12,015 image links fetched before the file was built, and 3,566 of them broken or blocked

What arrived

A catalogue of 1,342 products from a musical instrument retailer. For the pictures, we went to the pages we had matched each product to, on retailers’ and makers’ sites, and found 12,015 image links between them. An image link is the address of a photo on someone else’s server, and a store’s importer downloads each one when the file goes in.

What we found

A link that worked when a page was built can be dead by the time an importer runs, or refused by the site that hosts it. We fetched all 12,015 links before the file was built, and 3,566 came back broken or blocked, nearly a third of them.

What it showed

Nearly a third of the pictures we had found could not be fetched, and many of the failures were whole sites turning away every request. A spreadsheet shows none of that. A picture that opens in your browser is no proof that an importer can fetch it, so the check is made before the file is built, not left to the import.

How we counted

Every image link we found for this catalogue has a fetch result from that check. Of 12,015 links, 3,566 came back as an error, blocked or dead. That figure covers this one catalogue, not a total across jobs.

What it means for your file

Fetch every image link before you import, not after, and from a server as well as in a browser. A dead or blocked link fails at import time, one product at a time, and the first anyone hears of it is a product page missing the photo it should have.

Google Shopping image requirements covers the size rules a working link still has to meet.