Plans: All plans · In the app: Sidebar → Analytics → Shopping
Shopping never feeds your visibility score, share of voice or sentiment. A product in the carousel is a different signal from a brand named in the answer.
How it works
- ChatGPT only, read from the answers Quotable already collects. There is no model filter, and nothing extra to turn on.
- Carousels appear for buying-intent prompts, such as “best cordless vacuum for pet hair” or “top rated air fryer”. If you see No product carousels yet, add prompts like these.
- Counts are placements in the daily sample, never sales or impressions. One product listed by three retailers on one day is three slots.
- Slot is the position in the carousel, where 1 is first.
Overview
Panels:
- Shopping over time: how often answers carried a carousel, and how often one of your products held the first slot.
- Share of shelf: every product slot in the range, divided by the brand that owns it: you, up to five named competitors, other tracked competitors and everyone else. Also shown over time.
- Shopping by topic: shopping rate, answers, products and your share per topic.
- Movement (gaining and losing against the previous period) and New in the carousel.
- Products: each product with Times seen, Share, Win rate (how often it held slot 1), Price, Rating, Slot and Shops. Open a product for its details.
- Competitive quadrant (share of shelf against average slot), Slot stability (best to worst slot), and Beaten by (products shown beside yours that took a better slot).
Attributes
When ChatGPT draws a comparison table beside an answer, Quotable reads it.- Who you are compared with: every brand ChatGPT put in a table with you, including untracked brands.
- Attributes: each dimension with Seen, You, Field (the mean across every brand in the table), Gap and Leader. Ratings and facts carry a score (stars 1 to 5, a tick 1, a cross 0); characteristics do not. A negative gap is a dimension where ChatGPT rates you below the brands it shows you with.
Searches
- Shopping searches: the product searches ChatGPT ran behind its carousels, with Times run, Prompts and Days. These are separate from the query fan-outs behind regular answers.
- Terms the model added: words in those searches that appear in none of the prompts that triggered them. They show what ChatGPT adds to a buyer’s question.
Retailers
- Direct to consumer: where your product slots send the shopper: Your own store, Marketplaces, Resellers or Shop not named.
- Who gets the sale: the shops that carry your products in the carousel.
- Shops that appear together, and a Retailers table with Carries yours for each shop.
Prompts
- Prompts you are missing: prompts where a competitor’s product was recommended and none of yours, ranked by how reliably the prompt shows products and how much of the carousel the competitor holds. Prompts you held in the previous period are marked Lost.
- Prompts: the most recent carousel for each prompt.
Catalog
Connect your product catalog to see which of your products ChatGPT shows. This tab works before any shopping data exists.- Catalog feed: a Google Merchant XML feed, a Shopify product export, or any public CSV or TSV URL. It is re-read once a day and replaces the catalog. Up to 20,000 products are kept. A feed not refreshed for three days is marked stale. Use Sync now or Disconnect feed.
- Upload: a
.csv,.tsvor.txtfile up to 12 MB.
What to do with it
- Never shown products with strong demand are your first content and feed fixes.
- Priced differently means ChatGPT shows a price that does not match yours. Check your feed and the retailers it cites.
- Negative attribute gaps tell you which claims to strengthen on product pages and in reviews.
Next steps
Ads Tracker
Paid placements beside ChatGPT answers.
Write prompts
Add buying-intent prompts that trigger carousels.