Does the product URL open the variant your feed describes?
Check that a product link selects the same size, capacity or pack described in the feed. A URL can load successfully while opening another variant, leaving the shopper with a different offer.
Record the final page state after redirects and variant selection, then compare it with the exported record.
A valid URL can still lead to the wrong offer
Imagine a synthetic feed row for a blue, 750 ml bottle sold as a single item. Its URL resolves successfully, but the shop opens the smaller red bottle by default. The visitor can eventually select the intended item. That does not make the feed row's initial destination unambiguous.
The failure is semantic: the retrieved product and the displayed selection describe different things. A redirect check alone will miss it. So will a screenshot of the product family if the selected option and selling unit are outside the capture.
Record the selected variant after the final navigation, not merely the URL before the request. Include the market and any visitor state that affects what is shown. A returning customer's saved choice and a first-time visitor's default can produce different observations.
| Feed claim in this fictional example | What to inspect on arrival |
|---|---|
| Blue | Selected colour, not the first thumbnail alone |
| 750 ml | Selected capacity and variant identifier |
| One bottle | Selling unit and included contents |
| A price in a specified currency | Price attached to this selection in that market |
| Available | Availability attached to the same selection |
A product-family URL may be legitimate for a particular implementation. The question is whether the destination preserves the meaning of the submitted item under the destination platform's rules. Google's product-variant documentation distinguishes variant modelling and URL arrangements; it should not be reduced to a universal demand for one particular routing pattern.
Follow the route with the browser state written down
Start with the exact exported URL. Record redirects, the final address and the option selected when the page settles. Then inspect the page's product identifier and structured data where available. Those checks answer different questions: routing identifies the destination, the interface shows the buyer-facing selection, and markup reveals the structured representation.
If the route contains a variant parameter, verify what happens to it during redirects. Losing the parameter may be harmless when the path already identifies the selection, or decisive when the default page shows another item. Test the consequence rather than judging the URL by appearance.
For a market-sensitive shop, repeat the specific affected case with the relevant market selection. Do not change language, country and account state simultaneously and then attribute every difference to translation. Keep the variable you are investigating identifiable.
Decide which layer owns the correction
An exporter that supplies the wrong route needs a mapping fix. A correct route that loses its parameter needs routing attention. A correct selection with stale markup needs a template or data-delivery correction. Rewriting the feed title will not solve those other faults.
Preserve a compact before-and-after record for the failing item. The same route, market and selection should be checked after the correction. This gives the next release a useful regression case and prevents a screenshot of an unrelated working variant from closing the incident.
Distinguish product changes from page defects
A sold-out variant can be a legitimate catalogue change. A page that silently switches to another variant is a different issue. The first requires data freshness analysis; the second requires examining navigation and selection behaviour.
Keep these cases separate in the incident record. Include the observed value and time rather than writing only “price mismatch” or “wrong product”.
Inspect the selected offer up to the point relevant to your test. Use an authorised test environment if verifying the purchase control would create an order.
Cover the routes likely to break
A representative sample should include a direct link, a redirected link and a product with several similar variants. Prioritise recently changed templates and rules over repeatedly checking the easiest page.
Include this handoff from retrieved information to the shop in your production checks. Check the destination with the same care as the feed.
