Skip to main content
FHAVI

Free resource

Fifteen minutes a week, before the disapproval email

Most feed problems are visible days before Google reports them. A price moves, a product goes out of stock, a landing page starts redirecting — each one shows up in a report you already have, if anyone reads it. This is the weekly read.

Read this first

Score, do not just look

Opening Merchant Center and thinking “looks fine” is not a check. A scorecard exists so that next week's answer is comparable to this week's, and so that a slow drift — 3% of items disapproved, then 8%, then 20% — becomes visible as a number rather than a surprise.

Score each line 0, 1 or 2: 0 means broken or unknown, 1 means present but imperfect, 2 means clean. Fifteen minutes, once a week, same day.

Group 1

Feed delivery

01

Is the feed arriving intact?

  1. Last fetch succeeded, and the timestamp is within your scheduled interval — not “yesterday” when it should be hourly.
  2. Items processed equals items submitted. A gap means rows are being dropped before they are even judged.
  3. No duplicate item ids across the feed. Duplicates cause one item to overwrite another and produce phantom disapparances.
  4. Required attributes present for every item: id, title, description, link, image link, availability, price, brand or identifier decision.
  5. The feed is the source of truth for price and stock, or it is synchronised with the store — not updated by hand at intervals nobody tracks.

Group 2

Item health

02

Are items allowed to show?

  1. Disapproved item count, and the change since last week. Record the number; the trend is the signal.
  2. Top three disapproval reasons this week, by count — not the total, the reasons.
  3. Approved-but-not-showing count, which is usually a bid, budget or eligibility issue rather than data.
  4. Out-of-stock items still marked available. This is the single most common self-inflicted disapproval.
  5. Percentage of items with a price that differs from the landing page. Anything above zero is a defect; anything above 2% is a process problem.
03

Is the data trustworthy?

  1. Identifier coverage matches the share of the catalogue that is genuinely branded — no invented GTINs, no GTINs shared across variants.
  2. Images are the product on a clean background, with no watermark, logo, text overlay or supplier tag visible.
  3. Titles describe the item in your own words, with no borrowed brand names and no keyword stuffing.
  4. Descriptions are not the same paragraph pasted across the catalogue.
  5. Sale prices are either genuinely running or removed — a permanent sale price is a data quality flag.

Group 3

The destination, not the feed

04

Do the links land?

  1. A sample of ten product links returns 200 and shows that product at that price — checked in a private window, not logged in.
  2. No redirect chains on product links; the feed link should be the canonical page.
  3. Mobile rendering works, because most Shopping clicks are mobile.
  4. Checkout is reachable and completes for the target country.
  5. Shipping settings in Merchant Center match what checkout charges — a mismatch here is a suspension trigger, not a feed nit.

Recording it

What to keep week to week

One row per week, seven columns: date, items submitted, items approved, items disapproved, top disapproval reason, price mismatch percentage, and the score out of 44.

Two weeks is enough to see a trend, four is enough to prove a fix. Without the record, every disapproval feels sudden, and every fix feels unverifiable — which is how merchants end up rewriting the same titles twice and still failing review.

Keep going

The rest of the library

Every checklist here is free and ungated. The research pages carry the reasoning behind them.

All resources · Research library

Feed failing every week and nobody owns it?

That is usually a process problem, not a data problem. Tell us the catalogue size and what changes most often, and we will scope it.

Prefer email? Write to md@fhavi.com