Of 1,094 topic categories analyzed in ChatGPT between January and June 2026, only 15.2% had a clear brand owner. Semrush defined "clear owner" strictly: a brand had to appear in at least four of five related buyer-intent prompts and lead the runner-up by 5 percentage points or more. The rest? 31.2% had an emerging leader. 53.7% were completely unsettled, with no brand showing up in even three of five prompts.
If you're running demand gen for a B2B SaaS company, those numbers should change how you think about AI-assisted discovery. And they should make you skeptical of anyone selling "SEO for ChatGPT" as the whole answer.
SEO Signals Correlate With Ownership About Half the Time
Semrush compared category owners to runners-up on three traditional SEO metrics: branded search volume, organic traffic, and Authority Score. Owners had higher branded search volume 55.7% of the time, higher organic traffic 48.4%, and a higher Authority Score 52.5%. Only branded search volume reached statistical significance.
That's roughly a coin toss for two of three metrics. Kevin Indig, founder of Growth Memo and a collaborator on the study, put it plainly: "Traditional SEO metrics aren't enough to explain who owns a topic. While they play their role, there's more to it."
So what is "more to it"? The data points toward something less comfortable for teams that have invested heavily in on-site content: third-party evidence density.
ChatGPT Trusts Third Parties More Than Your Website
Pepper's analysis of B2B software recommendations found that about 88% of sources ChatGPT cites come from outside the vendor's own site. Review aggregators, comparison pages, community discussions. MarketScale's benchmark tells a similar story: 71% of B2B vendor citations trace to earned media versus 29% to owned content.
This reframes the competitive problem. You can have a technically excellent site, strong domain authority, and high organic traffic — and still not show up consistently when a buyer asks ChatGPT "what's the best tool for X." Because ChatGPT isn't crawling your blog the way Google indexes it. The model is synthesizing evidence from wherever credible third-party proof exists: G2 reviews, Capterra comparisons, Reddit threads, editorial coverage.
MaxAEO's research argues that what looks like big-brand bias in ChatGPT is better explained by evidence density and perceived answer risk. Brands with more public proof (reviews, citations, comparisons) get surfaced more often. That's not favoritism; it's risk reduction. The model defaults to answers it can back up.
The practical implication for smaller brands: you can compete, but not by optimizing your own site harder. You compete by building credible third-party proof across the places ChatGPT actually pulls from.
High-Demand Topics Are Harder to Own
Here's where sequencing matters. Semrush split categories by AI search demand. In the top half (which accounted for 98% of AI search volume in the sample), only 11.3% had a clear owner. In the lower-demand half, 19% did.
The categories with the most buyer attention are the least settled. That's both opportunity and warning. Opportunity because most categories genuinely don't have an incumbent. Warning because high-demand topics attract more competition and require more sustained evidence to hold.
One encouraging signal: once a brand achieves clear ownership, it tends to keep it. Clear owners retained first place in 90.4% of month-over-month comparisons. Brands with narrow leads (averaging 1.3 percentage points) were far more likely to lose position than those with wider leads (2.9 points). Ownership compounds, but only if the margin is wide enough to absorb noise.
What This Changes for Your Measurement Stack
If you're tracking AI visibility the same way you track organic search, you're measuring the wrong thing. Winning one prompt isn't ownership. Semrush found that only 21% of the most-cited domains in a category were also the most-mentioned brand in ChatGPT's answer text. Citations and mentions didn't align.
Topic ownership means consistent surfacing across a cluster of related buyer questions: definitions, comparisons, alternatives, use cases, buying decisions. That's five angles on the same category, not one keyword ranking.
The measurement shift: track prompt clusters, not individual queries. Track brand mentions in answer text, not just citation links. And track third-party coverage density across review sites, comparison pages, and community threads, because that's where the model is drawing confidence.
The Trade-Off You're Accepting
Investing in third-party evidence density is slower and less controllable than publishing blog posts on your own domain. You can't A/B test a G2 review the way you test a landing page. You can't control what Reddit says about your product. The feedback loops are longer and noisier.
But the 15.2% stat tells you something uncomfortable: the controllable stuff (your site, your content, your SEO) only gets you partway there. The other half of the equation lives in places you influence but don't own.
53.7% of ChatGPT topic categories are unsettled. No dominant brand. For demand gen teams willing to build evidence where it actually counts, that's not a gap in the data. That's the opening.
