Roughly 67 of your 291 products are missing enough required fields that an agent skips them rather than ranks them. Agents do not guess — a null field is a disqualification.
Each pillar scores out of 100 over the checks that feed it — they are cut from the same audit as the composite above and do not sum to it. Acid marks the weakest, which is where a fix returns the most.
Whether conventional search engines can crawl, parse, and rank the catalogue.
Whether the catalogue can be lifted as a direct answer to a shopper's question.
Whether generative engines will cite this store when composing a recommendation.
Whether an AI system can correctly parse and act on the record without guessing. Includes transactability as a named sub-score.
100transactabilityUnder ~180 characters there is nothing for an agent to extract facts from.
Material, dimensions, composition. The fields agents filter and compare on.
Every variant needs a stable identifier or an agent cannot reference what it buys.
Blank type means your product joins no comparison set at all.
Three or more per product. How use-case and occasion queries find you.
No weight means no delivery estimate - you drop out of speed-filtered results.
Multimodal agents read images. No image is a hard skip on visual surfaces.
Default Title tells an agent nothing about size, colour or configuration.
Brand attribution. Feeds brand queries and Merchant Center matching.
JSON-LD Product found. Offers ✓ · availability ✓ · shipping & returns ✕. Shipping and return terms in schema are what let an agent promise a delivery date.
Barcodes are not published in a storefront's public product feed, so a free scan cannot check them — this is excluded from your score rather than counted against it. GTIN is how an agent knows two listings are the same physical product. Connecting the store reads it directly from the Admin API.
No AI crawlers disallowed. Nothing locked out at the door.
Starter · under 500 skus · $99/mo
An engine that cannot find a quotable answer summarises the page in its own words instead. The shopper still gets the answer; the store loses the attribution, the citation and the click that follows it.
Whether conventional search engines can crawl, parse, and rank the catalogue.
A product with no type joins no comparison set — it is not ranked badly, it is not in the running. Where several URLs serve the same item with no canonical to settle it, an index picks one and drops the rest, so variants a shopper searches for are simply not there.
Rank Sniper drafts a description of the right length from fields the product already carries and stages it for one-click apply. This is the text a search result shows under your title; with none, the engine writes its own from whatever page text it can find.
Exactly one H1 per product page is a theme-template change, so Rank Sniper hands you the markup and the file rather than making it. It re-checks on the next scan. Two H1s — or none — leave an index guessing what the page is about.
What fixing it changesSetting a product type and three or more tags puts each product into a comparison set instead of leaving it uncategorised. A valid sitemap.xml and one self-referencing canonical per page remove the duplicate-URL ambiguity that keeps variants out of an index.
Whether the catalogue can be lifted as a direct answer to a shopper's question.
An engine that cannot find a quotable answer summarises the page in its own words instead. The shopper still gets the answer; the store loses the attribution, the citation and the click that follows it.
Rank Sniper writes the FAQPage markup, but the questions and answers are yours: it drafts them from your existing support copy and stages them for review. Rank Sniper does not publish a question written in your voice that you have not seen.
Rank Sniper restructures a description into the lists and tables its content already implies, and stages it for one-click apply. A specification an engine can lift as a table is quotable; the same facts buried in a paragraph are not.
Rank Sniper reorders the opening of the description so it starts with a declarative sentence, and stages it for you to apply. An answer engine reads the first sentence to decide whether the page answers the question at all.
What fixing it changesA description past ~180 characters that opens with a declarative sentence and carries a list or table gives an answer engine something it can quote. FAQPage or HowTo markup turns existing support copy into a citable answer rather than prose an engine has to summarise.
Whether generative engines will cite this store when composing a recommendation.
A disallowed agent is a hard exclusion, not a ranking penalty — the store is absent from the answer rather than low in it, and no amount of catalogue quality compensates. Where there is no vendor and no date, a model has nothing to attribute a claim to and prefers a source that does.
A quotable claim is a number with a unit — gsm, litres, centimetres. Rank Sniper stages the sentence once you supply the figure; it will not invent a measurement about your product. Numbers with units are what a model repeats verbatim instead of paraphrasing.
Rank Sniper proposes the materials, standards and provenance already named in your copy as explicit entities and stages them for review; where the fact is not in your copy it asks rather than guesses. Named entities are what connect your product to what the shopper actually asked for.
What fixing it changesPublishing llms.txt states your terms to generative crawlers instead of leaving them to infer. Naming a vendor and keeping updated_at current give a model something to attribute the claim to and a reason to prefer the record over a stale one. Unblocking an agent in robots.txt is the one change that removes a hard exclusion rather than improving a ranking.
Whether an AI system can correctly parse and act on the record without guessing. Includes transactability as a named sub-score.
Contradictory markup is worse than absent markup: a parser that finds two prices on one page cannot pick one, and an agent that cannot resolve price or availability cannot transact at all. Without a SKU there is no stable way to refer to what is being bought, so the record can be described but not acted on.
Rank Sniper extracts the material, dimensions and composition already written in your descriptions and publishes them as structured attributes. Nothing is invented — a fact that is not in your copy is not published.
What fixing it changesA SKU on every variant, offers carrying both price and priceCurrency, and an explicit availability value are what let an agent quote and transact rather than merely describe. Shipping weight is what turns a listing into a delivery estimate. Resolving a schema conflict matters more than filling a blank field: contradictory markup is worse for a parser than absent markup.
How these get fixedRank Sniper generates every fix. Fixes to the ranksniper namespace, the JSON-LD and the feed publish automatically. Fixes to your own product fields apply in one click — in bulk — and revert byte-for-byte. Where a change lives in your theme or your store settings, Rank Sniper writes it for you and checks it on the next scan.
Monitoring is the useful part. Rank Sniper re-scans allbirds.com and emails you when the number moves, so a theme update that strips your markup does not go unnoticed for a quarter.
free without an address: the score, the verdict, every category percentage and the top 3 offenders — all already shown above