Principle
A commercial metric is a definition applied to a dataset. Where the definition changes and the name does not, a time series that appears continuous is in fact two different measurements plotted on the same axis, and the resulting trend describes an accounting decision rather than a commercial reality.
Behaviour
Definitions drift for entirely ordinary reasons. A new revenue operations leader arrives and reasonably concludes that pipeline should exclude opportunities below a certain probability. A finance team adjusts what counts as a qualified opportunity to align with a new reporting standard. A CRM migration silently changes how stage entry dates are recorded. Each change is defensible and each is implemented without a corresponding change to the metric's name.
What management then observes is a trend. Pipeline coverage improves. Sales cycle shortens. Win rate rises. These improvements are real in the sense that the numbers moved, and entirely artefactual in the sense that nothing about the commercial organisation changed. The improvement is the definitional adjustment, arriving in the data as though it were performance.
The pattern is most damaging where the definitional change coincides with a period of genuine underperformance, because the artefact masks the deterioration. A business whose true win rate declined by four points in a year, while the definition of a qualified opportunity tightened enough to raise the measured win rate by five, reports an improving win rate throughout. Nobody constructed this deception. It assembled itself.
A time series that appears continuous is in fact two different measurements plotted on the same axis.
Evidence
For each headline commercial metric, request the written definition as it stood at the beginning of the reporting period and as it stands now. In most organisations the earlier definition cannot be produced, which is itself the finding: a metric whose historical definition is unrecoverable cannot support a historical trend.
Look for discontinuities. A step change in any commercial metric that coincides with a leadership change, a systems migration, or a fiscal year boundary should be assumed definitional until proven otherwise. Genuine commercial improvement is gradual and messy. Definitional improvement is clean and arrives on a specific date.
Reconstruct one metric independently from raw transaction data across the full period, applying a single consistent definition throughout. Compare the reconstructed series against the reported one. The divergence between them is the accumulated effect of every definitional change nobody recorded, and it is frequently larger than the trend the reported series purports to demonstrate.
Psychology
Nobody changes a definition in order to flatter a number. They change it because the old definition was genuinely imperfect and the new one is genuinely better, and the improvement in the metric that follows is experienced as vindication of the change rather than as an artefact of it.
Announcing a definitional change is also mildly costly and produces no benefit. It requires explaining that prior reporting was flawed, invites questions about what else might be, and complicates a board narrative for no gain. The rational course is to implement the better definition and allow the improved number to speak for itself, which is precisely what happens, thousands of times, across every reporting system in existence.
Commercial Risk
Historical trend is the primary evidence in most commercial diligence, and trend requires definitional stability that is almost never verified. Where a definition has drifted, the trend supporting the investment thesis may be measuring the accumulated preferences of successive revenue operations leaders rather than the trajectory of the business.
The risk is asymmetric in a specific way. Definitional changes that flatter a metric survive, because nobody investigates an improving number. Definitional changes that damage a metric are questioned immediately and frequently reversed. The surviving population of definitional changes is therefore systematically biased toward those that improved the reported figures, and the aggregate effect on a multi-year trend is consistently in one direction.
Investment Committee Note