We scanned 65 aftermarket catalogs. Four were clean.
We wanted to know how common broken product data actually is, so we checked.
Between 27 July and 11 August 2026 we ran the same automated check across 65 automotive aftermarket sites: brands and manufacturers selling their own parts, the kind of catalog that also feeds Google Shopping, Amazon, and distributor portals. The check reads public product pages and looks at the structured data behind them, the machine-readable block that channels and crawlers use instead of the visible page.
The result that surprised us was not what we found on the broken catalogs. It was how few catalogs we could read at all.
Most catalogs could not be assessed
Of 65 sites:
| Outcome | Sites | Share |
|---|---|---|
| Served readable product data | 18 | 28% |
| Reachable, but no machine-readable product data | 28 | 43% |
| Could not be loaded from our servers at all | 19 | 29% |
The middle row is the interesting one. Those 28 sites loaded fine. They have product pages, prices, photographs, an add-to-cart button. A human can shop them. But there is no structured data behind the page, so a machine reading it gets nothing.
That is not a small formatting gap. Channels do not read your page the way a customer does. They read the markup. A product page with no structured data is a product that has to be described again, by hand, for every channel you sell on, and re-described every time something changes.
The 19 we could not load are a different story and we are careful not to draw conclusions from them. Bot protection, geo-restriction, and aggressive rate limiting all look identical to our scanner. Those sites may have immaculate data. We simply cannot say.
What we found on the 18 we could read
Across 18 sites we checked 128 individual product pages, roughly seven per site.
| Result | Sites |
|---|---|
| Nothing wrong | 4 |
| At least one blocking issue | 7 |
| Warnings only | 7 |
Four clean catalogs out of 18. We name them, because clean work deserves it and because it shows the check is not rigged to find fault: deatschwerks.com, csfrace.com, corsaperformance.com, and hotchkis.net all came back with nothing to report.
Here is every issue we found, counted across all 128 pages:
| Issue | Pages | Severity |
|---|---|---|
| Part number present, but no barcode | 73 | Warning |
| Price missing from the markup | 17 | Blocker |
| Brand missing | 9 | Blocker |
| Stock status missing | 8 | Warning |
| No product identifier at all | 4 | Blocker |
| Description too short to be useful | 4 | Warning |
| Description missing entirely | 2 | Warning |
The single most common problem is also the most ambiguous
Seventy-three pages carried a manufacturer part number but no barcode. That was the most frequent finding by a wide margin, and it is the one we are least willing to be dramatic about.
Google accepts an MPN on its own in some product categories and suppresses the listing in others. Whether it hurts you depends on which categories your products sit in. We flag it as a warning rather than a blocker for exactly that reason.
The blockers are less ambiguous. Thirty of the 128 pages carried at least one.
Seventeen showed a price to a shopper but carried no price in the markup, which is the failure mode nobody catches internally, because the page looks perfect. Nine were missing a brand, which every major channel requires for a branded product, so anything without one tends to be the first thing held back. Four carried no product identifier of any kind: no barcode, no part number, nothing. A product with no identifier is the one every channel struggles to match to anything.
What this does and does not show
This is a small sample and we would rather say so than oversell it.
The 65 sites were not chosen at random. We picked them to match the kind of business we work with: manufacturers running their own storefront, large enough that broken listings cost real money. A random sample of e-commerce would look different.
Seven product pages per site is a sample, not an audit. A site we marked clean could have problems on the 900 pages we did not open.
Most importantly, this measures public product pages, not the feeds those companies actually send to their channels. Those are different pipelines. A company could have thin markup on its own site and a perfectly maintained Google Shopping feed. We have no visibility into the second thing and are not claiming otherwise.
What it does show is that the visible half is frequently broken, and that the people running these catalogs almost certainly do not know it. Nothing tells you. No email, no alert. The page keeps looking right.
How the check works
We fetch the sitemap, sample product URLs across it, and parse the structured data on each page. Every finding maps to a published channel requirement rather than to our opinion, and each one is either a blocker, meaning a major channel will reject or suppress the product, or a warning, meaning it depends on category or channel.
We do not log in, submit forms, or touch anything that is not publicly served. Nothing here required cooperation from any of the 65 companies.
If you want the same check run against your own catalog, the short version is free and takes a domain: run the scan. The full audit is what we sell. Either way we tell you what we find, including when the answer is that nothing is wrong. Four of these eighteen got exactly that.