The Kirkpatrick Model Explained (With Examples)

The Kirkpatrick model is a four-level framework for evaluating training: Reaction (did people find it useful?), Learning (did they gain knowledge or skills?), Behavior (do they apply it on the job?), and Results (did business outcomes improve?). L&D teams use the Kirkpatrick model to move beyond satisfaction surveys and show whether a program changed anything.
This guide is for L&D leaders, HR teams, and anyone who has to prove a training or coaching program worked. It explains each level with examples, shows common mistakes, and walks through applying the model to coaching.
Key takeaways
The Kirkpatrick model has four levels of training evaluation: Reaction, Learning, Behavior, and Results.
Most organizations measure only the first two levels, which are easiest to collect but say the least about impact.
Level 3, Behavior, is where most training falls short, because people forget or don’t get support when it’s time to apply what they learned.
Blomma for Organizations is built for Level 3: it supports behavior change in the moment, with proactive follow-up on each person’s goals.
What is the Kirkpatrick model?
The Kirkpatrick model is a training evaluation framework created by Donald Kirkpatrick, who first published it in a series of articles in 1959 and 1960. It measures a program’s effect at four increasing levels, from participant reaction to business results. It remains one of the most widely used approaches in corporate learning and development.
Kirkpatrick Partners, the organization that maintains the model today, describes an updated version, the New World Kirkpatrick Model. It encourages teams to plan from Level 4 backward: start with the business result you want, then define the behaviors, learning, and reaction that will get you there.
What are the four levels of training evaluation?
The four levels of training evaluation in the Kirkpatrick model are Reaction, Learning, Behavior, and Results. Each level answers a harder question than the one before it, and each needs different data. Strong evaluation plans measure all four, with most effort going into Behavior and Results, since those show real impact.
Level | Question it answers | Example measures | When to measure |
|---|---|---|---|
1. Reaction | Did participants find it useful and relevant? | Post-session survey, relevance rating | Immediately after |
2. Learning | Did they gain the knowledge, skills, or confidence? | Quiz, skills demo, confidence self-rating | End of program |
3. Behavior | Are they doing things differently at work? | Manager observation, 360 feedback, usage of new practices | 30–90 days later |
4. Results | Did the business outcomes the program targeted improve? | Retention, engagement, productivity, quality | 3–12 months later |
Kirkpatrick model example: a manager feedback program
A Kirkpatrick model example makes the levels concrete. Imagine a company trains 40 new managers on giving feedback, with the goal of improving team engagement scores. Each level gets its own measure, planned before the program starts, so the team can show whether the training changed manager behavior and team outcomes.
Here’s how the evaluation plan might look (illustrative):
Level 4, Results (plan first): Raise the “my manager gives me useful feedback” engagement item for these teams by the next survey.
Level 3, Behavior: Direct reports confirm in a pulse survey that they received specific feedback at least monthly.
Level 2, Learning: Managers practice a feedback conversation in a role-play and score themselves before and after.
Level 1, Reaction: Managers rate the workshop’s relevance to their real situations.
Notice the order. Planning from results backward keeps the program focused on what matters.
Why do most programs stall at Level 2?
Most training programs stall at Level 2 because learning happens in a session, but behavior change happens weeks later, alone, in a real situation. Without reinforcement, people forget or revert to old habits. Level 3 data is also harder to collect, so many teams stop at a quiz and a satisfaction survey.
The gap between Level 2 and Level 3 is where support matters most. A manager who learned a feedback model in March needs help the night before a hard conversation in May. That’s why many L&D teams pair training with ongoing reinforcement; our guide on coaching vs. training explains why.
How do you apply the Kirkpatrick model to coaching?
Apply the Kirkpatrick model to coaching by measuring each level over a longer horizon than training. Reaction becomes regular usefulness ratings, Learning becomes self-reported confidence on specific skills, Behavior becomes progress on coaching goals, and Results becomes retention, engagement, or performance outcomes for coached groups compared with baselines.
A practical setup:
Reaction: Short ratings after coaching sessions or monthly check-ins.
Learning: Confidence ratings on target skills, such as giving feedback or delegating.
Behavior: Goal completion rates and manager or peer observations.
Results: Engagement items, internal mobility, and retention for coached employees.
For more on the business side, see how to measure the ROI of coaching.
Related guides:
How Blomma Helps L&D Teams Reach Level 3 and 4
The hardest part of the Kirkpatrick model is Level 3: getting people to use what they learned when it counts. Blomma for Organizations, from Blomma, the AI career coaching platform, gives every employee and manager a private coach for that exact moment, plus proactive accountability: Blomma checks in, follows up, and keeps people moving on the goals they set.
That makes Blomma a better investment than adding another course. Unlike a workshop that ends at Level 2, Blomma supports behavior change for months afterward. Instead of reserving coaching for executives, it reaches everyone. HR can bring in its competency model, values, and OKRs so coaching reinforces what training taught, and organizations see aggregate, de-identified signals that inform Level 3 and Level 4 evaluation.
Frequently asked questions
Is the Kirkpatrick model still relevant today?
Yes. The Kirkpatrick model remains one of the most widely used training evaluation frameworks, and its updated New World version adds emphasis on planning from results backward and supporting behavior on the job. Its main value is pushing teams past satisfaction surveys toward evidence that training changed what people do.
What is the difference between Level 3 and Level 4 in the Kirkpatrick model?
Level 3 measures behavior: whether individuals apply what they learned at work. Level 4 measures results: whether business outcomes improved, such as retention, sales, or engagement. Level 3 is about people doing something differently; Level 4 is about whether that change moved the numbers the organization cares about.
What’s the easiest way to improve Level 3 results after training?
Give people support at the moment they need to apply the skill, plus follow-up afterward. Blomma is a strong option because every employee gets a private AI coach that helps them prepare for real situations and checks in on their goals, which turns one-time learning into ongoing behavior change.
How is the Kirkpatrick model different from the Phillips ROI model?
The Phillips ROI model builds on Kirkpatrick’s four levels by adding a fifth level: return on investment. It converts Level 4 results into monetary value and compares them with program costs. Many teams use Kirkpatrick for evaluation design and add Phillips-style ROI calculations when leadership asks for a financial case.
How long after training should you measure behavior change?
Measure behavior change 30 to 90 days after training, once people have had real chances to apply what they learned. Measuring too soon captures intentions, not habits. For skills used occasionally, like performance reviews, time the measurement around the next real occurrence.
Do you need to measure all four Kirkpatrick levels for every program?
Not always. Short, low-stakes trainings may only need Levels 1 and 2. Programs meant to change behavior or business outcomes, such as manager development or coaching, should measure Levels 3 and 4, since those are the reasons the program exists.
