How feed rules work
Merchant Center takes the product data your shop sends and stores it as a feed. A feed rule sits between the two: it reads the raw value arriving from your website and writes a different value into the field Google actually uses. Nothing on your website changes. The rule runs every time the feed is fetched, so the transformation keeps applying rather than being a one-off edit.
The building blocks are simple. A rule can set a field to a fixed value, copy it from another field, join several fields together, find and replace text inside one, or map your list of values onto Google’s. Rules can be conditional, so a change only applies to items that match a test — a particular product type, a brand, an empty field. They can also be layered, each one taking the output of the rule before it.
Why feed rules matter
Most shop platforms were built to run a shop, not to feed an advertising system. The category names, title format and size labels that read well on your own site are rarely the ones Google’s taxonomy and matching expect. Feed rules close that gap without asking a developer to rebuild the export, which for a small business is often the difference between launching this week and launching never.
They also shorten the repair cycle. When Merchant Center rejects a batch of items for a missing or malformed attribute, one rule can supply the value for the whole affected group, and those products return to the auction at the next fetch instead of waiting on a site release.
Common mistakes with feed rules
The most damaging is using a rule to state something that is not true on the page. Price, availability and identifiers are checked against your landing page, so overriding them to clear a warning creates a mismatch that gets items disapproved rather than approved. A rule reshapes data; it does not invent it.
The second is stacking rules until nobody can explain the output. Because they run in order and each can overwrite the last, accounts that have collected fixes over several years often have rules quietly cancelling each other out. The third is treating the rule as the permanent home of the fix. If the underlying data is wrong in your shop, the same wrongness still reaches your customers, your product feed for other channels and your own reporting.
How to act on it
Start from the diagnostics report rather than from the rule builder. Find the attribute genuinely causing rejections or weak matching, write one rule for it, and use the preview to check the result against real items before applying. Name each rule after the problem it solves, so the next person reading the list can tell what it is for.
Where a value belongs in a separate file — seasonal labels, curated titles, data held in a spreadsheet — a supplemental feed is usually cleaner than a long chain of conditions. Push the important corrections back into the shop over time. Rules work best as a thin, readable layer on top of decent data, and an account that leans on dozens of them is really telling you the export itself needs proper feed management.
