AI Overviews cited smaller domains more often in paired buyer queries
A recount of 2,309 AI Overview source rows checks how citations changed when B2B software questions included buyer context.
Google AI Overviews cited domains with fewer than 1,000 referring domains more often when B2B software questions included buyer context, according to GrowthSpree’s November study. Its researchers ran 200 short questions and 200 longer versions in the US market from November 1 to 11. The longer versions added details such as team size, budget, or an existing tool. GrowthSpree’s study and source files
The result holds when each topic gets equal weight
SearchEngineAnswer recounted the study’s public AI Overview CSV using the author’s classification for domains below the threshold. Those domains accounted for 434 of 1,137 sources on short questions (38.2%) and 677 of 1,172 on longer questions (57.8%). The pooled difference is 19.6 percentage points.
Some answers listed more sources than others, so we also gave each complete topic pair equal weight. We calculated the share within each answer, then averaged across the 194 topics with source rows for both versions. The result was 36.6% for short questions and 56.4% for longer ones, a 19.8-point gap. This is a reanalysis of the published data. We did not collect new search results.
The threshold describes the study’s domain-level link measure. It cannot tell you how many sites link to an individual cited page. The public domain table also flagged some sources as absent from the link index. The recount preserves the study’s original classifications.
Test the buyer detail your page actually answers
Start with a short comparison query for one category. Add a single constraint that your documentation addresses, such as compatibility with a named integration. Run both versions in the same market, save the source URLs, and check whether the cited pages answer the constraint. Add a second version with a different constraint if you want to see which pages recur.
GrowthSpree changed both question length and buyer context. The study cannot isolate the effect of extra words or establish that adding backlinks changes citation odds. Its results support testing specific buyer situations and recording which source pages surface. Related analysis of brand popularity and AI citations
Keep learning
Continue this topic
Audit citations
Locate the failing answer layer
Compare evidence
Separate mentions, citations and visits
Measure referrals
Check AI search visits in GA4
Research
Ask a question or join the discussion