The ADKAR Model Explained: Five Steps of Individual Change

The ADKAR model lists five things a person needs before a change at work becomes their normal way of working: Awareness of why it is happening, Desire to take part, Knowledge of how to change, Ability to do it in practice, and Reinforcement that keeps it in place. Prosci, the US firm behind it, treats the five as a sequence and uses the model to find the first one a person lacks.¹

Finding that first gap is the practical use of the model. When a new system goes live and half the team still works around it, ADKAR asks which of the five each person is missing, so the help matches the gap, and Prosci’s advice is to focus support on the earliest unmet element.¹ The model belongs to the people side of project management, where a project can deliver on time and still fail because nobody uses what it built.

What the ADKAR model is, and who built it

The ADKAR model is Prosci’s framework for individual change. It rests on the idea that an organization changes only when its people do, one person at a time, and it names the five outcomes each person has to reach. Prosci describes it as a model for individual change and adds that it is not a complete change management methodology.¹

Jeff Hiatt, a former Bell Labs engineer and program manager, founded Prosci in 1994.² By the company’s account, he developed ADKAR in 1996, first published it in a 1999 white paper called The Perfect Change, and released the first book on it in 2006 after nearly a decade of research by the firm.³ Prosci, based in Colorado, sells certification, training and consulting built on its methods.² The company that wrote the model is therefore also the main source of claims that it works, and most evidence for ADKAR comes from Prosci’s own surveys, a point the evidence section returns to.

The focus on individuals follows from Prosci’s argument that organizational change is the accumulation of many individual changes.³ A company can switch on a new expense system on a Monday. The change has happened only when each employee files their next claim in the new tool instead of emailing a spreadsheet to finance.

Write your change plan in terms of what specific people will do differently, and by when. A launch date tells you when the system is ready; it says nothing about when the people are.

What are the 5 stages of ADKAR?

The five ADKAR stages are Awareness, Desire, Knowledge, Ability and Reinforcement. Prosci defines them as understanding why the change is needed, choosing to take part in and support it, knowing how to change, putting the new skills and behaviors into practice, and keeping the new way going until old habits stop pulling people back.¹

1. 1Awarenesswhy the change is needed
2. 2Desirechoosing to take part
3. 3Knowledgehow to change
4. 4Abilitydoing it in real work
5. 5Reinforcementkeeping it from slipping

The order carries the logic. Training someone who does not see why the change matters teaches steps they have no reason to use. Rewarding a behavior that people cannot yet perform rewards nobody. Prosci calls the elements sequential, though not always linear: people move at different speeds and can drop back, for instance when something learned in training weakens their Desire.⁴

Two pairs are easy to confuse. Awareness is knowing that a change is coming and why; Desire is a personal decision to go along with it, and a manager cannot order it into being. Knowledge is knowing what to do; Ability is doing it under real conditions. A call-center agent can pass the course on a new customer database and still freeze when a caller is waiting and the screen layout is unfamiliar. Coaching at the desk and practice on real cases are aimed at that gap, while another classroom session would repeat what the agent already knows.

The table pairs each element with what it tends to look like when missing, which are our own examples, and with the kind of support a 2026 case study from New South Wales, Australia, matched to it when cancer services rolled out a new patient-reported outcomes system.

How do you use the ADKAR model to find where a change is stuck?

Prosci’s method is to rate each of the five elements for a person on a scale from 1 to 5, then treat the first element scored 3 or less as the barrier point and work on it before anything later in the sequence. Prosci also suggests repeating the assessment at several milestones, to see whether a barrier is moving forward, staying put or slipping back.⁴

1. Awareness: understands why the change is needed
2. Desire: chooses to take part and support it
3. Knowledge: knows how to change: here the first gap, the barrier point
4. Ability: can do the new work in practice
5. Reinforcement: keeps the new way and does not slip back

The scores can come from a short conversation or a few written statements per element, such as “I understand why we are changing” for Awareness. A 2018 survey of 320 nurses in hospitals in Tehran, Iran, used four statements per element to gauge readiness for a revised nursing record, a format a team lead could copy.⁶

A worked example: a new ticketing system

In this invented example, a 12-person IT support team moves from a shared email inbox to a ticketing tool. Six weeks after launch, email requests are still arriving and some tickets sit untouched. The team lead asks three people to rate themselves from 1 to 5 on each element.

• A senior technician scores Awareness 4, Desire 2. The barrier is Desire: the tool feels like extra clicks with no benefit to him. More training will not move him. A conversation about what he gains, such as no longer losing requests when a colleague is on leave, has a better chance.
• A newer team member scores high on the first three and 2 on Ability. She knows the steps but cannot work the queue fast enough on busy mornings. Sitting beside her for an hour, or fixing a slow view in the tool, targets the real gap.
• The fastest responder scores 4 or 5 on everything except Reinforcement, at 2. She uses the tool, but the team lead still answers requests sent by email, so the old route keeps working. The fix sits with the lead: stop accepting email requests and show the team the weekly count of resolved tickets.

Each person is stuck at a different point, so a single team-wide refresher course would have helped one of them at most. Solving the newer member’s ability problem for the senior technician, who is stuck at Desire, would only produce a trained person who still does not want to use the tool.

When most of a team shares the same barrier point, treat it as a team problem. Prosci reports that participants in its practitioner surveys consistently name a lack of Awareness as the main reason people resist change, which makes “do people know why?” a sensible first question.⁴

Find the ADKAR barrier point

• Name the change in terms of what each person will do differently
• Ask each person to rate Awareness, Desire, Knowledge, Ability and Reinforcement from 1 to 5
• Find the first element scored 3 or less for each person
• Choose support that fits that element, not the one after it
• Look for a barrier most of the team shares
• Repeat the ratings at the next milestone

What does the evidence say about ADKAR?

Evidence for ADKAR is thin and mostly comes from Prosci. The firm’s own practitioner surveys link good change management in general to projects meeting their objectives, and independent studies are mostly single case reports and one-off surveys. We found no controlled study comparing ADKAR with another approach.⁷⁸

Prosci’s best-known figure pools several editions of its Best Practices in Change Management surveys, which ask practitioners to rate their own change management and their project’s results.⁹

The study

Limited evidence

Seven times the success rate? Prosci's pooled practitioner surveys (reported 2023)

Of respondents who rated their change management as excellent, 88 percent said their project met or exceeded its objectives, compared with 13 percent of respondents who rated it poor. Prosci presents this as a sevenfold difference.⁷

The survey measures change management in general, not ADKAR, and respondents graded both their own change work and their project’s results, so a team pleased with its project may also rate its methods highly. Some comparison groups were small; in one chart of US respondents the poor group had 5 people.⁷ The figure suggests that attention to the people side goes with better outcomes; it cannot show that ADKAR caused them.

Independent research is scattered. A search of the OpenAlex research index in November 2026 found about 140 works with ADKAR in the title, many of them describing its use in a single project.⁸ The Tehran nurse survey measured readiness at a single point, drew on a convenience sampleconvenience sample: A group of study participants chosen because they were easy to reach, such as students, app users or online volunteers, rather than drawn at random from the population of interest. Results from such samples may not hold for other people.Full entry in the glossary and did not follow whether the revised record was used.⁶ The New South Wales team described its rollout in a retrospective case studycase study: A detailed look at one organization, group, project or event, often with interviews and documents. It can show how something happened in context, but with no comparison group it cannot show that a change caused the result or that it would work elsewhere.Full entry in the glossary and reported success; a case study of one rollout has no comparison group to show what would have happened without ADKAR.⁵

ADKAR shares this gap with most popular change models. Jeroen Stouten, Denise Rousseau and David De Cremer, reviewing change research in a 2018 paper for the Academy of Management Annals, describe practitioners relying on popular change models that more often cite expert opinion as their foundation than scientific evidence.¹⁰

Use ADKAR as a structured set of questions that keeps you from training people who are not yet convinced. Treat any promise that it raises success rates as unproven.

How ADKAR sits beside Kotter’s eight steps and Lewin’s three stages

ADKAR describes what each person needs to change; John Kotter’s eight steps describe what leaders do to move a whole organization. Kotter first set out the approach in Harvard Business Review in 1995, drawing on more than 100 companies he had watched try to remake themselves.¹¹

His firm’s current list runs: create a sense of urgency, build a guiding coalition, form a strategic vision, enlist a volunteer army, enable action by removing barriers, generate short-term wins, sustain acceleration, and institute change.¹² The steps are about sponsorship, coalitions and momentum, which ADKAR leaves out. ADKAR, in turn, gives a way to check each person that Kotter’s list lacks. Kotter has also written on how leading differs from managing.

Kotter’s model has also been tested only in parts. A 2012 review by Steven Appelbaum and colleagues, which searched 15 years of research for support for the version in Kotter’s 1996 book, found backing for most individual steps and no evidence against the model, but no formal study of it as a whole. The authors concluded that its popularity comes more from its usable format than from scientific consensus.¹³

Kurt Lewin’s three stages, unfreeze, change and refreeze, are often presented as the classic foundation of change management. Stephen Cummings, Todd Bridgman and Kenneth Brown compared what Lewin wrote with how he is cited and argued in 2016 that he never developed such a model; it took shape after his death.¹⁴

Myth

Kurt Lewin's unfreeze-change-refreeze model is the tested foundation of change management.

Fact

A 2016 study of Lewin's writings argues he never set out such a model; it took shape after his death.

The levels fit together. A project sponsor can follow an organization-level sequence such as Kotter’s while team leads use ADKAR to check where each person stands. ADKAR asks what people need; the change curve describes how people may feel along the way.

Where ADKAR falls short

ADKAR reads every stalled change as a person who has not yet reached an outcome, and that is sometimes the wrong diagnosis. Because Prosci built it as a model for individual change, all five of its questions are about the person.¹

A low Ability score can mean the person needs coaching, or it can mean the new system is slow, the process around it is broken or the workload leaves no time to learn. In the ticketing example, the newer team member’s problem may have been a badly designed queue view. Before you coach a person, check whether five colleagues give the same low score for the same reason; if they do, fix the tool or the process.

Self-ratings bring their own risk. People may not admit low Desire to the manager who is pushing the change, so a 4 can hide a 2. Collect ratings privately, or have someone outside the reporting line ask, and check behavior as well, such as how many requests still arrive by email.

The bottom line

The model’s best use is diagnosis. It tells you which of five outcomes each person is missing, so you stop sending reluctant people to training and start with the reason they are stuck. Its evidence is mostly the vendor’s own surveys. For your next rollout, ask each person the five questions privately, find the first score of 3 or less, and pick the support that fits that step.