To measure knowledge retention, test the same critical knowledge at fixed points after training, such as right after the course and again at four, eight and twelve weeks, then compare the scores and link them to on-the-job behaviour. Useful KPIs include retention at 30 and 90 days, topic-level accuracy, application, mistakes and real-world results.

Most training evaluations ask whether participants liked the course. That is useful feedback, but it says almost nothing about what happens to the knowledge afterwards. For HR and L&D, the more relevant question is not “was the training evaluated?” but “what happens to the knowledge?” This guide explains how to answer it.

Why is measuring retention different from evaluating training?

A typical evaluation captures satisfaction and sometimes a final test. Both happen when memory is freshest. Retention measurement looks at the period afterwards, when the real work begins.

The forgetting curve is the reason this matters. Knowledge changes over time, and the speed of that change varies with the topic, the learner and whether the knowledge is used. Without measurement at later points, you simply do not know whether training results survive. A deeper discussion of this nuance appears in the forgetting curve in practice.

What are the main knowledge retention KPIs?

No single number captures retention. A balanced set combines knowledge, behaviour and results.

Knowledge-based KPIs

• Immediate score. Knowledge right after the training. This is your baseline.
• Retention at 30 days. The same or equivalent questions a month later.
• Retention at 90 days. A longer-term check, valuable for critical knowledge.
• Retention rate. Later score divided by baseline score, which shows how much knowledge is maintained.
• Topic-level accuracy. Scores per topic, revealing which content sticks and which fades.
• Improvement after intervention. The change between a pre-intervention measurement and a later one.

Behaviour-based KPIs

• Application. Do employees use the knowledge in their work, as reported by managers or observed directly?
• Errors and incidents. Does the number of mistakes related to the trained topic decrease?
• Procedure compliance. Are steps followed consistently?

Result-based KPIs

• Quality, safety or productivity indicators linked to the training goal.
• Time to competence for new employees.
• Support requests about topics that were trained.

A more detailed list is available in the article on 10 KPIs to measure knowledge retention.

When should you measure retention?

A practical structure uses four to five measurement points:

There is no universal timetable. Choose intervals according to how long knowledge needs to remain available. For fast-changing topics, shorter windows may make sense. For safety-critical knowledge, you may want measurements long after the course.

How should you design retention questions?

Measurement is only as good as its questions. A few principles help.

1. Test critical knowledge. Do not measure trivia just because it is easy to test.
2. Prefer recall and application. Recognition-based questions can overestimate retention, since recognising an answer is easier than producing it. Scenario questions and open prompts give a more realistic picture. See the explanation of retrieval practice.
3. Use equivalent questions. Repeating the identical question can measure memory of the question rather than the knowledge. Alternate wording or scenarios where possible.
4. Keep sessions short. Three to five minutes per measurement keeps participation high.
5. Include feedback. Measurement can double as learning when answers are explained.

What does retention measurement look like in practice?

Here is an illustrative example. A company trains 100 employees on a compliance procedure. The average knowledge score right after training is 90%. After 30 days it is 78%, after 60 days 74%, and after a short retention intervention it rises to 82%. These figures are an example, not research data, but they illustrate what becomes visible:

• the decline between training and day 30,
• the continued but slower decline afterwards,
• the partial recovery after follow-up,
• and the gap that remains, which indicates where additional attention is needed.

Break the numbers down by topic and you might find that one procedure drops sharply while another holds steady. That insight directs repetition to where it is needed, which is the logic of the seventh principle in the seven principles of knowledge retention.

How do you connect retention to business results?

Knowledge scores matter only if they relate to outcomes. To build the link:

1. Define the goal before the training. What should change in work: fewer errors, faster handling, better compliance?
2. Select a metric you already track. Use existing operational data rather than creating new reporting burdens.
3. Compare groups or periods. Look at changes before and after, or between groups with and without follow-up.
4. Be careful with causality. Many factors affect results. Present the connection as supporting evidence, not proof.

The article on measuring training ROI goes deeper into this, and the piece on knowledge retention after training explains what employees remember and why.

What mistakes should you avoid when measuring retention?

• Measuring only participation. Attendance and completion do not show retention.
• Using only recognition questions. They may flatter the results.
• Collecting data without acting. Retention data should lead to adjustments, such as repeating weak topics.
• Comparing without context. A drop in scores means different things for a rarely used procedure and a daily task.
• Overpromising precision. Retention measurement gives indications, not exact predictions.

What does measurement mean for training providers?

For training providers, retention data creates a new kind of deliverable. Instead of an evaluation sheet, the client receives a report showing how knowledge develops over time, and the provider can respond when topics need attention. This helps move the conversation from the cost of a training day to the value of the outcome, a theme explored in you’re not selling a training day.

Frequently asked questions

What is a good retention rate?

There is no universal benchmark. A good rate depends on the topic, the time since training and how critical the knowledge is. Compare against your own baselines and look at trends.

How often should we measure?

Fixed points such as after training and at weeks 4, 8 and 12 are common. Add later checks for knowledge that must last.

Do we need to test everyone?

Ideally yes, because individual results guide personalised repetition. If that is not possible, use representative samples while you build the process.

Can retention be measured without software?

Yes, with surveys or quizzes, but manual processes are hard to scale and to keep consistent. A dedicated platform can automate scheduling and reporting; see the Stiqqo method.

Key takeaways

• Retention measurement looks at knowledge after the training, not just at satisfaction.
• Combine knowledge, behaviour and result KPIs.
• Measure at fixed points, such as after training and at weeks 4, 8 and 12.
• Use recall and scenario questions, and act on the data.

Ready to see how knowledge develops after your trainings? Request a demo.