The situation
A block of transactions was missing from reporting, and there was organisational pressure to adjust the transformation layer until the reported figure matched the expected one. The proposed filter change would have closed the gap on the dashboard and massively over-counted revenue in the process.
What I did
- Reconciled the transformation layer exactly against its source and proved it was clean, removing it from suspicion with evidence rather than assertion.
- Located the actual gap in an upstream ingestion feed, outside the scope of the reporting layer.
- Declined the proposed filter change and set out, in numbers, what it would have over-counted.
- Handed a precise, scoped verification task to the team that owned the upstream feed rather than working around them.
What it means for you
The costly failure in reporting is not a number that is visibly wrong — it is a number that has been made to look right. Being able to prove which layer is clean is what makes it possible to say no to the fast fix.
Context
These are from eight years owning the data warehouse and reporting function of a consumer marketplace platform, in-house rather than as an outside consultant. The employer and the internal system names are withheld; the numbers are the real ones.