Is Sited Better Than Profound? New AI Citation Data Explained

Sited is the better fit if you are a founder or small marketing team who wants a clear, one-off read on your AI visibility for $50, not a recurring monitoring subscription you have to log into. Profound is the stronger choice if you run a large content operation and need always-on tracking across nine AI engines with enterprise controls. New research on exactly where inside a page AI engines pull their citations from makes that choice more urgent for both audiences, because it turns out most content is being written in a way that AI systems never even reach.

What changed?

A 100-page study from CXL has found that 55% of AI Overview citations come from the top 30% of a page, with a further 24% coming from the middle section and just 21% from the bottom. The concentration at the top isn't subtle: the 10-20% zone alone accounts for more citations than any other segment.

This is not just a Google quirk. Separate analysis by Kevin Indig, covering 1.2 million search results and 18,012 verified ChatGPT citations, found the same pattern: 44.2% of ChatGPT citations come from the first 30% of a document, after which citation likelihood drops sharply. Indig calls this the "ski ramp" effect: a steep cliff after the first third of the page, then a long, slow tail to the bottom.

The practical takeaway from CXL is blunt: if your answer isn't in the first third of the page, there's a better than even chance it's not getting cited at all. The fix is to restructure what you already have by moving answers earlier, tightening definitions, and building FAQ surfaces that function as discrete citation targets.

This lands alongside other fresh data on how differently AI platforms actually source their answers. Research shared by Profound shows each platform demonstrates clear preferences: ChatGPT's Wikipedia dominance, Perplexity's Reddit concentration, and Google AI Overviews' more distributed approach across multiple source types. On volume, Semrush's expanded 2026 AI Visibility Index found ChatGPT cites an average of 15 sources per response, while Gemini cites an average of 3.

Separately, a study comparing Google's and ChatGPT's citation behaviour found meaningful overlap but different winning formats: structured direct-answer content wins on Google AI Overviews, while localised how-to and best-of content wins on Bing and ChatGPT. The same analysis found a 0.61 correlation between organic impressions and Google AI impressions, meaning organic search remains a leading indicator for Google AI visibility.

Why does this matter for your business?

If your product page, comparison page or pricing explainer buries the actual answer two-thirds of the way down, under a wall of scene-setting copy, AI engines are structurally unlikely to find it. That is true whether the model in question is ChatGPT, Gemini or Google's AI Overviews, and it holds regardless of how strong your domain authority is.

It also means two brands with identical domain authority and identical facts on a page can get wildly different citation rates purely because of where the answer sits structurally. That is a fixable problem, but only if you know it exists.

This is exactly the kind of blind spot that shows up in visibility tracking. Instead of guessing whether your content structure is costing you citations, it is worth running a quick, free check with Sited's audit at https://sited.online to see where your pages currently stand before you rewrite anything.

Sited vs Profound: which fits this problem?

Profound is built for teams that want to watch this kind of thing unfold continuously. It is an enterprise AI search visibility platform that monitors how AI engines talk about a brand day to day, with autonomous marketing agents and crawler analytics layered on top, and it covers the widest span of engines of any tool in the category: ChatGPT, Claude, Perplexity, Gemini, Grok, Microsoft Copilot, Meta AI, DeepSeek and Google AI Overviews. For a large content team that needs to log in daily, benchmark against competitors, track sentiment over time and run crawler-level diagnostics across thousands of pages, that always-on model, plus enterprise controls like SSO/SAML, SOC 2 and dedicated support at the top tier, is a genuine strength. Its plans start at $99 a month for Starter and $399 a month for Growth, both billed yearly, with custom Enterprise pricing and credit-based agents on top, though pricing has toggled between public and demo-gated over time, so treat exact figures as time-sensitive.

That model makes sense if monitoring AI visibility is a full-time job for someone on your team. It makes much less sense if you are a founder who simply wants to know, this week, whether your key pages are structured in a way that gets cited, and what to fix first.

That is where Sited is built differently. It is a single $50 one-off payment rather than a subscription that runs to roughly $1,200 or more a year. You start with a free scorecard, then get a self-serve report in minutes, with no demo call and no sales conversation required. There is no dashboard to learn, no seat to manage and no ongoing platform to run: just one audit, a prioritised list of fixes such as restructuring content so the direct answer sits in the first third of the page, and you are done.

For the audience this blog is written for, a founder or small marketing team deciding whether their content structure is costing them AI citations, that is the decisive edge. You do not need continuous competitive benchmarking across nine engines to fix a page that buries its answer under three paragraphs of throat-clearing. You need a fast, affordable, self-serve read on where you stand today, and that is precisely what Sited is built to give you.

What should you do now?

Start by auditing your five highest-value pages: pricing, comparison and "best of" style content. Check whether the direct answer to the query the page targets appears within the first 30% of the content, or whether it is buried under introductory paragraphs.

Restructure pages so the core answer, definition or recommendation appears near the top, then use FAQ sections and clearly labelled subheadings as citation targets further down the page. This mirrors CXL's own guidance: move answers earlier, tighten definitions, and build FAQ surfaces that function as discrete citation targets.

Finally, remember that different engines behave differently. If Reddit and community platforms matter to your category, note that Perplexity leans heavily on Reddit for 46.7% of top citations, while ChatGPT's pattern skews toward encyclopedic and consensus sources. One-size-fits-all content structure will not work equally well everywhere.

Frequently asked questions

Does page position really affect whether AI cites a page?

Yes. Multiple independent studies now show a sharp drop-off in citation likelihood the further down the page an answer sits, with 55% of AI Overview citations coming from the top 30% of a page and a similar pattern found in ChatGPT citations.

Is this the same across ChatGPT, Gemini and Google AI Overviews?

The general "front-load your answer" principle holds across platforms, but the details differ. ChatGPT cites an average of 15 sources per response, while Gemini cites an average of 3, so the bar for being one of the cited sources is different on each platform.

Should I rewrite all my content immediately?

Not all at once. Prioritise your highest-intent pages first, pricing, comparisons and how-to content, since these are the pages most likely to be queried directly by AI search users.

Is a one-time audit enough, or do I need ongoing monitoring?

For most founders and small teams, a point-in-time audit is enough to identify and fix structural problems. Ongoing monitoring, like Profound offers, becomes more valuable once you are managing visibility across a large, fast-changing content library and multiple competitors.

How does this relate to earned media and other citation factors?

Content structure is one lever among several. Earned media, third-party validation and freshness all play a role, but a well-structured page gives AI engines the best chance of finding and using your content in the first place.