# Measurement: one useful change in team capability

Measure a useful change in team capability. Choose measurements that connect learning to work. Define them before the pilot so an encouraging result can be interpreted without changing the question after the fact.

Updated October 15, 2026

## Build a small measurement set

Use one primary outcome, one quality safeguard, and a measure of the effort required. For onboarding, these could be time to pass a practical readiness task, accuracy on that task, and manager support time. Completion of a lesson can help explain participation, but cannot establish that someone can perform the job.

## Write operational definitions

Specify when timing starts and stops, which people are included, how missing observations are handled, and who scores performance. Preserve the same task rubric across comparison groups. Report the number of participants alongside any percentage, and show the spread of results when a single average would hide substantial differences between learners.

## Make the comparison understandable

Choose a recent, reasonably comparable baseline and record differences in role, experience, workload, and coaching. If a reliable baseline is missing, spend the first period collecting one. A fictional example of six learners improving is useful for practicing the calculation; it is not evidence that Elvina caused an improvement in a real organization.

## Count effort without overstating savings

Suppose support time falls from four hours to three hours per learner across ten comparable learners. That suggests ten hours of manager capacity released within the measured period. Also record preparation, review, and maintenance effort. Converting capacity into cash savings requires a separate business assumption about how that time will actually be used.

## Report limitations with the result

Include changed staffing, different task difficulty, missing data, and negative outcomes. A small before-and-after pilot usually provides directional evidence rather than a firm causal estimate. In the intended Elvina workflow, these definitions should remain visible beside the result; verify any reporting functionality before relying on its calculations or coverage.