The Diagnostic That Change Management Skips
From the Change Management collection
In 2000 Michael Beer and Nitin Nohria published an essay in the Harvard Business Review observing that 70% of change efforts fail. The figure entered the management literature as settled fact. In 2011 Mark Hughes went looking for the original study and could not find one. The 70% estimate had been repeated by so many subsequent authors that it had acquired the weight of research, but it traced back to rough guesses and consultancy claims rather than to a defined sample, a control group, or a clear measure of failure.
The number matters less than what happened next: an entire industry organized itself around solving a problem it had never properly measured. The consulting methodology, the eight-step model, the certification program — each assumed a general failure rate and offered a general solution. What the research actually shows is more specific and more useful.
Change efforts break down in identifiable patterns, and the patterns trace to diagnosis. The most consequential diagnostic distinction in the field belongs to Ronald Heifetz, who in 1994 separated technical problems from adaptive challenges. A technical problem has a known solution and requires execution: install the software, restructure the reporting lines, change the policy. An adaptive challenge has no ready solution because the people involved are part of the problem — their values, assumptions, or habits must shift before any structural fix will hold.
The diagnostic matters because the tools are incompatible. Technical problems respond to planning, authority, and project management. Adaptive challenges respond to learning, experimentation, and the willingness to let people struggle with the discomfort of not yet knowing the answer. When an organization treats an adaptive challenge as a technical problem — and this is the default, because technical solutions are faster, more legible, and more comfortable for leaders — the new system is installed and nothing changes. The policy is updated and behavior continues. The merger closes and the cultures remain separate. The reorganization is completed and the coordination problems move to new seams.
A second diagnostic runs alongside the first. Watzlawick, Weakland and Fisch distinguished in 1974 between first-order change — doing more or less of the same thing within an existing framework — and second-order change, which alters the framework itself. The distinction explains a pattern that managers frequently encounter: resistance that looks irrational from a first-order perspective often makes perfect sense from a second-order one. The people involved correctly perceive that what is being asked is a redefinition, and they are resisting the redefinition, not the adjustment.
Kurt Lewin, whose work predates both frameworks, provided the most practical tool. His force-field analysis holds that any status quo is maintained by an equilibrium between driving forces (pushing for change) and restraining forces (holding things in place). The effective intervention is usually to weaken the restraints rather than to increase the pressure. This is counterintuitive. The instinct of most leaders is to push harder — more resources, more executive sponsorship, more urgency. Lewin's analysis suggests that identifying what holds the status quo in place — embedded routines, political interests, skill gaps, the fear of temporary incompetence — and reducing those forces is more effective and less likely to produce the rigidity that comes from increased pressure.
William Bridges added the human dimension. In 1991 he separated the change — the external event — from the transition, which is the internal psychological process people go through. A transition begins with an ending: letting go of the old competence, the old relationships, the old identity. It passes through a neutral zone where performance drops and confusion rises. It reaches a new beginning only when the person has made meaning of the change. Organizations schedule the change; the transition follows its own timeline. Most change timelines account for the project plan and ignore the transition, which is why a change that looks complete on the Gantt chart has not yet taken hold in the organization.
The consulting literature built on these foundations is individually useful and collectively misleading, because each model presents its framework as the framework. Kotter's eight steps, the most widely taught model, treats change as a planned, top-down sequence. Beer, Eisenstat and Spector's research found that the most effective changes started at the unit level, aligned to a business task, and worked upward — which sits uncomfortably beside Kotter's top-down sequence. Weick and Quinn argued that in turbulent environments, change is continuous and emergent rather than episodic and planned. None of these models is wrong. Each describes a different type of change under different conditions, and the failure is applying any one of them without first asking what type of change the organization faces.
That question — what type of change is this? — is the diagnostic that most change management programs skip. They jump to method: how to communicate, how to build a coalition, how to generate quick wins. Those are design questions, and they matter, but they come second. The first question is whether the organization faces a technical problem or an adaptive challenge, a first-order adjustment or a second-order shift, a planned intervention or an emergent pattern. The answer determines which tools apply, which models fit, and what timeline is realistic. Getting it wrong is the single most reliable source of change failure, and it produces the characteristic symptom: the change is implemented, the project is closed, and the organization has not changed.
The practical skill, then, is diagnosis before design. It requires asking what the change is actually asking people to do — not what the project plan says, but what the people in the middle of it experience. If they are being asked to use a new system, the problem is likely technical. If they are being asked to think differently about their work, their competence, or their identity, the problem is adaptive, and no amount of project management will resolve it. The distinction is not always clean, and many real changes contain both elements. But the diagnostic habit — pausing to classify before committing to a method — catches the most common error in the field and points toward a more honest assessment of what the change will take.