Churn means something different to each team
Customer success reports churn as the number of accounts that cancelled. Finance reports it as lost recurring revenue. The board pack uses a figure someone calculated last year with a method nobody remembers. Each is reasonable. Put side by side, they tell different stories.
The same happens with active customers, conversion rate, on-time delivery, utilisation and gross margin. Every department has a version, and most do not know the others exist.
Where the definitions drift
A KPI sounds like a single thing, but each one hides several choices. Is an active customer someone who bought in the last 30 days or the last 90? Does gross margin include delivery costs? Does on-time mean the promised date or the requested date? Is a customer who pauses a subscription churned?
| KPI | Choices hiding inside it |
|---|---|
| Active customer | Time window, what counts as activity, trial users in or out |
| Churn | Accounts or revenue, pauses, downgrades, the period used |
| Gross margin | Which costs are included, returns, discounts |
| On-time delivery | Promised or requested date, partial deliveries |
| Conversion rate | Which visits count, which conversions, bots excluded |
When nobody writes these choices down, each person building a report makes them fresh. Formulas get copied, tweaked and copied again. Within a year the same name covers several calculations.
What inconsistent KPIs cost
Trends become meaningless when the method changes between periods. Targets are argued over because the team being measured uses a different calculation. Bonuses tied to KPIs create disputes. Investors or lenders see one figure in a deck and another in the management accounts, and ask awkward questions.
Most of all, time goes on reconciling figures instead of acting on them.
There is a subtler effect as well. When a team can pick its own calculation, it tends to drift towards the one that flatters it. Nobody is being dishonest. It is just human, and a single written definition takes that temptation away.
How we pin the definitions down
- We collect every place each KPI appears: dashboards, spreadsheets, board packs, and the formulas behind them.
- We lay out the variations side by side and show where and why they differ, in plain language.
- Your leadership team chooses one definition per KPI, and we write it up as a short metric definition: name, formula, inclusions, exclusions, owner.
- We build each KPI once, in a shared data layer such as a database view, a dbt model or a Power BI semantic model, so it is calculated in one place only.
- Every dashboard and export reads the KPI from that layer instead of recalculating it. Spreadsheet users get a connected table rather than their own formula.
- We show the definition next to the number in reports, and changing a definition becomes a deliberate, recorded decision with the history recalculated.
Where a department truly needs a different view, such as revenue churn next to logo churn, both get their own names. The rule is one name, one meaning.
One number, one meaning
The board pack, the team dashboards and the finance spreadsheet show the same figure for the same KPI because they read it from the same place. When someone asks how it is calculated, the answer is a click away rather than a hunt through old formulas.
It also makes targets fairer. When a team's bonus depends on a KPI, the team knows exactly how it is calculated and cannot be surprised by a different method at year end. Disputes shift from how the number was worked out to what to do about it, which is a far more useful argument.
New starters learn the definitions from the glossary rather than inheriting someone's spreadsheet habits.
Warning signs
- Two reports show different values for a KPI with the same name
- Nobody can say for sure how a board-level metric is calculated
- Targets are disputed because of how they are measured
- Copies of the same formula differ slightly across spreadsheets