The situation
A European energy supplier is moving its customers onto a new platform in groups. Testing was functional-first: check the journey works on screen, then worry about the data underneath it.
The problem
- A journey can look right on screen while the data behind it is wrong, and data faults surface later, as wrong bills and complaints.
- The checks that did exist were run by hand, with no agreed priority, owner or schedule.
What I did
- Proposed and ran the kick-off that landed three decisions: reconciliation runs ahead of functional testing, the checks are automated in the data pipeline rather than run by hand, and they are owned long-term by the domain teams, not a central tester.
- Wrote the first catalogue: sixteen checks across reference data, mapping, assignment and account integrity, each with a priority from P1 to P3, a named owner and a due date.
- Now lead the task force that maintains the catalogue, so the checks keep pace as each new group of customers migrates.
The screens can look right while the data is wrong. Check the data first.