Limited performance and other warnings
The non-blocking Merchant Center warnings, such as limited by missing identifiers or attributes, capped, and similar, that let products serve but quietly suppress their reach, and why to fix them.
Not every Merchant Center problem stops a product serving. A whole class of warnings lets items keep showing while quietly capping their reach: limited by missing identifiers, limited by missing attributes, capped visibility. Because the product still appears, these get ignored for months, silently costing impressions and auctions. This guide covers the common warnings and why fixing them pays.
The warnings nobody fixes
Merchant Center grades problems by severity, and the dangerous middle grade is warning: the product still serves, but something is wrong. Because the item keeps appearing, warnings get filed under "not urgent" and sit there for months. That's the trap. A warning doesn't cost you a visible disapproval; it costs you reach, silently, on a product that looks fine in every count you're watching. The bill arrives as auctions you didn't know you were in and lost.
Understanding where warnings sit in the taxonomy helps: see the disapproval taxonomy and Diagnostics for how severity decides whether an item serves, serves-with-a-cap, or doesn't serve at all.
The common non-blocking warnings
Limited performance due to missing identifiers
The product serves, but without a valid GTIN (and brand where relevant) it can't be matched to Google's product catalogue, so it misses the richer placements, comparison surfaces and "popular product" treatments that matched items get. This is usually the highest-impact warning to clear because identifiers unlock the most. Fix by supplying valid identifiers; see product identifiers: GTIN, MPN and brand.
Limited performance due to missing attributes
Recommended-but-not-required attributes (color, size, gender, age_group, material, google_product_category at full depth) are absent, so Google can't match the product to as many shopper queries or refinement filters. The item serves for fewer searches than it could. Fix by completing the attributes the product's category expects.
Capped / limited visibility
The product is serving but visibility is capped, often because of a soft data-quality issue, thin content, or a category–attribute gap that stops short of disapproval. It shows, but throttled. Fix by resolving the underlying data gap the warning names.
Missing recommended fields
Softer than the above: descriptions too thin, no additional_image_link, sparse product detail. Each one slightly reduces eligibility or ranking. Individually minor; collectively a drag across the catalogue.
| Warning | What it costs | Fix |
|---|---|---|
| Limited performance, missing identifiers | Catalogue match, rich placements | Valid GTIN + brand |
| Limited performance, missing attributes | Query and filter matches | Complete category attributes |
| Capped / limited visibility | Throttled impressions | Resolve named data gap |
| Missing recommended fields | Small ranking/eligibility loss each | Fill descriptions, extra images, detail |
Deep dive Why suppressed reach is invisible, and how to make it visible
The reason warnings go unfixed is a measurement blind spot. A disapproval shows up as a product that was getting impressions and now gets zero: a visible cliff in the data. A limited-performance warning shows up as… nothing. The product keeps serving, keeps getting some impressions, keeps converting a little. There's no cliff, no alert, no line that drops. The cost is the difference between the reach it gets and the reach it could get, and that counterfactual never appears in any standard report. You can't lose what you never see, so nobody prioritises it.
This is why warnings quietly become the largest source of foregone performance in a mature catalogue. A store can clear every disapproval, sit at zero errors, and still have thousands of items each serving at 60% of their potential reach because of missing identifiers and thin attributes. Multiply a modest per-item suppression across a big catalogue and the aggregate dwarfs the occasional visible disapproval, but it's spread so evenly and invisibly that it never triggers an investigation.
Making it visible needs a measure of completeness against potential, not just a count of what's broken. That's the job a feed-quality score does: it weighs every field Google rewards across every product, so the gap between "serving" and "serving at full reach" becomes a single number you can watch and move. Warnings that were invisible as lost impressions become visible as score drag, ranked by how many products each affects and how heavily the field is weighted, so you fix the biggest suppressed-reach blocks first instead of chasing whatever disapproved loudest. See Feed Health Score explained. Because enrichment is non-destructive, filling the missing identifiers and attributes happens in a supplemental layer over your source data, and lifts every channel that reads the feed, not just Google.
Working the warnings
- Clear errors first, then treat the warning list as your suppressed-reach backlog: it's the next tier of value once nothing is blocked.
- Sort by impact, not by warning type. The biggest block of missing identifiers usually beats a hundred scattered thin-description notices.
- Fix at source, so the correction lifts every channel and holds across future feed fetches; don't hand-edit items in Merchant Center.
- Watch the trend. A rising warning count after a data import is a quiet regression worth tracing before it costs a season of reach. See the disapproval taxonomy and Diagnostics for reading states over time.
Warnings are the difference between a feed that's approved and a feed that's performing. Clearing disapprovals gets you serving; clearing warnings gets you serving everywhere you could.
Frequently asked questions
If the product is still showing, why fix the warning?
Because "showing" and "showing everywhere it could" are different things. A warning-flagged product serves in fewer places, wins fewer auctions, and appears on fewer surfaces than a clean one. You lose impressions you never see reported as lost, which is exactly why these warnings get ignored.
What's the difference between an error and a warning in Diagnostics?
An error disapproves the item: it stops serving. A warning lets it serve but flags a limitation or an escalating risk. Warnings are the quiet tax; errors are the visible block.
Which warning should I fix first?
The one affecting the most items and the heaviest-weighted field, usually missing identifiers, since identifier-complete products get matched to Google's catalogue and unlock the richest placements. Sort by affected-item count and fix the biggest block.