Measure more than course completion

Define success before development

Agree on the target task, quality standard and context with the person accountable for the work. Identify the current performance gap and how it is measured. Record a baseline where feasible. If the existing metric is unreliable, improving that measurement may be your first task.

Match the assessment to the behavior

If learners need to diagnose an issue, have them diagnose a new case. If they need to have a conversation, use a role-play with an explicit rubric. Recall questions can assess recall; they are weak evidence for performance on a complex task.

Build the rubric before writing the practice. State observable criteria, acceptable variation and critical errors. Use examples to calibrate reviewers so that two evaluators do not apply entirely different standards.

Collect evidence at the right times

During the experience, collect formative evidence to help learners improve. Immediately afterward, test independent performance. Later, sample real work or follow-up tasks to investigate transfer and retention. Choose the interval based on how often people use the skill.

Report with appropriate caution

Describe the sample, timing, measurement method and limitations. Completion tells you who finished; satisfaction tells you how people felt. Neither proves that behavior changed.

A before-and-after improvement is promising, but staffing, software, policy or incentives may also have changed. Use a comparison group or stronger evaluation design when the decision warrants it, and avoid claiming causation from a simple change in a dashboard.