FoundRateWe Scanned 82 Shopify Stores for ChatGPT Visibility. 87% Were Invisible.
Over the past week we ran 82 independent Shopify stores through FoundRate's visibility scanner: for each store, we generated the real buying questions a shopper would ask an AI assistant (built from that store's own catalog), asked ChatGPT, and recorded who got recommended. 246 buying questions in total. Here is what came back.
Finding 1: invisibility is the default
71 of the 82 stores (87%) did not appear in any of the answers to their own category's buying questions. These are not junk stores. The sample includes well-known indie brands with strong Google rankings, loyal followings, and six to eight figure revenues. When a shopper asks ChatGPT "best face moisturizer under $50" or "top rated coffee", the AI simply recommends someone else.
The uncomfortable part for merchants: nothing tells you this is happening. There is no dashboard, no analytics event, no referral entry. The sale is lost before your site is ever visited.
Finding 2: the marketplaces are eating the answers
Across 246 answers, the stores ChatGPT recommended most often were not stores at all in the indie sense. They were marketplaces and big-box retail:
| Recommended seller | Share of answers naming it |
|---|---|
| Amazon | 70% |
| Target | 30% |
| Walmart | 26% |
| Ulta (incl. Ulta Beauty) | 26% |
| Sephora | 22% |
When the AI does name specific brands, they are the category giants: Blue Bottle and Peet's in coffee, CeraVe and La Roche-Posay in skincare. The indie brand that actually makes the best product in its niche is structurally absent unless its product data gives the AI something to work with.
Finding 3: the product data explains a lot of it
Alongside each scan we checked the store's own catalog data, 3,407 products in total:
| Catalog issue | Share of products affected |
|---|---|
| Fewer than 3 product images | 38% |
| No product type set (invisible to category questions) | 24% |
| Description under 200 characters of substance | 19% |
An AI assistant recommends what it can read and verify. A product with no category, a two-line description and one photo gives it nothing. These are fixable in an afternoon in Shopify admin, which is the strange good news inside the bleak numbers: the bar is low because almost nobody is clearing it.
What to do about it
If you sell on Shopify, start with the AI visibility checklist: the twelve product-data and storefront fixes that determine whether an AI can recommend you. Then verify your Product JSON-LD is intact. And before you optimize anything, measure where you stand, because 13% of stores in our sample were already winning some questions and did not know that either.
Run the same scan we used for this study on your own store, free, in about a minute: foundrate.co/scan. No install, no signup.