New study puts a number on the visibility gap for small brands

What happened?

A preprint posted to arXiv in June 2026 analyzed more than 100,000 AI search responses covering 102 brands across ChatGPT, Gemini, Perplexity, Claude, and Grok, gathered between March and May 2026. The author is the founder of Ranqo, an AI visibility tracking company, and the data comes from that company platform. Sorting brands by public stature, the study found first-run visibility on category questions that did not name the brand fell from about 73 percent for household names, to 44 percent for mid-market brands, to about 11 percent for niche and small ones.

Why does it matter?

Eleven percent is the number to sit with. If you handle communications for a small organization, the study puts a figure on something you probably suspected: engines answer category questions with the names they already know. It also found that only about 3 percent of citations pointed at the tracked brand own site, while roughly 75 percent pointed at pages owned by other companies, and that ranked best-of listicles were the single most cited content format at about 36 percent of cited content pages. Where you get written about appears to matter as much as what you publish.

Should I do anything?

Read the source composition finding first, since it is the one most likely to change how you spend your time. Then pick one category question a person might realistically ask an AI engine to find an organization like yours, ask it, and write down which sources the answer cites. If those are roundups, directories, and third-party pages rather than organization websites, that tells you where your name needs to appear next. Treat this as one data point from one vendor dataset, not a reason to redo your site.

Learn more

• The paper: Generative Engine Optimization at Scale (arXiv)
• Our guide: understanding AI mentions
• Our guide: conduct an AI visibility audit