Only 16% of Digital Transformations Succeed Long-Term. Why?

Sixteen percent. That is the share of organizations that McKinsey found had improved performance AND sustained that improvement over time from a digital transformation initiative. Not 16% that launched successfully. Not 16% that hit their first-year targets. Sixteen percent that actually changed how they work and kept that change in place. The McKinsey research is not new, but the number has not moved much. Even in high-tech, media, and telecom, where companies are presumably the most digitally sophisticated, the success rate does not exceed 26%. And the broader transformation failure rate, across all types, sits around 70%.

I find the 16% figure more interesting than the 70% figure. The 70% tells you that most transformations fail. The 16% tells you something more specific: that the gap between initial improvement and sustained improvement is enormous. Organizations can make real progress. They just cannot hold it.

My read on why: most organizations confuse technology deployment with transformation, and that confusion means the things that actually sustain change never get funded or prioritized.

When an organization deploys a new ERP, or migrates to the cloud, or stands up a data platform, something real happens. Infrastructure changes. Workflows get updated. Reports look different. The budget spent was substantial and visible. The executive team can point at a go-live date. This looks like transformation because the technology changed. But if the processes that run on top of the technology are unchanged, if the behaviors that generate decisions are unchanged, if the culture around data and evidence is unchanged, then the technical change is not a transformation. It is a more expensive version of what came before.

The academic literature has been clear about this for a while. Vial (2019) defined digital transformation as a process where IT-enabled changes in products, processes, and organizational structures disrupt strategy, requiring organizational responses that produce value. That phrase "requiring organizational responses" is the one that most CIO roadmaps skip. The technology change is the visible part of the project. The organizational response is the part that no vendor can deliver and no consulting engagement can guarantee. It requires leaders who stay committed after the go-live, employees who change their actual behavior, and measurement systems that track the right things.

On executive commitment: this is where I think the failure pattern gets locked in early. Large transformation initiatives tend to have high executive attention at the start. The vision gets articulated. The budget gets approved. The steering committee meets monthly. Then the initiative enters the long middle, where the novelty has worn off, the quick wins have been announced, and the hard work of embedding new practices is still ongoing. Attention shifts. The executive sponsor gets a new priority. The steering committee starts meeting quarterly. The people doing the work lose their air cover. The transformation stalls.

The measurement problem compounds this. Most transformation programs measure activity: number of employees trained, number of processes digitized, percentage of infrastructure migrated. These are not meaningless, but they do not measure whether the transformation is taking hold. The useful measures are things like whether decisions are actually being made differently, whether the organization is producing outcomes it could not produce before, whether people at the working level have internalized new ways of doing things. Those measures are harder to collect, harder to present in a board deck, and harder to attribute to any single intervention. So organizations default to the easy metrics, declare early victory, and find out two years later that the change did not stick.

There is a tension I keep thinking about, particularly from an IS research angle. Agile delivery methods are the dominant paradigm for technology initiatives now. Ship fast, iterate, get feedback. That approach is genuinely useful for building software. But organizational change does not work on sprint cycles. Culture shifts slowly. The routines that govern how an organization makes decisions are sticky by design. They evolved to be stable. An agile delivery model can ship a new system in months. Getting the organization to actually use that system, in the way it was designed to be used, in a way that changes outcomes, takes years. The mismatch between the delivery cadence and the change cadence is, I think, one of the structural reasons transformations succeed technically and fail organizationally.

What would actually predict sustainable digital transformation outcomes is a research question I take seriously. From the IS literature, a few things seem like genuine predictors, not just correlates. Prior digital capabilities matter because organizations that have successfully absorbed prior technology changes have more absorptive capacity to draw on. Top management team alignment, not just sponsorship, matters because sustained change requires ongoing resource commitment and conflict resolution that only leadership can provide. Process redesign maturity matters because organizations that know how to redesign processes, not just adopt tools, are better positioned to integrate the human side of transformation with the technical side. And measurement framework design matters from the start, not as an afterthought, because what gets measured shapes where leadership attention goes.

None of those are things a technology vendor can sell you. They are organizational conditions that either exist or have to be built. The organizations in the 16% probably did not stumble into the success list by deploying better technology than the 84%. They probably had better conditions for absorbing and sustaining the change the technology made possible.

The 16% number will stay stubbornly low as long as the dominant mental model for "transformation" is "we bought new technology and trained people on it." Technology is a necessary condition. It has never been a sufficient one.