Go-to-market

Profound says ChatGPT Shopping switched from web search to merchant product feeds on 10 July, and one day's shift moved feed retrieval from 8 percent to 62 percent

September 9, 2026 at 3:15 PM PT

ChatGPT Shopping results with product feed cards

Image: Search Engine Journal

Why it mattersA brand that does not have a compliant product feed reaching ChatGPT is now competing for a shrinking slice of the results. The share of unique merchants shown fell 20 percent in the same window, and the top ten merchants took 41.8 percent of the visibility.

Profound, an AI-visibility analytics firm, published a study on 3 September that measured how ChatGPT chooses products in Shopping mode. It tracked 1,757,723 prompt runs from 687 customers across July, and found a step change on a single day: feed-integrated retrieval went from 8.26 percent of product recommendations to 61.54 percent on 10 July, the day after OpenAI shipped ChatGPT 5.6. Search Engine Journal covered the report on 9 September.

What the numbers say

Profound reports three shifts, all in the same direction. By 3 September, about 65 percent of Shopping recommendations came from feed-integrated sources rather than open-web search. The top ten merchants' share of visibility rose from 22.5 percent to 41.8 percent. And the count of unique merchants ChatGPT showed at all dropped from 13,524 to 10,607, a fall of more than 20 percent.

The winners and losers are counted in the same sample. Of 687 customers Profound tracks, 450 lost at least a third of their visibility around 10 July and 67 gained the equivalent. A regression Profound published against the same data explained 83 percent of the variation in visibility change by whether a merchant had a compliant product feed reaching ChatGPT.

Shopify integration accounted for about 35 percent of all feed-integrated retrieval in July. OpenAI has not confirmed a shopping-related change tied to the 5.6 release, so the attribution is Profound's, based on the date the shift happened.

What a merchant can act on

For anyone selling through ChatGPT, this is the same kind of switch that hit Google shopping years ago when free product listings started pulling from Merchant Center feeds rather than the crawl. The mechanism changed, the required work changed with it. A page that a crawler could read is no longer the unit of participation; a structured product feed is.

Two things follow. First, a brand that reaches ChatGPT Shopping only through its website is now competing for the roughly 35 percent of results that still come from open-web retrieval, against a field that includes the same feed-backed merchants plus everything the crawler finds. Second, the loss of visibility is concentrated: two thirds of tracked merchants lost at least a third of their share in one day, so nobody who depends on this channel can wait to see what happens next.

The compliant feed OpenAI accepts runs through its Agentic Commerce Protocol, which extended to product discovery in March. A team without that in place today has a specific piece of work to schedule this week. It sits closer to a Merchant Center or affiliate feed build than to a page rewrite, so the owner is likely a data or ecommerce engineer rather than an SEO writer.

Source

Search Engine Journal: ChatGPT Shopping results lean hard on product feeds, 9 September 2026, reporting on Profound's ChatGPT 5.6 Shopping Transformation, published 3 September 2026.

Reported by: Search Engine Journal

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.

More from AI News

Ahrefs data shows adding schema markup barely moves AI citations, and a rank tracker is measuring the wrong thing

Ahrefs research reported by Search Engine Journal finds adding JSON-LD schema lifted AI Mode citations by only 2.4% and ChatGPT by 2.2%, both statistically insignificant, on pages that already had at least 100 AI Overview citations.

Source: PressGo-to-market

Search Engine Journal argues Perplexity should be de-weighted in AI visibility tracking after its referral share fell to 4.31 percent in August

Greg Jarboe at Search Engine Journal argues Perplexity should be de-weighted rather than removed from AI visibility trackers, after StatCounter data shows Perplexity fell from a 7.91 percent AI chatbot referral share in June to 4.31 percent in August 2026.

Source: PressGo-to-market

ChatGPT.com passed 1 billion U.S. monthly visits and Bing lost half its traffic, Semrush data shows

Search Engine Land's Andy Chadwick pulled Semrush Traffic Analytics data for the top U.S. websites in July 2026, and found ChatGPT.com up 48.38% year over year to 1.09 billion monthly visits (rank 9), Bing down 50.43%, and enough attribution loss inside GA4 that a publisher's own numbers may no longer explain themselves.

Source: PressGo-to-market