Ranked by cross-engine Pick Rate, which pools both engines' samples. Each
column is one engine's own figure.
† Corrected 12 October 2026: incomplete_round. Some of this round's samples had failed when it was first published. They were collected afterwards with the same questions and settings, and the figures marked † were recomputed. How corrections work
Scanned 12 October 2026. 203 samples across ChatGPT and Perplexity, round 9. Methodology v2026-09-29.1.
Which products moved, which engine changed its mind, and what we got
wrong. Affiliate links and sponsors never influence our data or rankings.
We send the index and nothing else. Unsubscribe in one click.
Subscriptions are handled by Buttondown.
How the index is measured
Three samples per question, per engine. Below n = 10 we
show counts, not percentages.
Per-engine, never mixed. ChatGPT and Perplexity read
different sources and recommend different brands. A cross-engine
figure, where we show one, is labelled as one.
Every figure carries a provenance label: measured,
provisional, estimated, or unverified.
Perplexity and ChatGPT were asked the same 20 shopping questions. They named 66 brands between them and agreed on 15 — an overlap of 0.19. On the four questions that named brands themselves, agreement was total; on the open-ended ones it was close to zero.
Shoppers increasingly ask AI what to buy. Here are three practical ways to check whether ChatGPT, Perplexity, and Gemini recommend your Shopify store — and the exact fixes if they don't.
A practical, do-this-then-that checklist to make your Shopify store visible to AI assistants — crawler access, structured data, content patterns and measurement.