Docket for online shops
We audited 10 large online shops expecting to find broken product markup, and did not find it — zero had missing or invalid Product schema and zero had thin product pages, at a median score of 88.2. The problems that did show up are narrower and more dangerous than the ones an SEO pitch usually promises, and one of them can remove every rich result you have.
What we actually found
10 of 12 shops crawled cleanly on 2026-08-07, 20 pages each, mostly product pages. 2 refused the crawler or timed out and are excluded from every count below — a site that could not be read cannot fail a check, and counting it as a pass would flatter these numbers.
| Checked | Shops affected |
|---|---|
| Missing, invalid or incomplete Product schema | 0 of 10 |
| Product pages with almost no content | 0 of 10 |
| Indexing problems (noindex, canonical conflicts) | 1 of 10 |
| Star ratings Docket could not confirm were visible | 3 of 10 |
Scores ran from 55.0 to 91.4. These are well-resourced retailers with teams, so this is the easy case rather than a random sample of the web — but it is worth saying plainly that the standard pitch, that your product markup is quietly broken, did not survive contact with ten real shops.
The one that carries real risk
3 shops carry AggregateRating markup on pages where
Docket could not confirm, from the HTML alone, that a rating is actually shown to a visitor.
That is not an accusation — it is the check declining to make one. Google's
review
snippet guidance requires the rating to be visible on the page carrying the markup, and the
consequence of getting it wrong is a manual action that removes every rich result across the
whole site, not just the offending page.
Star widgets are very often drawn in JavaScript, so the served HTML has markup and no
visible rating while the rendered page is perfectly compliant. Docket used to call that a
violation and was wrong about a real shop; it now demands rendered evidence before it accuses,
and says "unconfirmed" otherwise. Run --render and the question is settled either
way.
Where a competitor is better for this
If what you need is to pull the price, the availability and the SKU off every product page and diff them against your feed, Screaming Frog does that properly with custom extraction and Docket does not do it at all. We decided against building it rather than shipping a worse version of a mature feature, and for inventory-scale auditing that decision means Screaming Frog is the right tool. Its crawl has no page ceiling either, which matters when a catalogue runs to six figures.
For keyword and competitor product research, Ahrefs and Semrush own search-volume indexes that Docket has no equivalent of and is not trying to build.
What Docket is for, on a shop
The four lanes a crawler does not cover, which on an ecommerce site are: whether an AI answer replaces the page a customer would have visited; whether your brand is consistent across a catalogue built by several teams over several years; whether the checkout path's copy actually converts; and whether your contact address can receive mail, which for a shop taking returns is not a small thing.
Plus the deploy gate. A catalogue site ships constantly, and a noindex reaching production is the mistake that costs the most and is noticed the latest.
Common questions
Do online shops usually have broken product schema?
Not in our sample. We audited 10 large retailers and none had missing, invalid or incomplete Product schema, and none had thin product pages. These are well-resourced companies rather than a random sample, but the common claim that your product markup is quietly broken did not survive contact with them.
What happens if my star ratings are not visible?
Google requires the rating to be visible on the page carrying the AggregateRating markup, and a manual action for breaking that removes rich results across the entire site rather than just the page at fault. The common false alarm is a star widget drawn in JavaScript: the served HTML has no visible rating and the rendered page is fine. Check the rendered page before believing any tool that accuses you of this.
Can Docket extract prices from every product page?
No. Screaming Frog does that with custom extraction and it is the right tool for auditing a catalogue against a feed. Docket deliberately does not compete there — it also has a page ceiling per crawl, which matters for a large catalogue.