The widely cited figure is that around 70% of change programmes fail. They fail for a reason most change management never addresses: they change the visible layer of an organisation and never touch the layer underneath.
Few statistics are repeated more often in management than the claim that roughly seventy per cent of change initiatives fail. Whatever the precise number, anyone who has lived through a few large transformations recognises the pattern: the reorganisation is announced, the new structure goes up, the town halls happen, the values are rewritten, the consultants leave — and eighteen months later the organisation is recognisably the same one, now with a new org chart and a faint collective cynicism about the next initiative. The change was managed, in the project sense. It just didn't take.
The reason is that most change management operates entirely on the visible layer of an organisation — its structures, processes, reporting lines, and stated values — while the organisation actually runs on something underneath that the programme never touches.
You can change everything visible about an organisation and change nothing real. You can redraw the divisions, rename the teams, publish a new set of values with admirable verbs, and rebuild the process maps — and the same people will arrive on Monday holding the same interpretations, making the same private assessments of each other, and carefully avoiding the same conversations they were avoiding before. The visible layer moved. The operating layer didn't. That operating layer — how people are actually being, and which conversations they are and aren't willing to have — is where the organisation's behaviour really comes from, and it is almost entirely untouched by a restructure.
Consider a familiar case. A company decides it needs to be “more innovative” and runs a serious transformation: new innovation function, new process, new physical space, posters about failing fast. A year on, nothing innovative is happening, because the thing actually blocking innovation was never structural. It was that proposing something risky in this organisation had, for years, been a quiet career mistake — the assessment everyone held was “stick your neck out and you'll regret it” — and no new function or poster changed that assessment. The unsayable remained unsayable. The transformation rearranged the furniture in a room whose actual problem was the conversation no one would have.
You can change everything visible about an organisation and change nothing real.
If the operating layer is made of how people are being and which conversations are possible, then real change has to happen there — and that's harder and less photogenic than a restructure, which is exactly why so few programmes attempt it. Two things move the operating layer.
The first is changing the conversations the organisation is actually having — specifically, surfacing the assumptions it has been running on without examining them. Most organisations are held in place by a set of shared, unspoken interpretations (“leadership doesn't really want bad news,” “that team can't be trusted,” “we don't have permission to question the strategy”) that drive behaviour far more powerfully than any process. Until those are brought into the open and genuinely worked through, they keep producing the old organisation no matter what the new structure says.
The second is shifting who leaders are being, because the conversations a team can have are bounded by the leader. A leader who, under pressure, becomes someone bad news bounces off will preside over a team that hides problems, regardless of how many times they've said “my door is always open.” Change the leader's way of being and you change what becomes sayable around them, and the operating layer starts to move with it.
If the operating layer is where change actually happens, why do so many leaders keep funding the visible one? Because the visible layer is legible and the operating layer isn't. A restructure can be drawn on a slide, costed, scheduled, and reported as progress; surfacing the assumptions an organisation runs on, and shifting how its leaders are being, cannot be put on a Gantt chart and produces no satisfying announcement. Under pressure to show that something is being done, leaders reach for the change that is visible precisely because it is visible — it looks like action. The deeper work looks like talking, and talking is easy to dismiss as not-doing, right up until you notice it's the only thing that ever moved the culture. Choosing the operating layer over the visible one takes a particular nerve: the willingness to do work that is harder to point to and slower to show, against the constant temptation of the photogenic reorganisation that changes nothing.
So before backing the next transformation, ask a blunt question: does it change anything other than the visible layer? If the entire programme consists of new structures, new processes, and new communications — if nothing in it touches the conversations people are willing to have or who leaders are being when it matters — then you are about to spend a great deal of money and goodwill rearranging the furniture, and you can predict the result. The organisations that actually change are the ones willing to do the slower, less legible work underneath the slides. The slides are not the organisation. The conversations are.