Historical Data Wasn’t Wrong. The Rules Changed.

A company decides that one of its most important KPIs no longer reflects the business well enough.

The formula may be outdated. The business model may have evolved. New information may now be available.

At first, the solution sounds simple:

Change the calculation.

But changing a KPI is rarely just a technical adjustment. It changes how the organisation interprets performance, compares periods and explains results.

That makes it a strategic business decision.

The initiative should therefore come from management and the business teams that use the KPI. They must understand what the new methodology measures, why it is more useful and how it will affect future decisions.

Before anything is implemented, the old and proposed methodologies should be applied to the same historical periods.

This bridge analysis shows how much of the difference comes from the new calculation and how much reflects an actual change in business performance.

Without that comparison, management may reject a real business signal as merely a consequence of the new formula—or interpret a methodological change as genuine growth or decline.

Once the new KPI is approved, the transition should not happen overnight.

Historical data must be prepared. Required data points must be available and validated. Definitions, ownership and calculation rules must be documented.

For meaningful comparison, the new KPI should normally be calculated for at least three previous years, or for another period that reflects the company’s planning and decision-making needs.

The change should also have a clear cut-off date aligned with a natural business cycle, such as the end of a fiscal year or season.

From that point, the new KPI becomes the official measure.

The previous KPI should not disappear, however.

Its definition, methodology, period of validity and management approval should remain archived. Versioning—supported, where appropriate, by a Slowly Changing Dimension Type 2 approach—preserves the historical context without forcing the organisation to continue using two active KPIs.

Data Governance has an important role throughout this process, but it should not decide which KPI the business must use.

Its responsibility is to assess data availability and quality, document the methodology, establish lineage and controls, and clearly present any risks or limitations.

If management chooses to proceed despite known gaps, that decision should also be formally documented, together with the assumptions, accepted risks and accountable decision-makers.

Historical data was not necessarily wrong.

It was calculated according to the rules that applied at the time.

When those rules change, the organisation’s responsibility is not to rewrite history carelessly.

It is to create a new, comparable version of it—without losing the evidence of how the business originally understood itself.