Checks 1–3: is the title, attribute set and description present and specific?
The first three checks on a product page are the fields buyers and search engines read first. The title must name the product type and the attributes buyers filter on, such as material, size or compatibility; a title that is only the product name fails. The attribute set must be populated for every filter the store offers; an empty attribute hides the product from that filter. The description must be unique to the SKU; a description copied from a sibling fails.
- Title: open the page and the search result preview; verify the type and the two or three filterable attributes appear.
- Attributes: filter the collection by each attribute; verify the product appears where it should.
- Description: search the store for the first sentence; verify it appears on one product page only.
Checks 4–6: are the images, alt text and schema usable?
The next three checks cover what image search, accessibility tools and rich results read. Images must show the product clearly, including a main image on a plain background where a marketplace or Shopping requires it. Alt text must describe the product, not repeat the file name. Product schema must validate and match the visible page: name, price, availability and identifiers must be the same in the markup and on the page.
- Images: verify a main image exists and any marketplace image rule the store sells under is met.
- Alt text: read the alt attribute of each image; verify it describes the product in plain words.
- Schema: run the page through a structured-data validator; verify it passes and the values match the page.
Checks 7–9: is the feed record clean and the price and availability accurate?
The feed and price checks cover the reasons Merchant Center rejects products. The feed record must carry no disapproval and no warning that limits visibility; the reason for any disapproval names the field to fix. Price must match between the feed and the page, including sale prices and currency. Availability must match stock: a page that says in stock while the feed says out of stock, or the reverse, fails.
- Feed: open Merchant Center diagnostics; verify the item has no disapproval and note the field for any warning.
- Price: compare the feed price with the page price at the same moment; verify they match to the cent.
- Availability: compare feed availability with the page and the inventory record; verify all three agree.
Checks 10–12: are variants, internal links and crawlability correct?
The last three checks catch the structural failures that keep an otherwise complete page from being found. Variants must resolve to one canonical product URL so sibling SKUs do not compete with each other in search. The product must be linked from at least one collection page the crawler can reach. The page itself must be crawlable: indexable, returning the right status code, and not blocked by a robots rule or a noindex tag left over from a launch.
- Canonical: view the page source; verify the canonical tag points at the product URL and variants share it.
- Internal links: verify the product appears in a collection that is itself linked from navigation.
- Crawlability: check the URL in Search Console; verify it is indexed and not excluded by a rule.
How should the checklist be run across a whole catalog?
The checklist should be run on the top SKUs by revenue first, then on each collection, because failures repeat within a collection far more than across the catalog. A collection with missing alt text is usually missing it on every product; a collection with a price mismatch usually has one feed rule causing all of them. Grouping by collection turns a hundred failures into a handful of fixes.
Record the result as a pass or fail per check per SKU, and keep it. The record is the baseline the next audit is compared with, and it is the evidence for which fixes were worth the most. The Product Catalog FTE keeps this record in the Catalog tab and updates it daily.
What does the Product Catalog FTE add to a manual audit?
The Product Catalog FTE runs all twelve checks on every SKU every day, ranks failures by revenue at stake, and proposes each fix in Approvals with its source, so the audit stops being a quarterly project and becomes a daily state. A manual audit finds the failures; the FTE finds them, drafts the fix from the catalog or a supplier sheet, verifies feed fixes against the same rule, and waits for approval.
What it does not add is judgment about copy quality or a guess to fill a gap. A page that fails the description check because the sibling copy was duplicated gets a rewrite proposal built from real fields; a page missing a GTIN gets a request to the owner, not an invented number.