A note on accuracy. Shopify does not publish the internal list of checks its readiness scan runs, so this page does not pretend to reproduce it. What follows are the underlying catalog problems any such scan is detecting — each one measurable on your public storefront, and each one something we can show you by name.
Missing product identifiers
What the finding means: your products do not carry a GTIN (the barcode number), or a brand, or a per-variant SKU.
Why it costs sales: a GTIN is how an agent knows your "Aurora Candle 250g" is the same object as the one it has seen on three other sites. Without it your product cannot enter a comparison at all — not ranked badly, absent. A missing brand fails every query that names a brand.
What fixing it involves: populating the barcode field per variant and the vendor field per product. This is bulk data entry across every SKU, which is exactly what a CSV import is for.
No structured data on product pages
What the finding means: your product pages do not publish machine-readable product data, or publish it incompletely.
Why it costs sales: this is the block an agent reads first. If it is absent, the agent is left parsing your theme's HTML and guessing. Price, availability, currency, condition and return policy all live here. An incomplete block is a common and quiet failure: the page looks fine and the agent still cannot read the offer.
What fixing it involves: a theme-level template change, plus making sure the underlying product fields it renders are actually populated. The second half is the part a fix file handles.
Unresolved or wrong product category
What the finding means: your products are not mapped to Shopify's Standard Product Taxonomy, or are mapped to something too generic.
Why it costs sales: the category decides which comparison your product is even eligible for. Wrong category, wrong shelf. Worth knowing: a corrected category is genuinely valuable and cannot be verified from outside, because Shopify exposes assigned categories on no public surface. We tell you that rather than quietly bill you for an unverifiable improvement.
What fixing it involves: resolving each product against the live taxonomy — 10,000-plus categories — and setting the category field per product.
Thin per-category attributes
What the finding means: your products lack the typed detail fields their category defines — roast level for coffee, wax type for candles, sleeve length for shirts.
Why it costs sales: this is the one that decides most agent recommendations, and the one almost nobody has done. It is what lets a product survive "waterproof, size 11, under $150". The first stores we measured scored 0.5% and 3.9% attribute coverage. Every tool in this space assigns a category and stops; the attributes are where the actual constraint lives.
What fixing it involves: for each product, filling the attributes its category defines, using only the values Shopify's closed vocabulary allows for that attribute. Guessing a value Shopify does not recognise means the import silently drops it, which is why this needs generated-and-validated data rather than hand entry.
Descriptions that answer nothing
What the finding means: a description too short or too vague to answer a shopper's actual question, or a product with several variants whose options are still Shopify's Title / Default Title placeholder.
Why it costs sales: a placeholder option name means the agent cannot tell your variants apart, so it cannot honour "the large one in navy".
No llms.txt or agents.md
What the finding means: your domain publishes neither of the two files an AI agent looks for to learn where your catalog lives and how it may be used.
Why it costs sales: these are cheap and most stores skip them. It is the lowest-effort item on this page.
The order to do this in
- Identifiers first. Cheapest, highest leverage, no theme work.
- Category next. Nothing downstream is right if the shelf is wrong.
- Attributes third. The biggest single block of value, and the one that needs generated data.
- Structured data. Needs a developer or a theme that already does it.
- llms.txt and agents.md. Twenty minutes.
See it against your own catalog
The free scan names real products from your store and the specific thing each one is missing. No account, no email required to see the results.