Your dashboard agrees with your spreadsheet. Both can be wrong.

The comfortable check

Most reporting checks compare one of your figures with another. The dashboard against the spreadsheet. This month’s report against the workbook. When they match, everyone relaxes.

But if both read from the same data and rest on the same assumptions, a match only proves they agree. They can agree and both be wrong. Your own match stats can say you dominated the game. The scoreboard is what counts.

Tying to your own table is consistency, not correctness.

We learnt this the hard way

On one piece of client work, every check that compared our figures with our own tables passed. The client still found errors. Each one was the kind an internal check cannot see, because the check shared the mistake.

Where the errors hide

In our experience they fall into four groups.

1. The same thing spelt two ways. A product code with a suffix in one system and without it in another. The totals look complete while some of the sales sit under a name nothing reads.
2. A rule borrowed from another channel. A fee or tax rule that is right for one marketplace, applied to another where it is not.
3. A different window. Order date in one report, settlement date in another. “Last 30 days” meaning two different things.
4. A gap in the model. A cost nobody added: refunds, storage, a fee that only appears on the settlement.

None of these shows up when you compare your figures with themselves.

Check against the source

The fix is simple to say. Compare the number with the place the money actually comes from: the marketplace settlement, the bank statement, the platform’s own report. Do it on a schedule, not once.

• Pick the three figures you act on every week.
• For each one, name the source system that would settle an argument.
• Compare them every month. When they differ, fix the cause upstream, not with an adjusting line at the bottom.

Consistency is useful. It is just not the same thing as being right.