Stall library · Evaluate
Variants are not described in structured data
A variant data stall happens when a product sells several sizes or colours but the product page markup describes only one undifferentiated product, so an AI agent cannot tell which combination it is being offered or whether the one a shopper asked for exists. The agent knows the product is right and cannot confirm the specific thing the shopper wants, which is the point at which it stops recommending you.
Why the agent stops
Shopping instructions are usually specific. A shopper asks for a size, a colour, a capacity. A product page that presents one price and one availability for a grid of twenty combinations forces the agent to guess which of those facts applies to the combination in question.
Guessing here is unusually costly, because a wrong answer about size availability produces a failed purchase rather than a mild disappointment. Agents avoid it by preferring stores that state the grid.
How often we see this
Not stated here. A frequency means nothing without the cohort it was counted over, and this page is about one stall rather than one category. Where a category has enough stores measured at every stage, the share of them showing this stall is published on its benchmark page, with the methodology beside it.
See it for yourself
- Pick a product with at least two options, for example size and colour.
- Fetch the page with curl and read the JSON-LD. Look for hasVariant entries, each with its own offer.
- If only one offer is present, check whether it reflects the cheapest variant, the default variant, or something else. Whichever it is, an agent cannot tell.
- Compare with your product feed. A feed that lists the full grid while the page does not is the common shape of this stall.
What to change
- Add hasVariant entries to the Product markup, one per purchasable combination, each with its own offer carrying price and availability.
- Name the option axes explicitly so an agent can map a shopper request onto them.
- Give each variant a stable URL or an identifier so the choice can be referenced later in the journey.
- Keep the page markup and the feed in agreement about which combinations exist.
What Ottom looks at
Ottom cross checks the feed against the page. When the feed says a product has several variants and the page markup describes none, that specific disagreement is the finding, and the report names the product, how many combinations the feed claims, and which option axes they run along.
Questions
Is missing variant data the same as a broken size picker?
No, and the difference matters. This is about whether the options are described in the data. A hover only variant picker is about whether an option that is described can actually be chosen, and it stops the journey later, at the buy stage.
Do I need variant markup on the page if it is in the feed?
Yes, because agents do not all take the same path. An agent that arrived at a product page from a search result may never fetch your feed, and it should still be able to answer a question about size availability from the page it is on.