Finding the atomic problem inside a transformation programme
Corporates · May 2026 · by Ascendra Labs
Every forty-slide programme hides one sentence worth solving. A field guide to extracting it before the committee does.
Every transformation programme we have ever read contains one sentence worth solving. It is rarely on the first slide. It hides in the risk register, or in the appendix where someone quantified the leakage, or in a footnote written by the one analyst who watched the process actually run.
The programme exists because that sentence was too small to fund on its own. A reconciliation leak does not justify a steering committee; a finance transformation does. So the atomic problem gets wrapped in strategy language until it is large enough to survive governance, and by then it is far too large to ever finish. That is the quiet tragedy of the corporate programme: it has to be big to get approved, and being big is exactly what kills it.
The evidence for that is stronger than the famous number, and the famous number is worth a pause, because we would rather be accurate than dramatic. You have heard that seventy percent of transformations fail. It sits in nearly every consulting deck, usually with a straight face and no footnote. Follow it back and it dissolves. The earliest version comes from Hammer and Champy in 1993, who called their own figure an unscientific estimate and were only talking about reengineering. It spread through a 1995 Kotter article and a 2000 Beer and Nohria piece that opened with "about seventy percent of all change initiatives fail" and cited no one. A 2011 review in the Journal of Change Management went hunting for the study underneath and found what it called academic matryoshka dolls, sources citing sources that never reach any data. Even the McKinsey reports that repeat the number end up pointing back to Kotter.
So the headline is folklore. What is real, and worse, is the measured data beneath it. McKinsey's own survey work found that only about sixteen percent of digital transformations both improved performance and kept the gains. Their research on where the value goes, drawn from fifteen years of transformation data, is the part that should keep executives awake: even the transformations that succeed capture only about two-thirds of the money that was on the table, and much of the loss happens early, during target-setting and planning, before anyone builds or changes a thing. The value does not mostly leak in execution. It leaks in the strategy-language phase, the same phase where the atomic problem got dressed up and inflated.
Then there is the finding we keep coming back to, and it comes from Bain, not from us. In their 2024 study, where eighty-eight percent of transformations missed their original ambitions, Bain reported that ninety percent of the value in a transformation is created by less than five percent of the roles. Ninety percent of the value, five percent of the people. Bain also found that naming one accountable owner for the effort added about a quarter more of the planned value, and that the transformations most likely to fail were the ones that buried their leaders in breadth.
Sit with what that means. The most pro-programme firms in the world, the ones who sell the big engagements, keep publishing data that describes an atomic-problem-shaped fix hiding inside the sprawl. Find the five percent. Name the owner. Go after the number that target-setting buried. Everything else in the programme is, in the cold light of the research, mostly ceremony.
Our method is extraction, and it is a discipline before it is a service. We read the programme the way an editor reads a manuscript, hunting for the one sentence with a number in it, an owner near it, and a system behind it. That sentence becomes the engagement. Everything else goes back on the shelf, unharmed, for the committee to keep.
We will be fair to breadth, because sometimes it is real. A post-merger integration, a solvency crisis, a genuine re-platforming: these are cases where a single problem cannot carry the weight, and coordination is the work rather than a disguise for it. The discipline is not to refuse coordination. It is to refuse to let coordination become the deliverable. Most of what lands on the table as a forty-slide programme is not one of those cases. It is one solvable problem wearing the costume of ten.
The committee can keep the programme. You keep the working system and the numbers. In our experience, the second thing quietly ends the first.
