Setup sorts your products into four groups by whose turn it is. Three of the four are ours. This page says which, and — where the honest answer is uncomfortable — says that too.
We matched the product to a brand and to a category we ask about, with enough confidence to act on it. Nothing is pending on either side.
“Ready” does not mean a badge is showing. It means nothing is in the way: if and when that category earns a badge, these products can carry it.
The match is in progress on our side. Three things put a product here:
All three resolve as measurement fills in. There is no button for you here, and if we showed you one it would be decoration.
Two signals disagreed about what this product is, or the independent signal was too weak to act on. That is on our side to resolve.
And today there is no way for you to fix it yourself.
We are saying this plainly because the alternative is worse. There is no correction control in the app: you cannot override our match. So a product in this group waits for us, and the honest thing is to say so rather than to imply you have an option you do not have. If one of your products is stuck here and it matters, write to us and a human will look at it.
Not a problem with the product. Our measurement runs by category, and a product outside the categories we ask about has nothing to be measured against.
Two reasons land here: the product does not map to any category in the programme, or the brand is not in the registry we measure. Both are statements about our coverage, and neither is a verdict on the product.
The independent signal is our own reading of your product photo: a vision model looks at the image and says which category it thinks the product is in. It is independent of your catalog text — which you write — and that is the entire point. Without it, a product could be moved into a category by editing a description.
Two things it is not. It is not the assistants’ answers: those decide whether a category is earned, and this never touches that. And it does not decide on its own — if it disagrees with your catalog text about what a product is, we activate neither, because one of the two is wrong and we cannot tell which.
So its only job is deciding which of your products can carry a claim the category already earned. Running it costs us money per product, which is why it runs once per photo and only where your brand has actually won the category.
There is no AI-recommendation signal at the level of an individual item: assistants recommend brands and models for a need, not SKUs. So a badge is issued for a (brand × category × market) and displays on every product you map to that category.
That is also why we never bill per product. Billing by catalogue size would be billing a fiction.