What does limited performance mean in Merchant Center?

Merchant Center separates issues by severity. A disapproval stops the product serving. An account-level issue can stop the whole catalogue. A warning of this kind leaves everything running while telling you that the item is less useful than it could be.

That distinction matters for how you triage. Disapprovals are urgent because revenue has stopped. Warnings are important because revenue is smaller than it should be, which is easier to ignore and often costs more in total across a large catalogue.

The mechanism behind the warning is usually matching. Google decides which requests a product is a good answer for using the data you supply, and an item missing the attribute that a shopper is conditioning on cannot be matched against that condition. It is not being penalised; it simply has less to be matched on.

Some of these warnings also affect which presentations a product qualifies for. Richer placements and comparison features draw on attributes such as images, highlights and shipping detail, and an item lacking them may be shown in a plainer form.

Missing identifiers are the most common. A product without a GTIN where one exists cannot be matched to the shared product record, so your offer stands alone rather than joining a page where shoppers compare sellers.

Thin attributes come next. An absent product type, a missing colour or size on an item where those matter, no material, no age group or gender in apparel. Each one is a condition your product cannot be matched against.

Image quality is a frequent flag: resolution below what is preferred, promotional text or logos overlaid on the main image, a placeholder, or an image that shows packaging rather than the product. Product panels are visual, so this one carries more weight than merchants expect.

Missing shipping and returns detail appears where the account settings do not cover a targeted region or where returns terms have not been configured. Shoppers ask about both, so an item that cannot answer is a weaker candidate.

Price competitiveness shows up where Google has enough comparable offers to benchmark against. It is information rather than an instruction: sometimes the right response is a price change, and sometimes it is to accept the position because your margin or your service justifies it.

Description and highlight gaps round out the list, particularly on items where the description is a single line inherited from a supplier.

Because the cost is invisible. A disapproved product shows a clear before and after in the reporting. A product with a warning quietly gets fewer impressions than it might have, and there is no line in any report saying how many.

The effect compounds with the way shoppers now ask questions. Long, conditional requests, whether typed into a search box or asked in an AI surface, are resolved against structured attributes. Every missing attribute is a set of requests your product cannot satisfy, and those are the requests with the highest intent.

There is a second-order cost too. Items with weak data tend to accumulate less performance history, which means automated bidding has less to work with on them, which tends to keep them quiet. A store can end up with a long tail of products that never get a fair test.

So the case for clearing warnings is not that any single fix will produce a visible jump. It is that the catalogue as a whole becomes matchable, and the benefit shows up as more of the range earning impressions rather than as a spike on one item.

Group by issue, not by product. The diagnostics view lets you see how many items share each problem, and one missing attribute usually explains a large block. Fixing the cause at source clears hundreds of warnings in a single fetch, whereas editing items one at a time never finishes.

Prioritise by revenue and impressions. Join the issue list to your performance data and start with the products that already earn, since an improvement there is measurable within weeks. The dormant long tail can wait for the systematic fix.

Fix in the source system rather than in a patch. If a supplier feed is missing colours, get colours into your product data so the storefront, the feed and any marketplace listing all improve together. Use a supplemental feed for overrides you genuinely cannot fix upstream, and treat those as temporary.

Then verify. Warnings clear on the next successful fetch and review, so check the count a couple of days later rather than assuming. Keep a record of the count by issue type over time, since that trend is the only honest measure of whether the work is progressing.

Accept that some warnings will not clear. A handmade product will not have a GTIN, and a premium item may always sit above the benchmark price. Mark those as known and reviewed so the list stays meaningful.