Definition
A product is absent from search results for one of four reasons, best tested in order: it was never indexed, it lacks a required attribute, its values are inconsistent, or a filter rule excluded it. The attribute is one the surface itself requires; the inconsistency matters when it splits relevance across duplicate records; and the exclusion comes from a facet or filter rule actively ruling the product out. Each cause produces the same symptom — nothing appears — but the fix for one does nothing for the others. Diagnosis is therefore a matter of ruling causes out from the retrieval pipeline inwards, not of guessing at relevance tuning.
Key points
- Not indexed — the record never reached the index. Nothing downstream can retrieve it, and no amount of copy or markup changes that.
- Missing required attributes — the record exists but omits a field the surface treats as mandatory, so it is rejected or held back rather than ranked low.
- Inconsistent values — the same product or the same fact is expressed several ways, splitting relevance across near-duplicates so no single record is strong enough to surface.
- Filtered out — the product is indexed and complete but excluded by a facet, an availability rule, a price constraint, or a category restriction the buyer's query implies.
How do you diagnose a missing product?
Work the causes in order. First confirm indexation: look the product up by its exact identifier or SKU with no filters applied — if it does not return, the problem is ingestion, not relevance. Second, check attribute coverage against the surface's required-field list and read the feed's rejection or warning report. Third, check value consistency: look for the same product under two records, or the same specification written two ways, which divides matching signal. Fourth, replay the buyer's query with facets cleared one at a time to see which filter removed it. Only after all four come back clean is relevance tuning the right lever.
Common pitfalls
- Starting with relevance tuning or synonyms when the record was never indexed, so weeks are spent adjusting a search engine that has nothing to retrieve.
- Ignoring feed disapproval and warning reports, which usually name the missing required field outright.
- Testing only with the internal product name, which the buyer never types, and concluding the product is findable.
FAQ
The product appears in site search but not in shopping results. Why?
Site search indexes your own record and will match on title text alone. External shopping surfaces apply their own required-field rules — identifier, category, availability, price — and drop items that fail them. Check the feed's rejection report first; the field it names is usually the entire explanation.
Could two records for the same product be the cause?
Yes, and it is a common one. Duplicated or near-duplicate records split reviews, links, and matching signal across variants, so neither version accumulates enough relevance to surface. Consolidating to one canonical record with correctly modeled variants often restores visibility without any other change.
Source
The fields Google's Merchant Center product data specification marks Required — whose absence means, in the specification's own words, a product "won't be able to serve in ads and free listings" — alongside those it marks It Depends and Optional, are enumerated there, making it the fastest way to check step two of the sequence above. The corresponding markup on the product page itself is covered by the Product structured data guidelines.