breadth is the natural answer.
On S-shaped channels, you have to commit a block of budget past the inflection before the channel earns its place, and then concentrate rather than spread.
Lay the harvest-versus-create cut on top. Harvesting channels (branded, retargeting, non-brand search) are your C-curves: fund the first dollars, then cap them early, because they saturate fast and most of the tail isn’t incremental, no matter how strong the attributed ROAS looks.
Prospecting channels (Meta, YouTube, LinkedIn, the expansion half of PMax) are your S-curves and your only real source of incremental growth: commit past the warm-up or don’t start, and judge them on incremental lift rather than attributed CPA, or you’ll kill the thing that was working.
Classic search rewards going wide. PMax, AI Max, and Meta prospecting reward going deep on fewer bets and giving each enough volume to clear the warm-up. Run an S-curve like a C-curve and you’ll starve it, read the underfunded result, and kill a channel that would’ve been one of your best.
Amazon is bringing transactions directly into advertising with a new format that allows consumers to discover products, ask questions and complete purchases entirely through a conversation with Alexa+, potentially shortening the path from ad impression to conversion.
What’s happening. Amazon today introduced Alexa+ Agentic Ads, a new advertising format designed to let customers move from seeing an ad to completing a purchase without ever leaving the Alexa experience.
The format launches with partners including Papa Johns for food ordering and artists Beck, Jill Scott and Omar Courtz for concert ticket sales. The experience is currently available on Echo Show devices.
Why we care. Alexa+ Agentic Ads remove the traditional handoff between an ad and a checkout page, allowing consumers to complete purchases directly within a conversation. For early adopters, that could lead to higher conversion rates, lower drop-off and a new way to capture high-intent customers at the exact moment they’re ready to act.
How it works. Unlike traditional digital ads that redirect users to a website or app, Alexa+ Agentic Ads keep the entire purchase journey inside a conversation.
Users can engage with an ad, ask questions, compare options, check availability and complete a transaction through natural language interactions with Alexa.
The goal: eliminate friction between interest and purchase.
Concert tickets become conversational commerce. Amazon is initially showcasing the format through live event promotions.
Fans who see an ad for an upcoming concert can ask Alexa about show details, review available seats, compare pricing and purchase tickets directly through the device. Purchased tickets are then delivered to their Ticketmaster account without requiring them to open another app or website.
The experience is designed to transform entertainment advertising from an awareness channel into a direct sales channel.
Food ordering gets the same treatment. The format also extends to restaurant ordering.
A customer looking for dinner ideas could encounter a Papa Johns ad and begin placing an order immediately. Because Alexa+ can draw on previous interactions and preferences, it may suggest favorite toppings or commonly ordered meals before completing the transaction.
The entire process—from ad exposure to order confirmation—takes place within the conversation.
What to watch. Alexa+ Agentic Ads could offer an early look at how AI assistants reshape digital advertising. If consumers become comfortable completing purchases inside conversations, brands may increasingly view AI assistants not just as discovery tools but as full-fledged commerce platforms.
Anthropic’s latest job posting has the SEO industry abuzz. They may as well have titled it Search Gawd. The truth is, it’s everywhere.
To be transparent, I’ve written this job description a few times and interviewed for it. I’ve yet to see any of these roles get filled, but I’ll come back to that in a minute.
Sometimes the title is Head of SEO. Sometimes it’s Director of AI Search, VP of Search, Director of SEO, AEO and GEO, or — wait for it — Agentic Commerce GEO Consultant.
Lots of titles. The assignment is basically the same: own technical SEO, understand paid search, shape content, partner with engineering and product, build measurement, prepare for AI-mediated discovery, explain it to leadership, and turn it into growth.
The predictable reaction is that this is a lot of jobs rolled into one. An entire agency behind a single employee badge. Fair, but it misses the point.
Companies have been looking for this person for years. Generative search is just forcing the issue.
Publicis / Starcom: VP, SEO (Performance Content).
Accenture: Agentic Commerce GEO Consultant.
SailPoint: AEO/GEO Manager.
AirOps: Senior SEO Manager spanning SGE, Perplexity, ChatGPT, Gemini.
Responsive: Senior Manager, Web Strategy — SEO, GEO, plus Next.js, React, Vercel, DNS.
Danaher, Experian Health, Amazon News: some version of SEO + AEO + GEO.
Anthropic: SEO Lead, $255K–$320K.
Different industries. Different price points. Same job, unwittingly all looking for the same person.
Even the titles are arguing with the job descriptions
Agency X is hiring a “Director, SEO/SEM” whose responsibilities contain no SEO — just paid search, SEM platforms, vendor management, and a team of seven.
Consulting firm Y is hiring a “Director, SEO/AIO,” where AIO appears to be an in-house acronym no one bothered to define.
An indy agency’s “VP/Director, SEO” lists paid search, paid social, and pharmaceutical marketing among the nice-to-haves.
A token research firm is hiring a “Director, SEO & AEO” whose responsibilities actually describe SEO and AEO work — rare enough to be worth mentioning.
If the company can’t agree on what the role is before posting it, the candidate has no chance of meeting expectations that were never written down.
The taxonomy says one thing. The JD says another. The recruiter screens for a third. The hiring manager interviews for a fourth. The ATS filters out anyone worth a shit.
Looking for the missing link
You need someone who can see across technical search, content, PR, product, engineering, analytics, performance media, and brand — and understand that those functions were never as independent as the org chart suggested.
Search has always exposed the seams. A technical problem can look like a content problem. A content problem can be a product problem. A visibility problem may be an authority problem, not an optimization problem. Paid search often surfaces a messaging problem before brand research does.
Generative discovery makes those dependencies impossible to ignore. When results become answers, SEO stops being a traffic function.
At the risk of going full Yoda to avoid AI-slop speak: found, information is, only if infrastructure allows it. Content makes it understood. Brand makes it trusted. Product turns discovery into use — or it doesn’t.
You’re not asking one person to execute every task. You’re asking one person to understand how the pieces connect. That person exists. Your chances of finding that person through a conventional scoring system are slim by design.
The résumé will not look the way you expect
The value of this candidate isn’t captured by years under an SEO title or a checklist of software. The value is judgment:
Knowing which technical issue matters and which is noise.
Recognizing when the content team can’t solve the content problem.
Knowing when to spend, when to automate, when to wait, and when to tell leadership to stop doing that.
That judgment is hard to capture on a résumé. The candidate may have moved through agencies, publishing, product, consulting, and operating roles. Their career may look less focused than a specialist’s. That’s precisely why they can do the job.
Your ATS will screen them out. Your recruiter will flag them as “non-linear.” Your hiring panel will note they haven’t held the title before. Well, the title didn’t exist before. No one can agree on what to call it.
You can see how this search is already going sideways.
A less charitable possibility
Some of these processes may be less about filling a role than learning from the people willing to interview for it.
Senior candidates diagnose. They explain how they’d structure the function, where the organization is weak, what the first 90 days should look like, which tools they’d buy, and which work they’d kill. Invite enough of them in, and a company can collect competing organizational models and strategic priorities without hiring any of them.
Perhaps that isn’t the intent. But when a role stays open for months, gets repeatedly reposted, changes title and scope, and produces interviews that feel more like advisory sessions, candidates are entitled to ask what the company is actually buying: talent acquisition or knowledge harvesting?
The solution isn’t a shorter job description
The breadth is real, so cutting half the bullets doesn’t make the work disappear. Decide what you want. Is it:
A specialist who will execute?
A leader who will build a team?
An executive who can connect search, content, product, brand, and performance?
A consultant who can tell you which one you need?
Those are different jobs. Pretending they’re one role and waiting for a unicorn isn’t a strategy.
A closing note, since you asked
I would, however, be very good at the job. So would a handful of others who’d get screened out for the same reason.
The Anthropic job? Not getting it.
Five years under a title that didn’t exist five years ago — I don’t have them. My résumé reads like the job spec itself, in exactly the shape an ATS is built to reject. It’s an easy system to game. So easy that anyone worth their salt knows how.
The missing link is real. Generative search didn’t create it; it just made it harder to ignore. Before you hire someone to connect these systems, make sure your company can recognize them, hire them, and let them do the job.
The company that figures out how to recognize the candidate—not just write the job description—quietly wins the next decade while everyone else argues on LinkedIn about whether GEO is a word.
Google is widening its financial services advertiser verification program across much of Europe, adding new compliance requirements that could prevent unverified advertisers from running financial services ads in 24 European Economic Area (EEA) markets starting this summer.
What’s happening. Google announced new financial services verification requirements for advertisers promoting financial products and services in 24 additional EEA countries. The policy will begin rolling out on July 23rd.
The affected markets include Austria, Belgium, Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Finland, Greece, Hungary, Iceland, Latvia, Liechtenstein, Lithuania, Luxembourg, Malta, the Netherlands, Norway, Poland, Romania, Slovakia, Slovenia and Sweden.
Advertisers operating in financial services categories deemed in scope by Google will be required to complete verification if notified by the company.
Why we care. Failure to complete Google’s new financial services verification process could result in their ads no longer being eligible to run across 24 EEA markets. The update affects not only banks, lenders, insurers and investment firms, but also agencies managing campaigns on their behalf. Affected advertisers will need to secure regulatory verification through G2 and Google to avoid disruptions to campaign delivery, lead generation and revenue.
The big picture. The move is part of Google’s broader effort to combat financial fraud, improve advertiser transparency and ensure consumers see ads from legitimate, regulated financial providers.
Affected advertisers will receive in-platform notifications warning that “Ad performance may be impacted by financial services verification policy” and directing them to complete the required checks.
Failure to comply could result in advertisers losing the ability to serve financial services ads in the affected countries.
How verification works. The updated process involves two steps:
Complete verification through Google’s third-party compliance partner, G2.
Submit a financial services verification application to Google using the verification code provided by G2.
During the G2 review process, advertisers will be asked to provide information including:
The type of financial services offered
Regulatory licensing status
Registration numbers
Evidence of authorization by the relevant financial regulator, or proof of exemption where applicable
Agencies also affected. The requirements extend beyond direct advertisers. Advertising agencies and account managers running campaigns on behalf of financial services clients will also need to obtain verification for affected advertiser accounts.
This means both brands and the agencies representing them may need to complete compliance checks before campaigns can continue running.
A key distinction: third-party advertisers. Google is drawing a line between regulated financial institutions and third-party promoters. Advertisers promoting financial services with approval from a verified financial institution, but who are not themselves directly authorized by a regulator, cannot apply independently.
Instead, these “Approved Third Party Advertisers” must be verified through a sponsoring First Party or Authorized Advertiser, which must submit the verification request on their behalf.
Which services may be impacted? Google says verification requests may apply to advertisers promoting financial services categories including, banking, credit cards, credit and lending products, and more.
It is not an exhaustive list and may evolve over time as Google updates its policies.
What to watch. Financial brands targeting European consumers should review their compliance status now, as delays in verification could disrupt campaign delivery once enforcement begins later this year.
For agencies managing multiple financial services clients, the administrative burden may be significant, particularly as verification requirements increasingly become a prerequisite for advertising access across regulated sectors.
Google is changing how target-based bid strategies behave when campaigns are constrained by budget, aiming to make performance more consistent with advertiser targets even as budgets fluctuate.
The update will take effect on August 17th, with a new Bid Target Adjustment Tool becoming available on July 6.
What’s happening. Google says campaigns using target-based bidding strategies, such as Target CPA, will more closely align with their configured targets when budget limitations exist.
The company is also introducing a Bid Target Adjustment Tool that will allow advertisers to review and modify targets before the changes take effect.
Why we care. Campaigns that have been outperforming their target CPA or ROAS goals may no longer continue doing so automatically after the update. Google’s changes are designed to make budget-constrained campaigns adhere more closely to their stated targets, which could alter performance and efficiency if targets haven’t been reviewed recently.
For example, a campaign using a Target CPA of $10 that is currently achieving a $5 CPA could see performance move closer to the $10 target unless the advertiser updates the target setting.
The new Bid Target Adjustment Tool gives advertisers a chance to proactively update bidding goals before the August rollout. For some advertisers, failing to adjust targets could mean paying more per conversion or seeing performance shift toward Google’s target rather than the campaign’s historical results.
Why Google is making the change. According to Google, the update is intended to reduce volatility and create more predictable performance when advertisers increase, decrease or otherwise adjust campaign budgets.
The company says the new tool will help advertisers align bidding targets more closely with actual business objectives before enforcement begins.
What advertisers should do. Google is encouraging advertisers to review campaigns that use target-based bidding strategies and evaluate whether existing targets still reflect desired outcomes.
Advertisers will receive notifications in their Google Ads accounts before the rollout and can use the Bid Target Adjustment Tool to identify campaigns that may be affected.
Between the lines. The update could have significant implications for advertisers whose campaigns are consistently outperforming their targets. In some cases, maintaining current performance levels may require lowering target settings rather than leaving them unchanged.
Rankings, traffic, and conversions still matter. But they don’t tell you whether buyers can find you, understand you, and feel confident enough to choose you.
That’s a bigger challenge now that people move between search engines, AI assistants, social platforms, marketplaces, review sites, and private communities before making a decision.
Viewed through that lens, search performance comes down to three things: presence, interpretation, and momentum.
Are you present where demand forms?
Are you being understood?
Is anything compounding?
1. Are you present where demand forms?
Is your brand showing up in the places where demand starts, not just where it converts?
This goes far beyond rankings or impression share to ask whether the brand appears when people are exploring the category:
Asking early questions.
Comparing options.
Reading reviews.
Checking marketplaces.
Watching creators.
Trying to understand the problem in their own words.
If a brand only appears once someone already knows its name, it’s arriving late. It may look efficient because branded demand converts well. But commercially, it means the brand is depending on other forces to create demand before it turns up to harvest it.
A pattern we see repeatedly is brands mistaking weak presence for weak conversion. Across 196 brands we tracked over 12 months, the same shape kept appearing.
Branded search healthy, CPA respectable, but presence sitting in the bottom half of the competitive set. The brand was converting people who already knew it while missing the moments where the category was being explored by everyone else.
Travel illustrates this most clearly. It’s a category where presence is the dominant driver of market share because people often shop for holidays before they have a brand in mind.
If a travel brand is absent from those early discovery moments, it never enters the consideration set. CRO can’t fix that. Ask what share of category discovery moments you’re actually present in.
When branded conversion is strong but unbranded presence is weak, the growth opportunity sits upstream:
Review sites.
Marketplaces.
Creator content.
Social search.
Long-tail non-brand queries.
That’s where the category is being decided.
If presence is weak, interpretation won’t save you. But presence alone isn’t enough — being found is only useful if what people find makes sense.
When your brand does show up, is it being understood in a way that helps you get chosen?
People search to find something and reduce doubt. The language gives it away:
“Is it worth it?”
“Best alternative to X.”
“What do real customers think?”
“For people like me.”
It’s the buyer showing you the anxiety in the decision. Many brands answer these questions badly or too late.
The ad says one thing, the organic result says another, the reviews raise a concern, the landing page is generic, and the AI answer gives a technically accurate but underwhelming summary. The customer is left to connect the dots themselves.
AI makes this harder because the answer is increasingly compressed and increasingly unstable. Brands aren’t just fighting for a blue link anymore. They are fighting to be included, described, and trusted inside an answer that may look completely different next month.
In practice, that means AI search visibility should not be judged solely by traffic volume.
Publishers such as Reuters and The Guardian receive less than 1% of referral traffic from AI platforms despite being frequently cited, but The Washington Post found that visitors arriving from AI platforms converted to subscriptions at four to five times the rate of traditional search visitors.
The audience AI search delivers can be smaller and significantly more valuable. That only holds if the brand is being described accurately and compellingly enough to send the right people.
Our own research across categories adds a layer of nuance worth sitting with: LLM visibility isn’t what most brands think it is. LLM visibility correlates with market share at +0.19 on average. That’s weaker than many brands assume.
But the average hides a category split that matters. In fashion, the correlation is +0.58. In travel, +0.43. In finance and general retail, it inverts: −0.26 and −0.25. In those latter two categories, the brands appearing most often in AI answers are the ones losing share.
Reading the underlying signals, we see that AI systems are describing those brands in ways that don’t help them be chosen. The attributes being surfaced are wrong, outdated, or framed in terms of challenger comparisons that the established brand can’t win.
Audit your AI citations. Run the prompts your category buyers actually run, in the platforms they use, and read what is being said about you.
If the framing is wrong, the fix isn’t paid media. It’s the source signals AI systems pull from — editorial coverage, structured content, and the third-party comparisons in your category. That’s the work that changes how the answer reads next month, and the month after that.
Solving interpretation without presence means you’re explaining yourself to people who can’t find you. But solving both without momentum means you’re winning the same ground again every month.
Is your brand becoming easier to find, trust, and choose over time — or does every sale still need to be bought?
Most measurement gets too short-term to answer this question honestly, and marketers know it. Search shows whether compounding is happening.
Is branded search growing without heavy brand spend?
Is direct traffic strengthening?
Is organic content continuing to bring people in without fresh media spend?
Is review volume building?
These are signs that the brand is accumulating memory, trust, and proof — that the system is starting to work harder without needing to be pushed every time.
The opposite is equally visible. Paid dependency increases. Organic demand softens. Branded search only moves when campaigns are live. People arrive, compare, but don’t choose. The brand is active, but nothing is building.
The first thing to look at when performance feels stuck is the gap between a brand’s discoverability and its actual market share. The size and direction of that gap is usually enough to diagnose which problem you are dealing with.
A brand whose demand rank outperforms its discoverability rank is running on borrowed time. The numbers look strong, but the upstream signals aren’t building. In our data, that gap tends to close within three months. The play is to reinvest now before the lag catches up.
A brand whose discoverability outperforms its demand rank has something building that hasn’t surfaced yet. The instinct is to keep optimizing the conversion layer. Usually, the right call is the opposite: hold, let the upstream work compound, then drill into where the gap actually lives.
Weak presence with healthy momentum means the funnel is working, but the top of it is empty. Invest in category visibility, not conversion.
A strong presence with weak interpretation means visibility isn’t the problem. Fix how the brand is described in search, reviews, and AI answers before spending more on media.
Weak momentum with both presence and interpretation intact usually means proof is missing: reviews, share of voice, and word of mouth need building before acquisition spend pays back.
Almost no brand sits cleanly in one bucket. But knowing which gap to fund and which to let run is usually enough to act on.
The challenge is knowing which of these three problems you’re dealing with before you start optimizing the wrong thing. That’s often harder than it sounds because brand and performance are managed by separate teams, with separate budgets, and separate definitions of success.
Customers don’t experience the brand in silos. They experience the search result, the review, the ad, the AI answer, the marketplace listing, the creator mention, and the thing their friend said in the group chat as one decision environment.
Presence feeds interpretation. Interpretation feeds momentum. Momentum reduces the cost of presence. Understanding where that loop breaks is often the fastest way to identify the constraint holding growth back.
Reddit is rolling out a suite of new advertising products powered by its Community Intelligence engine, betting that authentic user conversations have become one of the most influential forces in consumer decision-making.
What’s happening. Reddit is introducing new ad products designed to help brands tap into the platform’s vast repository of consumer conversations, which now includes more than 25 billion posts and comments.
The company says the new tools will help advertisers create more relevant campaigns, accelerate purchase decisions and improve performance by leveraging insights derived directly from Reddit communities.
Why we care. Reddit is giving brands new ways to turn authentic community conversations into ad creative, shopping experiences and performance signals.
As more consumers use Reddit to research and validate purchase decisions—including AI-generated recommendations—these tools could help brands reach high-intent audiences with messaging rooted in real user sentiment.
New creative tools. Reddit is launching several new AI-powered creative features.
Free-form ad generator (Beta)
The tool automatically combines information from a brand’s website with relevant Reddit conversations to create ads inspired by Reddit’s long-form post format.
The goal is to help advertisers build campaigns that feel more native to the platform.
Tailored creative assets (Beta)
Using AI, Reddit identifies relevant communities and audience segments, then generates customized headlines and image variations designed to improve ad relevance.
The feature is currently available for Max Campaigns.
Redditor Highlights (GA)
Now generally available, the feature allows eligible brands to incorporate real Reddit conversations directly into their ads.
Ads can display summaries of community sentiment alongside related organic posts, helping brands showcase authentic customer perspectives.
Shopping Listing Ads (Alpha). Reddit is also entering the shopping ads space with a new ad type. The new format matches products from participating advertisers to relevant conversations and displays them in a carousel format.
The experience mirrors how users already compare products and seek recommendations within Reddit discussions.
Performance gets an AI boost. Alongside the new ad formats, Reddit says it continues to invest heavily in machine learning across its advertising platform.
They highlighted early testing results from a new six-second engaged video view optimization goal, which delivered:
A 130% increase in view-through rates.
A 71% improvement in video completion rates.
Reddit says its combination of machine learning and Community Intelligence enables more contextual ad targeting and content understanding across the platform.
A 2023 Google patent describes how AI systems could build an understanding of businesses, brands, products, and other entities from websites and public data.
The filing outlines a process for extracting information, identifying relationships, and synthesizing what Google calls a “deep, holistic characterization” of an entity.
If systems like this become more influential in search, SEO may increasingly involve helping Google understand the entity behind your content, not just the content itself.
The shift from documents to entities
Google has spent more than two decades helping users find information published on webpages. Whether through traditional search results, featured snippets, or AI-generated answers, the process has generally started with understanding documents.
As Google’s search products become more conversational and recommendation-driven, understanding individual documents may no longer be enough.
Before an AI system can recommend a business, compare products, explain a brand, or suggest a service provider, it must first understand the entity behind the content.
At first glance, the patent may seem like another content extraction system. Search engines have been extracting information from webpages for years. However, Google describes a broader objective.
According to the filing:
“The techniques described throughout this specification enable artificial intelligence to generate and enhance a deep, holistic characterization of a particular entity.”
Google defines an entity broadly, including people, companies, businesses, places, objects, and concepts.
Rather than simply identifying facts or indexing content, the system is designed to interpret information, identify relationships, generate summaries, and develop an understanding of the entity represented by that information.
Google’s patent describes a system that collects information from websites and public sources, processes it with an AI system, and generates an understanding of an entity.
How Google’s patent creates an understanding of an entity
At a high level, the patent describes a system for collecting information from multiple sources, interpreting that information, and synthesizing an understanding of an entity.
A simplified interpretation of the process described in Google’s patent. Information from webpages and other sources is collected, interpreted, enriched with additional context, and used to develop an understanding of an entity.
Step 1: Identify the entity
The process begins by identifying a domain and an associated entity. The system then gathers information from webpages associated with that domain and processes it using an artificial intelligence system that includes a large language model (LLM).
Step 2: Interpret the information
Rather than simply extracting facts from individual pages, the system is designed to generate what the patent calls a characterization of the entity.
Google explains that this characterization is “an interpretation of the extracted first content and extracted second content rather than a verbatim duplication of the extracted content.”
In other words, the system goes beyond collecting information. It interprets that information and forms conclusions about the entity behind it.
Step 3: Extract attributes and relationships
The patent further explains that the AI system can analyze webpages to extract information such as an entity’s presence, age, principles, services, reputation, social media sentiment, and relationships between different elements associated with the organization.
These signals help the system move beyond understanding individual webpages toward understanding the entity itself.
Step 4: Supplement with third-party information
Importantly, the patent isn’t limited to information found on a company’s own website. Google notes:
“The artificial intelligence systems may use online maps data, job listing data, business information, or other suitable third-party data as additional or augmenting input to provide context for generating the characterization that is output by the artificial intelligence system.”
Taken together, the goal appears to be to build a more complete understanding of the entity than could be obtained from any single webpage.
How the patent represents entities
The system is designed to organize information about an entity into a format that can be interpreted, expanded, and used by other systems.
Entity summaries
After collecting information from webpages and other sources, the patent describes generating an entity summary. The examples provided in the filing aren’t page summaries. Instead, they read more like descriptions of a company’s identity, positioning, values, and characteristics.
One example included in the patent describes a hypothetical company’s brand identity, noting associations with simplicity, accessibility, trust, innovation, and social responsibility.
“Example Search Co’s brand identity is one of simplicity, clarity, and accessibility. The company’s logo, a colorful, sans-serif E, is instantly recognizable and easy to remember. The color palette is also simple, with a focus on blue and green, which are associated with trust and reliability. Example Search Co’s typography is also clear and easy to read, even at small sizes. The overall tone of Example Search Co’s brand identity is friendly and approachable. The company’s marketing materials often feature simple, humorous illustrations that help to make Example Search Co’s products and services more relatable to users. Example Search Co. also emphasizes its commitment to making information accessible to everyone, regardless of their background or technical expertise.”
Another example presents those same concepts as a set of key attributes rather than a narrative summary.
“Here are some key aspects of Example Search Co’s brand identity: – Trustworthiness: Example Search Co. is known for its reliable and trustworthy search engine. The company also has a strong commitment to privacy and security. – Innovation: Example Search Co. is constantly innovating and releasing new products and services. The company is known for its ability to anticipate user needs and deliver innovative solutions. – Accessibility: Example Search Co’s products and services are designed to be accessible to everyone, regardless of their background or technical expertise. – Social responsibility: Example Search Co. is committed to using its technology to make a positive impact on the world. The company has a number of initiatives in place to promote sustainability, diversity, and inclusion.”
What’s important here is the overall format. The system takes information distributed across multiple sources, transforms it into an interpretation of the entity, and synthesizes it into a higher-level understanding of the entity.
Entity graphs
Google builds this understanding through hierarchical graph structures. According to the patent, the generated characterization can include:
“[A] hierarchical graph structure that includes at least one parent node representing a first attribute of the characterization and at least one leaf node representing a second attribute of the characterization.”
The accompanying figures from the patent provide a better sense of what this means in practice.
How the system organizes business attributes and relationships into a hierarchical graph structure.
The figure above shows an example graph generated for a service-based company.
The figure below provides a similar example for a product-based company. In both cases, the system organizes information into connected relationships rather than isolated facts.
A similar graph structure for products, connecting attributes, features, categories, and related concepts.
Instead of just knowing that a business offers a service, the system associates that service with audiences, locations, reputation signals, differentiators, and other related attributes.
Instead of only identifying a product, the system can also connect it to features, categories, use cases, and related offerings.
Entity models
The patent begins to resemble an entity modeling system more than a content extraction system.
Extracting information answers one question: What information appears on this website?
Entity modeling answers a different question: What do we understand about this business?
That difference becomes apparent when you look at the types of information Google says the system can analyze.
The patent specifically references extracting information related to an entity’s presence, age, principles, services, reputation, social media sentiment, and relationships between different elements associated with the business. It also discusses incorporating information from external sources such as maps data, user reviews, business information, and job listings.
Taken together, these aren’t just website attributes. They’re also signals that help define an entity’s identity.
The result is a model that appears capable of answering broader questions about an organization than traditional extraction systems were designed to address.
Rather than identifying products, services, or facts, the system develops a contextual understanding of who the entity is, what it does, how it’s perceived, and how it relates to other entities.
This is where the patent becomes particularly interesting for SEO.
Understanding information regardless of format
Google has spent years building systems that help machines understand information on the web. Structured data, schema markup, product feeds, business listings, and knowledge graphs all exist, in part, to make information easier to organize, interpret, and connect.
One aspect the patent emphasizes repeatedly is the ability to extract information that wasn’t specifically structured for machine consumption.
The patent explains that the AI system can extract content that has “not been structured for parsing by the artificial intelligence system” and can process information from webpages that haven’t been organized according to the requirements of traditional content extraction systems.
Google identifies this as one of the primary advantages of the approach.
According to the filing, existing content extractors are often limited to content that follows predefined structures, while the proposed system can extract and interpret information “irrespective of its format.” Rather than reproducing extracted text, the system can generate new content that interprets and synthesizes the information it finds.
The patent suggests Google is exploring ways to use this capability to build a more complete understanding of an entity. That understanding isn’t limited to information found on a company’s own website.
The patent explicitly discusses supplementing website content with information from maps data, business information, job listings, and other third-party sources.
Taken together, the process begins to resemble an entity analysis system rather than a webpage analysis system. The website remains vitally important, but it’s no longer the only source of truth. Instead, the website becomes one of several inputs used to construct an understanding of the entity behind it.
As AI-powered search experiences become more focused on answering questions, making recommendations, and helping users evaluate options, the quality of those outputs depends on the quality of the system’s understanding.
Before an AI system can recommend a business, summarize a brand, compare products, or explain why one option may be a better fit than another, it first needs a model of the entities involved. The patent describes one possible approach for creating that model.
From webpages to entities: What this means for SEO
Patents don’t tell us exactly how Google will use a technology. Many patents never become products, and even when they do, the implementation often looks different from what is described in the filing.
What patents can do is reveal how Google is thinking about a problem. In this case, the problem appears to be understanding entities.
That may sound familiar because entity understanding isn’t a new concept within Google Search. Google’s Knowledge Graph, introduced more than a decade ago, was built around connecting entities and relationships.
More recently, Google’s emphasis on E-E-A-T, product reviews, business information, and reputation signals has reflected a similar objective: understanding not just what a page says, but who is behind it and whether that source can be trusted.
LLMs expand Google’s ability to understand entities
What makes this patent worth examining is the role large language models now play in that process.
This patent describes a process in which an AI system can:
Analyze websites and public information.
Interpret the information it finds.
Synthesize an understanding of an entity without requiring that information to be presented in a specific format.
That capability becomes increasingly important as Google’s search experiences move beyond document retrieval.
Consider what is required for a system like AI Overviews to answer a question about a company, product, or service. The system must first determine what that entity is, what it offers, who it serves, how it differs from alternatives, and whether it is relevant to the user’s query.
The same challenge exists in AI Mode, Gemini, and recommendation-driven experiences such as Ask Maps. Before an AI system can recommend an entity, it must first understand it.
That idea appears throughout the patent. Google repeatedly describes collecting information from multiple sources, generating summaries, organizing attributes into relationships, and developing an understanding of the entity as a whole.
The patent explains that the system can identify characteristics such as services, reputation, principles, social sentiment, and relationships between different elements associated with the entity.
From webpages to entities: content, reviews, profiles, and other signals contribute to how AI systems understand and recommend businesses, products, and organizations.
Webpages become evidence
Through an SEO lens, this suggests a change in how webpages may function.
Traditionally, webpages have been optimized to rank for queries. A service page targets a service keyword. A category page targets a product category. A location page targets a geographic market. Those objectives remain important.
However, if systems like the one described in this patent become more influential, webpages may increasingly serve a second purpose. They become evidence used to construct an understanding of the entity behind them.
A service page does more than target a keyword. It helps establish what services a business offers.
A case study does more than attract traffic. It demonstrates experience and expertise.
A team page helps identify the people behind the organization.
Customer reviews contribute information about reputation.
Press coverage, social media, and industry references provide additional signals that reinforce or challenge the system’s developing understanding.
This is one reason the patent’s emphasis on multiple data sources is so interesting. The filing doesn’t describe building an understanding from a single webpage. It describes combining information from websites, maps data, business information, job listings, and other public sources to create a more complete picture of the entity.
Visibility may increasingly depend on entity understanding
The implication here is that visibility may increasingly depend on how effectively Google understands the entity associated with those keywords. That becomes especially important in environments where users are no longer choosing from a list of 10 blue links.
When an AI system is summarizing options, making recommendations, or narrowing choices on behalf of a user, the quality of its understanding becomes a critical factor in determining which entities are surfaced and how they are described.
The challenge for SEO may no longer be limited to helping Google understand a page. It may increasingly involve helping Google understand who you are.
How brands can influence entity understanding
If Google’s goal is to synthesize an understanding of a business from its website and other public sources, the practical question becomes: What can organizations do to help shape that understanding?
The patent suggests that entity understanding emerges from the accumulation and interpretation of information across multiple sources rather than any single webpage, profile, or signal.
While the patent doesn’t provide optimization recommendations, it does point to several areas businesses should pay attention to.
Maintain consistency across sources
The patent repeatedly references using information from multiple sources to generate a characterization of an entity.
Because that characterization is “an interpretation of the extracted first and second content rather than a verbatim duplication of the extracted content,” consistency becomes increasingly important.
Review how your business is described across:
Your website.
Business profiles and listings.
Social media accounts.
Press coverage.
Recruiting and job postings.
Industry directories.
The goal isn’t identical wording everywhere. The goal is to ensure AI systems encounter a consistent understanding of who you are, what you do, and who you serve.
Define the attributes you want associated with your brand
The patent’s example entity summaries focus on characteristics such as trustworthiness, innovation, accessibility, and social responsibility.
Ask yourself:
What do we want to be known for?
What differentiates us from competitors?
What attributes should be associated with our brand?
Examples might include:
Enterprise software: security, compliance, and scalability.
Ecommerce: quality, value, and sustainability.
Local services: expertise, responsiveness, and reputation.
The clearer those differentiators are communicated, the easier they become for AI systems to identify and associate with the entity.
Support claims with evidence
The patent describes building an understanding of an entity from multiple sources. That means claims alone may carry less weight than evidence that reinforces those claims.
Examples of supporting evidence include:
Customer reviews.
Case studies.
Testimonials.
Press coverage.
Industry citations.
Awards and certifications.
Author profiles and expertise signals.
The goal isn’t simply publishing more content. The goal is providing evidence that supports the attributes you want associated with your entity.
Strengthen entity relationships
One of the more interesting aspects of the patent is its use of hierarchical graphs to organize relationships between different attributes and concepts.
Businesses should make it easy for search engines and AI systems to understand relationships between:
Products and services.
Locations and service areas.
Audiences and use cases.
Brands and people.
Organizations and industries.
The easier those relationships are to identify, the easier it becomes for AI systems to understand where an entity fits and when it should be recommended.
Audit your entity footprint
A useful exercise is to ask:
If an AI system had to describe our company using information from our website, reviews, profiles, listings, and third-party mentions, what would it say?
The answer may reveal gaps, inconsistencies, or missed opportunities that are difficult to identify when looking at individual pages in isolation.
As AI-powered search becomes increasingly focused on understanding and recommending entities, that broader view of your digital presence may become just as important as traditional page-level optimization.
What this means for enterprise, ecommerce, and local businesses
One of the strengths of this patent is that it isn’t limited to a particular type of entity. Google’s definition is intentionally broad, encompassing businesses, organizations, products, places, concepts, and people.
That breadth suggests the framework could potentially be applied across many different search experiences and industries. The challenges associated with entity understanding are likely to vary depending on the type of business being analyzed.
Enterprise and B2B organizations
Enterprise organizations often face a consistency challenge. Information about the business may be distributed across product pages, investor relations content, press releases, partner websites, recruiting materials, analyst reports, and social media channels. Different departments frequently describe the organization in different ways.
If AI systems are synthesizing an understanding of the entity from multiple sources, consider:
Is our positioning consistent across channels?
Would an AI system describe our company the same way regardless of the source it analyzed?
Are our core differentiators clearly communicated and reinforced?
As AI systems increasingly interpret information across channels, maintaining a coherent entity identity may become just as important as maintaining a consistent brand identity.
Ecommerce and product-focused businesses
The patent’s product-related examples suggest that entity understanding may extend beyond organizations to individual products.
Users often ask questions that require evaluation rather than retrieval. Rather than just searching for a product, they’re asking which product is best for a specific use case, budget, audience, or situation.
For ecommerce brands, consider:
Are product attributes clearly defined?
Are category and product relationships easy to understand?
Do reviews reinforce product strengths and use cases?
Is supporting content helping explain who a product is for and when it should be recommended?
Product information architecture, reviews, category relationships, and supporting content may all contribute to how products are understood and surfaced in AI-driven experiences.
Local businesses
Local businesses often face a reputational and specialization challenge.
Many of the attributes referenced in the patent align closely with signals already used in local search, including services, reputation, social sentiment, and business information.
For local businesses, consider:
Is your expertise clearly communicated?
Do reviews reinforce the services and specialties you want to be known for?
Are service areas consistently represented across sources?
Does your website, Google Business Profile, and third-party presence tell the same story?
A local business is more than a collection of service pages. It is an entity associated with specific services, locations, expertise, reviews, and reputation signals gathered from across the web.
The common thread
Across enterprise, ecommerce, and local search, the challenges are similar. Before Google can recommend an entity, compare an entity, or explain an entity, it must first understand that entity. The patent provides one of the clearest examples yet of how that understanding might be built.
Patents aren’t product announcements. Google files thousands of patents, and many never become user-facing features.
The most useful way to view this patent isn’t as a roadmap for a future ranking algorithm, but as a window into how Google is approaching the challenge of understanding entities in the age of LLMs.
Throughout the filing, Google repeatedly returns to the same objective: using AI to collect information from websites and public sources, interpret that information, and synthesize an understanding of an entity.
In Google’s own words, the techniques described in the patent enable artificial intelligence to “extract content from a website or domain and other public sources to synthesize an understanding of a particular entity.”
That objective aligns closely with the direction of Google’s newer search experiences. AI Overviews, AI Mode, Ask Maps, and other AI-powered systems all depend on understanding the businesses, products, organizations, and concepts they reference. They evaluate, summarize, compare, and recommend entities.
For SEOs, that may be the most important takeaway. Historically, SEO has focused on helping Google understand webpages.
Patents like this suggest that the next challenge is helping Google understand the entity behind them. That understanding may influence who gets surfaced, who gets cited, and ultimately, who gets chosen.
SEO attribution has always been messy. Unlike paid search, organic search lacks the same tracking granularity, has a lag between the work and the results, and often depends on signals you can’t control, such as rankings.
To make matters worse, attribution is even more of a black box today, with AI-generated answers dominating SERPs and LLMs that don’t always link back to your site or pass referrer strings.
Businesses have never really cared about that complexity. They care about the return they’re getting from their marketing dollars.
The good news?
SEOs can still tell a compelling ROI story, but it takes more nuance, more data digging, and more math than ever before. This article walks through key considerations as you build your next SEO ROI story.
The historical formula for calculating ROI
SEO ROI has traditionally been calculated using variations of the same formula:
ROI = ((Incremental organic revenue − SEO costs) / SEO costs) x 100
It’s clean, fits on a single slide for executives, and makes sense for a long time. Before the rise of generative AI, driving incremental traffic — and therefore revenue — was the north star for most SEO campaigns.
However, sharp increases in zero-click searches and major attribution gaps from LLMs have upended traditional models.
In many cases, organic traffic may be flat or even declining, even as overall visibility increases through impressions, AI Overview rankings, or prompt visibility.
Looking only at organic metrics and incremental gains tells only part of the story. To show SEO’s true value, we need to rethink the formula.
Here are three ways to build a more complete SEO ROI model.
1. Take credit for all organic revenue, not just incremental gains
With 60% of searches ending without a click, and that number continuing to grow, a huge part of SEO’s value today is defensive. That means maintaining and protecting traffic that would otherwise erode due to various factors.
The formula above doesn’t account for that at all. For example, would you judge a goalkeeper’s performance by how many goals they’ve scored?
The same is increasingly true of SEO. Only counting what’s new erases everything you preserved.
In a flat or declining landscape, holding the line is a major win, yet it’s completely ignored when you focus only on incremental gains.
The starting point for this ROI story shouldn’t be incremental organic revenue alone. It should be all organic revenue. That’s the entire asset SEO is responsible for maintaining and defending.
This may be a tough sell for many website owners, but it’s the truth. And if you go this route, there’s one major caveat.
Segment out brand vs. non-brand clicks
Claiming all organic revenue is disingenuous if branded growth is driving most of the performance.
Branded traffic is influenced by many factors outside SEO’s control, including PR campaigns, paid media, product, word of mouth, and more. When someone Googles your website by name, SEO rarely created that demand. It simply captured it.
SEO’s real lever is non-branded search. Before taking credit for total organic revenue, you need to segment it accordingly.
Since this can’t be done cleanly in Google Analytics, start by pulling the branded-versus-non-branded split from Google Search Console. Then apply that split to your total organic revenue using a weighted model.
Here’s some real-world data, for example:
Branded traffic accounts for about 70% of total clicks and is declining, while non-brand traffic accounts for roughly 30% and is growing. That split tells a value story on its own because non-branded growth offsetting branded decline is exactly what good SEO should produce. Before taking credit for total organic revenue, though, we need to create a blended weight.
Let’s say SEO gets 10% credit for branded traffic and 100% credit for non-branded traffic. Treat these as starting points and calibrate them for each client. The calculation would be:
(70% brand x 10% weight) + (30% non-brand x 100% weight) = 7% + 30% = 37% blended attribution weight
Apply that 37% weight to total organic revenue. If the site generates $100,000 in organic revenue per month, SEO gets credit for $37,000, not the full $100,000.
That’s likely far higher than the revenue you’d attribute to incremental gains alone. Because you’ve openly discounted credit you don’t deserve, the model is more defensible and shows stakeholders that you understand its limitations.
2. Account for assisted conversions and first-click influence
Last-click attribution buries SEO. That’s nothing new, but it’s even more relevant today.
Organic is often the first touchpoint in a user’s journey. Today, that might mean only an impression, with no measurable click at all.
Remove that influence, and SEO can look like a minor contributor to revenue it actually initiated.
SEO should take credit for the conversions it assists, even when another channel closes them. Here’s an example:
Organic dominates all three stages of touchpoint credit, but a last-click-only model rewards only the final click.
The catch is that GA4 doesn’t surface a clean assisted-conversion value the way Universal Analytics did. Precisely calculating assisted conversions requires exporting path data to BigQuery and deriving a true fractional value for each channel.
However, data-driven attribution in GA4 provides a defensible shortcut. Google already assigns each channel fractional credit based on its influence on conversions. We can use organic’s early- and mid-touch credit as a proxy for the assist value that last-click attribution ignores.
1,345.69 (early) + 687.34 (mid) = 2,033.03 in conversion credit
From there, multiply by the value of a conversion, using $100 as an illustration.
2,033.03 x $100 (conversion value) = $203,303
The same brand-versus-non-brand logic technically applies to assists. Since GA4 doesn’t cleanly split assist credit by query type, we left out late-touch credit. That’s where branded behavior tends to concentrate, and excluding it removes much of the credit that would otherwise be discounted.
Even as a directional number, the data proves the point: Organic is providing real value. Relying on last-click attribution alone leaves that ROI out of your story.
3. Measure SEO content impact across other channels
SEO-led content doesn’t stay within the organic channel. The same research, briefs, and articles your team produces can be repurposed across multiple channels:
Social channels run campaigns and posts built on organic content.
Email runs drip campaigns fueled by blog content and resources.
Looking only at organic revenue ignores all the downstream value generated by SEO efforts.
I recently looked at a client where we’d just started publishing new articles and refreshing existing ones. After only one month, we were already seeing some of that content being used and generating conversions in other channels.
Sure, 29 calls and five qualified leads aren’t a huge number. But it will grow over time, and those are conversions SEO shouldn’t ignore.
Similar to assisted conversions, drilling down to an actual dollar amount requires some fancy math, but it’s possible.
Start by:
Finding and cataloging which SEO-led pages are used across different channels.
Calculating the percentage of conversions generated by those SEO-led pages.
Applying that percentage to the total conversion value of each channel.
In practice, it could look something like this:
500 conversions (paid search) x $100 (conversion value) x 5% (percentage of conversions from SEO-led pages) = $2,500 in downstream value
Even if the numbers feel small, this is another way SEO can rightfully claim revenue and help justify the overall cost of the campaign.
SEO can and should take credit for value beyond incremental organic revenue. Your exact methodology may differ, so work with your most data-savvy team members to get it right. The general concepts are what matter:
Take credit for all organic performance, but don’t take credit for every branded click as if SEO created that demand.
Look at assisted conversions and other attribution models. Don’t evaluate SEO within the organic silo alone.
Take credit when SEO content is used by other channels. Don’t ignore the downstream impact it can have.
Get creative when solving the ROI puzzle. Don’t let an outdated formula undersell your work.
The classic ROI formula isn’t wrong. It’s incomplete.
As search evolves, the way we measure ROI should evolve with it.
Platform changes, AI-driven SERPs, and shifting measurement models are forcing search and performance marketers to rethink their skills more frequently.
What worked six months ago may not work today, and the gap between current best practices and outdated knowledge keeps widening.
That’s why continuous learning now directly affects SEO performance. The organizations that adapt fastest don’t treat learning as a separate activity. They build it into how they test, share knowledge, and make decisions.
Why search and performance marketing skills expire quickly
Search skills have a shorter shelf life than most people realize. I’ve sat in meetings where approaches that were solid 18 months ago were actively working against performance.
Platform updates, automation changes, and shifts in user behavior can turn effective tactics into outdated ones faster than most expect. Without ongoing learning, it’s easy to fall behind current best practices.
Misinterpreting data, overrelying on automation, or using outdated SEO methods can all weaken results. To keep pace, you need to adapt to changes driven by AI Overviews, evolving SERP features, and increasing zero-click experiences.
AI reduces execution time, but it increases the need to validate outputs, particularly in reporting and prioritization. As automation becomes more capable, the value shifts from execution to interpretation, prioritization, and decision-making.
If you rely on AI outputs without validation, you risk inaccurate reporting, weak content decisions, and poor prioritization. Prioritizing decisions over activity shows up in trade-offs, validation of automated outputs, cross-channel performance interpretation, and commercial decision-making.
As AI adoption outpaces structured training, gaps between tool use and real capability become more visible. The challenge isn’t operating tools efficiently. It’s turning outputs into decisions.
In this environment, learning is less about mastering tools and more about applying sound judgment. Most people aren’t limited by access to learning. They’re limited by the assumption that what they already know is still good enough.
Skill decay and the rise of systems thinking
One of the biggest mistakes I see is assuming knowledge stays relevant longer than it does. Skills can become outdated surprisingly quickly when platforms, reporting, and user behavior are changing at once.
As platforms evolve and delivery pressure increases, gaps form between what the job requires and what people know. Those gaps become especially visible during platform updates, reporting changes, and shifts in search behavior. They’re also more likely when knowledge sits with individuals instead of documented systems.
That’s why systems thinking matters more than isolated tool knowledge.
High-performing organizations focus on how disciplines connect:
SEO, paid media, analytics, and content operate as one system.
Technical work is tied to commercial impact.
Prioritization is driven by outcomes, not activity volume.
Platform updates are interpreted at the system level, not the task level.
You also need to learn across adjacent disciplines because performance issues rarely sit within a single channel.
Tools such as Semrush, Ahrefs, Screaming Frog, and Sitebulb remain important, but they don’t prevent skill decay on their own. The key difference is how well you interpret what the tools show.
If you learned SEO primarily through legacy keyword tactics, adapting to entity-based search, AI Overviews, and changing SERP layouts becomes much harder once learning stops.
To reduce knowledge loss, build simple reinforcement habits: review campaign performance regularly, share platform updates internally, and document what tests reveal so learning carries forward instead of staying with one person.
Staying current requires more than consuming information. You need processes that turn new insights into better decisions.
Build depth in core SEO tools
SEO tools are often used for basic tasks despite far broader functionality. Tools such as Semrush, Ahrefs, Screaming Frog, and Sitebulb are typically used for only a fraction of their capability.
I’ve often found that investing time in deeper product knowledge delivers faster gains than adding another tool to the stack.
That deeper knowledge shows up in audits that take half the time, diagnoses that don’t rely on a third party, and analysis that moves things forward rather than restating what the tool already surfaced.
Use certifications to build cross-channel understanding
Some of the most effective people I’ve worked with understand far more than SEO. They understand how paid media, analytics, and measurement fit together, which makes collaboration and prioritization much easier.
Training across Google Ads helps you understand how paid and organic search interact, along with bidding behaviors, visibility dynamics, and data structures across channels.
This broader view supports better decision-making and reduces siloed thinking.
Google Skillshop certifications are also useful for building broader platform knowledge, particularly across Google Ads and Google Analytics.
Turning conference insights into something usable
Industry events create value when the learning continues after you leave.
At our agency, insights from conferences are shared directly in our Teams channel alongside publicly available slide decks, so everyone benefits regardless of who attended. Anything worth exploring gets tested in live environments rather than filed away.
That loop — share, test, reflect — is what turns conference insights into meaningful changes in how you work.
Combine learning with experimentation
As part of our internal testing process, anything worth exploring gets tried on our own site first. We monitor it over weeks or months, depending on what we’re testing, before any recommendation touches a client account.
If the results are positive, it goes onto the client roadmap. If something is already well supported by industry evidence, we’ll move faster and factor it into client work sooner.
That approach grounds recommendations in either our own evidence or a strong body of industry data.
Measure the impact of learning
The clearest signs of progress tend to be operational.
Onboarding takes less time when knowledge is documented and shared consistently.
Reporting becomes more reliable when you understand what you’re measuring and why.
Prioritization improves when you have enough context to make confident decisions rather than defaulting to activity.
When these habits are working, you’ll notice it in the quality of conversations, decisions, and outcomes.
AI is accelerating the pace of change in search. Skills evolve faster, and success depends increasingly on judgment, adaptation, and decision-making.
If you’re falling behind, it’s rarely because you lack tools or data. More often, it’s because you’re relying on knowledge that no longer reflects the current landscape.
The best people in search don’t assume yesterday’s knowledge still applies. They stay curious, keep learning, and adapt as the landscape changes.
In May 2025, Google brought 25 of us into a closed-door room at I/O to talk about the post-click SERP. The instruction we left with was short: create non-commoditized content.
Here’s the uncomfortable part:
For 15+ years, we never commoditized content. We commoditized the sale. We built billions of pages that took a messy human problem, targeted whatever keyword the person compressed it into, and answered with some version of “buying now is the best choice.”
We did it well for our brands, and we built a sales-first web.
Oops.
AI search is the bill for that cognitive debt finally coming due. For years, we skipped past the buyer’s real thinking and answered “buy now” instead of their actual questions.
Paying that debt off is a link building problem. Not just hyperlinks, but the links between sources, roles, risks, and decisions.
A co-citation gap analysis maps those links, showing which sources AI search trusts for each buyer role and where your content is missing from the decision.
In this piece, we’ll show you how to run it: map the sources AI search reads and cites, find the role your content doesn’t support yet, and build the asset that closes the gap.
Moving from anchor text to anchor context
We’ve been running co-citation analysis on the link graph for 15 years at Citation Labs.
In 2011, I published a six-step co-citation method for link builders: find the pages that curate a topic, count which sources they cite together, and reverse-engineer what made those pages worth citing.
Instead of asking which pages mention a topic together, you’re asking which sources an AI assistant trusts when each buyer role asks about the same decision, and which role your content fails to support.
That missing decision support is the co-citation gap.
How to run a co-citation gap analysis by hand
A co-citation gap analysis counts what AI search reads and cites across same-phase, same-problem prompts for different buyer roles.
The overlaps and absences show which decisions your content doesn’t support yet.
You don’t need software for this, just one buyer decision, the committee around it, a set of prompts, an AI tool that shows its work, and a spreadsheet.
1. Map the committee, and write down each role’s fear
List everyone who has to say yes before the thing you sell gets bought — the real deciders, not the org chart.
For example, for a funded biotech that’s choosing a logo, the committee is the CEO, in-house counsel, ops, and marketing.
Next to each, write what that person is afraid of. That fear is the question they’ll bring to an assistant, so write it in first person:
CEO: Do we look like a serious, fundable company?
Counsel: Is this name or mark going to get us sued or forced to rebrand?
Ops: Will this survive production (from a favicon to signage to a 96-well plate label)?
Marketing: Will this perform (get recognized, stand out cold to investors and recruits)?
Four roles, four fears, one decision.
Note: Solo purchases work the same way. The “committee” is one buyer’s competing concerns. The unit isn’t the committee, it’s the choice point.
2. Write one prompt per role
Create one controlled prompt per role, all based on the same buying decision. Keep the scenario fixed and change only the role. That makes the role’s domain of practice the variable.
Follow these five rules:
One shared scenario. Keep the situation fixed; change only who’s asking.
First person, in the role’s voice. Model their words, worries, and perspective.
Put the role in a specific situation. Generic prompts get generic answers. Specific situations trigger retrieval, and retrieval is the data.
Bundle the role’s real concerns. Real deciders carry a cluster, not one tidy question.
Name no brands. Brand names pollute the citation set.
Here’s the CEO prompt from our “make a logo” example:
“I’m the founder-CEO of a biotech that closed a Series A, briefing a design firm on a new logo. I want a point of view to show them, not a blank page. What do credible, well-funded biotech brands share visually? Which startups nailed a post-raise rebrand? Which botched it? What do investors and pharma partners read into a young brand? What should a non-designer like me use to rough out a few options?”
Then add one kill-switch prompt per role: “What’s the one thing here that, if we get it wrong, we can’t undo? What would make me say no?”
This surfaces the hard veto role.
Run more than one phrasing. The signal is what repeats, not what appears once.
3. Capture what the assistant searches, reads, and cites
Use an assistant that exposes its sources. For each role’s run, open the activity or sources panel and copy three fields into your spreadsheet:
Sub-queries generated
Pages read
Pages cited
Tag each read page by type as you go:
Forum/hub (Reddit, aggregators, etc.)
Primary/official (a regulator, a standards body, etc.)
Vendor (services and solutions)
The mix shows what kind of evidence each role trusts.
In our example, nearly a third of everything read was Reddit, and over 40% was high-volume hubs.
Your pages will sort into three states:
If it’s read and cited, it was consulted and used.
If it’s read but not cited, it was consulted but dropped. That’s a content problem.
If it’s not read, it was never consulted. That’s a discoverability problem.
The dropped-but-read pages are where the cheapest wins hide.
4. Build the citation matrix
Now, turn your sort list into a matrix — this is the co-citation analysis.
Make one row per unique cited URL, one column per role, and one count column showing how many roles cited it.
Note: One row per cited URL
Here’s what the matrix looked like in our “make a logo” example:
Sort by count. You’ll see which sources are shared, which are role-exclusive, and which roles have almost no overlap.
5. Find the role with the veto power
With the matrix sorted, look for three structures.
The shared core is everything cited by 2+ roles. If it’s nearly empty, the committee is disjointed (you serve it seat by seat). If it’s substantial, the committee is convergent (you win the commons and the gatekeeper).
An isolated decider is a decisive role whose cited sources are mostly role-exclusive. An empty edge is two must-agree roles that share nothing.
Then set priority: whose “no” is final, and whose sources overlap least with everyone else’s?
That’s your veto × isolate seat. Build content there first.
Capture this in a spreadsheet you can update and rerun:
6. Add the phase axis
Run the analysis again, but move the same decision forward in time. Keep each role the same while changing the scenario:
Choosing → Rolling Out → Getting Value and Renewing
The cited set shifts at every stage while the owner of the answer hands off as the buyer moves.
Note where you’re present in one stage but missing in others.
Those gaps are asset targets.
7. Create a content and outreach plan
Your matrix gives you a prioritized content strategy:
Gatekeeper: The veto-isolate seat, decisive and unserved.
Empty edges: Where two must-agree roles share nothing.
Shared core: Sources cited by multiple roles.
Phase gaps: Stages where your brand appears, then disappears, by role.
The new work is placement, and the model already told you where. The sub-queries show the domains it trusts for each role.
In our example, the CEO run ran site:businesswire.com, site:fiercebiotech.com, and site:prnewswire.com; the ops run ran site:developers.google.com/search/docs
Nearly every role appended “official” to chase primary sources.
That’s your placement and outreach list — written by the AI assistant. Earn your evidence on the surfaces it already uses when answering that role.
And remember the unit you’re building in: anchor context, not anchor text.
The asset’s job isn’t to show a popular option. It shows: who this helps, what it solves, when it fits, and why it belongs — for this role, at this choice point.
Rebuild the cited column and compare it to the baseline. Look for three movements: your brand appears where it didn’t before, placed sources enter the gatekeeper’s answer, or an empty edge starts to fill.
It’s tedious by hand, but the loop is always the same: baseline, build, re-run, compare.
The cited set is the scoreboard.
If you’d like a faster way to run this, reach out to Citation Labs. We’ll share our prompt stack with you.
How to turn the matrix into decisions
Once the matrix is sorted, review the shape of the citation set.
If a few sources are cited by most of the committee, it’s convergent. There’s shared ground to win. If the top is nearly bare, it’s disjointed. Every role is in its own world, and generic “bottom of funnel” content won’t carry the decision.
In our make-a-logo example, exactly one source was cited by multiple roles: Canva’s brand-kit page. At the read level, Reddit, Wikipedia, and arXiv showed up across roles, but almost none of that survived into what got cited.
Now find the seat that shares almost nothing with anyone.
That’s the gap.
In our example run, it was Counsel: 14 cited sources, none shared with another role, all from legal, regulatory, and trademark sources.
Lowest competition on the map. Highest leverage.
You may also find an empty edge: two roles that both have to say yes but cite nothing in common. Their criteria collide with no content in between. Each empty edge is a bridge asset waiting to be built.
Don’t be surprised by who the gatekeeper is. In the committees we’ve mapped so far, the veto-isolate has consistently been compliance, security, or legal. The org chart underweights them, but the citation map doesn’t. It shows the seat that can stop the decision and has the least content support.
That’s where you build first.
Then check the phase re-run. When you move the committee from choosing to rolling out to getting value, the citation set shifts. Most brands focus on “choosing” and ignore everything after.
Also, the decision doesn’t end at the sale. It runs through rollout, adoption, renewal, and the next internal justification.
The move that pays is to drag the late “no” upstream, so the veto lands as a redirect rather than a demolition.
For our logo committee, “gatekeeper first” became a Founder’s Preliminary Trademark Clearance Brief: a one-page brief the founder fills out before Counsel reviews a name or mark. It captures proposed assets, commercial context, preliminary checks, and specific questions for Counsel.
Watch what that single page does:
It gives the CEO something Counsel can review before the work goes too far. The veto surfaces before money gets spent. And both sides avoid the silent standoff: “Why do they keep blocking this?” versus “FFS, did they even check?”
The brief makes them check (in a form that Counsel can act on) before either of them gets angry.
That’s anchor text becoming anchor context: “here’s exactly what this decider needs, at the moment they need it, in the form that lets them say yes.”
Google Ads is automatically enabling conversion-based customer lists for eligible advertisers starting, with data processing scheduled to begin on Aug. 18.
The update applies to advertisers already using both Enhanced Conversions and Customer Match but who have not yet activated conversion-based customer lists.
Why we care. As privacy changes continue to reshape digital advertising, Google is increasingly encouraging advertisers to rely on first-party data. Conversion-based customer lists provide another way to build audiences using customer data already collected through conversions.
The feature could help advertisers create more relevant audience segments and improve campaign performance without requiring additional implementation work.
The details. Eligible advertisers do not need to take any action. Beginning Aug. 18, Google will start processing data and automatically make conversion-based customer lists available within affected accounts.
Advertisers can then choose whether to attach those audiences to campaigns and ad groups as part of their targeting strategy.
The catch. Advertisers who do not want the feature enabled can opt out before Aug. 18 by disabling conversion-based customer lists within their account settings.
After that date, Google will begin processing data and generating the lists automatically.
First spotted. This update was spotted by JXT Group Founder Menachem Ani, who shared the comms he recevied about it on X.
Google Ads is changing how Smart Bidding strategies are labeled, separating target-based bidding strategies from volume-based bidding strategies.
Starting this month, “Maximize conversions with a Target CPA” will once again be called Target CPA, while “Maximize conversion value with a Target ROAS” will return to Target ROAS.
Why we care. The change is designed to make it clearer whether a campaign is optimising for maximum volume or attempting to hit a specific performance target.
The details.
Maximize Conversions remains available for advertisers focused on driving as many conversions as possible within budget.
Maximize Conversion Value remains available for advertisers focused on generating the highest conversion value possible within budget.
What isn’t changing. The update is purely organisational.
Google says there are:
No changes to bidding behaviour
No changes to campaign performance
No changes required from advertisers
Campaigns will continue to bid exactly as they do today.
For API users. Google is also aligning the interface more closely with how bidding strategies are represented in the Google Ads API.
Developers should review integrations, reporting tools, and campaign creation workflows to ensure they correctly recognise standalone TARGET_CPA and TARGET_ROAS strategy types.
Google is encouraging API users to monitor future updates related to:
The BiddingStrategyType enum
Standalone TargetCpa and TargetRoas messages
Optional target settings within MaximizeConversions and MaximizeConversionValue
One of the hardest lessons in PPC has nothing to do with bidding strategies, keywords, or campaign structure. It’s knowing when to walk away from a client.
On a recent episode of PPC Live The Podcast, performance marketing strategist Laura Abreu shared how taking on the wrong client early in her career became one of her most valuable professional lessons.
When your gut is telling you something
Laura’s first client was launching an ecommerce store selling beauty products from well-known brands. On the surface, it seemed like a great opportunity, but something felt off.
The products were available elsewhere at the same price, giving customers little reason to buy from an unknown retailer. Despite her concerns, Laura ignored her instincts and accepted the project anyway.
Great marketing can’t fix a weak business model
The team tried everything. Search campaigns, Meta ads, seasonal offers, product bundles, PR activity, and customer testimonials.
After three months of testing and optimisation, they hadn’t generated a single sale. The issue wasn’t the marketing. The business simply hadn’t established a compelling reason for customers to choose them over established competitors.
The importance of market validation
Many business owners believe hiring a marketer will automatically create growth. In reality, marketing amplifies demand—it doesn’t create it.
Today, Laura asks prospective clients whether they’ve tested the market, generated sales, and gathered customer feedback before investing in advertising. If the foundations aren’t there, paid media won’t solve the problem.
Pretty creative doesn’t equal performance
One of the biggest mistakes marketers make is judging creative based on personal preference rather than data.
The team invested heavily in creating beautiful visuals, but attractive creative alone wasn’t enough to drive sales. Customers don’t buy because an ad looks good; they buy because the offer resonates with their needs.
The emotional cost of a bad client
The failed project affected Laura far beyond the campaign results. As many marketers do, she tied her self-worth to the outcome.
The experience damaged her confidence so much that she stopped taking PPC clients for a period of time. Looking back, she realised she was carrying responsibility for a business problem that advertising could never have fixed.
Why expectations matter
One lesson Laura now applies with every client is setting expectations early and clearly.
Rather than promising immediate growth, she positions advertising as a way to test assumptions, validate demand, and uncover opportunities. This creates more honest conversations and avoids unrealistic expectations from the outset.
Why Laura doesn’t work with friends or family
Perhaps the strongest lesson from the experience is a rule she follows to this day: she doesn’t work with friends or family.
Maintaining professional distance allows her to stay objective, make decisions based on data, and avoid the emotional complications that can arise when personal relationships and business become intertwined.
Reputation is more valuable than revenue
When campaigns don’t go as planned, Laura believes honesty is non-negotiable.
Whether that means admitting mistakes, offering additional support, or refunding fees where appropriate, protecting your reputation is more important than protecting your ego. In an industry built on referrals, trust is everything.
Common mistakes Laura sees in PPC accounts
Having audited accounts across multiple markets, Laura says one of the biggest mistakes marketers make is treating campaigns as “set and forget” assets. She often finds underperforming creatives left running for months, ad copy that hasn’t been refreshed, and winning ads that aren’t being scaled effectively.
She also sees businesses creating unnecessary friction in lead generation campaigns. Long-form copy, overly complex forms, and sending users to external landing pages instead of testing native lead forms can all reduce conversion rates. In her experience, simpler journeys often deliver better results.
How Laura thinks marketers should use AI
Laura sees AI as a powerful tool for automating repetitive tasks rather than replacing marketers. She recommends using it to monitor performance, automate alerts, and streamline workflows so practitioners can spend more time on strategy and client communication.
At the same time, she warns against relying blindly on AI-generated outputs. Poor-quality ad descriptions and generic messaging can hurt performance, so human oversight remains essential. The marketers who succeed will be those who combine AI efficiency with strong strategic thinking.
OpenAI is quietly expanding its advertising infrastructure, giving UK businesses early access to a self-serve Ads Manager for ChatGPT — a signal that the company is building out the tools needed to scale advertising on its rapidly growing AI platform.
What’s happening. OpenAI has begun rolling out Ads Manager Beta to businesses in the UK, according to an email sent to advertisers announcing availability of the platform.
The self-serve interface allows businesses to create advertising accounts and begin exploring campaign management tools with minimal setup friction.
How it works. The Ads Manager dashboard is organized into four core areas: campaigns, tools, billing and settings.
The interface is designed to be familiar to digital marketers, with user management and campaign controls accessible through a simplified navigation structure.
For agencies. OpenAI is advising agencies and freelancers not to create accounts on behalf of clients.
Instead, clients should:
Create their own Ads Manager account.
Navigate to Settings → Users → Invites.
Invite agency partners with appropriate permission levels.
Once invited, users receive an email prompting them to accept access and can then switch between client accounts within the platform.
The catch. Unlike Google Ads’ Manager Account (MCC) structure, advertisers cannot currently view or manage multiple accounts simultaneously from a centralized interface. Users can switch between accounts, but each account must be accessed individually.
Why we care. Access to Ads Manager gives UK brands and agencies the opportunity to understand the interface, workflows, and account structure before broader adoption begins.
By removing requirements such as upfront billing information and simplifying account creation, OpenAI is lowering barriers for marketers who want to test and familiarize themselves with ChatGPT’s emerging advertising ecosystem.
What to watch. The UK rollout offers one of the clearest indications yet that OpenAI is moving beyond experimentation and toward a scalable advertising platform.
The next questions for advertisers will be less about account setup and more about inventory, targeting capabilities, measurement tools, and how ads ultimately appear inside ChatGPT conversations.
For now, marketers are getting their first hands-on look at the infrastructure that could underpin OpenAI’s future advertising business.
First spotted. The update was spotted by Head of Paid Media at Evoluted, Chris Ridley, when he shared the comms he received on LinkedIn.
Google is bringing generative AI directly into Google Ad Manager with the launch of Ask Ad Manager, a new Gemini-powered assistant designed to help publishers analyze performance, troubleshoot issues and navigate the platform using natural language.
The beta launches this month as Google pushes deeper into AI-powered ad operations.
What’s happening. Ask Ad Manager is a conversational AI agent built specifically for publishers using Google Ad Manager.
Unlike traditional reporting tools, publishers can ask questions in plain language and receive personalized answers, recommendations and reports based on their own Ad Manager data.
Google says the tool is designed to help users move from analysis to action faster by reducing the time spent generating reports, diagnosing problems and navigating the platform.
What it can do:
Troubleshoot delivery issues.
Instead of manually pulling reports to investigate underperforming line items, publishers can ask the AI agent questions and receive guidance on potential causes and next steps.
Generate reports on demand.
Users can request custom metrics, benchmarks and performance reports through a simple prompt rather than building multiple reports manually.
Navigate Ad Manager faster.
Ask Ad Manager can direct users to relevant pages within the platform and automatically apply the appropriate filters and settings based on the conversation.
Why we care. For publishers managing large inventories and complex campaigns, the ability to quickly surface insights and diagnose issues could reduce operational workload and accelerate decision-making.
The feature also reflects a growing shift across ad tech toward AI agents that can perform tasks and streamline workflows instead of simply generating information.
Looking ahead. Google says Ask Ad Manager is just the beginning of a broader move toward what it calls a more “agentic” future for advertising operations.
The company plans to introduce additional AI capabilities throughout the year, including developer tools such as REST APIs and an MCP server to support workflow automation and integrations.
Google is also developing specialized agents that could help publishers and agencies discover inventory, negotiate deals and execute campaigns more efficiently.
Bottom line. Ask Ad Manager brings Gemini-powered assistance directly into Google Ad Manager, giving publishers a new way to access insights, resolve issues and manage advertising operations through natural language prompts.
USA Today Co. is using AI-assisted shell files to publish breaking sports coverage faster. The strategy is designed to capture search traffic before Google’s AI Overviews summarize the news.
The publisher tested the approach during the 2026 Winter Olympics and is now using it for coverage of the 2026 FIFA World Cup, Digiday reported.
USA Today pre-writes breaking stories. The USA Today network, which includes the flagship site and more than 200 local publications, creates automated shell files for likely breaking news events. AI pulls subheads, photos, and links from the publisher’s archive. Editors turn that material into ready-to-publish files, allowing reporters to add new details, update the headline, and publish quickly.
“We’re trying not to be as reliant on SEO strategy. Pre-writes are huge,” Alicia DelGallo, USA Today Sports editorial director, told Digiday.
The search window is shrinking. Publishers have long pre-written stories to move faster in Google Search. AI Overviews have increased the pressure.
DelGallo said USA Today wants to publish while search interest is still rising, before Google has enough information to generate an AI Overview.
Barry Adams, founder of Polemic Digital, told Digiday he has seen AI Overviews appear for news events within about four hours and no later than half a day, though he said there is no firm data yet.
Olympics coverage drove 116 million views. USA Today Co. said its national and local network generated 116 million page views from Winter Olympics coverage between Jan. 1 and Feb. 28. The flagship USA Today site drew 91 million page views, up 82% from the 2022 Winter Olympics.
DelGallo said the shell-file system helped the publisher move quickly on breaking Olympics coverage, including Lindsey Vonn’s crash.
Why we care. AI Overviews can compress breaking news into answers within hours. Publishing first improves your chances of capturing search demand before Google answers the query itself.
World Cup gets the playbook. USA Today Co. is now using the system for World Cup coverage, with five shell files ready each day. The publisher is also investing in original reporting. It has reporters in all 16 host cities and a dedicated World Cup hub.
DelGallo said the newsroom wants stories that don’t read like generic search content. That means stronger byline authority, more on-the-ground reporting, and angles readers can’t find elsewhere.
Traffic may still fall short. USA Today Co. has 40 million monthly unique visitors to its sports content and expects a World Cup traffic boost, especially with the U.S. co-hosting the tournament. DelGallo said USA Today still expects “massive audience” spikes from the World Cup. But she said AI Overviews have likely lowered the traffic ceiling compared with a year ago.
Google Ads is rolling out a beta that allows advertisers to connect additional data sources directly to website conversion actions, giving marketers a new way to supplement tag-based measurement with backend conversion data.
The feature enables advertisers to combine conversion signals collected through Google tags with transaction data from systems such as CRMs, order databases and ecommerce platforms.
What’s new. Advertisers can now attach an additional data source to an existing website conversion action through Google Ads Data Manager or the Data Manager API.
The beta is designed to supplement — not replace — website tagging by allowing advertisers to send conversion data from backend systems into the same conversion action used for campaign measurement and optimization.
Why we care. The new beta helps fill conversion measurement gaps by combining Google tag data with first-party data from backend systems like CRMs and order databases. This can recover conversions that may be missed due to browser restrictions, privacy settings, or ad blockers, giving advertisers a more complete view of campaign performance.
Why Google launched it. According to Google, combining tag-based measurement with backend conversion data can help advertisers create a more complete picture of conversions and improve campaign performance.
The company says the feature can help:
Recover conversions that may not be captured by website tags.
Improve measurement resilience.
Provide more comprehensive data for automated bidding.
Simplify data integration through Data Manager.
How it works. The system combines website conversion data collected through Google tags with conversion records uploaded from an advertiser’s backend systems.
To prevent duplicate reporting, Google uses transaction IDs to identify and deduplicate conversions between the tag and the additional data source within the same conversion action.
What advertisers need to know. The beta is currently limited to website conversion actions that use Google tag or Google Tag Manager implementations.
It is not available for:
Google Analytics imported conversions.
URL-based conversion actions.
Google recommends adding an additional data source to an existing conversion action rather than creating a new one to avoid potential double-counting across campaign goals.
Data requirements. Every upload must include:
Transaction ID.
Conversion date and time.
Advertisers must also provide at least one attribution identifier, such as hashed customer information or a Google click identifier.
Google recommends uploading conversion data as quickly as possible and ensuring uploaded conversion values match the same currency format used by website tags.
Bottom line. The beta marks Google’s latest effort to strengthen conversion measurement by bringing backend transaction data directly into Google Ads. As advertisers look for more complete performance data, the new capability offers a streamlined way to supplement website measurement with first-party business data.
AI is changing how Americans search for information. A new Pew Research Center report found that 60% read AI-generated summaries at the top of search results, and about 40% use chatbots to find information.
AI-generated answers now appear across both traditional search results and dedicated chatbot platforms, including tools like ChatGPT, Gemini, and Copilot, Pew found.
AI summaries reach most searchers. Six in 10 U.S. adults said they’ve read AI summaries at the top of search results, Pew found. Three in 10 said they haven’t.
Another 10% were unsure, suggesting some users don’t clearly recognize AI summaries in search results.
Men were slightly more likely than women to report reading them, 63% versus 57%. Adults 65 and older were the least likely age group to read them.
Chatbots are search tools. About half of U.S. adults now use AI chatbots, up from about one-third in 2024. Roughly one in four adults use them daily.
Searching for information was the most common chatbot use Pew measured. About 40% of U.S. adults use chatbots for search, ahead of entertainment, image and video creation, medical advice, fitness information, news, emotional support, and companionship.
Work was close behind. Among employed adults, 38% said they use chatbots for job-related tasks.
ChatGPT dominates. ChatGPT remains the most widely used chatbot by a wide margin. Pew found that 44% of U.S. adults now use ChatGPT, up from 34% last year and more than double the share measured in 2023.
Gemini ranked second, with about a quarter of adults using it. Copilot and Meta AI followed.
Grok, Claude, and Character.ai had much smaller reach. About one in 10 adults or fewer said they had used each tool.
Why we care. People now find information through traditional results, AI summaries, and chatbot answers. A traditional search ranking may not reflect every place people now find answers.
About the data. Pew Research Center surveyed 5,119 U.S. adults from Feb. 17-23, 2026, through its nationally representative American Trends Panel. The full-sample margin of error was plus or minus 1.6 percentage points.
Google AI Overviews cited self-promotional “best” listicles while excluding the brands behind them from recommendations in 69% of cases, according to a new analysis of B2B software queries by Lily Ray.
Brands have used self-serving listicles to influence AI search results, but Ray found Google often cited those pages while recommending competitors instead.
By the numbers. Ray analyzed 100 B2B “best [category] software” queries in Google AI Overviews across three dates: April 15, May 15, and June 8.
Of the 80 prompts that triggered an AI Overview, self-promotional listicles were cited 323 times.
In 224 cases, Google cited a brand’s own page but didn’t recommend that brand.
Competitors get recommended. Ray documented several cases where Google cited a brand’s “best” listicle while recommending better-known competitors.
For “best LMS for selling courses,” Google cited Oasis LMS but did not recommend it. Instead, it recommended Kajabi, Thinkific, LearnWorlds, and Teachable — all of which are named in the Oasis LMS article.
Similar patterns appeared in queries for help desk, task management, survey, CRM, and SEO software.
Stronger brands still appeared. Brands that already led their categories, were widely mentioned by third-party sources, and had stronger link profiles were more likely to appear in AI Overview recommendations, according to Ray.
The data showed a consistent split between citations and recommendations. A brand’s page could appear as a source while competitors received the recommendation.
Organic visibility fell. Ray also reported organic search declines for many sites that relied heavily on self-promotional listicles.
The declines began around Jan. 20, across dozens of sites she analyzed. Many also scaled other SEO- and GEO-focused content formats, including AI-generated articles, comparison pages, and large volumes of “best” pages ranking their own brand first.
Those declines continued and accelerated during Google’s May 2026 core update, according to Ray.
Review sites gained citations. Ray found Google relied heavily on third-party and user-generated-content sites for “best” queries, with Reddit citations increasing sharply in recent months.
Forbes, Reddit, and YouTube were among the most-cited domains in AI Overview responses containing “best.”
Why we care. A citation is not a recommendation. Your content can appear in an AI answer while helping competitors capture the visibility that matters most.
Search Engine Land also reported that the tactic may create legal risk under the FTC’s Consumer Review Rule when company-controlled content is presented as independent reviews, reviews are not based on real use, or material relationships are not clearly disclosed.
About the data. Using Ahrefs Brand Radar, Ray collected AI Overview answer text and cited sources for 100 B2B “best [category] software” queries at three checkpoints between April and June. The analysis measured two outcomes: whether a self-promotional listicle was cited and whether the brand behind it was recommended.
PPC budgeting in 2026 isn’t just about setting spend levels. It’s about knowing when to adjust budgets, when to scale campaigns, and how the data feeding Google’s automation influences those decisions.
Google’s automation systems have always followed the signals you give them. In 2026, they follow them faster and with more confidence than before, which means clean signal architecture matters more than ever.
The fundamentals of budget management haven’t changed. What has changed is how quickly a poorly architected account can waste budget.
Two budget mechanics you need to understand right now
Before you adjust targets, audiences, or bid strategies, make sure you understand how these two budget controls work.
The ad scheduling pacing change
Google now paces all campaigns with ad scheduling toward the full 30.4x monthly billing cap, regardless of how many days your ads actually run. Before this change, a $100 daily budget on a weekday-only campaign targeted roughly $2,200 in monthly spend across 22 active days.
Now it targets $3,040, compressed into those same weekdays. The billing ceiling hasn’t changed. The system pursues it more aggressively within your active windows.
If your campaigns use ad scheduling, recalculate your daily budget based on your intended monthly spend rather than active days: divide your monthly target by 30.4 and set that as your daily limit. A $2,200 monthly target becomes a $72 daily budget. Campaigns running 24/7 aren’t affected.
Available for Demand Gen, Search, Standard Shopping, Performance Max, and YouTube campaigns, campaign total budgets let you set a fixed spend ceiling for a defined period rather than managing a daily limit.
For Search, Standard Shopping, and PMax, the window is three to 90 days. For Demand Gen and YouTube, it can run up to a year.
Unlike daily budgets, there’s no daily spending cap. The system can front-load or back-load spend within the flight to hit the total, which makes these useful for promotions and product launches, but worth monitoring closely when run alongside always-on campaigns.
Budget type can’t be changed after campaign creation, so the decision is final at setup.
What actually controls how Google Ads spends your budget
Efficiency targets usually constrain spend before budgets do
Smart Bidding treats your efficiency target as the primary constraint and your daily budget as the secondary one.
If you set a $50 tCPA and market conditions are returning leads at $80,the system restricts bids rather than generating conversions above your target. The daily budget cap never gets hit because the efficiency target is stopping spend first. What looks like a budget problem is usually a target problem.
When the gap between target and market reality is that wide, set your initial target closer to where the market is actually converting. Let the system accumulate conversion data and establish what efficiency looks like for your account, then gradually tighten toward your real goal.
The 10%-20% margin above target is a fine-tuning tool. It gives Smart Bidding enough room to find conversion opportunities when you’re already close to where you want to be, not when you’re $30 away.
Performance Max decides where your budget goes
Performance Max automatically distributes budget across Search, Shopping, Display, YouTube, and Discover. You set the total. Google decides the split.
Without brand exclusions, PMax will serve branded queries that would have converted through Search campaigns at a lower cost, which inflates its apparent efficiency while increasing your overall costs.
Campaign-level negative keyword lists for PMax have been available since January 2025, with the per-campaign limit expanded to 10,000 in March 2025. If your PMax campaigns predate that rollout, audit whether you have categorical exclusion lists built at the campaign level.
Jobs, salary, free, login, reviews, and any vertical-specific non-customer queries should be in there before the campaign launches, not added reactively from the search term report.
AI Max expands where your ads can appear
AI Max for Search, generally available since April, expands query matching beyond your keyword list, generates ad copy from your existing assets, and adjusts landing page targeting dynamically.
The budget risk is query drift: spend that was concentrated on your defined keywords now competes with AI-generated matches. AI Max provides search term reporting, which makes monitoring tractable. Review it closely during the first 60 days and proactively build categorical negatives.
The signal problem that makes budget allocation fail
An insurance broker running Smart Bidding toward form completions saw conversion volume rise 416% year over year while revenue stayed flat. The conversion action was firing on form starts, not form submissions.
The system had found the most efficient path to form page interactions and was scaling it confidently. A significant portion of those interactions were Cyrillic-language spam submissions from outside the service area. The dashboard was green. The pipeline was empty.
This is the core mechanism behind most budget waste in lead generation: identical conversion values across all form fills leave Smart Bidding with no basis to distinguish a qualified lead from a bounced session.
The system optimizes for volume and finds the cheapest path to completions. It follows its instructions precisely. The instructions are the problem.
Primary conversions should be high-intent, high-value actions that directly train Smart Bidding. Secondary conversions, such as newsletter signups, page views, and soft engagement, belong in reporting but should not influence bidding. Getting this distinction right is more consequential for budget efficiency than any adjustment to bid strategy.
Journey-aware bidding, currently in beta for Search campaigns on Target CPA, addresses the delayed-conversion problem that compounds this issue for B2B accounts.
Instead of optimizing only toward front-end actions, the system learns from the full lead-to-sale funnel — form submissions through closed deals — using intermediate stages as learning signals without counting them as biddable conversions.
The feature requires first-party CRM data, connected via Offline Conversion Import or Enhanced Conversions for Leads, to function. Without that pipeline data, there’s nothing for the system to learn beyond the form fills it was already optimizing toward.
For accounts not yet in the beta, extending your conversion window to 90 days and evaluating performance over 60- to 90-day periods is the right workaround.
First-party data as budget guidance
Customer Match is the most direct way to tell automation what valuable traffic looks like. Google enforces a 540-day maximum membership duration for Customer Match lists, effective April 2025. Any record not refreshed within that window expires, which shrinks your list over time without regular uploads or a continuous CRM sync.
The most effective use of Customer Match for budget allocation is to exclude before expanding.
Apply your existing customer list as an exclusion on acquisition campaigns so the acquisition budget reaches new customers rather than people who are already buying from you.
Run retention separately, with its own budget, targets, and messaging. Mixing both in the same campaign with identical conversion goals produces a blended signal. Smart Bidding typically settles on the segment that converts most cheaply, which is rarely the most valuable one.
Note that using Customer Match for targeting and bid adjustments requires at least 90 days of account history and $50,000 in lifetime spend. Exclusions are available to all compliant accounts regardless of spend history.
For always-on daily budget campaigns, the 10-20% weekly increase guidance still applies. For campaigns using ad scheduling, work in monthly targets and divide by 30.4 rather than scaling daily limits.
Smart Bidding Exploration is now in open beta for Performance Max, with Shopping expansion announced at GML 2026. On Search campaigns, it generates, on average, 27% more unique converting users by pursuing queries the account wasn’t previously winning, temporarily relaxing efficiency targets to test new conversion sources. Short-term CPA or ROAS fluctuations during the exploration phase are expected. Evaluate on a 60-day window before drawing conclusions.
Demand-led pacing, announced at GML 2026 and rolling out for Search and Shopping campaigns, dynamically shifts daily spend toward periods of predicted higher consumer demand within your existing budget parameters. It’s a complement to daily budget management, not a replacement. Monitor your account for rollout availability.
For B2B accounts, scale on 60- to 90-day evaluation windows, not 30-day ones. Short windows systematically undervalue campaigns with long sales cycles by cutting spend before the attribution data has time to accumulate.
Adobe announced a new solution to help businesses ensure their brands are visible, trusted, and chosen across AI surfaces.
Called Adobe Brand Visibility, the product is part of Adobe CX Enterprise, an agentic AI system designed to simplify customer lifecycle management — from acquisition and prospect engagement to conversion and long-term loyalty.
AI traffic is exploding. The use of LLMs to identify and research products and services marks a significant shift for both marketers and consumers. Alongside the announcement, Adobe released data showing substantial growth in LLM usage. AI traffic to U.S. retail sites surged 1,324% between October 2024 and May 2026. In the travel sector, AI traffic increased 2,215% over the same period.
“We used to get back the same thing (a SERP page with links on it). Now, the answers appear to be random, but they aren’t at scale. But companies don’t have tools to do it,” Loni Stark, vice president of strategy and product, Adobe, told MarTech.
Measuring brand visibility in AI search. Adobe Brand Visibility is Adobe’s first generative engine optimization (GEO) product since its acquisition of Semrush in May. It combines Adobe LLM Optimizer with Semrush’s AI Optimization tool.
Adobe Brand Visibility draws on nearly 300 million real-world AI search prompts, which Adobe says is the largest global database of its kind, helping teams identify which prompts they’re winning or losing.
Combined with Adobe’s first-party signals from owned channels, the platform gives marketers a view of how their brands appear across ChatGPT, Google AI Mode, Microsoft Copilot, and Perplexity. Metrics include mention frequency, audience reach, competitive share of voice, and content gaps. AI agents then surface prioritized recommendations, enabling teams to deploy updates quickly and measure their impact directly in the platform.
Competitive intelligence. Adobe Brand Visibility includes competitive brand comparison tools that let marketers benchmark against competitors, identify where their brands are cited, track brand mentions, and analyze historical trends.
The platform also includes SEO intelligence, reflecting the continued importance of SEO fundamentals in AI search visibility. Powered by Semrush data spanning 28.5 billion keywords and 43 trillion backlinks collected over 17 years, it shows where existing search authority should be generating AI citations and where content investments can close gaps across both channels.
There’s still much to learn about how LLMs work and how brands can improve visibility, but Star is confident Adobe is well positioned to lead in this space.
“Adobe had owned data. Semrush had data and trends. We don’t have all of the answers, but we have the best data,” Stark said.
Ask ChatGPT or Gemini to “review my on-page SEO,” and you’ll get a perfectly reasonable answer.
Reasonable. Generic. Boring. Uninspired. And almost identical to the answer your competitors get when they ask the same question.
That’s the problem with AI out of the box. It’s a generalist. It knows a little about everything and nothing about you — your business, your customers, your market, or the way you do SEO. The questions are loosely framed and inevitably come back with general answers.
The good news is that’s also the opportunity. The same tools that produce generic answers can become specialist assistants that encode your knowledge, process, and standards. No code required.
Building one is simpler than most people think. With tools like GPTs, Gems, and Claude Projects, you can package your SEO process into a reusable assistant that helps identify opportunities, automate repetitive tasks, and apply your expertise consistently.
Why generic AI gives generic answers
You don’t need a computer science degree here, but a basic understanding of how AI works helps explain the benefits of this approach.
Large language models are prediction engines. They’ve been trained on a huge slice of the internet and human knowledge, and when you ask a question, they predict the most plausible response based on everything they’ve seen.
In other words, by default, you get something close to the internet’s average opinion on a topic.
For SEO, the internet’s average opinion is … fine. It’s the same advice repeated across a million articles. Check your title tags. Improve your content. Build some links. Blah.
What the model doesn’t know is anything about your specific situation:
Your business, services, and commercial priorities.
Your marketplace and competitors.
Your customers and the problems they’re trying to solve.
Your way of working — the checklists, thresholds, and judgment calls you’ve refined over years.
The output is only as contextual as the input. Give it nothing, and you get the average. Give it your knowledge, and you get something far more useful.
It’s a computing problem as old as computing itself: garbage in, garbage out (GIGO).
There are a few ways to add that missing context, in increasing order of effort:
Better prompts: Include context in your question: who you are, what the business does, who the customer is, and what good looks like. This works, but you end up pasting the same 500-word preamble into every chat. It’s tedious and easy to skip when you’re busy, which impacts the quality of the output.
Custom instructions and knowledge files: Most AI platforms now let you save a set of standing instructions and upload reference documents. The AI reads these every time, so you set the context once and it persists.
Simple AI apps: Package those instructions and documents into a named, reusable tool with a specific job. This is where GPTs and Gems come in.
Actual software: Use AI coding tools to build real scripts and applications when you need automation beyond a chat interface.
The great thing is that the jump from a “big prompt” to a “simple app” is smaller than it sounds.
The skill is the same: clearly describing the job, the process, and the standards. If you can write a good brief for a junior team member, or a standard operating procedure (SOP), you can absolutely build one of these.
This is an important point because most people assume this is far more complicated than it is, and that assumption is holding them back. You don’t need to be a developer to do this.
The development of custom tools is no longer a heavily technical job. It’s becoming more of a creative endeavor enabled by these new AI tools and the simple, descriptive way of building apps.
If you can document your process, you can build an AI app.
The platforms: GPTs, Gems, Claude, and Replit
A quick tour of the main options for building simple AI apps:
GPTs (ChatGPT): Custom versions of ChatGPT with their own instructions, knowledge files, and capabilities. They’re shareable via the GPT Store, which is handy if you want to publish a tool for clients or your audience.
Gems (Gemini): Google’s equivalent. Custom versions of Gemini with instructions and knowledge files, with the obvious appeal for SEOs living in the Google ecosystem alongside Search Console, Analytics, Drive, and Sheets.
Claude Projects (Claude): Anthropic’s take. Project-level instructions and knowledge with a large context window, so it can hold a lot of your documentation in mind at once. My personal favorite at the moment.
Replit: A browser-based platform where you describe an app in plain English, and AI builds and deploys actual working software. Use this when a chat interface isn’t enough and you want a real tool with a real interface processing real data.
Claude Code: An agentic coding tool from Anthropic where you delegate coding tasks in plain language, and it writes, runs, and fixes the code. It’s brilliant for building scripts that crunch large exports — say, processing a 100,000-row Search Console export that would choke a chat window.
For most SEO and marketing professionals dealing with day-to-day optimization work, the sweet spot is the first tier: GPTs, Gems, or Claude Projects. They take minutes to build, require no code, and capture 80% of the value.
I’ll use Gemini Gems for the worked example below, as it’s the closest to home for those of us who live in Google’s world. The principles transfer directly to GPTs and Claude, and if you want to build something a little more advanced, have a play with Replit.
Google Gemini interface (on the Gems page)
Why not use existing SEO tools?
Standard SEO tools are brilliant at what they do — crawling, rank tracking, and link data. I use them every day. But they share a weakness: They’re generic by design, while your business is totally unique (or at least it should be). They have to work for every business in every industry, so they can’t know what matters to you. Everyone sees the same scores, the same recommendations, and the same “issues,” many of which don’t matter for your situation.
The tools are also largely focused on analysis and opportunity. The kinds of tools you can build with AI are more focused on the actual work.
Vanilla AI has the same problem from a different direction. Hugely capable, zero context.
The strength of building your own simple AI tools is personalization:
Your business: The AI knows your services, priorities, and commercial goals.
Your marketplace: It understands your competitors, customers, and niche.
Your knowledge: It applies your process — the way you’ve learned to do this work over the years — rather than the internet’s average.
That last point is the big one. After 30 years of doing this, my honest take is that the value isn’t the AI. The value is the knowledge and process you encode into it. Your experience is what matters — the AI is just your superpower.
What should you automate?
A simple rule: Automate repetitive tasks. Good candidates are tasks that are:
Repetitive: You do them the same way, over and over.
Process-driven: You could write the steps down for a junior team member to follow.
Data-heavy: They involve staring at exports and spotting patterns — exactly what machines are good at and humans get bored with and subsequently do poorly.
Reviewing Search Console data ticks all three boxes. So do first-pass on-page reviews, log file triage, internal link analysis, and monthly reporting prep.
What you don’t automate is judgment: strategy, prioritization against business goals, and the final call on what actually ships. The AI does the legwork and surfaces the candidates. You decide.
Example: Search Console quick-wins Gem
Let’s build a simple tool to help you mine Google Search Console for content ideas and easy wins.
I wrote “How to unlock easy wins in Google Search Console” two years ago, covering the creaky old human way of doing it. Let’s automate it to free up time for the really valuable creative work.
Note: This is a purposely simple example that’s ideal for AI and automation because the task is repetitive and the data is free.
Step 1: Define the job
Write one sentence describing what the tool does:
“Review Google Search Console performance data and identify prioritized quick-win opportunities, with specific recommended actions for each.”
Simple enough.
Step 2: Document your process
This is the important bit, and it’s where you have to think about the process.
What do you actually do here? What process do you follow? What easy wins and opportunities are you looking for?
Striking-distance keywords: Queries ranking just off page one (or just off the top positions) with meaningful impressions. Small improvements here can have an outsized impact.
High impressions, low CTR: You’re visible but not winning the click — usually a title and meta description problem, or a SERP feature is eating your lunch.
Declining queries and pages: Anything trending down versus the previous period that deserves attention before it becomes a problem.
Query-page mismatches: Queries landing on the wrong page, or multiple pages competing for the same query.
Unexpected queries: Things you rank for accidentally that hint at content opportunities.
For each of these, also note the thresholds and judgment calls. What counts as “meaningful impressions” — 100? 500? What CTR is “low” for position 3 versus position 8?
This is your experience being made explicit, possibly for the first time.
Step 3: Write the Gem instructions
Now open Gemini, create a new Gem, and translate that process into instructions. A solid structure is:
Role: Who the Gem is.
Task: What it does with the data it’s given.
Process: The steps, checks, and thresholds — your documented process from Step 2.
Output: The exact format you want back.
Guardrails: What it should never do.
Here’s an abridged example to adapt:
Role: You are an experienced SEO analyst. You are methodical, skeptical, and prioritize commercial impact over vanity metrics.
Task: I will provide an export of Google Search Console performance data (queries and/or pages, with clicks, impressions, CTR, and position). Review it and identify quick-win opportunities.
Process: Check for, in priority order:
Striking-distance queries — average position 5–15 with 100+ impressions.
High-impression, low-CTR queries — flag where CTR is significantly below what you’d expect for that position.
Pages or queries declining versus the comparison period.
Multiple pages ranking for the same query.
Output: A prioritized table with opportunity, query/page, current metrics, recommended action, and expected impact (high/medium/low). Maximum 15 rows. Quality over quantity.
Below the table, provide a short plain-English summary of the three actions I should take first.
Guardrails: Only use the data provided. Never invent queries, pages, or metrics. If the data is insufficient to assess something, say so. Ask clarifying questions if the export format is unclear.
That guardrails section matters more than people realize. “Only use the data provided” is your main defense against the AI confidently inventing things.
Gems can reference uploaded knowledge files. This is where you fine-tune things and add depth without bloating the instructions.
Examples include:
Your on-page optimization checklist (for when the Gem recommends title or content changes).
Your title and meta description guidelines, so suggested rewrites follow your standards.
A short brand and business context document — who the client is, what they sell, and which products or services are commercial priorities.
This lets the Gem prioritize opportunities that matter, not just opportunities that exist. That’s especially important when reviewing Search Console data, as most sites show up for a wide range of searches that aren’t aligned with the client’s core goals.
Step 5: Save it
It really is that simple. Hit save, and you’ve created an AI app.
Step 6: Feed it data and test
Export your performance data from Search Console (Performance report > Export, or via the API or Sheets if you want more rows), then start a chat with your Gem and upload the file.
Browse to the Performance report and click Export in the upper-right corner. In this example, I use Google Sheets to keep everything in the Google ecosystem.
Then upload the file and ask for the output you want.
The Gem’s output — a prioritized quick-wins table for a real site
If at first you don’t succeed: The first output here wasn’t terribly useful for this site.
The recommendations didn’t align with the client’s goals. I had to revisit my third knowledge file regarding the business’s commercial goals and priorities.
After refining that document and running the analysis again, the suggestions became much more useful.
Step 7: Iterate like you would with a junior team member
The first version will get things wrong. That’s expected, and it’s actually the useful part.
A bad recommendation is a way to identify what could be improved. Whatever your answer is, that’s a rule that was missing from the instructions. Add it, and the Gem gets a little closer to working the way you do.
Treat it like a new team member. Review its work, correct it, and update the brief. After a few rounds, you’ll have something that delivers a genuinely useful first pass in seconds — and a documented process that’s valuable in its own right.
A note of caution
Some honesty before you let this loose on client work:
AI gets things wrong: Confidently. Always verify recommendations against the actual data before acting, and never let AI output go straight to a client without review.
Mind the data: GSC exports are business data. Check the privacy and data settings on whatever platform you use, especially when client information is involved, and make sure your approach aligns with any agreements you have in place.
It’s a first pass, not a final answer: The tool surfaces candidates. You supply the judgment. The moment you stop checking is the moment you make a mistake.
More simple SEO tools to build
Once you’ve built one, the pattern repeats. Same recipe — role, task, process, output, guardrails, and knowledge files — different job.
Any manual task you do repeatedly is a good candidate for this type of tooling. Examples include:
Keyword research assistant. Feed it seed terms and keyword exports. It clusters by intent and maps keywords to your site structure using your intent categories and customer personas.
On-page optimization reviewer. Paste a URL’s content and target query. It reviews the page against your checklist and suggests improvements in your preferred style.
Technical SEO triage. Feed it crawl exports. It prioritizes issues based on actual impact for your site rather than default tool severity scores.
Link opportunity finder. Feed it competitor backlink exports. It identifies realistic, relevant prospects based on your criteria and drafts outreach angles.
Content strategist. Load it with your personas and content strategy frameworks. It generates briefs and ideas anchored to real customer problems rather than generic topics.
Analytics insight reviewer. Feed it GA4 exports. It summarizes what changed, why it might have changed, and what’s worth investigating in plain English.
Search Console opportunity finder. The example we just built, easily extended into variants for content decay, cannibalization, or indexing reviews.
Each of these is an afternoon’s work.
The constraint isn’t technical. It’s whether you’ve documented your process clearly enough to hand it over. If not, this is a good opportunity to systemize your business and accelerate the work with a simple app.
Content operations can run on instinct at a small scale. With a strong editorial team, a handful of trusted writers, and an understanding of voice, there’s usually enough discipline to keep the calendar moving.
But some businesses aren’t built that way. For media rollups, large affiliate networks, entertainment properties, sports brands, and other content-led businesses, publishing at triple-digit volumes per day makes sense.
In some cases, it’s necessary to survive because content is the operating model rather than a marketing function, as it is in many B2B organizations.
At that scale, content strategies don’t break because of content. More often, they break because economics, systems, and editorial judgment stop speaking to each other.
That B2B distinction is important. If you sell a niche manufacturing ERP, you simply don’t need that scale of content. There’s not enough to publish. You’d be burning cash and operating outside the market.
Some categories have the depth and audience appetite required to sustain hundreds of daily articles. Sports is an obvious example. There are games, trades, injuries, recaps, rankings, interviews, opinion pieces, explainers, storylines, and the list goes on.
A business like The Athletic can support significant publishing volume because audience demand is real, while the revenue model includes subscriptions, direct sales, programmatic display, affiliate revenue, and likely other sources under the hood.
In Q2 2025, The Athletic generated $54 million in revenue, according to its last standalone financial report. Of that, 64% came from subscriptions, 26% from advertising, and 10% from affiliate and licensing revenue.
When most revenue comes from people actively choosing to pay, editorial quality is no longer a judgment call. It’s the most important commercial requirement. Economics, systems, and editorial judgment are forced to speak the same language.
Other models are more fragile. The clearest example is when monetization is driven primarily by programmatic display measured by RPM (say, more than 70% of revenue), with content rewritten from existing coverage or produced around short-term search and social opportunities, where margins require high output and very low production costs.
The formula is simple:
Revenue = (Pageviews ÷ 1,000) × RPM
Profit = ((Pageviews ÷ 1,000) × RPM) − Production Cost
So if a website earns 4,000 pageviews per article at a $16 RPM, it generates $64 in revenue.
Subtract production costs. The margin gets thin fast.
To generate meaningful profit, the organization has little choice but to publish hundreds of articles per day while doing everything it can to maintain quality, discoverability, and audience trust.
That’s where these content strategies break.
A content model that breaks under its own weight
More content can look like more revenue. But the spreadsheet tells only a fraction of the story.
Numbers don’t show editorial quality, whether thinner work is being produced to feed the machine, or whether monetization decisions are inadvertently weakening the asset.
Data surfaces where that drift starts. Points captured within a CMS include:
Content types.
Categories.
Tags.
Author and editor attribution.
Cross-referenced with sessions, pageviews, pageviews per session, session duration, RPM, source/medium, and other metrics.
That lets analysts drill into content types by source, category, and tag, while providing visibility into top performers, opportunities to optimize the ad stack by content type, and more.
Here are some simple scenarios that highlight what that looks like in practice:
An analyst runs a pivot table on an entertainment property and notices higher pageviews from Google Discover per article among list content in the reality television category tagged to a specific show. Since traffic equals more revenue, the conclusion is to write more lists about that show.
An analyst notices RPM is lower on features than lists, even though average word counts are the same. The reason is that the ad stack serves programmatic display after each image, and features have four times fewer images than lists. Since images drive higher RPM, the conclusion is to increase the number of images in features or reduce the number of published features in favor of more lists.
Fairly simple stuff on the surface. However, this is where judgment becomes the difference between a healthy operation and one that’s quietly eating itself.
Scaling these operations past 100 writers is mainly a question of whether the business has the systems, data, and judgment required to keep the operation from collapsing under its own volume.
It’s worth noting that 100 writers is rarely just 100 writers. For many of these businesses, it’s 100 writers across a dozen properties, which is actually more than 1,000 writers when you account for the full footprint.
Independent publishers don’t typically hit that scale because the infrastructure requires a level of investment they most likely don’t have access to.
That infrastructure includes clearly defined communication structures for editors, project management ownership, and comprehensive guides covering writing, linking, imagery, social, and CMS usage.
Without them, standards can degrade unpredictably across properties, and editors lose the ability to diagnose why or quickly point people toward resources when putting out fires.
On the data side, granularity is a must. Without consistent tagging and categorization built into the CMS from the start, analytics can become too fuzzy to act on.
Performance needs to be attributable at every level, rolled up into a P&L for each property, and then rolled up again across the conglomerate.
Technical infrastructure is essential as well, often in ways editorial teams wouldn’t expect.
If you consider how to get images into Google Discover, for example, it requires CDN delivery within specific guidelines. That’s more of an engineering problem than an editorial one. User roles and permissions across CMS and revenue dashboards are another example, along with the development resources required to implement the CMS architecture needed for data capture and reporting in the first place.
Proprietary systems can also be beneficial depending on a business’s scale. If you’re a rollup with a dozen properties operating on one or two CMS templates, it’s much easier to make bulk optimizations or accelerate the integration of newly acquired properties.
Channel distribution isn’t static either. Platform value to publishers shifts. Think about when Facebook stopped sharing news links in Canada. It changes the economics of whether a platform is worth optimizing for. Consistent monitoring and testing need to be built in.
The judgment that keeps it from collapsing
The systems above create favorable conditions, but they don’t guarantee sound judgment.
Let’s revisit one of the examples above:
The ad stack serves programmatic display after each image. Editorial guidelines require one image per entry in a list. This generates higher RPM across Google Discover traffic for lists with 20 thin entries at 1,000 words than for a well-constructed feature.
If you’re looking only at the spreadsheet, you’d favor doing as much of that as possible. That’s tempting, especially if employers incentivize target RPMs or sessions per article as KPIs tied to bonus compensation.
However, thin content at volume isn’t ideal for organic visibility. Once readers and search engines encounter too much low-quality output, the traffic disappears.
You’d essentially optimize for short-term yield, reinforce that behavior through employee bonuses, and damage the asset in the process.
Or another example:
An editor notices that updating a datePublished timestamp drives a short-term bump in traffic. The conclusion is to roll out timestamp updates across hundreds of pages.
The problem is that doing it at scale without substantive edits and strict guidelines may create distrust. That’s the judgment call.
Three things need to be held in tension: economic logic, infrastructure and systems, and the judgment not to sacrifice long-term gains for short-term wins.
While that sounds like common sense, these responsibilities are often owned by different people who don’t speak the same language.
Finding a way to bridge that gap is the most important challenge in a scaled content operation. Diversified revenue streams like The Athletic’s help enforce that alignment.
Otherwise, your content strategy will probably fail when you scale past 100 writers. And the examples above are just two of hundreds of scenarios where the spreadsheet points one way, and the right decision points another.
Looking to take the next step in your search marketing career?
Below, you will find the latest SEO, PPC, and digital marketing jobs at brands and agencies. We also include positions from previous weeks that are still open.
At Wave, we help small businesses to thrive so the heart of our communities beats stronger. We work in an environment buzzing with creative energy and inspiration. No matter where you are or how you get the job done, you have what you need to be successful and connected. The mark of true success at […]
About The Role Hi, I’m Jason Eliason and I lead the Jane Marketing Services team. I’ve been at Jane for two years, and what we’re building here is genuinely unusual – a team embedded inside a healthcare SaaS product, delivering SEO, Google Ads, Social, and Website services exclusively to the clinics that run their practices […]
What You’ll Own Own SEO strategy across StealthGPT product pages, blog, free tools, comparison pages, and programmatic landing pages. Build keyword maps around high-intent AI writing, AI humanizer, AI detector, SEO writer, and competitor-alternative searches. Create and manage content briefs for landing pages, articles, free tools, refreshes, and comparison pages. Improve page copy, titles, metadata, […]
Botify’s leading agentic AI search technology and seasoned experts ensure every brand has the power to be found, both in traditional and AI search. With one powerful platform, brands achieve visibility, relevance, and greater control across Google, Bing, ChatGPT, Perplexity, and more. Botify’s technology powers agentic workflows, AI-driven recommendations, and automated cross-platform indexation and deployment. […]
ABOUT THE ROLE We’re looking for a Growth Marketer to own the entire lead gen cycle. You’ll be the one turning heads and converting them into qualified leads (MQLs), pipeline opportunities (SQLs), and new revenue. This role is focused on building and scaling non-traditional lead gen paths that reach customers where they actually hang out. […]
VP / Head of Search & AI Visibility Location: United States (Remote / Hybrid Preferred) Reports To – President/Founder Company: Milestone Inc. (direct hire) Term: Full-time About Milestone Milestone Inc. is a leading Digital Experience Software and Services company dedicated to providing comprehensive solutions across all touch points that enhance customer engagement and drive business growth. […]
Clarity is the Global Growth Consultancy for B2B technology brands. As a senior-led consultancy, we align leadership, markets, and execution to turn complex growth ambitions into commercial momentum. Operating from hubs in London, New York, Amsterdam, and Sydney, our 100+ global team helps leaders navigate high-stakes growth tensions across Enterprise Tech, FinTech, Cybersecurity, and HealthTech. […]
We’re a content and organic discovery agency that helps brands show up in the right places — Reddit, YouTube, editorial, AI search. Our team is small, the work is real, and everyone here actually cares about doing it well. We’re looking for a Client Account Manager to be the connective tissue between our clients and […]
ABOUT NOGIGIDDY NoGigiddy is a digital platform built for gig workers, side hustlers, and anyone building an income outside the traditional 9-to-5. We connect our community with real earning opportunities — remote jobs, surveys, gig platforms, and financial tools — all in one place, free to access, no gatekeeping. We built what we wish had […]
Animalz is a content marketing agency that partners with B2B SaaS companies, venture capital firms, and other tech organizations to drive long-term, sustainable growth through high-quality content. Our fully remote team of strategists and content marketers delivers content strategies tailored to each customer’s goals and context. We pride ourselves on our deep interest and understanding […]
About Veracity: Veracity is the first doctor-developed, clinically proven metabolic health system designed to fix what’s actually wrong—not just patch symptoms. Supported by a team of 11 doctors and our Chief Science Officer with a PhD in cellular biology, previously the creator of Nutrafol, and backed by leading investors including Maveron, Melitas, L Catterton, and […]
Bright Data Ltd. is seeking a results-driven Marketing Manager – Americas to own regional pipeline growth across AI and non-AI offers. This hybrid position requires building and executing country-specific marketing plans, improving lead quality, and collaborating with various stakeholders. The ideal candidate will have proven experience in B2B marketing and a strong understanding of pipeline […]
Zocdoc is seeking a Strategic Finance Senior Manager to support its marketing teams in New York. This role is essential for driving growth and improving investment efficiency through data analytics and actionable insights. The ideal candidate should have over 8 years of experience in finance or analytics, excellent proficiency in Excel, and a strong understanding […]
Mammoth Brands, based in New York, is looking for a Growth Marketing Manager to enhance customer acquisition across Harry’s and Flamingo brands. This role emphasizes data-driven strategies for profitable growth through managing digital marketing campaigns on platforms like Meta, TikTok, and Reddit. The successful candidate will collaborate closely with internal teams and external partners to […]
Mitchell Martin is seeking an E-commerce Advertising Specialist for a hybrid role based in New York, NY. This full-time position involves optimizing advertising strategies for enterprise clients on major eCommerce platforms. The ideal candidate will develop and manage advertising strategies, analyze campaign performance, and collaborate with clients to meet business goals. Candidates should have a […]
Lead organic strategy for a portfolio of clients, building annual and quarterly roadmaps that account for both traditional SEO performance and Generative Engine Optimization (GEO) visibility across AI-driven search environments.
Serve as the primary point of contact for clients, guiding strategic conversations around organic growth, AI Overviews, LLM-powered discovery tools, and how evolving search behavior impacts brand visibility and competitive positioning.
Act as the lead interpreter of search performance and visibility quality across both traditional and AI-led discovery experiences
Maintain a broad expertise across the evolving landscape of LLMs and generative engines, including Google Gemini, OpenAI GPT, Anthropic Claude, and other leading frontier models
Write and publish high-quality, search-optimized content on a consistent cadence, including (but not limited to) blogs, thought leadership, comparison pages, FAQs, infographics, and other digital content assets.
Create structured, answer-first content designed to be surfaced in AI-generated responses and LLM-driven experiences.
Own and scale Talkiatry’s paid search program end-to-end, including forecasting, budgeting, pacing, bidding strategies, account structure, and creative testing for Google, Bing, and ZocDoc.
Develop and execute a rigorous testing roadmap, including ad copy, keyword strategy, landing page variants, and automation/algorithmic controls; quantify impact using sound experimental design.
Login
Exclude first-party audiences.
Get visibility into your ad placements.
Project spend with budget reporting.
Analyze audience demographics.
