How to measure training effectiveness

The four levels

The most widely used framework is Donald Kirkpatrick's four levels of training evaluation, first described in the 1950s:

1. Reaction: did people find the training useful and engaging?
2. Learning: did they gain the knowledge and skills it aimed to teach?
3. Behaviour: are they applying it at work?
4. Results: did it affect the business outcomes it was meant to improve?

Each level is harder to measure than the one before, and most organisations stop at level 1. The goal isn't to measure everything perfectly; it's to measure enough to make better decisions.

Level 1: reaction

Measure reaction with a short feedback survey at the end of the session, while people are still in the room. Five questions are enough: an overall rating, how useful it will be in their work, a rating of the trainer, a ranking of the topics, and one word to describe how they feel. Run it live in the last two minutes and response rates stay close to 100%. Our training feedback template uses exactly this structure.

Level 2: learning

Learning is where most of the value is, and it's easy to measure well with a pre and post test:

• Write 10 to 20 questions covering the training's goals, and tag each one with a skill.
• Run the quiz before the session as a baseline.
• Run the same quiz at the end, and again a few weeks later to check retention.
• Compare the results per skill and per person.

Scenario questions (“A customer says the product is too expensive. What's the best first reply?”) test application better than recall questions.

Level 3: behaviour

Behaviour change takes weeks to show. Useful signals include a short follow-up quiz 30 days later, manager observations using a simple checklist, and self-ratings of how often people use the new skill. Keep it light; a two-minute check-in beats a long survey nobody completes.

Level 4: results

Link the training to one or two business measures before it starts, such as customer satisfaction scores after service training or error rates after a process training, and compare them before and after. Be honest about other factors that could explain the change.

Putting it in one report

The report leadership wants fits on one page: attendance, the reaction score, the learning gain per skill, who still needs support, and one or two business signals. In QG Playground, the skills report shows a heatmap of people against skills, proficiency bands, and movement from the first run to the latest, and projects follow a batch across every quiz.