Adoption
Part of Customer education for B2B products
Measuring whether training changed product usage
Compare training exposure with eligible task use, check event definitions and other changes, and report what the evidence can support.
Measure the task training was meant to change among people who had reason and permission to perform it. Record relevant use before training and again after a suitable task cycle, then check whether completed work served the customer's purpose. An increase after training merits examination; timing alone does not prove causation.
Define the training question
Write down who received which training, which task should change, and when eligible work could occur. In a hypothetical approvals product, nominated managers might be trained to submit eligible decisions through the agreed route. Attendance records exposure; a valid submitted decision records product behaviour. The operations owner may still need to confirm the result was usable.
Identify the eligible population and unit before calculating a rate. Some trained managers may have no decision to make during the review period. Other active users may have no role in that task.
Decide whether the question concerns people, accounts or task opportunities, and use the same unit and task definition across comparisons. If eligibility or opportunities cannot be identified, report counts with that limitation instead of claiming a completion rate.
Keep exposure and behaviour separate
A measurement record should define:
| Field | What to record |
|---|---|
| Training exposure | Who attended or received the resource, when, and which version. |
| Behaviour | The task-start event and what qualifies as a valid completion. |
| Eligibility | Access, role and an opportunity to perform the task, where observable. |
| Comparison periods | Windows that let the task arise and are reasonably comparable. |
| Data checks | Event definition, duplicate activity, role and account mapping, and known gaps. |
Do not substitute an analytics dashboard's active-user percentage for the task's eligible population. In Amplitude Event Segmentation, for example, Active % uses active users as its denominator, subject to the chart's filters; that may differ from people eligible for the trained task. Guide views, lesson completions and quiz results describe exposure or learning activity, not completed customer work.
Interpret the change
If more eligible managers complete decisions afterwards, check what else changed. Were permissions granted, request volumes different, a defect fixed or a policy revised? Compare the same roles and task definition. Mark a break if instrumentation changed. Ask whether people who did not receive training encountered the same change.
Where a fair comparison is feasible, agree the groups, training timing and outcome before delivery. A later-training group may help, provided both groups had similar opportunities and one group's learning does not readily spill over to the other. A controlled comparison can strengthen an inference about cause.
Make a decision from the evidence
Report eligible people or opportunities where known, the exposed group, recorded starts and completions, the period, the customer's interpretation and the main data gaps. If people attended but did not attempt the task, ask whether it arose and whether access was ready.
If they started but stopped at one step, inspect that step. If use increased but the work still failed its purpose, investigate the workflow or route the fault to its owner.
Review again when another relevant task cycle has occurred. State what changed, what remains uncertain and what action follows.
