Kolbs Experiential Cycle
A four-step write-up after one attempt that turns what happened into a changed next attempt.
Kolbs Experiential Cycle
Kolbs is a write-up done after one attempt. It turns what happened into a changed next attempt. The four steps name that movement in ordinary words: what happened, what it meant, what rule that suggests, and what changes next time. The product is a different next attempt. A better understanding of the last attempt is not the product on its own.
Core takeaways
- Kolbs is a write-up after one attempt. Its product is a changed next attempt.
- The four steps are experience, reflection, abstraction, and experimentation. Each step uses the one before it.
- Reflection collects information about the process. Rumination repeats feelings and causes without a next change, and it blocks action.
- The abstraction is a hypothesis read from the reflection. The experiment tests it, and that experiment becomes the next cycle’s experience.
- Early cycles can take hours. With practice they take fifteen to twenty minutes, and every cycle is capped at thirty minutes.
- Several cycles are needed on one problem. The cycles stop when the problem goes away.
The four steps
Experience. The experience is one line: what was tried, and how it went. The data has to be real, and it has to describe a process rather than an outcome. “The test went badly” is a result. The experience is sitting the exam or preparing for it. There is one exception: when the reaction to the result is what needs understanding, the reaction is the experience. The line is specific, isolated, targeted, and recent. A whole week or a whole subject cannot be reflected on at the depth the later steps need. So a big experience is split into its processes, and they are run one at a time, starting with the process most likely to produce an improvement.
Reflection. Reflection collects information about the process. It is different from rumination. Rumination is repeated focus on feelings and causes without a next change; it predicts worse mood and blocks action. Reflection also fails in two other ways. Done as a checklist, it collects too little. Done at length, it can still only restate feelings and leave the process unexamined. The kind of information collected decides the kind of repair produced, so one-sided input produces a confident repair to the wrong thing. A feeling, such as discomfort, anxiety, or the sense that a step was harder than it should be, is often easier to recall than the process itself. The feeling points to a moment in the process, and that moment is what gets examined. A fixed question set, answered in full including the questions that feel repetitive, reduces the tendency to collect only what supports the explanation already held. The Kolbs Template in this vault is that set.
Abstraction. Abstraction is a hypothesis built from observation: a short prediction taken from the reflection, and probably not yet right. It comes from the reflection, and not from what was already believed. If no pattern appears, the reflection was thin; go back and add to the reflection instead of writing more under abstraction. When a cycle feels vague, one residue question from Red Teaming is enough: why did this happen, and what deeper condition made that possible? Repeat the question only until the answer is something that can be acted on. The first explanation is usually not the useful one.
Experimentation. Experimentation designs the next attempt from that hypothesis. “Try harder” changes nothing. It is allowed once, the first time a problem is noticed. After that, something concrete changes: the environment, the pacing, the order, the number of pieces, or the trigger. Plan one to three experiments, ideally one or two, because fixing one root issue commonly resolves several others. The exception is a one-off change that costs no effort and stays fixed once made. This step fails in four ways: no next action; “try harder” a second time; too many experiments; an experiment connected to nothing the abstraction found. When an experiment goes nowhere, the usual cause is an earlier step: the abstraction was not built from observation.
What to run a cycle on
A cycle needs a real attempt that produced data. The attempt does not need to be dramatic. It needs enough friction, emotion, result, or surprise to teach something. The first line of the write-up is what was tried and how it went. Deliberately testing something expected to fail is a valid reason to run a cycle.
| Good subject for a cycle | Poor subject for a cycle |
|---|---|
| A session that started slowly | Abstract planning with no attempt |
| A session that went better than expected | Venting with no next experiment |
| A technique that looked correct and did not work | Too many variables at once |
| A moment where a [[wiki/Concepts/The Shortcut Problem | shortcut]] appeared |
| A repeated failure mode | A long reflection where a quick note would do |
| A meaningful conversation | |
| A workout that revealed a bottleneck | |
| A language session where attention rose or collapsed | |
| An agent workflow that produced a useful or a bad result |
An established script that keeps failing is a different subject from a single attempt. That subject belongs to How to Unlearn Old or Bad Habits Efficiently. A session in which a weak frame showed is a suitable experience for Interleaving for Complex Problem Solving.
When to run a full cycle and when to write a quick note
| Full cycle | Quick note |
|---|---|
| Emotional weight, repeated friction, or a real performance lesson | The lesson is obvious and the next experiment is already clear |
| Repeated shortcut; strong emotional reaction; technique failure; relationship repair or conflict | Minor logistical issue; simple reminder; one-off low-stakes slip; obvious next action |
Run the cycle at the end of the day where possible, and at least twice a week. Never compress a week into one cycle, because each session is a separate experience with its own variables.
An example of one cycle, run on a language session:
Experience: a language session, attention held for twenty minutes and then collapsed.
Reflection: collapse arrived when the next item needed a word that would not come; frustration, then a switch to easier items.
Abstraction: switching down protects the session clock and prevents the retrieval that would have rebuilt the missing word.
Experiment: on the next miss, stay with the item for one more retrieval before switching.
Next cycle's experience: that stay-with-it attempt, and how it went.Which problem gets the cycle
A cycle cannot be run on every problem every day, and a beginner has dozens of candidate directions. So something has to choose which problem gets the thirty minutes. Marginal Gains chooses the improvement that is worth making next. Kolbs runs the cycle that makes that improvement. Together they do four jobs: Marginal Gains stops improvement from being random, and it names the target; Kolbs stops reflection from staying passive, and it improves the method. The few areas that get time first under Priority 0+1 are the ones where a cycle is worth its half hour. Kolbs belongs under Self-Management and is used as the Self-Regulation loop.
Why one write-up is not enough
The unit of the method is a run of cycles. A single write-up is one part of that run. The experiment planned in one cycle is the experience recorded in the next. The abstraction is a hypothesis and the experiment is its test, and the first test usually does not fix the problem. Several cycles are normally needed to see what the problem is, and more to repair it. The same hypothesis is retested until the problem goes away. Stop when the problem goes away. Do not stop at a set number of cycles.
The common failure looks like this: a cycle every day, each step done correctly, a fresh problem every time, and nothing building from one cycle to the next. The next experiment also gives cleaner feedback, but only when it comes back as the next day’s experience. A typed log that is kept makes the previous cycle findable, and that is what lets one cycle connect to the next. The Kolbs Template already asks whether this cycle reflects on an experiment from a previous one.
Where each step fails
An experiment that goes nowhere usually points to an abstraction that was not built from observation. A weak step limits every step after it.
| Step | What the failure looks like | What fixes it |
|---|---|---|
| Experience | A week in one sentence; an outcome in place of a process; reflection with no attempt behind it | One recent process, the most useful process first |
| Reflection | Repeating feelings with the next attempt unchanged; a checklist; long and still only feelings | Examine the process, then write a hypothesis and an experiment |
| Abstraction | Observations with no principle; a guess from theory; more reflection written under the abstraction heading | Name the pattern the record shows, or go back and add to the reflection |
| Experimentation | No next action; “try harder” a second time; too many experiments; unconnected to the findings | One or two concrete changes drawn from the hypothesis |
How long it takes and what to expect
Every step depends on self-awareness, and that limits the method: limited self-awareness limits what can be seen as worth reflecting on. Reflecting feels tedious or confusing while that awareness is still small. Awareness grows over weeks of repeated sessions, about four to twelve weeks, and it does not grow in one sitting. That skill is Metacognition: The Control Layer.
Early cycles can take hours. With practice they settle at fifteen to twenty minutes. Cap a cycle at thirty minutes, leave steps unfinished, and let speed come with practice. A capped cycle done regularly produces more than a complete cycle done rarely. The cap keeps the practice going while awareness is still growing. It does not measure how good the cycle was.
A working cycle goes from what happened, to why it happened, to what changes next. Four signs show that a cycle is working: the experience gets more specific as it is written; emotions and triggers become usable information; a pattern appears instead of a single event; the next experiment is smaller and clearer than the problem that started it. The first cycles diagnose the problem and do not fix it. Several cycles are normally needed before a repair works.
Stop when the write-up records the experience and does not change the next attempt. At that point it has become journaling. Where insight into one’s own process is still small, the write-up stays shallow and produces little. The repair is to build metacognition. More cycles do not repair it.
The four steps are a routine for practice. They are not a claim that the mind moves through four stages in order. Critics say the cycle is too linear and gives too little weight to social context.
What the write-up trains
After long practice the loop runs during the experience itself. The noticing happens live, and the write-up gets shorter. Across cycles, the writer gains a clearer diagnosis, a smaller next experiment, better awareness of mistakes and conditions, a more accurate model of cause and effect, and a practical next action. The write-up is a support used while that way of noticing is being trained. The end state is running the loop with no write-up at all.
How to practice this
- At the end of the day, write one line: what was tried, and how it went. Notice whether the line describes a process or only a result. “The test went badly” is a result; sitting the exam or preparing for it is the experience.
- Answer the Kolbs Template questions in full, including the ones that feel repetitive. Notice when a feeling comes to mind before the process does. Follow the feeling back to the moment in the process and write that moment down.
- Write one short prediction that the reflection supports. Notice whether it came from the reflection or from what you already believed. If no pattern appears, go back and add to the reflection.
- Plan one or two concrete changes for the next attempt: environment, pacing, order, number of pieces, or the trigger. Notice whether the plan only says “try harder”. That is allowed once, the first time a problem is noticed.
- Stop at thirty minutes, even with steps unfinished. Notice whether the next experiment is smaller and clearer than the problem you started with.
- In the next cycle, record that experiment as the experience. Notice whether the same problem is still there. Retest the same hypothesis until the problem goes away.
Related pages
- Upgrading Your Dimensions: dimensions as the things a run of cycles is aimed at upgrading.
- Mindset. A mindset trains on the same loop: name the problem it answers, meet that problem often enough to practise, and arrange the encounter deliberately when it is rare.
Open questions
How many cycles on one hypothesis before the hypothesis itself is the thing to change.
Sources
- Kolb, D. A. 1984. Experiential Learning: Experience as the Source of Learning and Development. Prentice-Hall. Also 2015 2nd ed. Four-stage cycle this version is built on.
- Kolb, D. A. & Fry, R. 1975. Earlier statement of the cycle.
- Nolen-Hoeksema, S. 1991; 2000. Journal of Abnormal Psychology. Rumination as repetitive focus on symptoms and causes without problem-solving.
- Pashler, H., McDaniel, M., Rohrer, D. & Bjork, R. 2008. Psychological Science in the Public Interest 9(3), 105–119. With Coffield, F., Moseley, D., Hall, E. & Ecclestone, K. 2004, and Newton, P. M. & Miah, M. 2017, Frontiers in Psychology 8:444. Why a learning-styles half is not taught here.
- Bergsteiner, H., Avery, G. C. & Neumann, R. 2010. Studies in Continuing Education 32(1), 29–46. With Jarvis, P. 1987, and Holman, D., Pavlica, K. & Thorpe, R. 1997, Management Learning 28(2). Linearity and social-context critiques.
- Schenck, J. & Cruickshank, J. 2015. Journal of Experiential Education 38(1). Update the cycle without the styles.
- Schön, D. A. 1983. The Reflective Practitioner. Reflection-in-action as the live end state.
- Gibbs, G. 1988. Learning by Doing. Six-stage teaching cousin.