Compare response HTML with the rendered product
Compare the same product URL’s response and rendered content for identity, price, stock, links and canonical preference. Hold variant, market and account state constant. The differences identify which delivery or rendering layer needs investigation; they do not by themselves establish what Google indexed or why a page is absent.
Worked example — illustrative data.
| Item | Response HTML | Rendered page |
|---|---|---|
| Product name | Empty application shell | Arc desk lamp |
| Price and stock | Not present | USD 85; available |
| Product links | No collection links | Collection and alternative links |
| Canonical | Generic root URL | Product-specific URL |
Use one exact URL and capture both views under the same test conditions. List missing text, links and metadata separately so the fix targets the right layer. Then check failed resource requests, rendering errors and any access rules that affect scripts. Inspect a real search-engine rendering view when authorized and available; do not rename a local browser render as a Google test. Recheck the purchase selection after any rendering change.
Decision: For this fictional shell, investigate product-data delivery and canonical generation first. Do not announce that Google cannot index the store based on this comparison alone.
Regional pricing can create a legitimate capture difference: Personalized stock or regional pricing can make two captures differ for legitimate reasons. Hold location, currency, account state and selected variant constant before treating the mismatch as a defect.
Continue with Guides for stores, software teams and local businesses, Technical SEO services and Define acceptance tests for an SEO fix.
For JavaScript rendering in Google Search, see Google JavaScript SEO documentation (checked November 2026).