The evidence has been available for nearly three decades. McKinsey first documented the 70% failure rate in 1996. The number hasn't improved. The question worth asking is not why change is hard — it's why organisations keep getting it wrong in the same ways.
The most common answer you'll hear is resistance. People resist change. They're comfortable with the status quo. They don't understand why the change is necessary. If only we could communicate better, change would happen.
This is one of the most persistent and damaging myths in organisational life. And it's expensive. Organisations invest billions annually in change programmes built on the assumption that the problem is communication and buy-in. They produce better presentations, run town halls, send more emails from the CEO. And then they're surprised when nothing changes.
The diagnosis problem
Most change programmes begin in the wrong place. A leader identifies a problem — declining performance, cultural dysfunction, strategic misalignment — and reaches for a structural solution. A reorganisation. A new system. A change in process. These are visible, measurable, and easy to put in a programme plan.
They are also, consistently, the wrong starting point.
"The presenting problem and the real problem are rarely the same thing. Thirty years of working in complex organisations hasn't changed that observation once." — Stephen Dixon-Mould FCIPD
The academic literature on organisational change is unambiguous. Kotter's foundational work identified eight reasons change fails — and the majority relate to human behaviour, not structural design. Beer and Nohria distinguished between Theory E change (hard, economic, top-down) and Theory O change (soft, organisational capability, bottom-up) and found that attempts to combine them without managing the tension between them almost always fail. Prosci's longitudinal research consistently shows that people-side readiness is the single greatest predictor of change outcomes — not the quality of the technical solution.
Yet organisations continue to design change programmes around the technical solution and treat the human dimension as a communications problem to be managed afterwards.
Why the same mistakes keep happening
The more interesting question is not why change fails — the evidence on that is clear — but why organisations keep making the same mistakes despite the evidence.
There are three structural reasons this happens.
1. Structural interventions are easier to commission
A restructure has a clear brief, a visible output, and a measurable timeline. You can put it in a board paper, track it against milestones, and declare it complete. Behavioural change has none of these properties. It's slower, messier, harder to quantify in the short term, and — critically — it requires leaders to examine their own behaviour. That's genuinely uncomfortable, and most organisations don't create the conditions that make it possible.
2. Consultancies are incentivised to deliver outputs
The professional services model is built around deliverables — reports, frameworks, redesigned operating models. These are what get invoiced. The outcome — whether anything actually changed — is harder to attribute, harder to measure, and typically occurs after the engagement has ended. There is no commercial incentive for the advisory firm to stay until the change is real. So they don't.
The output trap
A 40-page discovery report is not a diagnosis. A framework presented at an away day is not behaviour change. An implementation plan is not implementation. The gap between what gets delivered and what actually changes is where most change value is lost — and most organisations don't have a systematic way of closing it.
3. C-Suite behaviour is both the cause and the cure
This is the most important and least discussed dynamic in organisational change. Leaders commission change programmes and then largely exempt themselves from them. They attend the launch event, send the message, and then return to operating exactly as before — while expecting the organisation beneath them to behave differently.
The research is unequivocal on this. C-Suite behaviour is the single biggest predictor of change success or failure. When leaders model the new behaviours — visibly, consistently, even when it's uncomfortable — change spreads. When they don't, it doesn't, regardless of the quality of the programme beneath them.
This is not a criticism of leaders. It reflects the fact that most change programmes are not designed to support executive behaviour change. They're designed to change everyone else.
What the 30% do differently
The research on successful change — Kotter, Prosci, McKinsey's own follow-up work, the Hay Group's studies on leadership and change — points to the same set of underlying conditions. They are not complicated. But they require something most organisations find genuinely difficult: the discipline to start in the right place rather than the convenient one.
They diagnose before they design
Successful change programmes invest real time in understanding where the real constraint is — not just the presenting problem. The diagnosis shapes the intervention. Not the other way around.
They start with behaviour, not structure
Structure follows behaviour. Organisations that reverse this — designing the structure and hoping behaviour follows — consistently underperform those that understand the behavioural conditions first.
C-Suite actively leads, not sponsors
There is a meaningful difference between a leader who sponsors a change programme and one who actively models the new behaviours in their own decision-making. The latter is rare. It's also the primary determinant of success.
Kotter's research found that organisations where senior leaders were genuinely and visibly committed to change — not in a communications sense, but in terms of their own behaviour — were significantly more likely to achieve their objectives. The word "genuinely" matters. Employees are exceptionally good at distinguishing between leaders who believe in a change and those who are going through the motions of endorsing it.
"Change fails not because the strategy was wrong. It fails because behaviour was treated as an afterthought — something that would follow naturally once the system changed. It rarely does." — Stephen Dixon-Mould FCIPD
The behavioural science case
The academic grounding for this is well-established. BJ Fogg's Behaviour Model identifies three conditions that must be present simultaneously for behaviour to change: sufficient motivation, sufficient ability, and the right prompt at the right moment. Remove any one of these and the behaviour doesn't occur — regardless of how well the change has been communicated or how logical the case for change is.
Kahneman's work on System 1 and System 2 thinking is equally relevant. The vast majority of human behaviour — including workplace behaviour — is governed by System 1: fast, automatic, habitual, emotionally driven. Change programmes are almost universally designed to engage System 2: slow, deliberate, rational. They present a logical case for change and expect logical behaviour in response. This is not how behaviour works.
Effective change design works with System 1, not against it. It changes the environment in which decisions are made — the defaults, the prompts, the social norms — so that the new behaviour is the path of least resistance rather than the path of most effort. This is what nudge theory and behavioural economics have been demonstrating for decades. It is still largely absent from mainstream organisational change practice.
The operating model dimension
There is one more element that most behavioural approaches miss. Behaviour doesn't exist in a vacuum — it exists inside an operating model. And the operating model either supports or undermines the behaviours you're trying to create.
An organisation that wants its leaders to make faster decisions but runs them through a six-layer approval process is designing for the opposite of what it says it wants. An organisation that values collaboration but structures incentives around individual performance will get individual performance. The structure and the behaviour must be aligned, or one will win — and it's usually the structure.
This is why the most effective change interventions hold two lenses simultaneously: the behavioural and the structural. They design for how people actually act — not how the org chart assumes they will — while also ensuring that the operating model creates the conditions for those behaviours to persist and spread.
Most advisory approaches have one lens. Behavioural specialists understand behaviour. Organisational design specialists understand structure. The firms that get change right are those that can hold both — and that is a genuinely rare capability.
What this means in practice
The practical implication of all of this is relatively simple, even if the execution is not. Effective change requires three things to happen in the right sequence.
First, find where the problem actually lives. Not the presenting problem — the real one. This requires honest diagnosis, the willingness to hear uncomfortable things, and the discipline not to jump to solutions before the problem is properly understood.
Second, design for actual behaviour. Use what behavioural science tells us about how people make decisions, form habits, and respond to social environments. Design the intervention for System 1, not System 2. Change the defaults. Change the environment. Change what's easy.
Third, align the operating model. Make sure the structure, incentives, and decision-making architecture support the behaviours you're trying to create — not undermine them. This means being willing to redesign things that have been in place for a long time, which is always uncomfortable and always worth it.
The honest summary
The 70% failure rate is not a mystery. It reflects the consistent gap between how change is designed and what the research tells us about how behaviour actually works. The organisations that close that gap — that start with an honest diagnosis, design for real behaviour, and ensure their leaders are actively modelling the change rather than sponsoring it from a distance — are the ones that make it into the 30%.
None of this is easy. But it is knowable. The evidence has been there for thirty years.
This piece draws on: Kotter, J.P. (1995) 'Leading Change: Why Transformation Efforts Fail', Harvard Business Review; Beer, M. & Nohria, N. (2000) 'Cracking the Code of Change', HBR; Prosci Change Management Research (2018–2024); Fogg, B.J. (2019) Tiny Habits; Kahneman, D. (2011) Thinking, Fast and Slow; McKinsey & Company (1996, 2009, 2015) Organisational Change research series. Academic grounding informed by current MSc in Organisational & Business Psychology, University of Wolverhampton.