Kotter's 8-step change model: how it works and where it needs help

Kotter’s 8-step model is a leadership sequence for large organisational change. In its best-known form, from John Kotter’s 1996 book Leading Change, it runs from creating urgency and building a guiding coalition, through forming and communicating a vision, removing obstacles and securing short-term wins, to consolidating gains and anchoring new approaches in the culture. It has lasted partly because it names what Kotter saw stalling change programmes: complacency, a weak steering group, an unclear direction and a premature declaration of victory.

The model is most useful when each step is treated as a question to investigate rather than a box to tick. Can the people who must change see the need in evidence they can check? Does the steering group include people who understand the work? Are the early wins genuine? Have the routines, measures and decisions that reward the old way actually changed?

This guide follows a fictional housing association through the steps and compares the model with ADKAR. Our overview of change management treats models like these as lenses, not recipes.

Where the eight steps came from

The model began as an account of why change fails. Kotter’s 1995 Harvard Business Review article, “Leading Change: Why Transformation Efforts Fail”, opens with his observation of more than 100 companies trying to remake themselves under banners such as total quality management, reengineering and restructuring. That is an account built on observation, not a controlled comparison of organisations that did and did not follow the steps.

The 1996 book turned those lessons into eight stages. Pollack and Pollack’s study of the model in practice quotes Kotter’s emphasis on the “importance of sequence” and his warning that skipping even a single step “almost always creates problems”. They also show how many later writers read the model as a linear sequence.

Kotter returned to it in a 2012 Harvard Business Review article and the 2014 book Accelerate. Harvard Business School’s summary of the book describes a “dual operating system”: the hierarchy continues to run the everyday business, while an agile network alongside it pursues future opportunities. People are not assigned to the network; they volunteer.

Kotter Inc.’s own comparison of the two versions is explicit. The 1996 process handles episodic change “in finite and sequential ways”, driven by “a small, powerful core group” within the hierarchy. The 2014 process runs the steps “concurrently and continuously” and relies on “a large volunteer army from up, down, and across the organization”. The firm presents both versions as valid for different contexts.

Step	Leading Change (1996)	Current Kotter Inc. wording
1	Establishing a sense of urgency	Create a sense of urgency
2	Creating the guiding coalition	Build a guiding coalition
3	Developing a vision and strategy	Form a strategic vision
4	Communicating the change vision	Enlist a volunteer army
5	Empowering employees for broad-based action	Enable action by removing barriers
6	Generating short-term wins	Generate short-term wins
7	Consolidating gains and producing more change	Sustain acceleration
8	Anchoring new approaches in the culture	Institute change

The right-hand column follows Kotter Inc.’s current description of the eight steps. The largest shifts are in steps 4 and 7: communication becomes recruiting volunteers, and consolidation becomes a deliberate push after the first successes. The later material also frames urgency around a “big opportunity” as well as a threat.

A fictional example: three offices, one repair route

Consider a fictional regional housing association with three local offices: one in a market town, one on the coast and one in a rural valley. Tenants report repairs by phone, email or at the counter, and each office logs them differently. The town office uses the repairs module of the housing system. The coastal office keeps a shared spreadsheet. The valley office uses the system plus a paper diary of tenants who need a call from someone they know. A heating contractor covering all three areas receives jobs in three formats.

The director of housing services wants one shared process and one system: a single reporting route, triage by a central repairs coordinator, and one job record that office staff and contractors can all see. The counters stay open. A tenant panel reviews service changes. The association, its people and events are invented to illustrate the model; none of it is a case study or a measured result.

Each step as a question to investigate

Kotter’s steps name what needs attention, not how to achieve it locally. Framing each as a question keeps a team investigating rather than performing the step. The table is our own working aid, not a Kotter tool.

Step	Question it raises	What the housing team checks	Common failure
Urgency	Can those who must change check the evidence for it?	Recent repairs traced from first call to completion	A fear-based message that starts rumours
Coalition	Does the steering group understand the work and hold enough authority?	Whether coordinators, office staff, a contractor and a tenant take part	A board of senior managers and suppliers
Vision	Could someone test whether the future has arrived?	“One reference, one owner, one status” for every repair	A slogan nobody can check
Communicate and enlist	Who is taking part, and who cannot afford to?	Volunteers and questions from all three offices	Participation only from the least busy office
Remove barriers	What makes the new way harder than the old?	Missing fields, contractor systems, workarounds	Treating objections as attitude problems
Short-term wins	Is the early result real and meaningful?	Whether a tenant can get a status from any office	Counting logins as success
Sustain	What still needs to change after the first success?	The other offices’ different arrangements	Declaring victory after the first site
Anchor	What still rewards the old way?	Meeting agendas, reports, induction, complaint routing	A culture statement with unchanged incentives

Urgency built from evidence people can check

Kotter treats urgency as the foundation. In a passage Pollack and Pollack quote from Leading Change, he argues that when complacency is high, few people are even interested in working on the change problem. The difficulty is how urgency gets produced.

A common shortcut is the “burning platform”: persuade people that staying put is more dangerous than moving. McLaren, van der Hoorn and Fein argue that vilifying the status quo is combative, adds to employee anxiety and stress, and is futile given people’s status quo bias. They propose stressing the pressing importance of the future state instead. Theirs is a theoretical argument, not a field test. Rousseau and ten Have, writing on evidence-based change, warn that creating urgency, “a frequent tactic in top/down change”, often backfires when people are already stretched.

In our invented example, the director’s first briefing warned that rising complaints and regulatory scrutiny left “no option”. Within a week, coastal staff were telling each other that their counter would close, and problems stopped surfacing in team meetings. Where people expect blame, evidence gets quietly tidied, which is why psychological safety matters to this step.

The team replaced the message with a shared inquiry. Each office traced a handful of recent repairs from first contact to completion: how often the tenant chased, where details were retyped, and whether the contractor received the job once, twice or not at all. Tenant panel members added their experience, and staff could check every case against their own records. The trace also showed that the valley office’s diary often got vulnerable tenants a quicker call-back. The evidence now showed a problem people recognised and a practice worth protecting. The director confirmed in writing that no counter would close.

Urgency is not the same as readiness. People can agree that repairs go missing and still lack the time, access or authority to work differently. Our guide to assessing change readiness keeps those questions apart.

A coalition that includes people who know the work

The 1996 version asks for a powerful core group. Kotter Inc.’s 2018 overview asks for more: members from multiple layers of the hierarchy and many functions. Rousseau and ten Have encourage change leadership at every level, from executives and middle managers to front-line workers. The programme Pollack and Pollack studied needed several guiding coalitions, although Kotter’s writing tends to refer to one.

The association’s first project board held the director, the IT manager, a finance lead and the system supplier. Nobody on it had triaged a leaking boiler late on a Friday. The board added the repairs coordinator, one housing assistant from each office, the heating contractor’s scheduler and a tenant panel member. Stakeholder mapping showed who was still missing, including evening callers and tenants who never use email.

Before the first meeting, the director and the coalition agreed what the group could decide: triage rules and the order in which offices would move. Contractor contracts and staffing sat elsewhere, and everyone knew it.

Vision, communication and the volunteer question

Kotter Inc. describes the vision step as clarifying how the future will differ from the past. A useful test: could someone check that it has arrived? The coalition settled on: “Every repair has one reference, one owner and a status the tenant can ask about at any office.” A tenant could test that; “modern, joined-up repairs” could not.

Where the 1996 version communicates the vision, the 2014 version enlists volunteers who have a real choice and permission to step forward. That has a cost. Harvard Business School’s summary reports that volunteers found ways to do their normal jobs in less time, and that one senior executive feared for his revenue plan. Volunteering depends on spare capacity.

The town office offered three volunteers. The coastal office, the most short-staffed and the one handling most emergency call-outs, offered none. Rather than reading that as resistance, the coordinator arranged cover so a coastal assistant could take part. The team also ran a cross-office session using World Café rounds to collect questions. Each question went into a dated log with an owner, as a practical change communication plan recommends.

Barriers and short-term wins that are real

Removing barriers means finding what makes the new way harder than the old one. Kotter’s 2018 material lists silos, pressure to hit numbers and legacy rules among common barriers.

The valley office kept using its diary, although the plan retired it. The coalition asked why instead of issuing a reminder. The new system had no field for “needs a call from a named person”, and the contractor’s scheduling tool rejected the new job reference. Both were design problems. Our guide to resistance to change explains how to treat objections like this as information. The team added the field and left the diary in use until it worked.

Short-term wins show that the change is possible, and they are easy to stage. Kotter Inc. says an effective win is “unambiguous, visible, and tangible” and matters to people beyond those who produced it. Rousseau and ten Have describe small wins as local, often bottom-up experiments in which some failures are expected.

A launch-week draft announced that every town-office repair was now logged in the new system. True, but it counted activity, and some jobs had been retyped from phone notes. The coalition chose wins others could check. Tenant panel members phoned the coastal office about jobs logged in town and received a status. The contractor’s scheduler confirmed that one job list had replaced three email formats. The team also reported what was not yet working: out-of-hours emergency jobs still had no shared status.

How the steps interact

The steps depend on one another. Urgency without a credible coalition produces anxiety rather than action. A vision without barrier removal teaches people that announcements mean little. Genuine wins renew urgency and recruit volunteers. Anchoring works only once people have practised the new way long enough to see that it helps.

In practice the order is less tidy, and Kotter did acknowledge that earlier stages may need revisiting. Pollack and Pollack describe groups that started at different times and moved at different speeds. Ten months in, managers five levels below the chief executive still did not understand the urgency while others were celebrating wins, so the process was restarted with them. From above, the programme looked linear. Underneath, several instances of the process were running at once, which is close to what the 2014 version recommends.

After the town office went live, the board wanted to move both remaining offices together. But the coastal office was still at the urgency and coalition stage, and its out-of-hours arrangements differed. The team ran a premortem that imagined the coastal switch had failed. The most plausible story was an emergency leak at night reaching the wrong contractor. The date stayed provisional until that route had been tested.

Anchoring: change the routines that reinforce behaviour

Kotter’s 2018 material argues that culture changes after behaviour has changed and people can see the connection between new behaviours and better performance, “deep into a transformation, not at the beginning”. Rousseau and ten Have describe institutionalisation in concrete terms: aligning hiring, performance management and accounting systems with the change, and retiring legacy systems that keep the old routine alive.

For the association, anchoring meant finding what still rewarded the old way. The monthly report ranked offices by calls answered locally. The weekly repairs meeting reviewed three office lists. New assistants learned the spreadsheet by shadowing. Complaints went to office managers rather than to the job owner.

Each of these was changed or scheduled for change. The meeting now reviews the shared queue, the report shows jobs without an owner, and induction teaches the shared route. The spreadsheet became read-only after the final office moved. The most influential change was small: when a complaint arrives, the director asks who owns the job in the shared record. The questions leaders ask show staff what really matters. Our guide to understanding organisational culture helps teams investigate unwritten rules like these.

Kotter or ADKAR: organisation-level or person-level

Kotter’s steps describe what leaders and a coalition do to move an organisation. Prosci describes ADKAR as “a model for individual change, not a complete change management methodology”. Its elements are Awareness, Desire, Knowledge, Ability and Reinforcement. Prosci’s own comparison calls Kotter’s model top-down and leadership-driven. That is a competitor’s framing, and it fits the 1996 version better than the volunteer-based revision.

The difference is the unit of attention. Kotter asks what the organisation must do next; ADKAR asks where one person is stuck. The two meet at the barrier step. The valley assistant understood the need and knew the new system. A missing field blocked her ability to use it for vulnerable tenants, and her reluctance came from protecting them. ADKAR locates that gap in one person’s transition. Kotter’s barrier step is where the organisation removes it.

Use Kotter’s steps to plan and steer a change involving several groups. Use ADKAR when a particular person or role is stuck and you need to know why. Combine them where useful, but do not merge them into an improvised score.

Limitations and what the evidence shows

The evidence for the model as a whole is thin. Appelbaum, Habashy, Malo and Shafiq reviewed fifteen years of literature step by step. They found support for most individual steps but “no formal studies” covering the entire model, and suggested its popularity owes more to its direct, usable format than to scientific consensus. They found no evidence against it and recommended it as an implementation planning tool used alongside other approaches.

Stouten, Rousseau and De Cremer note that popular practitioner change models more often cite expert opinion than scientific evidence. Setting those models against organisational research, they summarise what is known, contested, untested and underused, and identify ten evidence-based steps. Pollack and Pollack’s paper is a single action-research case: it shows how the model was adapted in one organisation, not whether it works in general.

The main critiques are consistent. The 1996 version is leader-centric and reads as linear. The 2014 version answers some of that with volunteers and concurrency, but we found little independent research testing it, and it relies on discretionary effort that is not equally available across a stretched workforce.

Kotter’s descriptions establish what the model says, not that it works. His 2018 material introduces the urgency step by claiming that most transformation efforts fail. Check claims like that before repeating them, as our analysis of the 70% change failure myth explains.

Finally, the model cannot tell you whether the change itself is right. Tracing repairs in our fictional association might have shown that the real constraint was contractor capacity, not logging. The steps organise the leadership work; they do not replace diagnosing the problem.