Step 1 (days 1–2): record the baseline and write the remit
The first step with a Product Catalog FTE is to record the numbers the month will be judged against and to write the remit in the Setup tab. The baseline is catalog completeness by collection, open Merchant Center disapprovals, and search impressions for the collections that matter, all read from the store, Merchant Center and Search Console as they stand. Nothing is rewritten yet.
The remit names the brand voice, the claims the store never makes, the sources a rewrite may use, and the collections the FTE must not touch. It also names who approves, which is usually the owner or the person who runs merchandising. The Product Catalog FTE never asks for what onboarding already captured, such as the store URL or the platform, so the remit is short.
Step 2 (days 3–7): review the catalog audit in the Catalog tab
The second step is the catalog audit, which the Product Catalog FTE runs on connection and reports in the Catalog tab. Every SKU is scored on the six completeness groups, duplicates are grouped by sibling, and feed errors are listed with the failing field. The audit ranks the gaps by revenue at stake using the store's own sales data, with the arithmetic shown.
The owner's job in this step is to read the ranking and set the order of work. Accept it, move a launch collection to the top, or exclude a collection that is being retired. The owner also answers the data requests the audit raises, such as a missing supplier sheet, because those requests decide how many gaps can actually be filled.
Step 3 (days 8–21): approve the first rewrites in Approvals
The third step is the first batches of proposals in the Approvals tab. The Product Catalog FTE proposes rewrites for the top-ranked SKUs in batches sized for one review sitting. Each proposal shows the title, attributes and description before and after, and a map from every claim to the catalog field, supplier sheet or approved copy it came from. Alt text and schema changes are proposed alongside.
The owner approves, edits or sends back with a note. Send-backs are the most valuable signal in the first month: a note that says the voice is too formal, or that a phrase must never appear, becomes a remit rule the FTE applies from then on. Reviewing a batch takes minutes because the evidence is on the screen, not in another tab.
Step 4 (days 8–30, daily): clear feed disapprovals from the morning brief
The fourth step runs in parallel from the second week: every morning brief in the Today tab lists new Merchant Center disapprovals and warnings, the field each one failed on, and the fix the Product Catalog FTE proposes. Because the fix is re-checked against the same rule before it is proposed, approving it means approving something that already passes.
Disapprovals that need data the store does not hold, most often a GTIN, are listed as requests with the SKU named. The brief also shows the count of disapprovals open and closed since yesterday, so the owner can see the backlog shrinking without opening Merchant Center.
Step 5 (day 30): hold the first review in Performance and decide autonomy
The fifth step is the monthly review in the Performance tab. It compares the measured numbers against the baseline: completeness by collection, disapprovals open and cleared, rewrites live, and search impressions for rewritten pages from Search Console with the caveat that ranking has many causes. It also shows the approval record: how many proposals were approved, edited or sent back, and why.
The autonomy decision follows the record. A class of change that was approved unedited every time, most often alt text and schema completion, can be promoted to run without a gate. Copy rewrites usually stay gated. Any promotion can be reversed later from Setup, and the log continues either way.
What does done look like after 30 days?
Done after thirty days with a Product Catalog FTE looks like a measured completeness gain on the SKUs that carry the most revenue, a disapproval backlog cleared or reduced to the items waiting on supplier data, and an approval record long enough to make an honest autonomy decision. It also looks like a remit that has grown by the rules the send-backs taught it.
Done does not look like a conversion claim. A rewritten page's effect on conversion is a test, not a comparison, and that test belongs to a Conversion FTE or to the store's own testing tool. The Product Catalog FTE reports what it changed and what it measured; the review is honest because it stops there.