DEMONSTRATION / FICTIONAL DATA
See the work
behind the repair.
This is a practice dataset, not a customer case study. It shows how we record evidence, propose a change and leave uncertainty visible.
5Records compared
3Supported corrections
1Variant still unverified
What changed in the proposed copy
| Product | Issue | Before | Proposed |
|---|---|---|---|
| Black desk lamp | Product link | Old product address | Confirmed variant address |
| White ceramic mug | Price | 15.00 GBP | 12.00 GBP |
| Small planter | Availability | in_stock | out_of_stock |
| A5 notebook | No discrepancy | Reference matches | Unchanged |
| Blue ceramic mug | Missing evidence | Variant unverified | Unchanged; ask merchant |
Why the blue mug stays open
The white mug’s price is not evidence for a different variant. Without a confirmed reference for the blue mug, changing it would be a guess. The handover asks for that missing evidence.
Open the evidence
These files contain invented products and deliberately non-working .invalid URLs. They contain no customer information.
What this example does — and doesn’t — demonstrate
It demonstrates a local data comparison and an editable proposed correction. It does not test a live store or establish Google approval. A real repair additionally needs source inspection, merchant confirmation, agreed access and a recheck of the regenerated feed.
Discuss a similar issue