Skip to content
Snapshot, 17 Oct 2026
SnapshotStudy 011, Cited, not chosen · frozen 17 Oct 2026 · not updated

Comparison pages

pointsin how often a brand is recommended, once AI starts citing a page of its own. An exploratory estimate, not proof.

Flat before, a step up the week the page is first cited

Change in recommendation rate against the same brands’ never-cited prompts, by week.

Weeks before first citation stay near zero; the week AI first cites the page the rate is +3.6 points, with a likely range after of +1.5 to +4.3.
Change in recommendation rate by week, around the week AI first cites the page
Weeks from first citationPoints
-8−0.7
-7−0.5
-6−0.5
-5+0.1
-4−0.6
-3−0.3
-2−0.5
-10.0
0+3.6
1+2.7
2+2.4
3+3.0
4+3.1
5+3.2
6+3.4
7+3.3
8+2.4

Nearly a third of the pages AI cites for a brand name a rival

Citation is evidence, recommendation is the verdict. Medians per brand, from the first run in June 2026.

Cited pages naming a rivalMedian per brandUnique pages
Citations naming a rivalMedian per brandEvery appearance
The brand’s own pagesOf what AI cites about itMedian per brand
Pages first citedMay to Aug 2026For 180 brands

Bigger where AI searches the web. Nothing when it answers from memory.

Each estimate with its likely range, in points of recommendation rate.

EstimatePointsLikely range
All brands+3.0+1.7 to +4.3
Event-study version+2.9+1.5 to +4.3
Week of first citation+3.6+2.3 to +4.8
Challenger brands+4.2+1.8 to +6.8
Answers built on web search+6.0+3.8 to +8.5
Answers from memory0.0−0.9 to +0.8
All brands: +3.0 points, likely +1.7 to +4.3. Event-study version: +2.9 points, likely +1.5 to +4.3. Week of first citation: +3.6 points, likely +2.3 to +4.8. Challenger brands: +4.2 points, likely +1.8 to +6.8. Answers built on web search: +6.0 points, likely +3.8 to +8.5. Answers from memory: 0.0 points, likely −0.9 to +0.8

Common questions

Is being cited by AI the same as being recommended?

No. A model can quote your page as a source and still recommend a competitor as its top pick. In this study, owned pages are a median 4.6% of what AI cites about a brand, and nearly a third of cited pages name a rival. Citation is evidence; recommendation is the verdict.

Why does AI cite my page but recommend a competitor?

Comparison and "best of" pages are the pages AI reaches for, and those pages name rivals by design. So the page you wrote to win becomes the evidence the model uses to list everyone else. The less visible a brand already is, the more its citations leak into a competitor recommendation.

Do comparison pages actually help you get recommended?

The October 2026 rerun found a 3.0-point exploratory estimate around the first observed citation of an owned page, and 3.6 points in the citation week itself. It is not treated as causal: publication timing was unavailable, and in the first run a publish-anchored estimate crossed zero and placebo estimates were positive.

What did the engine split show?

The exploratory estimate was larger for answers built on web search (6.0 points) than for answers from memory (0.0). That pattern is suggestive, but it does not establish that the page caused the recommendation change.

Quasi-experimental, not a randomized trial. The effect is dated to first citation, not publication. Controls are the same brands’ prompts whose own pages were never cited as of 16 Oct 2026. Brand-level rows are withheld.

Snapshot, 17 Oct 20261,050,171 answer reports, 180 brandsAnswers 2 Mar to 11 Oct 2026Exploratory estimates, not proofTrakkr Data · CC BY 4.0