How to Audit Product JSON-LD Without Guessing

Start with evidence, not a pass/fail badge

Extract every Product and Offer object from the rendered page and the original HTML. Keep the raw values so a developer can reproduce the finding.

Compare the fields that change a purchase decision

At minimum, compare name, URL, image, brand, SKU or GTIN, price, price currency, availability, seller, and variant attributes with the visible page.

A syntactically valid object can still be commercially wrong. A stale price or incorrect availability is more serious than a missing optional field.

Test multiple catalog states

Audit an in-stock item, an out-of-stock item, a discounted item, a variable product, and a product with multiple images. Template bugs often appear only in one state.

Create a repair order

Fix incorrect price and availability first, then identifiers and variants, then completeness improvements. Rank repairs by customer and revenue risk.

Primary references

• Schema.org Offer
• Google product structured data