That is why these arguments drag on. Nobody is obviously making the number up. Each team has a source, a spreadsheet, or a dashboard behind it. Leadership ends up arbitrating reality instead of deciding what to do next.
The four usual causes
- 01Different source systems. One report starts in the CRM, another in finance, another in an ops sheet assembled later.
- 02Different definitions. Revenue might mean booked, invoiced, recognised, or collected depending on who owns the metric.
- 03Different cut-off timing. Reports run at different moments and capture different states of the business.
- 04Manual handling in the middle. Exporting, pasting, adjusting, and versioning by hand creates drift even when the source logic was sound.
Why 'clean all the data' is usually the wrong first reaction
A broad cleanup sounds responsible. In practice, it often delays the thing the business actually needs: one reconciled output for one decision that keeps getting blocked. If revenue by customer is the recurring fight, fix that path first. If the monthly management report is the recurring fight, fix that path first.
Trust comes back faster when one disputed metric becomes reliable than when a company launches another open-ended data initiative.
When one number keeps derailing management time, the best first move is usually a Mini Proof-of-Work on that one number — with explicit definitions, matched inputs, and one output everyone can inspect.
What a practical fix looks like
Start by tracing the disputed metric end to end. Which systems feed it? Which fields are doing the matching? Which business definition is leadership actually trying to use? Where are the manual steps? Which timing assumptions matter? That gives you enough shape for a narrow proof.
A useful proof does not promise to fix every report in the company. It proves that one recurring reporting argument can become a clean, explicit, reviewable output. That is how you create the case for a broader build without guessing.