# Measured to Fail > The number everyone quotes about transformations failing says more about how we measure the work than about whether the work ever actually worked. Published: 2026-07-09 · Updated: 2026-06-26 · Topics: Digital Transformation, Data & Analytics, Growth Leadership, Enterprise, Performance Marketing · Author: Alan Wizemann There is a statistic that gets quoted in almost every conversation about digital transformation, the one that says seventy percent of them fail, and I have heard it repeated so many times, in so many rooms, that it has stopped functioning as a fact and started functioning as a kind of permission slip for not trying. I do not believe the number. My honest read is that it is inflated, and that the real figure is closer to half. My own failure rate across the transformations I have led has been very low, though I am not going to pretend that owes nothing to luck and to having strong executive and board support behind me when it mattered. That last part is not false modesty, and I mean it as a real caveat. The work is hard enough that the difference between a transformation that lands and one that stalls is very often whether the people above you stayed in the boat with you the whole way through. But the part of that seventy percent figure that actually bothers me is not whether it is forty or fifty or seventy. It is that most of what gets counted as a failure is, when you look closely, a measurement problem. The trouble lives in both what we choose to measure and in what we even agree to call a transformation in the first place. I think about a peer I admire, a chief digital officer at a large and very legacy European brand, who was running three transformations at once: a data one, a "digital" one that was really performance marketing and sales wearing a more fashionable name, and an organizational one that touched how the teams worked. On paper, all three were branded failures (every single one of them). The only reason they were called failures is that someone had attached hockey-stick growth and efficiency numbers to them at the very beginning, the work missed those numbers, and the verdict was written from there – even though the actual work succeeded, and the things those teams genuinely needed to happen did happen. That is the trap, and it is set early, usually before the first line of real work is ever done. The banner numbers go up because they have to go up. They are how you win board confidence and investor money (the number is the ask, really), they are the headline on the slide that gets the budget approved, and then the entire effort gets measured for the rest of its life against figures it was never the silver bullet for. I have a line I keep coming back to. If you kept the last slide from every business deck ever made, we would all be running trillion-dollar companies, because the last slide always shows the same beautiful exponential curve climbing off the top of the page. Almost none of those curves were ever proven out or actually built. Here is the thing nobody says out loud in the meeting where the goals get set, the part everyone in the room already privately knows. A transformation is not the silver bullet for "grow thirty percent in two years," and it cannot be made into one no matter how the slide is worded. Growth like that comes from removing the things standing in the way of it, optimizing the few things that genuinely drive it, and fixing the talent and the process that have been quietly holding it back, and those are questions businesses very rarely answer truthfully, because the honest answer requires admitting that something was wrong before the transformation was ever announced. So instead the transformation gets handed the whole number, the entire thirty percent, as if a new data platform or a re-pointed marketing stack were going to conjure demand that the underlying business could not. When the number does not show up on schedule, the transformation takes the blame for a gap it was never built to close. I want to be careful here, because real failures absolutely do exist and I am not interested in explaining all of them away. They are just a different animal than the headline suggests. The genuine failure looks like this: you transform the data layer, you do that part beautifully, and then every application that depended on the old layer breaks because the teams that owned those applications were left out of the scope from the start. That is a real failure, and it is painful. But it is a systems and scope failure, a failure of who got included and what got considered, not the tidy "the transformation failed" story that ends up in the post-mortem and the trade press. The work succeeded and the planning around the work did not, and those are not the same thing, though they get filed under the same word. So when I hear the seventy percent again, and I will, probably this week, what I actually hear is a question we mostly avoid asking. Were these efforts measured against what they were built to do, or against the marketing that was needed to fund them in the first place? Those are very different questions, and only one of them is fair to the people who did the work. The discipline I keep trying to hold onto, in my own programs and in how I talk about other people's, is to measure the work and not the marketing wrapped around it, because the work is the only part that was ever real, and it is the only part worth learning anything from. --- Canonical: https://alanwizemann.com/articles/measured-to-fail