Insights
Kurt Lewin’s three stages are one of the oldest ideas in change management and still one of the most useful. Here is what unfreeze, change and refreeze mean in practice, the debate about where the model really came from, and a force field analysis template you can use straight away.
By Christopher Fagan, Managing Partner, PEF Consulting Partners
Delivered for and alongside
















Lewin’s change model describes organisational change in three stages: unfreeze, where the current way of working is loosened and people accept it must change; change, where the new way is introduced and learned; and refreeze, where the new way is embedded so it holds. It is usually paired with force field analysis.
The Lewin change model is simple enough to draw on a whiteboard, which is both why it has lasted and why it gets dismissed. Read properly, it makes a point many programmes still miss: you cannot install a new way of working on top of an old one that is still firmly in place.
In this article
Kurt Lewin was a psychologist who left Germany for the United States in 1933. He did foundational work on group dynamics, action research and field theory, the idea that behaviour is shaped by the forces acting on a person or group at a given moment.
The three stages are usually traced to his article Frontiers in Group Dynamics, published in the first issue of the journal Human Relations in 1947, the year he died. There he described a successful change as unfreezing the present level, moving to a new level and then freezing group life at the new level. Lewin wrote of freezing; refreeze is the label most textbooks use today.
You will also see the model called Lewin’s 3 stage model, the three step model or simply the unfreeze change refreeze model. The names differ but the idea is the same.
Scholars disagree about how much of the neat version is really Lewin’s. Writing in Human Relations in 2016, Stephen Cummings, Todd Bridgman and Kenneth Brown argued that Lewin mentioned the three steps only briefly, and that the tidy model taught today was largely assembled and popularised by others after his death. Bernard Burnes and others have defended Lewin’s authorship, arguing that the steps were one part of a wider approach that also took in field theory, group dynamics and action research.
For practitioners, the useful lesson from both sides is the same. The three stages work best alongside the rest of Lewin’s thinking, especially force field analysis, rather than as a standalone recipe.
The aim is to make the current way of working open to change. That means showing people, with evidence they trust, why the current position is not good enough, and involving them in working out what needs to change. It also means dealing with what keeps the current state stable: habits, incentives, reporting lines and the very reasonable fear of looking incompetent while learning something new.
Edgar Schein, who built on Lewin’s work, drew a useful distinction here. People change when the anxiety of staying put outweighs the anxiety of learning something new, and the better way to tip that balance is usually to make learning feel safe rather than to make the status quo feel frightening.
Practical unfreezing usually includes:
This is the move itself: the new process, system, structure or behaviour. People need clear direction, training, support and time, and the programme needs fast feedback so problems are fixed while people are still willing to try. Expect performance to dip here; the change curve describes what that usually looks like.
The new way is made stable so it does not slide back. Policies, job descriptions, performance measures and incentives are updated, old systems and reports are switched off, and ownership passes from the programme to the business. In a fast moving organisation, think of refreeze as making the new way the default rather than fixing it forever.
Force field analysis comes from Lewin’s field theory. Any current situation is held in place by two sets of forces: driving forces pushing for change and restraining forces holding things as they are. When the two are roughly balanced nothing moves, however many announcements are made.
To move, you can strengthen the driving forces, weaken the restraining ones, or both. A point usually traced back to Lewin is that weakening restraints tends to work better, because simply pushing harder raises tension and provokes more resistance. In practice that means removing barriers before adding pressure.
Running one takes an hour or two with the right people in the room:
The completed example below is for a hypothetical utility moving its field engineers from paper job sheets to a mobile app. Copy the structure, replace the rows with your own forces and keep the scoring simple.
| Force | Driving or restraining | Strength (1 to 5) | Action | Owner |
|---|---|---|---|---|
| Late and lost job sheets delay customer billing | Driving | 4 | Show each depot its own billing delays from the last quarter | Finance director |
| Engineers want to stop driving back to the depot with paperwork | Driving | 3 | Measure the time saved in the pilot and share it with every depot | Operations manager |
| Customers asking for photos and proof of visit | Driving | 2 | Share customer feedback from the pilot area | Customer service lead |
| Concern that the app will be used to track individuals | Restraining | 4 | Agree and publish a clear data use policy with staff representatives | HR director |
| Poor mobile signal in rural areas | Restraining | 4 | Test offline working properly before rollout | IT lead |
| Some engineers unsure about using a phone app | Restraining | 3 | Pair them with a peer champion and offer extra hands on sessions | Depot supervisors |
| Supervisor targets reward jobs closed, not data quality | Restraining | 3 | Change the measure for the rollout period | Operations director |
In this example the driving forces total 9 and the restraining forces total 14. However good the app is, the change will stall unless the restraining forces come down, and the two strongest are about trust and practicality rather than technology.
Using the same hypothetical utility, here is how the three stages might run. None of it is complicated; the discipline is in doing all three.
Unfreeze. The operations director shares the billing delay figures depot by depot, and engineers from each depot help choose between two shortlisted apps. The data use policy is agreed with staff representatives before anyone is asked to install anything. The aim is to make the case real and to remove the biggest fear early.
Change. One depot pilots the app for six weeks. Peer champions handle most questions, offline working is fixed after the first fortnight, and supervisor measures are changed for the pilot period. Feedback goes to the supplier weekly, and two small changes are made before the wider rollout.
Refreeze. At each depot, paper job sheets are withdrawn on a fixed date after go live rather than kept as a fallback. Supervisor objectives are rewritten to include data quality, app data feeds the weekly operations review, and new engineers learn the app from their first day.
Lewin’s change model is at its most useful for discrete changes with a clear before and after: a new system, a new structure, a relocation or a new policy. It is also a good way to explain to a sponsor why preparation and embedding matter as much as the launch.
It is less helpful where change is continuous, such as product teams releasing every fortnight, because the idea of freezing anything can sound wrong. It does not tell you how to run each stage in detail. On a large transformation with several overlapping changes, different parts of the organisation will also be in different stages at once.
For a fuller leadership sequence, Kotter’s 8 step change model goes much further, and for individual adoption Prosci’s ADKAR model is more precise. Many practitioners use Lewin as the frame and those models for the detail.
Lewin’s model is a frame, not a plan. Turn the actions from a force field analysis into named, dated tasks in your change management plan, then use the three stages to check that the plan allows enough time for preparation and embedding. For the bigger picture, see our guide to what change management is.
Separations show the model at its most literal. Christopher Fagan was Project Director for the National Grid separation, a programme touching more than 22 million managed assets. A separated organisation only stands on its own once its own systems, contracts and ways of working are in place and the old dependencies are switched off, which is refreeze in its plainest form. Our carve out and separation playbook covers that side in more depth.
Our change management and adoption team works alongside your people through all three stages, not just the launch. That includes the refreeze work that many programmes leave until after the team has gone.
The three stages are unfreeze, change and refreeze. Unfreezing makes the current way of working open to change by showing why it cannot continue and dealing with what holds it in place. Changing introduces the new process, system or behaviour. Refreezing embeds the new way through policies, measures, incentives and switching off the old way, so it does not slide back.
Force field analysis is a planning tool based on Kurt Lewin’s field theory. You list the driving forces pushing for a change and the restraining forces holding the current situation in place, score each for strength, and plan actions to strengthen the drivers or weaken the restraints. Weakening restraints usually works better than adding pressure, which tends to increase resistance.
Partly. Lewin described unfreezing, moving and freezing in his 1947 article Frontiers in Group Dynamics, published the year he died. Some scholars, notably Cummings, Bridgman and Brown in 2016, argue the neat three step model was largely assembled and popularised by others afterwards. Others, including Bernard Burnes, defend it as genuinely Lewin’s and part of a wider body of work.
Yes, for changes with a clear before and after, such as a new system, structure or policy. Its core point, that the old way must be loosened before a new way will hold, is still widely ignored. It is less suited to organisations that change continuously, where refreezing is better read as making the new way the default rather than fixing it in place.
Lewin’s model is a simple three stage frame for any change: loosen the old way, make the move, then embed the new way. Kotter’s eight steps are a more detailed leadership sequence, covering urgency, the guiding coalition, vision, communication, barriers, short term wins, momentum and culture. Kotter’s steps can be read as a detailed version of Lewin’s three stages.
Related reading: Kotter’s 8 step change model, The ADKAR model, The change curve and What is change management?.
We help organisations prepare for change, make the move and embed it so it sticks, working alongside your team rather than handing over a plan. Tell us what you are changing and we will give you a straight view of what is holding it back.
Book a confidential conversation
Sources: Kurt Lewin, Frontiers in Group Dynamics, Human Relations, 1947; Stephen Cummings, Todd Bridgman and Kenneth G. Brown, Unfreezing change as three steps: Rethinking Kurt Lewin’s legacy for change management, Human Relations, 2016; Bernard Burnes, Kurt Lewin and the Planned Approach to Change, Journal of Management Studies, 2004
Written by Christopher Fagan
Managing Partner, PEF Consulting Partners
Christopher has led systems and transformation delivery for 19 years, including the National Grid separation, the £9bn Energy Bills Support Scheme and the BP Pulse EV charging rollout.
Where next