Not sure where to start? Take a guided tour.
AI Site Grader
AI Query Scanner
llms.txt Generator
In short
Four findings, strongest first- +3.6pp
rise in recommendation rate in the week a brand’s own page is first cited. It holds at about +2.9 over the next eight weeks. Real, but modest.
Chapter 02 - +6.0pp
in answers built on web search. Answers from memory did not move (0.0).
Chapter 03 - 4.6%
median share of a brand’s AI citations that point at its own pages, in the first run (June 2026). The rest is everyone else.
Chapter 01 - 28.9%
median share of cited pages that name at least one competitor, in the first run. Your sources can sell someone else.
Chapter 04
AI can quote your page and still recommend your rival.
A citation means the AI read your page. A recommendation means it picked you. We separated the two across 1.05M AI answers, then asked the harder question: when you publish your own comparison page, does your chance of being recommended actually go up?
Short answer: yes, but more modestly than the hype suggests, and only where you would expect. A newly cited page lifts recommendations by about 3.6 points in the week it is first cited, and the lift holds near 3 points after. Challengers gain a little more, and the effect lives in answers built on web search. Below: the gap, the proof, and the catch most brands miss.
Cited a lot, chosen rarely.
Start with the problem a comparison page is meant to fix. Plenty of brands get quoted as a source and still miss the shortlist, challengers most of all. In the panel, each brand’s overall visibility sits next to the share of its cited prompts where it was a source but never made the top three picks.
Do comparison pages work?
So you publish the page. Does it actually move anything? We found 970 moments between May and August 2026 where a brand's own page was first cited, then compared each with the same brand's prompts that were never cited, in a staggered difference-in-differences design. The answer is yes, but modestly: the rate steps up in the citation week and holds.
Each point is the average treatment effect (ATT): how much a brand's weekly recommendation rate moved, compared with the same brands' prompts that were never cited, relative to the week its page was first cited (week 0). The flat pre-period is the parallel-trends check; the step at week 0 is the page doing the work. Shaded area marks the weeks after the page appears.
How we know it’s real.
A jump right after citation could just be coincidence, so the data carries its own control. Some answers are built on a live web search; others come from memory and cannot see a brand-new page. If the page is really doing the work, only the answers that can read it should move. That is exactly what happens.
Answers built on web search
+6.01ppAnswers from memory
0.00ppYour page sells your rivals.
Here is where brands hand the gains back. The same comparison page that helps you also names your competitors, and AI reads those names as a vetted list of who else to recommend.
In the first run (June 2026), a median 28.9% of a brand’s cited pages gave the AI at least one rival to suggest, and 38.0% of its citation appearances did.
What to do.
Four moves the data supports. None is a guarantee; each is grounded in the panel above.
Yes, build the comparison page
It works, just modestly, and most for challengers with room to climb. Expect to be cited a while before you are chosen. (+4.2pp challenger lift.)
Put it where AI can read it
The lift comes from answers built on a live web search. Answers from memory did not move. If your page is not easy to retrieve, it does nothing. (+6.0pp answers built on web search.)
Stop naming your rivals
Every competitor you list becomes the evidence AI uses to recommend them. Make the case for you, not a directory of everyone. (77% worst offender.)
Track chosen, not just cited
A citation is a foot in the door, not a win. Measure whether you are the pick, not just whether you were quoted. (~30% cited pages name a rival.)
