Kolb's experiential learning cycle

The American educational theorist David A. Kolb set out the cycle in full in his 1984 book Experiential Learning: Experience as the Source of Learning and Development, building on earlier work with Roger Fry in the 1970s. Kolb was not inventing from scratch. He was synthesising Kurt Lewin's action research, John Dewey's writing on experience and education, and Jean Piaget's account of how understanding develops through interaction with the world.

The reason it has outlived most training models is that it explains a failure everyone has seen: a team goes through something significant, everyone agrees it was useful, and six weeks later nothing about how they work has changed. The cycle tells you exactly where that team stopped.

The four stages of Kolb's learning cycle

The stages feed each other. Active experimentation produces a new concrete experience, which is why Kolb drew it as a spiral rather than a closed loop. Each pass should leave you somewhere slightly different from where you started.

Concrete experience: it has to be real enough to react to

The first stage is direct, unmediated involvement. Not a case study about someone else's problem. Something the learner is inside, with an outcome they care about.

The test is emotional, not intellectual. If nobody feels mildly uncomfortable, irritated, competitive or exposed, the experience was a demonstration and there will be nothing much to reflect on. This is the stage most corporate training gets wrong: a simulation so safe and so scripted that it produces no genuine reaction, followed by a debrief that has nothing to work with.

• Give the group a real constraint: a deadline, incomplete information, a conflicting objective between two people.
• Let the thing go wrong. Rescuing a group mid-task protects your session and destroys the raw material.
• Make sure everyone is a participant. Observers get a weaker experience and, later, a weaker insight.

Reflective observation: slow down before explaining

Reflection is where most cycles collapse, usually because the facilitator rushes to interpretation. Someone offers a neat conclusion in the first two minutes, the room agrees, and the other eleven observations in the room never get said.

Ask what happened before you ask why it happened. The moment a group starts explaining, it stops noticing.

Good reflection is deliberately multi-perspective. The person who took charge and the person who went quiet were in the same event and had entirely different experiences of it. Both accounts are data.

Questions that open reflection

• What happened, in order, with no interpretation attached?
• What was the moment things changed direction, and what was said just before it?
• What did you decide not to say, and what stopped you?
• What did that look like from the other side of the table?

This stage depends entirely on whether people will say true things out loud, which makes psychological safety a precondition rather than a nice-to-have. It also depends on active listening from the facilitator: the useful observation is often the half-sentence someone mutters and abandons.

Abstract conceptualisation: the stage that gets skipped

Here the group converts what it noticed into something portable. Not "we should communicate better" but "when two people hold conflicting information, the one with less authority waits to be asked, and nobody asks." A rule you could carry into a different room.

Kolb's cycle is often misread as an argument that doing beats reading. It is not. Abstract conceptualisation is a full quarter of the model, and it is where theory belongs. Introduce a framework here, after the experience, and people have somewhere to attach it. Introduce it first and it is trivia.

1. Ask the group to name the pattern in one sentence, then push until the sentence is specific enough to be wrong.
2. Offer the relevant model or research at this point, not at the start.
3. Write the conclusions down. Verbal insight evaporates by the following Tuesday.

Active experimentation: no commitment, no cycle

The last stage is testing the new rule where it costs something. Without it, a session produces articulate people who behave exactly as they did before.

• Convert each conclusion into a named behaviour, an owner and a first occasion: who does what, in which meeting, this week.
• Choose experiments small enough to run before the memory fades. One changed behaviour that survives beats five that were never attempted.
• Book the review in the same breath as the commitment. An experiment nobody checks is a wish.

The follow-up matters more than the session. When you review, you are not marking homework, you are starting the next turn of the spiral: what happened when you tried it becomes the new concrete experience.

Kolb's four learning styles

Kolb built two axes into the model: how people take hold of experience (concretely or abstractly) and how they transform it (by reflecting or by acting). Combining them gives the four styles he described in his Learning Style Inventory.

Read as a description of habit, this is genuinely useful for facilitators. It explains why the person who wants the framework before the exercise and the person who wants to start immediately are both frustrated by the same session, and it argues for running all four stages rather than the one you personally enjoy.

Honey and Mumford's variant

In the 1980s Peter Honey and Alan Mumford reworked Kolb's styles for a management audience, renaming them Activist, Reflector, Theorist and Pragmatist — one temperament per stage of the cycle. An Activist wants the task, a Reflector wants time, a Theorist wants the model, a Pragmatist wants to know what it changes on Monday. The labels travel better in a workplace than Kolb's originals, which is why most corporate questionnaires descend from Honey and Mumford rather than from the Learning Style Inventory itself. The same caution applies to both: use the styles to check your session serves all four appetites, not to sort people into boxes.

Where the model breaks down

1. The learning-styles half has not held up well. The psychometric quality of Kolb's inventory has been criticised for decades, and reviews of learning-styles research have repeatedly failed to support the matching hypothesis, the idea that teaching someone in their preferred style improves what they learn. Use the styles to widen your design, never to sort people or to justify giving someone only the quarter of the cycle they like.
2. Real learning is not a tidy circle. Donald Schön's work on reflection-in-action describes practitioners reflecting mid-task rather than afterwards, and experienced people routinely run several loops at once at different scales. The four stages are a checklist of what must be present, not a queue.
3. It is a model of an individual. The cycle says nothing about the group the learner sits in: whose reflection is treated as authoritative, who is safe enough to admit a mistake, what the organisation punishes. In practice those factors decide whether stage two happens at all.
4. Experience is not automatically educational. Dewey made this point before Kolb did. Badly designed experience teaches the wrong rule with great conviction, and a confident wrong rule is harder to shift than ignorance.

Held lightly, the cycle still earns its place, because it diagnoses the two most common training failures in one glance. Experience without reflection is a nice day out. Concepts without experimentation are a well-attended lecture. Both are common, and both are expensive.

Running a full cycle in a single session

1. Set up (short). Rules and constraints only. Save the theory.
2. Run the experience. Long enough for the group's real habits to appear rather than their best behaviour.
3. Reflect. Facts first, then perspectives, then feelings. Get every voice in before anyone concludes anything.
4. Conceptualise. Name the pattern, test it, introduce the framework that sharpens it.
5. Commit. One behaviour each, with an occasion and a date.
6. Review. A fortnight later, ask what happened. That conversation is the next cycle beginning.

Roughly speaking, the experience should take less time than everything that follows it. Sessions that invert that ratio feel excellent and change nothing.

A worked example: one full turn of the cycle

Picture a project team in a forty-minute simulation: they must deliver a build under a deadline while briefing a second team that holds half the information. It goes badly — the brief is rushed, the second team builds the wrong thing, and with five minutes left both sides are quietly blaming each other. That is the concrete experience, and the irritation in the room is the raw material.

In reflective observation the facilitator walks the group back through the timeline. The handover took ninety seconds of a forty-minute exercise. The receiving team asked one question and got a hurried answer, so they stopped asking. In abstract conceptualisation the group names the rule: under time pressure, we cut the conversation that decides whether the time is wasted. Someone connects it to last quarter's handover to the support team — the same shape. In active experimentation the team leaves with one change: every handover this month gets a ten-minute briefing with questions before work starts, owned by the person handing over. The review two weeks later, on real handovers, begins the next turn of the spiral.

Using the cycle beyond the training room

Nothing in the model restricts it to workshops. It is a design check for any moment where you want experience to leave a residue:

• Retrospectives. A retrospective is stages two and three run on a sprint's worth of experience; it fails exactly where the cycle predicts, when actions are named but never tested.
• Coaching. The GROW model's Reality–Options–Will sequence walks a coachee through reflection, conceptualisation and experimentation on their own experience.
• Onboarding. New joiners accumulate concrete experience at a ferocious rate; a weekly review that asks what they noticed and what rule they are forming turns onboarding into an accelerated cycle instead of a firehose.
• Debriefs. Every debrief is the middle of the cycle made explicit — which is why an activity without one teaches so little.

This is the whole argument for experiential learning as a method rather than a format. Put a team into a short, real challenge and you get all four stages in an afternoon: the behaviour appears under pressure, the group sees it together, they name the rule themselves, and they leave with something to try on Monday. Watch which stage a team resists and you learn something else useful. Groups that will not reflect are usually telling you about the safety of the room, not about the exercise.

Tour De Force runs this as live, experiential training for teams worldwide — online and in person. Talk to us, or play The Weekly Challenge from our gamified learning apps to see the method in ten minutes.