Somewhere between 2014 and 2020, "digital transformation" became the universal answer to every strategic question in industrial businesses. Margins compressing? Digital transformation. Talent leaving? Digital transformation. Competitor pulling ahead? Digital transformation. The phrase achieved a level of meaning-by-overuse that has made it almost impossible to talk about clearly, while simultaneously becoming the largest single category of capital expenditure for most industrial companies of any size. The contradiction is striking: a category nobody can define is absorbing more capital than any other operating investment.
What is interesting is not that companies are spending the money. Most of the underlying investments are necessary. The interesting question is why the success rate of these programs is so consistently poor. The numbers have been published for a decade. Roughly seventy percent of digital transformations fail to meet their stated objectives. The failure rate has not improved despite an enormous market for consultants, technologies, methodologies, and certifications dedicated to improving it. The pattern is too durable to be a series of unlucky implementations. It is a structural problem.
The structural problem is that "digital transformation" has been treated as a technology category, with technology owners, technology budgets, and technology metrics. It is not a technology category. It is a change-management category that uses technology as one of its instruments. The companies that succeed treat it that way. The companies that fail spend years discovering, expensively, that they were running the wrong kind of program.
- An estimated 70 percent of digital transformation initiatives fail to achieve their business objectives, a number that has remained essentially flat from 2014 to 2024 (McKinsey, BCG, Bain, all converging on this range).
- The average industrial company spends 4 to 6 percent of revenue on technology, of which roughly half is now labeled "transformation" (Gartner CIO Survey, 2024).
- Among the 30 percent of transformations that succeed, the single most consistent predictor is CEO involvement in change management activities, not technology decisions (BCG Digital Acceleration Index).
The costume
Watch what a digital transformation program looks like in most industrial companies. It has a chief digital officer, or a chief transformation officer, or a chief data officer. It has a budget line that is mostly software licenses and systems integration. It has a steering committee that reviews technology architecture decisions. It has a portfolio of projects, each named after the system being deployed. It reports progress in terms of milestones achieved on the technology delivery roadmap.
Now watch what is actually changing in the company while this is happening. The customer is still being quoted the same way. The order is still being processed the same way. The salesperson is still using the same playbook. The engineer is still making the same decisions with the same tools. The technology is being deployed alongside the existing operating model. The operating model has not been touched, because the program was not designed to touch it.
This is the costume problem. The program looks like a transformation. It has the title. It has the budget. It has the executive sponsor. What it does not have is any plan to change how work gets done. The implicit theory is that deploying the technology will, eventually, cause the work to change. The theory has been tested for fifteen years across thousands of companies. It does not appear to work.
The chart is the entire argument in one image. Companies are spending fifteen percent of their transformation budgets on the categories that explain seventy-one percent of the variation in success. They are spending seventy percent on the categories that explain twenty-nine percent. The asymmetry is not subtle, and it is not a measurement artifact. It is the structural reason most transformations fail.
What change management actually does
The phrase "change management" has been so abused by management consulting that it has lost most of its meaning. To recover it, it is worth saying what change management actually does in a successful transformation.
First, it redefines what work looks like. Before any technology goes live, the change program has documented the new workflow, identified the people whose work will change, and confirmed that the changed work is doable, desirable, and rewarded. The technology arrives into a workflow that has already been designed around it.
Second, it builds the capability to do the new work. This is mostly training, but not the kind of training that produces a certification and is forgotten. It is structured practice, observed coaching, embedded experts who sit with the team during the transition, and explicit support for the period of reduced productivity that always accompanies new work.
Third, it aligns the incentives. The team is being asked to change how they work. Their compensation, performance metrics, and career path need to reward the new behavior, not punish it. Companies that fail to do this find that their teams (rationally) continue to behave the old way while the technology sits unused.
Fourth, it sustains the change. Most transformations regress within 18 months of "completion" because the organizational gravity pulls behavior back to the old equilibrium. The change-management discipline includes ongoing reinforcement, measurement, and intervention in the years after the initial deployment. Without this, the company has paid for a temporary improvement.
A transformation that nobody can describe in terms of what people do differently is not a transformation. It is a technology procurement program with aspirational language.
The CEO test
The most reliable predictor of transformation success in the research is CEO engagement with the change activities, not the technology activities. This is awkward, because most CEOs delegate transformation to a transformation lead and check in monthly on technology milestones. The successful patterns are different.
CEOs who run successful transformations spend their time on the human side. They communicate the change repeatedly. They visibly use the new tools themselves. They make personnel decisions that signal the change is real. They allocate the change-management budget personally rather than delegating it. They show up at training sessions. They make the change-management lead a direct report, not a sub-function of the CIO.
This is hard. It is also the thing that distinguishes the transformations that work from the ones that do not. A CEO who delegates transformation to a transformation function has signaled, implicitly, that the change is a technology matter rather than a leadership matter. The organization will read this signal correctly and respond accordingly.
What success actually looks like
The transformations that produce durable change have a recognizable shape. They are not faster than failed transformations; if anything, they are slower in the first year. They produce visible operational changes earlier, often before the technology has fully landed. They are owned by line leaders, not by transformation leaders. They are measured against business outcomes (revenue, margin, retention, productivity), not against technology milestones.
The vocabulary inside successful transformations is also different. The conversations are about how work is being done, not about which system is being deployed. The lessons learned are about behavior, not about software. The risks being managed are about adoption and capability, not about integration and deployment.
If you sit in on a transformation steering committee and the conversation is dominated by technology architecture, integration milestones, and vendor management, you are watching a program that will probably fail. If the conversation is dominated by adoption metrics, capability building, role redesign, and behavioral change, you are watching one that has a chance.
A transformation that does not change how anyone works is not a transformation. It is an upgrade. The companies that have understood this distinction are spending less and producing more. The companies that have not are funding a costume.