Google has released the August 2026 spam update and said this update will take a few days to roll out. This update applies globally and to all languages, Google added.
More details. Google posted about the update on its incident dashboard and wrote:
Released the August 2026 spam update, which applies globally and to all languages. The rollout may take a few days to complete.
Why we care. This is the third Google spam update announced in 2026, the last one was the June 2026 spam update. If your site was not impacted, then you are good to go – at least for now.
There will always be cases of sites not spamming Google that get hit by a spam update but hopefully that won’t be one of your sites.
Most brands still don’t have a clear picture of how much traffic AI Overviews are sending them. Google hasn’t given us a clean signal for AI Overview traffic in Search Console, making it difficult to know how much organic traffic is coming from AI Overviews, which content is driving it, and how accurately that traffic is being reported.
So we built our own tracking.
Since September 2025, I’ve been capturing AI Overview referral data for one of the brands I manage in the transportation industry. From September 2025 to June 2026, we recorded 51,200 tracked events across 1,661 cited snippets.
The data shows that AI Overview prominence in the SERP isn’t stable. It also reveals clear patterns in the types of content Google cites, how those citations change over time, and how much of that traffic gets misattributed in analytics.
Here’s what nine months of tracking revealed — and what it means for SEO reporting, content prioritization, and GEO.
The tracking setup — and why it works
The methodology is straightforward, but most teams haven’t implemented it.
When a user clicks on a cited snippet inside a Google AI Overview, Google sometimes appends a #:~:text= fragment to the destination URL. We created a custom dimension in GA4 that fires whenever a session lands with that fragment present.
It’s not a perfect signal, but it’s the most reliable one available without waiting for Google to expose this natively in Search Console. The fragment is already sitting in your GA4 data. You just need to surface it.
Once we had it captured, we grouped snippets into thematic categories and started analyzing patterns. A few things stood out immediately.
Concentration is high: The top-performing snippet alone drove 2,276 events. The average across all 1,661 snippets is 31. Like most things in SEO, a small number of pages are doing the bulk of the work.
Snippets have lifecycles: Some peak during a specific period and then fade — usually when query intent shifts seasonally or when the content becomes stale relative to what Google is now preferring to cite. Others emerge months after publication and keep climbing. This makes freshness and specificity more important than you’d initially assume.
Content prioritization gets sharper: The most useful output of having this data is that it makes decisions more grounded. When you map snippet volume and growth trajectory against your current content investment, the gaps become visible.
Be the brand customers find first.
Track, grow, and measure your visibility across Google, AI search, social, local, and every channel that influences buying decisions.
Start your free trial
Certain topic types got cited
In our case:
Transfer time content and pricing content are already being cited frequently with clear upward momentum. The play there is to refresh and expand the existing content.
Destination guides are underperforming relative to what the data suggests they could be driving.
Structured transport comparison tables, formatted as actual HTML tables, are punching well above their weight in terms of citation frequency — which tells you something about how AI Overviews are selecting content to surface.
The broader pattern is consistent with what most people working in GEO are finding: AI Overview citations favor content that is specific, structured, and directly answers a well-defined question. Times, prices, named routes, comparison formats. Not vague editorial.
22.4% of AI Overview traffic is being misattributed to Direct
This was the most striking finding.
When we started plotting the dataset on a GA4 exploration report, we noticed that a meaningful portion of traffic arriving via AI Overviews was being attributed to the Direct channel rather than Organic Search.
We know AI Overviews only exist within Google Search. Traffic from there should be attributed to organic. So we built misattribution tracking into our AI Overview dashboard and started measuring it on a monthly and weekly basis.
Across the full dataset — more than 50,000 AI Overview events over nine months — the average misattribution rate was 22.4%.
That’s 11,468 AI Overview events attributed to Direct instead of Organic. The range varied by month:
Worst month: May 2026 – 29.3% of AI Overview sessions were misattributed to Direct.
Best month: April 2026 – 16.8% were misattributed.
The implications for SEO reporting are significant. Depending on your total organic traffic volume and how much AI Overview referral traffic you’re generating, you could be materially underreporting organic performance without realizing it.
AI Overviews are driving 7.53% of organic sessions — but it’s volatile
Using the same first-party dataset, we built a percentage metric to track what share of our total organic sessions came from AI Overviews each period.
From September 2025 to June 2026, 7.53% of organic sessions came from AI Overviews.
But the distribution is not flat:
At its peak in February-March 2026, it reached 16-17% — meaning nearly one in six organic visitors arrived via an AI Overview.
More recently (as of writing), it has trended down to around 2-4%.
That volatility is notable. It suggests AI Overview prominence in the SERP is not stable — it fluctuates based on query type, Google’s confidence in available content, and possibly broader algorithmic changes. Brands treating AI Overview traffic as a fixed percentage of organic are likely miscalibrating their models.
The caveats you should know
Two important limitations with this approach:
The #:~:text= identifier isn’t exclusive to AI Overviews
The same fragment is used by Featured Snippets and People Also Ask results. So some portion of what we’re capturing may bleed from those formats.
We did run the exercise in Ahrefs to check our Featured Snippet exposure, and at the time of checking, we were included in fewer than 30 — largely because we don’t target terms that typically trigger Featured Snippets.
In practice, we believe AI Overviews are the dominant driver given the volume and the period of data collection, but it’s not a perfectly clean signal.
The metric compares events to sessions
Our custom dimension in GA4 is event-scoped, not session-scoped. The AI Overview share percentage therefore compares events against sessions, which isn’t an ideal comparison.
It’s better than nothing — and it’s directionally accurate — but worth knowing if you replicate this.
If AI can’t find you, customers won’t either.
Track your visibility across AI search, uncover missed opportunities, and grow your presence where customers are asking questions.
See your AI visibility
What this means practically
If you’re not tracking this yet, start. The setup isn’t complex, and the signal — imperfect as it is — is the best first-party data you’ll get until Google decides to surface this natively.
A few things to take away from nine months of data:
Structured, specific content wins. Times, prices, routes, named comparisons. Not editorial padding.
Freshness matters more than you think. Snippets have lifecycles. Stale content loses citation share.
Your organic traffic is likely underreported. If ~22% of AI Overview events are going to Direct, your channel reporting has a systematic gap.
The share is volatile. Don’t treat a peak month as a new baseline.
The #:~:text= fragment is sitting there in your GA4 data right now. You just need to surface it.
Special thanks to Simant Sah for helping collate the data and put together the report where the screenshots come from.
https://i0.wp.com/dubadosolutions.com/wp-content/uploads/2026/08/image-61-WW0DkN.png?fit=2048%2C812&ssl=18122048Dubado Solutionshttp://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.pngDubado Solutions2026-08-18 15:00:002026-08-18 15:00:00What 9 months of AI Overview data and 51,000+ tracked events reveal
Microsoft is giving advertisers a new server-side option for sending conversion data directly to its advertising platform, reducing their dependence on browser-based tracking.
What’s happening. Microsoft Advertising has published new documentation for its Conversions API (CAPI), a server-side measurement solution currently in beta.
CAPI lets advertisers send conversion and customer interaction data directly from their own systems to Microsoft Advertising through a single connection.
The details. Microsoft says the Conversions API can support multiple measurement scenarios, including online and offline events.
The potential benefits include:
Improving conversion measurement accuracy.
Increasing attribution coverage.
Sending both online and offline conversion events.
Reducing reliance on browser-only tracking methods.
Supporting multiple measurement use cases through one server-side integration.
How it works. Microsoft recommends using CAPI alongside Universal Event Tracking (UET) rather than treating it as a replacement.
UET continues to capture activity in the browser, while CAPI provides a server-to-server connection for sending event data directly to Microsoft Advertising.
Using the two together can give advertisers more complete measurement coverage when browser signals are unavailable or limited.
Why we care. CAPI gives Microsoft advertisers another way to preserve conversion signals and connect online advertising with actions that happen outside the browser, including offline events. For advertisers already investing in server-side measurement across other ad platforms, Microsoft’s solution could also make it easier to build a more consistent measurement strategy across channels.
Yes, but. CAPI remains in beta and isn’t available to every Microsoft Advertising account yet.
Advertisers will also need the technical resources to establish and maintain a server-side integration, making implementation more involved than simply installing a browser tag.
The big picture. Microsoft is following the broader advertising industry’s shift toward combining browser and server-side signals rather than relying exclusively on tracking pixels.
The new documentation shows that server-side measurement is moving from an advanced implementation option toward a standard part of advertisers’ measurement stacks.
Google is giving advertisers clearer rules for measuring whether YouTube and Demand Gen ads prompt people to search for their brands — while making several notable changes from the metric’s original rollout.
What’s happening. Google has updated its documentation for Branded Searches, a conversion type that measures when someone searches for an advertiser’s brand on Google or YouTube after seeing an ad.
What’s changed is how Google now defines its availability, attribution and reporting.
The details: Google’s updated guidance introduces or clarifies several important points:
7-day default window: Branded Searches now uses a seven-day default conversion window, compared with the 30-day view-through window described when the feature originally launched. Advertisers can adjust the window between one and 30 days.
YouTube and Demand Gen: Google’s current documentation lists YouTube and Demand Gen as eligible campaign types. Performance Max, which was included when Branded Searches was announced in 2025, is no longer listed.
Consideration goal: Google now formally categorizes Branded Searches under the Consideration goal.
Reporting, not bidding: Branded Searches is treated as a primary conversion action but isn’t available as a bidding optimization goal. The data appears in Results and All Conversions rather than the standard Conversions column.
Brand mapping still matters: Advertisers don’t need to set up a Search Lift experiment, but Google says brand mapping must be configured for measurement to work.
Google’s documentation says Branded Searches can be viewed at the campaign, ad group and asset levels, as well as through Report Editor.
Why we care. Branded search behavior can provide advertisers with a useful signal between an ad impression and a traditional conversion.
Someone may see a YouTube ad, remember the company and search for the brand days later without clicking the original ad. Branded Searches is designed to make that influence more visible inside Google Ads.
That can help advertisers assess whether upper-funnel campaigns are generating genuine brand interest instead of evaluating them solely on clicks and direct conversions.
The bottom line. This isn’t a new conversion type or a new seven-day attribution window. The noteworthy update is that Google’s current Branded Searches documentation lists YouTube and Demand Gen — but no longer Performance Max.
First spotted. This update was spotted by Hana Kobzova, founder of PPC News Feed.
For the entirety of Google Search Console’s life, it has spoken one simple language: your website. Pages, queries, clicks, impressions, all tied to a domain you own and verify.
On July 7, that changed. Google shipped Platform properties, a new property type that lets you verify a social or video account and get the same first-party search performance data GSC has always given website owners. Think clicks, impressions, CTR, average position, and the actual queries behind them — just pointed at your social presence instead.
Worth noting up front: This covers Search, Discover, and News. So a TikTok clip or a YouTube Short can now be tracked across all three surfaces the same way you’d track a blog post.
The mechanics of the launch — which platforms are supported, how verification works, and how this differs from Google’s separate Search profiles feature — are already well covered.
However, one detail deserves repeating because it’s time-sensitive: there’s no historical backfill. Data collection starts the moment you verify. That means every week you spend deciding whether this is useful is a week of query data you can’t get back.
Let’s take a look at what this means for brands and SEOs, some examples of what it looks like in practice, and what it means for marketing going forward.
What GSC Platform properties actually changes for brands
The obvious win is a single first-party lens that gives brands access to both owned and earned content, instead of having to stitch together GA4 and native platform dashboards with a hefty dollop of guesswork.
But Google’s own guidance points at something more specific than better reporting. It points at a workflow. Three parts of it are worth knowing about:
Query groups: The Insights report clusters the search terms driving traffic into top, trending-up, and trending-down groups. That’s a real input for planning captions, hashtags and next topics. Ultimately, you’re working from demand that already exists rather than guessing at it.
The 24-hour filter: If a post starts picking up search traffic within a day of going live, you can catch it while it’s still live and act by cross-promoting to another platform or timing a follow-up.
Format comparisons: Using URL-based comparison filters, you can put Shorts against long-form, or Reels against static posts, and get an actual answer to “should we be putting more into short-form?”
These same three mechanics also change how you should work with creators. None of those three workflow parts are limited to owned content decisions, despite the signals coming from owned data.
Be the brand customers find first.
Track, grow, and measure your visibility across Google, AI search, social, local, and every channel that influences buying decisions.
Start your free trial
How GSC Platform properties can reshape creator partnerships
Smart brands could already be considering running them through a creator program to answer the questions brands have historically had to take on faith.
Query groups show you which conversations you already have traction in, but that also means, by omission, they tell you which conversations you don’t have a presence in. These insights become the foundations of an honest brief for creator activity.
Rather than simply picking creators based on reach and hoping the relevance follows, you can point to specific query territory you’re absent from and brief against it — developing your presence within new conversations. Alternatively, you could double down within a conversation by partnering with creators who can champion your brand, further developing your presence and preference.
This thinking reframes creator spend as buying coverage of conversations you don’t own yet or want to show up more within, rather than taking a risk on buying an audience you hope overlaps with yours.
Format comparisons can also feed straight into the brief itself. For example, if your own data shows Shorts consistently earning search visibility while long-form doesn’t for your category, that’s a concrete production instruction for a creator rather than a stylistic preference.
And the 24-hour filter gives campaign measurement a dimension it didn’t have before. During a live campaign, you can see whether creator content is earning genuine search visibility or just in-app impressions. This could provide a better read on whether the work is generating durable discovery or a spike that vanishes with feed changes.
There’s a neat symmetry here, too. This data lets creators show that their content genuinely ranks for specific searches, rather than simply racking up views. Platform properties give both sides of the negotiation the same evidence base and metrics.
Creator conversations that once centered on follower counts can now focus on which queries each party actually shows up for. That’s a far stronger basis for partnerships and can strengthen search marketing across the social-search landscape.
An example of how GSC Platform properties could work in the wild
Picture a hypothetical running shoe brand called Tiger Feet. They connect their Instagram and YouTube properties to the GSC Platform properties, and a few weeks later, the Performance report shows something nobody expected: A Reel about lacing techniques for wide feet is pulling steady clicks from Google for variations of “how to lace running shoes wide feet,” while the website has no page targeting that query.
That’s three decisions in one data point:
There’s proven demand: You’re seeing real-world results for a topic nobody on the Tiger Feet content team had previously flagged, and you know it’s real because people are searching it and landing on your content.
You’re capturing it in the wrong format for conversion. A Reel earns the click but sells nothing. A product-linked guide on your site, with the video embedded, could.
You now know what to write and roughly how to angle it. Because Platform properties give you the exact query language, rather than a keyword tool’s approximation of it, you have a road map to where to focus your content next.
Run the same logic in reverse, and it’s just as useful: If a social post is already ranking well for a query, that’s a case for not commissioning a page that could end up competing with it.
That’s the shift. Social content stops being measured purely on engagement and starts being measured on demand, which is a completely different question and a much more commercially useful one.
Platform properties give you first-party data, not competitive visibility
Platform properties only work on accounts you can verify, which means you get your own data and nothing else. No competitor view. No share of voice. No category benchmark.
Every other discipline in search has some form of competitive visibility. You can see who’s outranking you organically, estimate what competitors are bidding on, audit their backlinks, and track their positions.
Here, you get a perfectly clear view of your own performance and no visibility elsewhere — so this data should be combined with SERP analysis and manual research rather than relying on it in isolation.
What are the practical consequences of this limitation? It means:
You can’t tell good from great. If an article pulls 400 clicks from Google in a month, is that strong for your category or embarrassing? There’s no external reference point, so you’re benchmarking against your own history and nothing more.
You’ll spot demand, not competition: Query groups will tell you people are searching a topic and finding you. They won’t tell you that three competitors are already better positioned for it, or that the space is wide open. That still needs conventional keyword and SERP research alongside it.
Absence of data isn’t absence of opportunity: A query showing nothing in your reports might mean no demand, or it might mean healthy demand that’s all going to someone else. The report can’t distinguish between the two, and it’s an easy mistake to make.
The useful framing here isn’t that it’s a new report for SEOs to get lost within. The way to approach this development is as if it’s a new content inventory that behaves like your existing one.
That’s because Platform properties speak the vocabulary SEOs already work in, which means social content can easily fold into existing workflows instead of sitting in a silo owned by another team — further affording brands the opportunity to break down silos at a time when discovery of brands relies on collaboration across organic media strategies.
What can you do as an SEO with Platform properties? Turns out, a lot:
Gap and overlap analysis: You can check whether a query is already being answered by the brand’s own social content (or ask creator partners to share insights on their answers on your behalf) before commissioning a page for it. You can also spot cases where a page and an article are quietly competing and determine whether or not this is a good or bad thing.
Test captions and titles like meta titles: You can annotate the date that you rewrote a TikTok caption or YouTube title and then compare performance on either side of it. Same discipline you’d apply to standard optimizations.
Apply structural thinking off-site: Comparing things like playlist performance or formats within a platform allows you to think about grouping and cannibalization for content that doesn’t even live on your domain at all.
There are two caveats worth bringing up with clients before this lands in a report.
Reporting carries the usual Search Console delays, so this isn’t a real-time dashboard despite the 24-hour view.
Coverage will vary by platform and content type while the rollout matures, so early numbers need to be treated as directional rather than definitive.
Social strategy has traditionally sat with the social team and been judged on engagement. Platform properties give SEOs a legitimate reason to be in that conversation. The content is now visible in the same tool, in the same terms they already report in and understand, and that they can develop effective strategies around.
What GSC Platform properties change for search marketing, generally
If you step back and look at this new rollout, it confirms something Google has been building toward for years: The results page stopped being a list of blue links a long time ago.
Google’s own documentation shows exactly where this content appears — short-video carousels, “latest posts” carousels, and “what people are saying” SERP features — all within Search and Discover results.
Platform properties are Google handing marketers an instrument for a part of the SERP that was already there and simply going unmeasured. This points to three things worth thinking about:
The search universe is expanding, and the SEO/social divide keeps dissolving: When both are measured in the same tool with the same metrics, it gets harder to defend treating them as separate disciplines with separate owners and separate budgets.
Feed-based discovery is becoming as strategically important as query-based search: In terms of the role search plays in a brand’s discovery, that shift has been visible for a while, and this is the first proper measurement tool for it.
Google is now building on the assumption that search behavior is cross-platform: Google isn’t just tolerating the fact that people search for things they saw on TikTok anymore. Instead, it’s actively shipping insights that acknowledge that it’s happening.
Own the conversation before your competitors.
See where your brand appears, where it doesn’t, and exactly how to win more visibility across search, AI, local, social, and every channel that matters.
Start your free trial
What you can do this week to take advantage of Platform properties
You should connect every active social account now. Don’t wait for a question that needs answering. With no backfill, any account you haven’t connected to is a question you won’t be able to answer later. Take advantage.
If you’re an SEO, rather than a social manager, get yourself into that reporting. Query groups and format comparisons are exactly the kind of data that should be shaping a content calendar. For most brands, that decision-making is happening without SEO as part of the conversation — the best brands will ensure that that changes.
I’ll admit that this isn’t a dramatic feature, on the face of it, although I, for one, am extremely excited about the potential this affords “search everywhere” strategy considerations. The new platform properties are Google quietly conceding that “search performance” was never really about the website. It was always about wherever people go looking for answers.
https://i0.wp.com/dubadosolutions.com/wp-content/uploads/2026/08/Google-Search-Console-now-connects-social-content-to-search-demand-6uDZO6.png?fit=1920%2C1080&ssl=110801920Dubado Solutionshttp://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.pngDubado Solutions2026-08-18 14:00:002026-08-18 14:00:00Google Search Console now connects social content to search demand
A Peec AI study of 5 million query fanouts collected between April 1 and April 21, 2026 reveals how AI platforms rewrite and expand user queries before executing searches.
ChatGPT consistently injects words like “best,” “reviews,” and the current year into queries, even when those terms were not in the original prompt.
ChatGPT’s query fanouts containing “reddit” grew from roughly 0.15 percent to 3.68 percent between January and May 2026.
ChatGPT uses Reciprocal Rank Fusion, meaning content that appears across multiple fanout searches scores higher than content that surfaces for only one.
Fanout analysis should now be a standard part of AEO audits alongside citation tracking.
Most marketers optimizing for AI search visibility are focused on the wrong layer. They are tracking citations, which is the output. What they should be studying is fanouts, which is the input that determines whether a citation is even possible.
A new analysis of five million query fanouts from ChatGPT, Perplexity, and Grok reveals how AI search platforms actually work behind the scenes, and what that means for how brands show up in AI-generated answers.
How Query Fanouts Actually Work
When a user asks ChatGPT a question, the model does not simply search for that exact phrase. It executes a set of related sub-queries behind the scenes, each exploring a different angle of the original prompt, then merges those results to build its response. That set of sub-queries is the fanout.
A question like “best project management tools for remote teams” might produce fanouts for “top project management software 2026,” “remote team collaboration features,” “project management pricing comparison,” and “enterprise versus small team project management tools,” all running simultaneously. The answer ChatGPT provides draws from all of those sources combined, not just from content that matches the original wording.
ChatGPT uses Reciprocal Rank Fusion to combine scores across these sub-queries. This means content that appears across multiple fanout searches is weighted more heavily than content that only surfaces for one. Covering a topic thoroughly from multiple angles increases your citation probability because content surfacing across more sub-queries scores higher under RRF.
What the Reddit Signal Actually Means
The most actionable finding from the fanout data is the rise of Reddit as a deliberate source. ChatGPT’s query fanouts explicitly referencing Reddit grew from approximately 0.15 percent to 3.68 percent between January and May 2026. This is not incidental. It reflects a pattern where AI systems seek out human-centered, experiential content that branded or editorial sources often do not provide.
Reddit provides something that polished brand content typically cannot: unfiltered customer sentiment, specific use-case discussions, and genuine peer-to-peer evaluations of products and services. AI systems appear to be recognizing that value explicitly, surfacing Reddit as a source for the kind of human validation that helps construct credible answers.
For brands, this creates two implications. First, your presence and reputation within relevant Reddit communities now directly influences what AI systems say about you. Second, the broader principle is that AI platforms are seeking out authentic experiential content as a distinct source type from authoritative or editorial content. Both matter, and they are sourced differently.
Fanouts Before Citations: The Right Audit Sequence
Most AEO work currently focuses on citation tracking: which AI platforms mention your brand, in what context, and with what sentiment. Citation tracking is valuable, but it is a lagging indicator. By the time a citation appears or fails to appear, the fanout decisions that determined it have already been made.
Fanout analysis is the leading indicator. It reveals which angles, source types, and content formats AI systems are actively looking for when a user asks a question in your category. Knowing that ChatGPT consistently injects “best,” “reviews,” and the current year into queries means you can build content that specifically addresses those retrieval patterns, beyond the surface-level keywords your audience uses.
The practical audit sequence should run from fanout to source type to citation. Start by identifying the fanout patterns for the high-intent queries in your category. Then assess which source types are being pulled for each fanout. Then determine where your content appears or fails to appear across those source types. The gap between where your content exists and where AI systems are looking is the optimization target.
What Reciprocal Rank Fusion Means for Content Planning
The Reciprocal Rank Fusion mechanism ChatGPT uses to combine fanout results has a direct implication for how content should be planned and structured.
Because content that appears across multiple fanout searches is scored higher than content that appears for only one, the brands most likely to earn AI citation are the ones that cover their core topics from multiple angles across multiple content assets. A single well-ranked pillar page is less effective under RRF than a cluster of interrelated content that addresses the same topic from different perspectives: a main guide, a comparison piece, a use-case breakdown, an FAQ, and a data-driven research piece.
This is not a new content strategy principle. Topical authority and content clustering have been standard SEO practice for years. What is new is the mechanism that rewards it. Under traditional ranking, a single highly authoritative page on a topic could outcompete a thinner content cluster. Under RRF, the cluster wins because it surfaces across more of the sub-queries AI executes when researching the topic.
For content planning, this means mapping your content against the likely fanout patterns for your most important queries, alongside the primary keyword. If ChatGPT consistently generates eight sub-queries when a user asks something in your category, and your brand has content that addresses two of them, your citation probability is structurally lower than a competitor that addresses six.
Structural Implications for Content Strategy
The fanout data reinforces several content strategy principles that have value independently of AI search, but are now especially important.
Comprehensive topic coverage matters more than individual page optimization. Because Reciprocal Rank Fusion weights content that appears across multiple fanouts, a brand that covers a topic from multiple angles including comparisons, use cases, reviews, and Q&A formats is more likely to earn citation than a brand that has one well-ranked page on the subject.
Listicles and comparison content are structurally favored. “Best” is the most commonly injected word in ChatGPT fanouts. Content positioned around “best for specific need” or structured as a comparison aligns directly with how AI systems rewrite queries before they execute them.
Third-party and community signals are a distinct optimization layer. Given the explicit Reddit preference emerging in fanout data, brands that generate genuine customer advocacy, encourage community participation, and maintain a real presence in relevant online discussions are building AI visibility through a channel that owned content alone cannot replicate.
FAQs
What is a query fanout?
A query fanout is the set of additional searches an AI system executes behind the scenes after receiving a user prompt. Rather than searching only for what the user typed, AI platforms rewrite and expand the query into multiple sub-queries that explore different angles of the original question.
Why does Reddit appear so frequently in ChatGPT fanouts?
Reddit provides a form of human-centered, peer-validated content that AI systems appear to treat as a distinct source type. Review-style, experiential content from community discussions provides a signal that branded editorial content typically does not, particularly around product sentiment, real-world use cases, and comparative user experiences.
How do I audit my fanout coverage?
Start by identifying the high-intent queries most relevant to your category. Then use a fanout analysis tool or inspect ChatGPT’s web search behavior to see which sub-queries it generates. Map your existing content against those sub-queries and identify where you have no representation. Those gaps are your content agenda.
Should I try to game Reddit to improve AI visibility?
No, and this will backfire. AI systems that are sourcing Reddit content for its authenticity will not reward manufactured or inauthentic participation. The value of Reddit as a source comes from genuine user experience, not brand-managed content. The right approach is to build products and services worth discussing, support customers effectively, and participate in relevant communities authentically.
Conclusion
The shift from citation tracking to fanout analysis is the next maturation step for generative engine optimization strategy. Citations tell you what AI said. Fanouts tell you what AI looked for. If you are already tracking AI brand visibility, fanout analysis is the natural next layer to add. Brands that understand the latter will be better positioned to influence the former.
The Reddit signal is the most actionable near-term finding. Authentic presence in relevant online communities is a measurable input to AI search visibility as well as a brand-building exercise. Building that presence consistently, through genuine customer advocacy and community engagement, is one of the highest-leverage investments a brand can make for AI discoverability in the current environment.
http://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.png00Dubado Solutionshttp://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.pngDubado Solutions2026-08-17 16:00:002026-08-17 16:00:00Inside ChatGPT’s Source Preferences: What Query Fanouts Reveal About AI Discoverability
Google’s AI Mode has surpassed one billion monthly users in its first year, with queries more than doubling every quarter.
The new Intelligent Search Box accepts text, images, files, videos, and Chrome tabs as inputs, powered by Gemini 3.5 Flash.
Information Agents run 24/7 in the background, monitoring the web on the behalf of users without requiring active searches.
Agentic Search allows users to build custom tools, dashboards, and trackers directly inside Search.
Brands that are already investing in structured, authoritative, intent-led content are best positioned for this shift.
Click-through rates for position one rankings have already collapsed from 27 percent to 11 percent, meaning rankings have never mattered more.
Is Google Search still “search” at all? At Google I/O 2026, Search VP Liz Reid said it plainly: “Google search is AI search.” That’s not a product update. That’s a declaration that the platform has fundamentally changed what it is and what it does.
For marketers, the implications are significant. However, they’re not as disruptive as the headlines suggest, provided you’ve already been doing the work that matters.
What Google Actually Announced
The I/O 2026 announcements covered three distinct capabilities rolling out across Google Search.
The first is the Intelligent Search Box, described by Liz Reid as the biggest upgrade to the search interface in over 25 years. The redesigned input field dynamically expands for complex queries, accepts multiple input types including files, images, and open Chrome tabs. It also replaces traditional autocomplete with AI-powered intent suggestions. AI Overviews and AI Mode are now unified inside a single interface.
The second is Information Agents, persistent background systems that monitor the web 24 hours a day on behalf of users. Rather than requiring a user to return and search again, an information agent continuously tracks changes across blogs, news sites, social posts, and real-time data sources, then sends a synthesized update when something relevant shifts. The practical parallel is Google Alerts, rebuilt with an LLM’s ability to reason about what it finds.
The third is Agentic Search, which allows users to build custom mini-apps, dashboards, and trackers from inside Search using Gemini Spark and Google Antigravity. A query about tracking market movements can produce a live monitoring tool, while questions about apartment hunting can generate a filter and alert system. Search is no longer returning links. It’s producing outputs.
What This Means for Organic Search
The honest read on these announcements is that traditional ranking behavior is already changing. SISTRIX data published in March 2026 documented a collapse in position-one click-through rates in Germany, from 27 percent to 11 percent. AI overviews already reach 2.5 billion monthly users, and AI Mode crossed one billion monthly users in its first year.
Users are already getting answers without clicking, and in the future, Information Agents will provide users with information without initiating a search at all. The interaction model is shifting from reactive to proactive on Google’s end.
For brands, this changes two things. First, appearing in AI-generated answers is now as important as ranking in traditional results. Second, being visible to information agents is a new form of visibility that did not exist before. Both depend on the same foundation: content that is clear, authoritative, structured, and genuinely useful.
Why Existing SEO Foundations Still Win
The temptation after an announcement like this is to treat everything as a disruption requiring new tactics but the reality is more grounding. The structures we have been building (e.g., entity authority, intent-matched content, technical accessibility, and strong E-E-A-T signals) are precisely what AI-native search rewards.
Agentic formats and generative user interfaces (UI) are new creative outputs, not replacements for content quality. An information agent monitoring a topic for a user will surface the sources that best answer the relevant questions. Those sources win by being authoritative, current, and clearly structured, which are the same qualities that drive strong traditional rankings.
What changes is the scope of optimization. Brands now need to think about how their content performs in AI-generated answers, not only in ranked results. That means writing content that answers questions directly, using structured data appropriately, and building topical authority deep enough to earn citation across a range of related queries.
The Click-Through Collapse Is Already Happening
Before the I/O announcements, the underlying data was already telling this story. Position-one click-through rates have fallen across a variety of industries as AI Overviews expanded. That is not a future projection. It is a current reality in every category where AI Overviews appear consistently.
The implication is not that ranking has become less important. It is that ranking has become more important while delivering less traffic per position. A brand that holds position one and earns a citation in the AI Overview above it is doing as well as possible in the current environment. A brand that holds position one but does not appear in AI features is exposed to ongoing CTR erosion without the additional visibility.
This creates a new dual-visibility mandate: perform in traditional organic results and earn citation in AI features simultaneously. These are related but not identical optimization problems. Content that ranks well and earns AI citation is an asset. Content that ranks well but fails to earn AI citation is a risk.
What to Do Now
Audit your highest-value pages for AI readability. Ask whether an AI system could produce an accurate, complete summary of your content from what you have published. If the answer is “no,” that’s a gap worth closing.
Identify opportunities in emerging agentic formats. Information agents will surface sources that consistently answer a category of questions well. Brands that build comprehensive, current, well-organized content around their core topics will be the ones those agents trust.
Prioritize non-commodity content. Google has been explicit on this point across multiple communications. Content that adds original data, unique expertise, or genuine perspective will outperform content that simply covers topics already well-served on the web.
Monitor how your brand appears in AI Mode responses, not just in traditional rankings. These are now two distinct visibility channels, and both require attention.
FAQs
Does this change my existing SEO strategy?
Mostly it extends it rather than replacing it. The content quality, entity authority, and technical accessibility standards that drive strong traditional rankings are the same foundations that determine AI citation. What is new is the need to monitor AI visibility specifically and optimize for answer-engine citation alongside traditional ranking.
When do Information Agents launch?
Information Agents are rolling out first for Google AI Pro and Ultra subscribers in the United States during the summer of 2026, with broader availability planned for 2027.
Does ranking still matter if AI answers the query directly?
Yes, and arguably more than before. AI systems cite sources from across the quality spectrum, but authoritative, well-structured pages from high-trust domains are significantly more likely to earn citations. And for transactional or navigational queries where users still click, ranking position continues to drive meaningful traffic.
What is Agentic Search and should brands care about it now?
Agentic Search lets users build custom tools inside Search using Gemini Spark and Google Antigravity. In its current early rollout, it is most relevant to brands watching where search interfaces are headed. In the medium term, it will matter most for brands whose audiences track time-sensitive information: pricing, inventory, news, market data.
Conclusion
These recent announcements confirm a trend that has been developing for several years. AI is the primary interaction layer. Traditional blue-link results are becoming one output format among many. Users will increasingly get answers, alerts, and tools without clicking anywhere at all.
Notably, though not everyone is excited about this shift. In fact, as anti-AI sentiment grows, some users are actively switching to search engines like DuckDuckGo. Monitoring customer usage of alternative search engines (and optimizing for them if needed) is key.
The brands that will win in this environment are already building the kind of content that earns citation, authority, and trust. The work has not changed. The surface it needs to perform on has expanded.
http://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.png00Dubado Solutionshttp://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.pngDubado Solutions2026-08-14 16:00:002026-08-14 16:00:00Google Search Is Becoming AI Search: What This Means for Your Brand
Your brand could be ranking #1 on Google, but still be invisible to AI.
Absent from conversations your customers are having with large language models (LLMs) about your category.
Or worse, showing up inaccurately, with outdated or incorrect information that’ll hurt your sales.
Without prompt tracking, you’d never know.
Prompt tracking (sometimes called LLM visibility tracking) is the practice of monitoring how your brand shows up in AI answers over time, through mentions or citations.
It’s different from traditional SEO rank tracking, which tells you where your URLs appear on search engine results pages (SERPS) for specific keywords.
With rank tracking, you ask, “How close are we to position 1?”
But LLMs don’t answer questions with a static list of 10 blue links.
They pull from massive amounts of information to generate a unique response every time, tailored to the user and the context of the conversation.
You can even ask the same question twice and get two different answers.
In this search experience, it matters less whether your brand is mentioned first, and more that it says true positive things about you to the right people — consistently, across many runs of similar prompts.
But without a system to monitor it, you’re flying blind.
Prompt tracking gives you the directional intelligence to spot AI visibility gaps — the queries you’re consistently not showing up for — and close them.
This guide shows you exactly how. You’ll walk away with a free tracking template, a step-by-step system, and two real-world expert setups you can steal.
Free template: Download our prompt tracking spreadsheet to start understanding your brand’s AI visibility across LLMs ASAP.
Why Brands Need to Track Prompts
According to a study from Orbit Media, 55% of US internet users rely on AI as their primary or frequent research tool. Thirty-two percent use it for product recommendations.
Translation: A growing share of buyers are learning about you in AI tools. Without ever visiting your website.
Overall AI visibility scores tell you whether you’re showing up. Prompt tracking tells you where and how.
It can help you understand:
The types of questions you’re showing up for and where they fall on the customer journey (ToFu, MoFu, or BoFu?)
The questions your competitors are pushing you out of (and your share of voice on important topics)
The sentiment around your brand mentions
The questions you’re getting cited for, but not recommended (sometimes called ghost ranking)
This level of data helps you spot specific trends and gaps in your AI visibility over time.
They show up for prompts at all three stages of the funnel.
With prompt tracking, Gong can focus on conversations most likely to drive revenue and stop spending time and money tracking ones that won’t.
If their mention rate stays low on key BoFu topics over time, that’s a signal to update on-site or third-party content.
They can also see which relevant topics competitors are owning while they’re absent.
For example, Gong’s Engage product helps with lead generation.
But Salesforce and Hubspot consistently own these prompts.
This tells Gong two things:
Their audience isn’t aware of their lead generation use case
They need more content around it to train AI platforms to mention them
Sentiment prompts like ‘Is Gong worth the price?’ and prompts that surface ghost ranking can reveal similar trends and gaps.
The important thing is to track data over time.
AI answers are non-deterministic. The same prompt can return different brands across different runs.
Regular, repeated tracking is how you get meaningful signals you can act on.
When LLM Prompt Tracking Isn’t Worth It (and When It Is)
Prompt tracking isn’t for everyone.
Plenty of businesses spend time and budget on it and still walk away with data they can’t use.
Prompt tracking is NOT worth it when:
Your audience isn’t using AI to find solutions in your category
Your site isn’t set up to be crawled by AI
You don’t publish content regularly
You’re not looking for competitive or brand narrative insights
You need a single KPI to report upward
You don’t have bandwidth to act on insights
TL;DR: Prompt tracking yields valuable insights — but only if you’re set up to act on them.
If you are set up, the returns can be significant.
Take Gong, the sales call intelligence tool. It’s a great candidate for prompt tracking.
It has a full content engine that publishes across multiple channels (blog, reports, video, media coverage, audio, and more).
They compete in a crowded category where comparison prompts are common.
And they serve a buyer (sales leaders) who increasingly uses AI to evaluate software.
(Seventy-one percent of B2B software buyers now rely on AI chatbots for product research, according to G2, up from 60% in 2025.)
On the other hand, a local HVAC company that gets all its leads from Google Business Profile (GBP) and word-of-mouth doesn’t need prompt tracking. At least not yet.
Even if AI is overlooking them, building a content engine from scratch just to fix that isn’t a realistic investment.
Building Your Prompt Set: What to Include and Why
Don’t try to track every possible prompt your customers could be using.
Instead, focus on prompts you actually care about getting mentioned or cited in.
They should map directly to your product offering, audience pain points, and moments close to purchase.
Types of Prompts to Include
Tracking these four types of prompts over time will yield the most helpful insights:
Evaluation prompts: “Best tool for x use case” and specific feature queries
Reputation prompts: “Is x product worth the price?”
Comparison prompts: “Alternatives to x product,” “x tool vs. x tool,” “best x tools”
Gap prompts: Priority topics your competitors are pushing you out of
The first three are focused on understanding how your product is being recommended in buying conversations.
The last one is about understanding your competitive landscape and where you could improve.
Pro tip: Don’t just track one prompt for each type. Looking at answers for a single prompt is just noise. Reviewing a cluster of prompts over time is a real signal.
Margaret Kapitany, Offsite SEO Lead at Hootsuite, shares how she focuses her prompt set:
The prompts worth tracking are the ones that most closely mirror how a potential buyer would actually ask their AI for help, especially close to a purchase decision. For me at Hootsuite, that means prompts that cover comparison, evaluation, and recommendation queries from a social media manager or CMO, phrased the way they’d talk to a colleague or trusted industry peer.
What that looks like in practice at Hootsuite:
Prompts worth tracking
Prompts not worth tracking
“How does Hootsuite compare to [competitor]?” (Comparison)
“What is social media management?” (Pure definition — won’t convert)
“Which social media management platforms integrate with Salesforce?” (Evaluation)
“Is Hootsuite a good company?” (Vanity — brand mention is baked in, nothing actionable)
Where to Find Prompts Worth Tracking
Find prompts wherever you normally go to learn about your audience.
To find prompts worth tracking, look at:
Keyword research: Commercial and Transactional intent queries like “best x software”
Google’s “People also ask” (PAA) boxes: Comparison and evaluation questions like “Best alternative to x” or “does x integrate with y”
Perplexity’s related questions: Similar to PAA, comparison and evaluation questions
Reddit, Quora, Facebook Groups in your industry: Repeated questions, especially ones that compare options or express frustration
Sales call transcripts: Repeated questions asked right before or during a purchase decision
Semrush prompt suggestions for your brand: Queries tied to buying decisions that you or your competitors are showing up for
Pro tip: For every question you uncover, do a quick gut check: “If someone asked an LLM this, would I want to see my brand show up? Would I be upset if it didn’t?” If the answer is “Yes, and yes,” keep it.
Let’s return to Gong as an example of how to find prompts.
Keyword research shows me the questions people are asking Google about my category, how popular they are, and the language they use.
If I use a tool like Semrush, I can filter to Commercial or Transactional intent keywords (the ones labeled “C” or “T”). And add the most popular and relevant ones to my prompt tracker.
Then, I can dig through Reddit forums my audience frequents to find repeated frustrations and buyer queries.
I would take “Sales enablement tech stack suggestions,” “AI tools for sales enablement.” I’d ignore the queries about SMBs or startups because those aren’t Gong’s target audience.
Next, I’d extract the comparison or evaluative prompts Perplexity surfaces when I prompt it with terms related to my business.
For Gong, I used “best conversational insights tool for sales” to find some good candidates:
Gong vs. Chorus (or any other competitor on this list)
Which conversational insights tool has the best ROI for enterprises?
Pro tip: When writing prompts, don’t agonize over exact wording the way you would with keywords. LLMs cluster semantically similar queries together. Track a few natural variants, like “best sales enablement software” and “top sales enablement tools.”
These sources are solid, but you’re still inferring and collecting them manually is slow.
The exact prompts your audience runs in LLMs and how often they’re using them, ranked by frequency.
This grounds your prompt set in real demand that’s always up to date.
You can be sure you’re tracking questions your users are actually asking.
From this list for Gong, I’d choose to track prompts under “AI-Driven Sales Enablement” and “Sales Coaching and Enablement Tools,” as they’re BoFu queries related to my product.
Organize By Product or Use Case
To build your first prompt set, start small.
All you need is 20-30 prompts over 4-6 broad categories that align with your product offering or use cases.
Ensure every category includes a mix of your four types of high-value prompts and add them as tags.
For a B2B SaaS company like Asana, this prompt tracking setup could look like the following:
Project management
Task management
Workflow automation
Team reporting
Best project management software for marketing teams (Evaluation)
Best task tracking tools for cross-functional teams (Evaluation)
Best workflow automation software for ops teams (Evaluation)
Best project reporting tools for enterprise teams (Evaluation)
Asana vs. Monday.com for project management (Comparison)
Does Asana actually improve team productivity? (Reputation)
Asana vs. ClickUp for workflow automation (Comparison)
Is Asana’s reporting good enough for large teams? (Reputation)
Project management software with Slack integration (Evaluation)
Best task management software for small teams (Gap)
Asana vs. Notion for managing marketing workflows (Comparison)
Asana vs. Smartsheet for project visibility (Comparison)
Pro tip: Use branded prompts only for comparison and reputation tracking, as they can inflate your visibility score. Keep the rest of your prompt set unbranded so you actually learn where you’re getting found, not just where you’re already known.
This setup lets you easily see which categories you’re winning and losing in over time.
Or, what types of answers you need to do a better job of showing up in.
Example: Hootsuite Prompt Set
You may decide to add more tags or organize your prompts in a different way as you expand.
For example, Margaret uses multiple different tags — not just categories — for her prompt set for Hootsuite.
We’ve built out prompts in three ways:
Funnel stages, with most of our attention on conversion
Our target industries, with tailored terminology
Intent type: comparative, evaluative, integrative (e.g. “works with X tool”), and problem-led (“how do I solve Y”)
Almost all of our prompts carry multiple tags, e.g., BoFu + Healthcare + Evaluative. That tagging is what lets me slice the data later.
When leadership asks for numbers, she can report on which industries Hootsuite is most visible in, or cross-reference visibility for BoFu prompts with direct traffic trends.
Margaret’s setup shows an important lesson: However you build your prompt set, structure it so you can answer the questions you’ll want to ask later.
How to Track Prompts: Step-by-Step Process
All you need to get started with prompt tracking is a spreadsheet and 30 minutes a week.
With our tracker, you can log the following for each prompt:
The prompt itself
Category
Tags (e.g., type, industry)
The LLM you’re testing it in
Whether your brand was mentioned (yes/no)
Whether you were cited (yes/no)
Sentiment of the mention (positive, neutral, negative)
Competitors mentioned
The date
Customize it to whatever makes sense for your business.
Step 2. Run Each Prompt Across Multiple LLMs
Different LLMs pull from different sources, so answers will be different across all of them.
Tracking prompts from only one LLM won’t give you a full picture of your brand’s presence in AI search.
Ideally track prompts in all major LLMs, including:
ChatGPT
Gemini
Perplexity
Claude
If you need to save time, review Presenc AI’s 2026 platform demographics report to identify which LLMs your audience actually uses and focus on those.
Pro tip: Make sure you’re using a temporary chat to run your prompts. Your regular chat window will serve you an answer that takes into account everything it knows about you from past conversations. You want to track what an LLM might recommend to anyone, not just you specifically.
Step 3. Run Each Prompt at Least Twice Per Session (Optional)
Answers will vary run to run. So if you have the time, run each prompt 2-3 times per session for more reliable data.
You’ll catch when your brand shows up one out of three times (a 33% mention rate).
If you don’t have time, don’t worry. You’ll still see trends over time.
Step 4. Log Competitor Mentions, Too
Competitor data is half the value of prompt tracking.
Seeing a competitor get mentioned consistently in a category you’re performing less well in is a signal.
Maybe they published a new comparison page. Or were included in an influential report.
Look into their strategy to see what they’ve done to improve and learn from them.
Step 5. Track Weekly. Action on Data Monthly.
Regular prompt monitoring on a weekly basis is often enough to catch shifts.
It gives LLMs enough time to crawl and learn from new or updated content.
But don’t make any decisions with only one week of data.
A single week of low visibility in a category could be a fluke. Four weeks of it is a trend worth acting on.
(We’ll cover what to do when you spot one in the “How to Read and Act on Prompt Data” section below).
In this example, Asana shows up with a low citation rate for two LLMs only one week out of four.
Rushing to fix that ASAP could turn out to be a waste of time.
Step 6. Upgrade to an Automated Prompt Tracking Tool
Manual LLM visibility tracking is a great way to validate that you can actually get some useful insights from the practice.
But if you want to grow your prompt set beyond 20-30 prompts, it’s going to start taking much longer.
A tool like Semrush can help you move faster and suggest actionable opportunities based on your data.
To get started, open the Visibility Overview dashboard, enter your domain, and click “Check AI Visibility.”
You’ll see a summary of how often LLMs mention your brand, which competitors are mentioned alongside you, and a breakdown across each LLM.
Scroll down for a list of prompts where you’re already getting mentioned. Click “Opportunities” to see your gap prompts.
Click “Monitor” on the prompts you want to include in your prompt set.
Pick the LLMs you want to monitor, paste in your prompts, and hit “Start Tracking.”
You can add tags by intent, topic, or campaign so you can slice the data later.
Further reading: See how the top AI visibility tools stack up on pricing, LLM coverage, and reporting features.
Example: How an Agency Tracks Prompts for E-Commerce Brands
When you’re tracking hundreds of prompts across multiple clients, a basic spreadsheet won’t be enough.
Jonny Nastor, Founder & Head of Strategy at Digital Commerce Partners, knows this firsthand.
He built a prompt tracking map based on the theory that most buyers ask LLMs about a specific job they need done or task to accomplish. Then filter by their specific situation (a.k.a. constraints).
For example, when shopping for smart doorbells, a buyer might search “best video doorbell with no monthly subscription fees.”
The job to be done is “best doorbell.” The constraint is cost (low fees or no subscription).
He calls it a “Constraint Map.” Every intersection of job and constraint in the map becomes a prompt.
He gets ideas for constraints from keyword modifier data. In Semrush, you can see these in the Keyword Magic Tool.
He also pairs each prompt with search volume data to roughly understand its popularity and prioritize accordingly.
He runs his prompts across ChatGPT, Perplexity, and Gemini automatically through their APIs to log how each one responds.
Depending on the client’s needs, he tracks one or all of the following AI visibility metrics for each prompt:
How many times the brand was cited by LLMs
The brand’s recommendation (or mention) rate
Instances of ghost ranking
For this client, it was only citations on Bing’s AI search.
For each metric, he watches for trends over time, not drawing any conclusions based on one week of data.
He also audits content readiness with a mix of the following (again, depending on client needs):
If a page exists to address the prompt
If that page links directly to the specific products that answer the prompt
If product details (attributes) like price, dimensions, compatibility, etc., are listed on the page
If the content on the page is extractable by an LLM (e.g., no bot-blocking settings or JavaScript-heavy rendering that prevent AI crawlers from accessing the page)
This tells him exactly what to fix so AI is more likely to recommend the brand.
This system surfaced ~52,000 monthly searches with zero AI coverage for one of Jonny’s clients. And a content strategy for the next quarter.
How to Read and Act on Prompt Data
Generative AI prompt tracking data might seem hard to trust at face value.
LLMs give variable answers even in temporary chats.
Platforms push silent model updates that tweak source weighting.
And training data bias is real. One study found LLMs often favor global brands over local ones, meaning you might be invisible in your category because of the model’s defaults rather than your content.
Margaret at Hootsuite found the changing outputs of LLMs surprising when she first started prompt tracking:
It was way more chaotic than I expected. The same prompt can return a totally different brand list on ChatGPT vs Gemini vs Perplexity. “AI visibility” isn’t just a single thing to optimize for: you’re effectively running parallel strategies, and they don’t transfer cleanly, and answers will fluctuate constantly.
This variance doesn’t mean prompt tracking data is useless.
But it is the reason you need to track trends over time instead of getting hung up on moment-specific snapshots.
Here are a few meaningful signals to watch for.
Increased (or Decreased) Frequency Over Time
A consistent rise in mentions or citations usually means your content efforts are working.
A consistent drop could mean a competitor is gaining ground or a key piece of content has gone stale.
Action to take: If the trend is up, note what you published in the weeks before the lift. That’s your playbook. Keep doing it.
If it’s down for four or more consecutive weeks, pick one fix for that category:
Refresh a stale piece with updated data
Publish a new comparison page targeting the prompts you’re losing
Pitch a third-party source that’s being cited in that space
Consistent Source Inclusion
If you repeatedly see the same third-party sources cited in your LLM visibility tracking data, stop and take note.
It means the LLMs trust these sources.
And third-party sources are powerful for AI visibility. Airops found 85% of brand mentions come from third-party pages.
For example, if G2, TechRadar, and Capterra keep appearing across your evaluation prompts, those are your priority pitches.
You can see your top “Cited Sources” and “Source Opportunities” (where you’re not being mentioned but your competitors are).
Action to take: Make a list of the sources that keep showing up. For each one, check if you’re already featured. If so, is your listing current and accurate?
Then, pick one you’re missing from and pitch it. A contributed piece, product review campaign, or quote request can all work.
Unless you’re correcting a mention, don’t try to pitch competitors’ sites. They’re not likely to accept.
And don’t only focus on categories you’re losing. Reinforcing visibility in categories you’re already winning is valuable too.
Movement from Mention to Citation
If you used to get mentioned and now only get cited as a source, you’ve slipped into what Jonny calls “ghost ranking” territory:
“Your content shows up in the citation panel, but the AI recommends a competitor.”
Like this example from Teva.
An increase in your “ghost ranking count” over weeks means it’s time to investigate.
Jonny knows this first hand:
Our prompt tracking showed our agency was being cited on agency-directory pages (Trustpilot, Clutch, Semrush) but ghost-ranked at an 83% rate. AI was using directory content as the source of truth, then recommending whichever agency had the densest, most-named presence.
Action to take: Find the sources showing up in the citation panel and audit your presence on each one. Fuller profiles, more reviews, and accurate product details all help convert a citation into a recommendation.
For Teva, this means pitching to be included in the cited articles by REI, backpacker.com, and Outdoor Gear Lab.
It also means updating their own pages (including the ones being cited) with more or newer details.
Then, watching to see if they get less ghost rankings over time.
Common Misreads to Watch for
When you first start generative AI prompt tracking, try to avoid getting tripped up by the following false conclusions.
“Our AI visibility dropped this week. Something’s wrong.”
One week of low visibility or mentions is likely natural variability.
Wait for four consecutive weeks before treating it as a trend worth acting on.
“Our score looks low. We need more content.”
A score can be low for multiple reasons, and the fix isn’t always new content.
Sometimes it’s getting included in more third-party sources or forum threads. Or getting more customer reviews.
If you’re just starting out, you may need to go back to your prompts and make sure they’re not too vague or top-of-funnel.
For example, a broad query like “mesh wifi” may return a definition answer rather than a list of brands.
“Our overall visibility improved after adding more prompts to our tracker. We’re doing something right.”
Adding more prompts to your prompt set will usually make it look like your AI visibility has increased. You’re getting mentioned in more prompts.
“Our overall visibility improved after adding more branded queries. We’re doing something right.”
A branded query is one that mentions your brand name. Of course you get mentioned in the answer.
That doesn’t tell you anything useful.
Limit tracking branded prompts to comparison and reputation prompts. And keep them in a separate cluster so you can filter them when measuring your overall AI visibility.
Weekly Prompt Tracking Workflow
Turning prompt tracking data into a strategy is where you prove the value.
Margaret starts by looking at topics where Hootsuite has low visibility.
Then I look at individual prompt answers to see “What does the internet think about us in this category, and how do we change that?” This usually translates into a few concrete questions for further research:
Where are the third-party listicles, comparison posts, and analyst write-ups that the model is pulling from, and are we represented accurately on them?
Do we have a first-party comparison or evaluation page that an LLM can confidently cite, written for the actual buyer in that vertical?
Do we have our customers (ex. case studies, reviews) reinforcing the same message?
From here, she can recommend actions like updating a case study or getting a mention in a third-party listicle.
Here’s a simple workflow you can use to turn your prompt monitoring routine into AI visibility gains over time.
Pro tip: Don’t expect overnight wins. LLMs take time to reflect new content. Watch for directional improvement over weeks and months. You’re not chasing a score; you’re watching whether your gaps are closing over time.
Step 1. Review prompt cluster trends weekly: What’s the visibility score for your BOFU prompts? Has it fallen for your healthcare cluster? Or a specific use case category?
Step 2. Spot recurring gaps: Note any clusters, categories, or topics that have been underperforming for four weeks or more.
Step 2a. Plan one fix per cluster: Use your content strategy brain to determine the most impactful fix.
That might be:
Publishing a new comparison page to improve a BOFU prompt cluster’s score
Updating an existing article with fresher data
Publishing a type of content you haven’t tried yet on this topic (e.g., video, podcast, social post)
Step 3. Identify recurring third-party sources: Note sources AI consistently cites across your categories. Reddit? LinkedIn? G2? YouTube creators? A trade publication?
Step 3a. Pitch one source you’re missing from: If accepted, you’ll build more off-site authority and increase your chances of being mentioned in the answers you care about.
Bonus resource: Pitching a journalist or news outlet? Use our Journalist Pitch Template, designed by PR experts, to get started quickly.
Start Winning AI Visibility with Prompt Tracking
Prompt tracking isn’t a scoreboard. It’s a compass.
The brands that get real value out of prompt tracking aren’t monitoring every possible prompt.
And they aren’t reacting to one bad week.
They focus on bottom-of-funnel prompts and follow the direction of the graph, not the dot.
http://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.png00Dubado Solutionshttp://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.pngDubado Solutions2026-08-13 21:23:042026-08-13 21:23:04Prompt Tracking: How to Find (and Fix) Your AI Visibility Gaps
The DELEGATE-52 study from Microsoft Research tested 19 LLMs on document editing tasks across 52 professional domains over 20 editing interactions.
Even top frontier LLMs at the time, including Gemini 3.1 Pro, Claude 4.6 Opus, and GPT 5.4, corrupted an average of 25 percent of document content by the end of long editing workflows.
Average degradation reached 50 percent across all 19 LLMs tested.
Errors are sparse but severe: a small number of consequential changes that read as grammatically correct rather than many small typos.
Giving LLMs a basic agentic harness with file tools made performance slightly worse (roughly 6 percent more degradation) while consuming two to five times more input tokens.
Python was the only domain where most LLMs cleared the study’s 98 percent accuracy threshold. Even the best-performing model reached that bar in only 11 of the 52 domains tested.
When large language models (LLMs) edit documents, they make a specific kind of mistake that can be more dangerous than a hallucination. It’s subtle enough to pass a casual review, damaging enough to matter, and systematic enough to compound across multiple editing sessions.
A Microsoft Research study published on April 17, 2026, puts hard numbers on this. The findings should change how every content team thinks about where AI belongs in the editing workflow.
What the Research Actually Found
Microsoft researchers built DELEGATE-52 to mimic how people use LLMs for document work. It didn’t focus on one-off edits, but long, multi-session workflows where an LLM handles a running sequence of revisions and refinements.
The team gave 19 LLMs professional documents spanning 52 domains — including coding, crystallography, music notation, accounting records, and recipes — then asked them to complete 20 editing interactions. Those domains cover both highly structured formats (code, database schemas) and natural-language writing (fiction, email), and the corruption showed up in both, which is what makes the pattern relevant to the prose-heavy documents content teams produce.
Frontier LLMs, the ones considered most capable, corrupted an average of 25 percent of document content by interaction 20. Non-frontier models performed worse, dragging the average for all 19 models to 50 percent. Python was the only domain where most models cleared the study’s 98 percent accuracy threshold. Even the best-performing model, Gemini 3.1 Pro, hit that bar in just 11 of the 52 domains tested.
The specific error pattern is what makes this finding operationally important. The study calls the errors “sparse but severe”: the LLMs made a small number of high-impact mistakes rather than lots of little ones. In the kinds of documents content teams work with, those are the errors editors already worry about most: a statistic shifted by a digit, a clause dropped mid-sentence, or a name or attribution subtly altered. These errors read as grammatically correct, so a standard proofreading pass might miss them. Catching them takes a reviewer who knows what the original said.
The agentic finding is equally significant. Wrapping the LLMs in a basic agentic harness with file tools (the kind of setup that’s supposed to make LLMs more capable) made performance roughly 6 percent worse on DELEGATE-52 while consuming two to five times more input tokens. The “agentic version will handle this” response to the findings does not hold up against the data.
Why This Matters More for Long-Form Content
The error pattern described in DELEGATE-52 is most dangerous in the content types where a misattributed figure or altered claim does real reputational damage. Think white papers, pillar pages, executive thought leadership, client case studies, research reports, and legal or compliance documentation.
These are precisely the formats where teams are most tempted to hand an LLM an entire document and ask it to “clean this up” or “polish this section.” The open-ended, multi-turn editing request is exactly the scenario DELEGATE-52 tested, and it’s exactly where these tools fail in ways that look fine on the surface.
For short, tightly scoped edits, the risk is much lower. The corruption is cumulative rather than uniform. It builds up interaction by interaction, and compounds with document length. After 20 interactions, 1,000-token documents held at roughly 91 percent accuracy, while 10,000-token documents dropped to about 60 percent.
A surgical edit to a specific paragraph, a defined claim, or a single section produces dramatically fewer errors than an open-ended “improve the whole document” instruction. The scope of the request and the size of the document directly determine the level of risk.
Three Workflow Changes That Reduce the Risk
The research points toward three concrete shifts in how you should use LLMs in content production workflows.
Use LLMs for surgical edits, not open-ended passes. LLM editing can be great for a specific paragraph, a defined claim, or a single section. Scoped requests are far safer than sweeping ones. The more latitude a model has to interpret what needs to change, the more opportunity it has to introduce subtle errors.
Weight human review toward the back half of the workflow. Current practice in most content teams treats the first draft as the high-scrutiny moment and later editing interactions as lower-stakes. The DELEGATE-52 findings reverse that logic. Errors compound silently from one turn to the next, so rounds two, three, and four carry more accumulated risk than round one. When researchers extended the test to 100 interactions, the degradation kept climbing, with no point at which the models stabilized. Review intensity should ramp up as a document accumulates LLM interactions, not wind down.
Add targeted QA checkpoints for the error types LLMs introduce. Standard proofreading catches typos, grammatical errors, and obvious factual claims. It may not catch a shifted number that reads correctly, a dropped clause that changes meaning without breaking grammar, or an attribution that’s been quietly changed. Any QA process for LLM-assisted content should hunt specifically in the danger zones: numbers, named attributions, data points, and quoted material.
Where the Stakes Are Highest
In low-stakes content, this failure mode is survivable. A shifted phrase in a social post or a minor structural change in a blog draft is an inconvenience. In specific content categories, though, the same error pattern carries significantly higher consequences.
Legal and compliance documentation is the clearest example. A dropped clause in a contract summary or an altered definition in a terms-of-service summary can create material legal exposure. Standard proofreading may not catch these errors, because they read as correct prose and slot neatly into the surrounding context.
Client-facing research and attribution is another high-risk category. White papers, case studies, and thought leadership pieces that attribute specific statistics or quotes to clients or data sources carry reputational risk when even one of those attributions is off. A client who sees their name attached to a data point they did not provide, or a study whose findings have been slightly modified, faces a trust breakdown that is difficult to reverse.
Executive and spokesperson content carries the same risk at a different level. LLM editing of speeches, op-eds, or public statements, iterated over multiple review rounds, can drift meaningfully from the executive’s original intent through a series of small changes that each seem harmless. That cumulative drift, measured over 10 to 20 editing interactions, is exactly what DELEGATE-52 quantified.
For all these content types, the practical rule from the research is that the longer an LLM works on a document, the more scrutiny the final version requires.
What This Does Not Mean
The research is not an argument for eliminating LLMs from content workflows. They deliver genuine value in research, drafting, structural suggestions, and early draft generation. AI adds value in content workflows where human judgment needs to stay in control, particularly when the raw material for the work comes from a human with real expertise and subject matter knowledge.
The finding is specifically about delegated editing, which means handing a model a document and asking it to handle the revision process autonomously across multiple sessions. That specific use case is where the degradation pattern emerges. Keeping a human with genuine editing judgment in control of every revision decision, with LLMs as drafting and suggestion tools rather than autonomous editors, avoids the problem the research identifies.
Remember that mistakes are not always visible in output. LLM-corrupted content looks fine. It passes grammar checks. It reads fluently. The damage only surfaces when someone who knows the original compares it directly against what the model produced.
FAQs
Does this apply to all AI models or just older ones?
The study tested the most capable frontier LLMs available at the time, including Gemini 3.1 Pro, Claude 4.6 Opus, and GPT 5.4. All of them showed the 25 percent degradation pattern. This is not a problem that disappears with more capable models based on current evidence.
What kinds of errors does AI introduce most often?
The study characterizes LLM errors as sparse but severe: a small number of consequential changes rather than several small ones. In practice for content work, this shows up as shifted numbers, dropped clauses, or subtly altered attributions. These are meaningful changes that can read as grammatically correct, which is what makes them difficult to catch in standard review.
Does giving AI access to tools (agentic use) improve accuracy?
No. When LLMs were wrapped in a basic agentic harness with file tools, performance was roughly 6 percent worse than the non-agentic baseline, and the models used two to five times more input tokens. The “agentic upgrade will fix it” response to this research is not supported by the data.
Is there any domain where AI editing is reliable?
Python was the only domain where most LLMs cleared the 98 percent accuracy threshold, and even the best-performing model reached that bar in only 11 of 52 domains. Natural-language tasks across professional domains showed consistent degradation.
How should I change my content workflow based on this?
Use LLMs for scoped, specific edits, such as a defined paragraph, a single claim, or a targeted section. Increase human review intensity at the back end of the workflow, since errors compound across turns. Add QA checkpoints that specifically hunt for the error types LLMs introduce, like shifted numbers, altered attributions, or dropped clauses.
Conclusion
The DELEGATE-52 findings confirm what experienced content editors have observed informally: LLM editing in extended workflows introduces errors that standard review processes are not designed to catch. The research makes the scale of that risk quantifiable.
An LLM should never be the final authority on a document. The risk is too high, and the errors are too subtle. There are real consequences for content that carries reputational weight. The right role for LLMs in content production is as a capable assistant with a human editor maintaining control of every consequential revision decision.
http://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.png00Dubado Solutionshttp://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.pngDubado Solutions2026-08-12 19:00:002026-08-12 19:00:00Large Language Model (LLM) Editing Quietly Corrupts Documents. Here’s What the Research Says
Audience-first SEO builds your keyword and content strategy around a specific, high-value audience instead of a broad topic or category.
The tactics stay familiar (keyword research, content gap analysis, TAM (total addressable market) analysis), but the organizing principle behind all three changes.
Programs using this approach see growth concentrated among the audiences that actually drive revenue, not just a bump in overall traffic.
Applying it starts with defining your priority audience before you ever look at keyword volume.
Integrated agencies have a built-in advantage here, since paid audience research already exists internally and can inform SEO strategy directly.
Audience-first SEO also sets up your Digital PR targeting, since the same research tells you where that audience already spends its time.
Paid media has always started with the audience. Every campaign begins with who you’re trying to reach before it gets to what you’re going to say. Paid teams buy audiences, not channels.
SEO has worked the opposite way for years. Programs get organized around topics and keyword clusters first, then hope the right people show up once the content ranks. Audience-first SEO closes that gap by building your organic program the same way paid media already builds its targeting.
That gap matters more now than it used to. Search is fragmented across traditional results, AI-generated answers, and social, and attention is split thin across all of it. The brands winning right now are the ones cutting through to the specific audiences that move their business.
This isn’t a new set of tactics so much as a shift in starting point, one the strongest SEO programs are already using, often without a name attached to it. This piece names that instinct and lays out how to apply it on purpose.
What Audience-First SEO Actually Means
Traditional SEO is organized around topics and categories. A single topic can serve audiences with very different needs, intent levels, and business value, and that ambiguity is where a lot of traffic volume hides the fact that the wrong people are showing up.
Audience-first SEO flips that starting point. You define the specific audience you want to reach, usually the audience most valuable to the business rather than the largest one, and you run keyword research, content gap analysis, and TAM analysis through that lens.
The tactics themselves are the same ones you already run. What changes is the organizing principle behind them. Instead of asking which topics carry the most search volume, you ask which audiences matter most to the business, then find the keywords and prompts tied to them.
Paid media has operated this way for years. Every campaign starts with defining who you’re trying to reach before any decision gets made about creative or messaging. Audience-first SEO borrows that discipline and applies it to organic, with the same rigor paid teams bring to targeting, segmentation, and qualification.
The distinction between audience and topic shows up quickly once you look for it. A topic like wealth management could serve a first-generation saver building an emergency fund, a business owner preparing for an exit, or an ultra-affluent family managing a multi-generational estate. Content written to serve all of them tends to serve none of them particularly well, because their questions and buying triggers don’t overlap much.
This is where user personas for SEO earn their keep. Once you know which audience segment you’re writing for, keyword targets get sharper and content gets easier to write, because you’re solving one person’s problem instead of hedging for several readers at once.
The same logic applies to how you think about audience size. A larger audience isn’t automatically the right one to chase. Weighing audience size vs. revenue potential is often what separates a program that grows traffic from one that grows the business, and audience-first SEO forces that comparison earlier in the process instead of leaving it as an afterthought once the content is already live.
What It Looks Like When It’s Working
Plenty of programs succeeding with SEO right now are already running a version of audience-first SEO, even if nobody on the team has put a name to it. They’ve made deliberate calls about which audiences to prioritize, run their keyword and content work through that lens, and the results compound in ways topic-first strategies rarely do.
In one case, NP Digital was working with a financial services brand competing in a crowded search category with significant total addressable volume. Rather than chasing the broadest possible traffic, the SEO team built TAM analysis and content gap work around specific high-value audience segments, prioritizing the individuals and families most valuable to the business over the segment with the largest raw search volume. That decision started with the kind of audience segmentation work most brands skip, ranking each segment by demand, competitive difficulty, and long-term business value before a single piece of content got written.
Within six months, organic leads from the client’s highest-value segment, ultra-affluent individuals and families, had already surpassed the client’s full prior fiscal year total, hitting nearly 114% of the previous year’s volume before the year was even half over. Aflluent leads followed close behind at close to 90% of the prior year’s total, and the estimated pipeline impact tied to that lead volume grew from $1.46 billion for the full prior year to $2.13 billion year-to-date.
Keyword rankings still played a role here, but only not focusing on casting the widest net possible. Total organic lead volume across the whole program sat at roughly 71% of the prior year’s pace over the same six-month window, well behind the pace of the priority segments. The gap between those two numbers is the point: audience-first SEO deliberately trades some volume in lower-value segments for outsized gains in the ones that matter most to the business.
The insight from that program: when you build SEO strategy around a specific audience instead of a broad category, you make sharper calls about which keywords to pursue and which content gaps to close first. Raw traffic volume becomes a far less useful signal. Audience quality becomes the measure that matters.
How to Apply Audience-First SEO in Practice
Getting audience-first SEO right does not require a new toolset. It requires changing the order of operations, and that shift plays out in three practical moves.
Start with audience definition, not keyword volume. Before opening a keyword tool, decide which audience the business needs most. This usually means ranking segments by value to the business rather than by size, since the segment with the most search volume is rarely the one with the most revenue potential. A brand serving both mass-market and high-net-worth customers, for example, might find that general wealth management searches carry the most volume, while a much smaller audience of high-net-worth individuals and families carries far more business value per lead.
Run TAM analysis by audience segment. Rather than sizing the total addressable market around a single blended topic, break it out by persona. Score each segment on search demand, ranking difficulty, and business value, then map where those three overlap. That mapping tends to surface a small number of segments where the opportunity is real: enough search volume to matter, low enough difficulty to be attainable, and high enough value to justify the investment. It’s also important to do deeper research here on these audiences you find most important. This can range from using specialized tools to looking at information sources like Reddit or customer interviews.
Build content gap analysis around audience needs, not topic coverage. Once you know your priority segments, look at where your content already ranks for those specific audiences and where it doesn’t, rather than running a generic gap analysis against a topic list. A gap analysis run this way often turns up a surprising amount of untapped opportunity sitting several pages deep in the rankings, simply because it was never built with a specific audience in mind.
Put together, these three moves turn audience-first SEO from a concept into a repeatable process your team can run quarter over quarter, and each one gets easier once the audience definition from the first step is locked in.
Paid media teams have always treated audience research as a starting point, not an afterthought. Tools like Reddit Ads and GWI (Global Web Index) give paid strategists a detailed picture of who an audience is, what they care about, where they consume content, and how they talk about their own needs. Google Ads’ Audience Insights and Insights Finder tools add another layer of that same intelligence, surfacing which segments actually convert and what related interests and trends they share. That research is directly applicable to SEO strategy, and brands running SEO and paid in separate silos are rarely putting it to use that way.
In an integrated agency, that audience intelligence flows into SEO from day one. Keyword strategy gets informed by what paid research already knows about the audience, and content positioning reflects how that audience actually talks about their problems rather than just how they type their searches.
Tools like Ubersuggest and AnswerThePublic add a further layer, surfacing the specific questions and language patterns an audience uses at different points in their journey. Combined with paid audience research, they help build a content strategy that is both search-optimized and genuinely useful to the people you’re trying to reach.
The audience opportunity matrix is one of the clearest outputs of this integrated approach. Mapping audience segments by search demand, keyword difficulty, and business value shows exactly where the strongest opportunities sit, and just as importantly, where they don’t.
That same audience-first lens applies to sizing the opportunity. A TAM analysis broken out by segment, rather than blended across an entire category, shows which audiences carry the most untapped search volume relative to their ranking difficulty and business value, giving the team a clear place to focus first.
Audience-First SEO and Digital PR: Publishing Where Your Audience Already Goes
Traditional link building is organized around domain authority and topical relevance. Audience-first Digital PR is organized around presence: where does this specific audience actually go for information, and are you showing up there? Those questions are related, but they lead to different outreach lists.
When audience research is done well, you already have that answer. Paid audience research tools, combined with SEO audience data, surface the publications, communities, and platforms where your target audience spends its time. That becomes the actual targeting brief for Digital PR outreach, rather than a generic list of high-authority sites in your industry vertical.
AI insight tools add one more layer here, showing which content is already getting cited for prompts relevant to your audience, so you know where AI-driven visibility is concentrated too.
This connection also strengthens the SEO program directly. Earned coverage on the sites your audience trusts builds authority signals that are harder to manufacture and more durable than links acquired through broad outreach. It also creates touchpoints with your audience outside of search, which builds brand recognition that shapes how they engage once they do find you organically. A reader who saw your brand mentioned on a site they already trust arrives at your organic content with a head start on credibility that a first-time visitor doesn’t have.
The integrated play here: audience research informs SEO content strategy and Digital PR targeting at the same time. One brief drives two channel strategies, both built around the same specific people, which also means the two teams stop duplicating research that already exists somewhere else in the building.
The Complement to Search Everywhere Optimization
Search Everywhere Optimization is the right ambition for brands with the resources to pursue visibility across every surface their audience uses: traditional search, AI-generated results, social, video, and beyond. Audience-first SEO answers the question that has to come first: which of those surfaces matter most for the specific audience you’re trying to reach, and what do they need to see from you there?
For brands that can’t optimize everywhere at once, audience-first SEO becomes the prioritization framework. Instead of spreading effort across broad topic coverage, you concentrate on the overlap between audience need, search opportunity, and business value. That’s a more defensible use of SEO budget and a much clearer brief for the team running the program.
The two approaches work together rather than against each other. Audience-first SEO sharpens the who and the where. Search Everywhere Optimization expands the how. A brand with clarity on its priority audiences is better positioned to pursue visibility across surfaces, because it already knows what those audiences need and where they’re looking for it. It’s also worth noting that Google Search Console now surfaces how some of your social content performs on Google, giving audience-first teams one more data point for connecting the dots across surfaces.
For a marketing leader deciding where to invest next quarter, that ordering matters. Trying to run Search Everywhere Optimization without a clear audience definition underneath it usually means spreading budget across surfaces evenly instead of weighting it toward the ones your priority audience actually uses, which is a slower and more expensive way to get to the same result.
FAQs
How do I identify my target audience for SEO?
Rank your existing customer segments by business value, not by search volume, then check which of those segments has real search demand behind it. From there, build out SEO audience personas for the segments that clear both bars.
How does SEO engage an audience?
SEO engages an audience when the keywords, content, and site experience all reflect what a specific group of people needs, rather than serving a broad topic to whoever happens to land on the page.
How do I target an audience in SEO?
Run your keyword research, content gap analysis, and TAM analysis through an audience lens instead of a topic lens, and let that definition shape which keywords and content gaps you prioritize.
Conclusion
The best SEO programs are already doing some version of audience-first SEO. The question is whether your team is doing it on purpose or backing into it by accident.
Naming the approach matters because it makes the process repeatable. Define your priority audience, run your TAM and content gap analysis through that lens, and let audience quality, not raw traffic, tell you whether the program is working. Start with finding your target audience if that first step still feels undefined, and build out from there.
That’s the kind of program NP Digital builds and runs for clients every day, connecting audience research from paid media into SEO strategy so organic traffic starts working harder for the people who actually matter to your business.
http://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.png00Dubado Solutionshttp://dubadosolutions.com/wp-content/uploads/2017/05/dubado-logo-1.pngDubado Solutions2026-08-12 16:00:002026-08-12 16:00:00Audience-First SEO: How to Rank by Putting Readers First
