How feed rules and overrides work
A product feed is the file or connection that carries your catalogue into an ad platform. Rules sit between that source and the platform’s copy of it. Each rule says something like: take the value in this column, and before you save it, change it. Common actions are find-and-replace, adding a prefix or suffix, mapping one of your values onto a value the platform recognises, or setting a default when a field is empty.
An override is the narrow version of the same idea — a value pinned for one specific product rather than a rule applied to everything. Both Meta’s Commerce Manager and Google’s Merchant Center offer this pattern, and in both the original store data is untouched. If you delete the rule, the imported value reverts on the next refresh.
Why feed rules and overrides matter
Most catalogue problems are not really product problems. The store holds the right information under the wrong label: sizes written in a way the platform does not parse, a condition field left blank, a brand name that varies between listings, a category taken from your own internal tree. Rules fix all of that at the point of import, which is far quicker than a developer ticket and does not risk breaking the website.
They also solve a genuine conflict of purpose. Your website copy is written for humans reading a page, but a product feed is read by a matching system that cares about brand, type and attributes in a predictable order. A rule lets each surface have the wording it needs from one shared source.
Where feed rules and overrides go wrong
The biggest risk is that a rule hides a defect instead of curing it. If titles arrive incomplete because the product database is incomplete, a rule that pastes the brand name onto the front makes the feed look healthier while the underlying data stays broken — and every other channel keeps inheriting the fault.
Rules also accumulate. Nobody documents them, staff change, and a year later the feed has a stack of overlapping instructions that nobody dares touch because the interaction between them is not obvious. Manual overrides are worse in one respect: they stay pinned when the real product changes, so a price fixed by hand keeps showing the old figure long after the store has moved on.
Getting it right
Ask one question before writing any rule: should this be fixed at the source? Anything that also affects your website, your search visibility or your other sales channels belongs in the store. Reserve rules for genuine platform translation — mapping your categories onto theirs, formatting a field the way the specification demands, filling a gap that only this platform requires.
Then keep them reviewable. Write down what each rule does and why, use rules rather than per-item overrides wherever possible, and check the imported result after the next refresh rather than assuming the change took. When the same fix keeps being needed across catalogue and dynamic product ads, that is your signal to repair the store data once and delete the rule.