Written for Shopify apparel and accessories stores. Sources checked October 17, 2026.
Start with one affected variant
Open an affected item from Merchant Center’s Needs attention view. Record its item ID, size, color, destination URL and availability in Final Attributes. Google’s availability troubleshooter specifically asks you to compare the same variant on the landing page.
Keep the original submitted row too. Final Attributes is the comparison point for what Merchant Center currently holds; an old export alone cannot establish that state. Write down the observation time, target country and last source update, so a later stock change does not look like contradictory evidence.
A fictional example: medium is available, small is not
This is a synthetic diagnostic, not a client result. The item TEE-NAVY-M is a navy medium tee. Its feed link opens the same product family, but selects navy small. Both stock statements could be accurate for their respective sizes.
| Evidence location | Variant described | Observed state |
|---|---|---|
| Merchant Center Final Attributes | Navy / medium | in_stock |
| Submitted URL, freshly opened | Navy / small selected | Sold out; add-to-cart disabled |
| Page JSON-LD Offer | SKU identifies navy / small | OutOfStock |
| After manually selecting medium | Navy / medium | Available; medium remains selected in cart and checkout |
Selection mismatch and stale inventory are different findings. Here, the evidence points to an incorrect initial selection. It does not justify changing medium to out of stock. If the exact medium variant instead opens correctly but is sold out, investigate stock mapping and update timing.
The example supports a URL or selection investigation, not a confirmed implementation cause. Check the full submitted link, redirects and theme behavior before assigning the repair. A valid HTTP response proves that a page exists; it does not prove that the intended option opens.
Follow the same item through four checks
1. Final Attributes and the active source
Compare the affected item with its current source row. Identify the integration that supplies availability and the link: for example, the store’s Google channel, a feed app or a scheduled export. Record any rule or additional source that changes those fields. If the latest export is correct but Final Attributes is older, inspect processing evidence before rewriting the mapping.
2. The exact landing URL
Open the complete submitted URL in a fresh browser session. Record the final URL after redirects, selected size and color, visible availability and purchase-button state. Keep the target country consistent. A remembered selection can conceal the default option that a new visitor sees.
3. The matching structured-data offer
Inspect the Product and Offer data published on the page, including JSON-LD where used. Use identifiers and variant context to locate the relevant offer; finding one InStock value somewhere in a multi-variant page is insufficient. If the visible selection and offer disagree, record both and involve whoever owns the theme or markup-generating app.
4. Cart and checkout
Where the item is purchasable, add the intended variant and verify its identity and availability through checkout without placing an order. Record a blocking stock or delivery message rather than treating an enabled button as proof of purchase availability. Google’s availability guidance covers consistency between the data source, page, structured data and checkout, including preorder and backorder handling.
Repair the source that owns the field
Write the proposed change beside its owner. The store operator confirms inventory truth; the feed implementer owns export mapping; a theme developer may own URL selection or structured data. An isolated downstream edit can disappear at the next synchronization, so establish the authoritative source before changing it.
For the fictional example, the next step is to test a link that opens medium reliably, then update the responsible link mapping if that resolves the selection problem. Preserve a before-and-after record and regenerate only the agreed scope. Recheck the same item, then representative neighboring variants to see whether the change behaves consistently.
Confirm that the source update was processed and monitor the affected item’s status. Technical consistency does not guarantee approval or a particular review time. If the evidence points to an account-level issue, this variant check is only one part of the investigation.
Keep a reproducible evidence record
For each checked item, save: item ID; size and color; country; full submitted and final URLs; observation time; source name and update time; Final Attributes; visible selection; matching structured-data offer; checkout result; confirmed discrepancy; proposed owner; and verification result. Keep hypotheses separate from observations.
If the discrepancy concerns price as well as stock, continue with our price and variant mismatch guide. The synthetic sample report shows how an observation becomes a scoped repair proposal.
Need someone to trace the discrepancy?
The $299 diagnostic covers up to 2,500 variant rows and 20 manual product URL checks. You receive prioritized findings with evidence and a proposed repair scope. Repairs are separately agreed; the diagnostic does not promise reinstatement.