Understanding the reasons behind transformation failure is crucial, as it can provide valuable insights for future efforts.
I used to quote the “70% of transformations fail” statistic.
Then I did the research. And I owe you a correction.
The phantom statistic
The 70% failure rate has been repeated so often it feels like a law of physics. McKinsey says it. BCG references it. Every consulting pitch deck opens with it. I’ve used it myself.
But here’s what I found when I traced it back to its source.
It starts with Hammer and Champy in 1993, who estimated that 50 to 70% of reengineering efforts “did not achieve dramatic results.” That’s not the same as failure. That’s not achieving dramatic results. There’s a difference, and it matters.
John Kotter picked it up in 1995 in the Harvard Business Review. He wrote that he had “watched more than 100 companies” and observed “a distinct tilt toward the lower end of the scale.” No database. No regression analysis. Just observation from a very smart man.
Beer and Nohria cemented it in 2000: “The brutal fact is that about 70% of all change initiatives fail.” Again, no new evidence. Just a confident claim citing the same lineage.
Then Mark Hughes at the University of Brighton did what nobody else bothered to do. In 2011, he performed a formal citation audit of the five most prominent 70% claims. He found “absence of valid and reliable empirical evidence” in every single one. The citations looped back on each other. He called it a “phantom statistic.”
Jones and colleagues went further in 2018 with a meta-analysis of 200 change case studies. Their finding? The majority were actually considered successful.
The 70% number is a ghost story the consulting industry tells to sell torches.
So what’s the real number?
This is where it gets interesting. Because the answer depends entirely on who you ask and how they define failure.
Martin Smith’s 2002 meta-analysis of 49 academic studies broke it down by type of change:
– Strategy implementation: roughly 42% failure rate
– Restructuring: 54%
– Mergers and acquisitions: 67%
– Culture change: 81%
A single number was always absurd. It collapses fundamentally different things into one bucket. Changing your org structure is not the same challenge as changing how people think and behave. Anyone who’s tried both knows that.
The big consultancies have their own data, and to be fair, it’s substantial. BCG analysed 1,700 transformations over 20 years using actual financial performance data, not surveys. McKinsey surveyed thousands of executives across multiple years. They converge on roughly the same picture: about a third of transformations succeed fully.
But here’s the part that matters. When you dig in, roughly 50% of all transformations land in a grey zone. Not failed. Not succeeded. Partial results. Value leakage. The thing technically delivered, but the organisation captured maybe half of what it was supposed to.
McKinsey’s own numbers say even “successful” transformations only capture 67% of their potential value. The unsuccessful ones capture 37%. And 55% of that value loss happens during or after implementation.
So the real distribution looks something like this:
– 30% succeed fully
– 50% partially succeed, with significant value leakage
– 20% fail outright
The industry has spent 30 years arguing about the wrong number while the real problem, that massive 50% in the middle, has been hiding in plain sight.
Three groups, three blind spots
I’ve spent months reviewing conference proceedings, academic journals, and consultant research from 2023 to 2025. What struck me wasn’t just the data. It was how completely the three groups studying failure are talking past each other.
The academics have proven the 70% statistic is bunk. They focus on systemic causes, political dynamics, and the psychology of optimism bias. Rigorous, important work. But by the time it’s published, practitioners have already moved on to the next crisis. The research arrives like an autopsy report for a patient who died three years ago.
The consultants have large-scale data showing transformation is hard, and they’ve built useful frameworks around it. BCG’s research identifies a Chief Transformation Officer as worth a 22 percentage point improvement in success rates. McKinsey’s data shows that when companies execute their full set of recommended actions, success rates hit 78%. Good stuff. But the consultants also have a commercial incentive to keep the fear alive. The “70% failure” narrative sells transformation offices and expensive advisory. It creates a burning platform that compels investment. Funny how the people selling the cure are also the ones diagnosing the disease.
The practitioners, the project and program managers who live in the delivery trenches, focus on what to do when it’s already going wrong. They distinguish between “troubled” projects that can be recovered and “failed” ones that are terminal. They talk about “project stress” as an early indicator and argue that current project controls are retrospective, designed to report failure only after it’s happened. These people know the smell of a burning program. But they’re dealing with the consequences of decisions that were made, or not made, long before they arrived.
Each group is solving a different version of the problem.
Academics ask: why does failure happen?
Consultants ask: how do we prevent failure?
Practitioners ask: what do we do when it’s happening right now?
Nobody is connecting the three.
The gap nobody wants to talk about
An entire industry has been built around transformation. Billions spent on consulting, certification, methodology, training. And yet there is almost no established discipline for what to do when a transformation is in trouble.
Think about that.
We have extensive literature on why transformations fail. We have detailed frameworks for how to prevent failure. What we don’t have is a playbook for the moment a senior executive sits in a steering committee, looks at a dashboard full of green traffic lights, walks down the corridor, and hears a completely different story from the people doing the work.
That moment. When you realise nobody is telling you the truth. When the reports say one thing and your gut says another. When the consultants who sold the transformation are the same ones telling you it’s on track.
The research shows this is exactly where the majority of value is lost. McKinsey’s data says 55% of value leakage happens during and after implementation. Not in the planning. Not in the business case. During delivery, when the thing is supposed to be working.
And the failure modes are getting harder to spot. Research into hybrid work environments has found that passive resistance and conflict avoidance are now dominant. People don’t push back. They just quietly disengage. Gartner calls it “quiet failure”: technically green on every dashboard, practically dead. You won’t know until the money is spent and the results aren’t there.
This is about to get worse with AI. S&P Global reports that 42% of companies abandoned most of their AI initiatives in 2025, up from 17% the year before. Same pattern. Overpromise in the pitch, succeed at pilot, fail at scale. Except AI transformations fail faster, which means the window to intervene is smaller.
What this actually means
If you’re a senior executive sponsoring a transformation, here’s the honest version.
The failure rate is not 70%. But it’s not good either. About a third deliver what they promised. A fifth fail completely. And roughly half end up in a grey zone, delivering enough to claim success in the board pack but not enough to justify the investment.
That 50% in the middle is where the real conversation should be happening. Those aren’t hopeless cases. Those are transformations where the value is leaking out during delivery, where the early warning signs were missed or ignored, where intervention at the right moment could have changed the outcome.
But the industry isn’t set up for intervention. It’s set up for two things: selling you a plan, and writing a case study about what went wrong afterwards. The bit in the middle, the rescue, barely exists as a discipline.
When your transformation starts to wobble, when the steering committee goes quiet, when the status reports feel too clean, when the corridor conversations don’t match the dashboard, the answer isn’t more methodology. It isn’t firing the project manager. And it definitely isn’t bringing in the same consultants who sold you the plan and asking them to tell you what went wrong with it.
The answer starts with honesty. What is actually broken. What is it costing you. And what are your real options.
Nobody wakes up and says “how can I be an absolute failure today.” The people working on your transformation are trying their best. It’s the system they’re working within, the governance, the reporting, the incentive structures, that usually needs fixing.
Transformations don’t have to fail as much as they do. The research is clear on that. But rescuing one that’s in trouble requires a fundamentally different capability from designing one that might succeed.
The industry has spent decades perfecting the design. It’s time someone got serious about the rescue.
**************************************************************************************************
- Hammer & Champy, Reengineering the Corporation (1993), estimate about reengineering outcomes.
- Kotter, “Leading Change: Why Transformation Efforts Fail” (1995), practitioner observation.
- Beer & Nohria, “Cracking the Code of Change” (HBR, 2000), popularised “70% fail” phrasing.
- Hughes (2011), literature review challenging the empirical basis of the “70%” claim.
- Smith (2002), summary of median success rates by change type (implied failure rates vary by category).
- McKinsey (2021), “Successful Transformations”, value capture and value loss during/after implementation.
- BCG (2020), “Flipping the Odds of Digital Transformation Success”, “70% fall short” framing for digital.
No responses yet