Better transformation starts with a better definition of the problem

One of the least examined but most consequential acts in transformation is deciding what the problem actually is. 

In many organisations, that stage is treated as obvious. A performance issue is visible, pressure builds to respond, and attention moves quickly to interventions, workstreams, and delivery plans. But the evidence base suggests that this early move is often where later weakness begins. The way a problem is defined shapes what is seen, what is ignored, what later appears fundable or credible, and which interventions are treated as reasonable responses. In that sense, problem definition is not an administrative prelude to transformation. It is one of its most strategically important acts.[1]

Why the problem-definition stage is often underestimated

A useful way to understand this is to recognise that different traditions mean different things by “defining the problem.” Evidence-based change management is concerned with the quality of the judgement itself. Rousseau’s work argues that better change decisions come from integrating multiple sources of evidence – scientific, organisational, stakeholder, and practitioner evidence – rather than relying on preference, hierarchy, or inherited assumptions alone.[1] The practical implication is that a problem is not well defined simply because it is strongly asserted. It is well defined when it is supported by a disciplined reading of different forms of evidence and when competing interpretations have been tested rather than assumed.

Problem framing work in public and social innovation adds something different. Van der Bijl-Brouwer treats problem framing as an expert practice because complex public problems do not arrive neatly bounded. They must be interpreted, structured, and made workable.[2] That matters here because many organisational problems are not initially seen in their most decision-relevant form.

• A waiting list problem may be framed as a demand issue when the more meaningful problem is variation in pathway flow.

• A theatre productivity problem may be framed as underutilisation when the more decision-relevant issue is how capacity, scheduling logic, staffing, and throughput interact.

The key point is that framing is not neutral. It determines the intervention space. Frame the problem too narrowly and the solution set is narrowed with it.[2]

Implementation and quality-improvement literature take a third perspective. The QUERI pre-implementation roadmap is explicit that organisations need to define the “quality gap” before moving into implementation planning.[4] That means describing the current problem, specifying who and what is affected, clarifying the desired improvement over current performance, and using organisational data to shape the work before implementation begins.[4] This is a more operational framing than general problem-framing theory, but it is useful precisely because it makes clear that implementation should begin from a defined gap rather than from a preferred intervention looking for justification.

Why defining the problem is harder than it looks

Taken together, these traditions point to a more demanding view of problem definition than is often seen in practice. It is not just a matter of assembling data, listening to stakeholders, or spotting a poor KPI. It is the disciplined act of converting a broad concern into a problem statement that is analytically credible, operationally meaningful, and useful for later decision-making. That is harder than it sounds, because the visible symptom is often not the most useful framing of the issue, and because organisations are rarely short of opinions about what the answer should be before the problem has been properly described.[1][2]

There is a simple day-to-day equivalent of this outside healthcare that many readers will probably recognise in some form. In my own family life, if our household starts running late in the morning, it is easy to jump straight to the apparent solution that everyone needs to get up earlier. That sounds like a clean answer, until the next morning shows the problem was not simply what time everyone got up. In practice, the issue may sit somewhere between two working parents trying to get out of the house on time, children moving into their teenage years but not yet consistently organised enough to manage themselves, the dog choosing that exact moment to bolt down the street, one part of the routine taking longer than expected, or small delays quietly compounding until the whole morning starts to wobble. Bags may not be packed, things may not be where they should be, and if the children happen to be ready a little earlier than usual, that may simply create enough spare time for a quick game of FIFA before leaving. “Start earlier” sounds like a solution, but it may only be responding to the symptom. The same logic applies in transformation:

visible pressure often produces quick answers, but not always ones that properly identify the bottlenecks and interacting issues that have compounded into the symptom now being seen.

Why organisations move to solutions too early

This is where premature solutioning becomes so damaging.

In practice, many programmes move to solution design before the problem has been defined in a way that is stable enough to support later decisions. That happens for understandable reasons: leaders want pace, pressure creates demand for visible action, and a proposed intervention can feel more tangible than a period of disciplined diagnosis. But the cost of that speed is often hidden. Once a preferred solution enters the room too early, evidence starts being used to support it rather than to challenge it.

Stakeholder engagement also changes character. Instead of beginning with a shared exploration of the problem, it becomes a softer exercise in securing acceptance for a direction that already feels chosen. That is one reason poorly timed solutions can feel imposed even when the broad direction is sensible.[1]

Human-centred design sharpens this point. Chen and colleagues argue that implementation science and human-centred design are complementary because each addresses a weakness in the other: implementation science contributes structure and generalisable discipline, while human-centred design surfaces how problems are experienced in real settings, including needs, frictions, and contextual constraints that may not appear in routine data.[3] For transformation work, that is highly relevant. Data may show that a problem exists, but not necessarily how it is encountered in practice or why an apparently sensible intervention may still be resisted, bypassed, or rendered impractical.

Good problem definition therefore requires more than analytic description. It requires an understanding of how the problem is lived within the system that will later have to change.[3]

Why good problem definition improves engagement and decision quality

This is also why defining the problem properly improves later engagement.

In my experience, implementation is stronger when engagement begins with a credible definition of the problem rather than an over-specified solution. Often the broad direction of the intervention may already be apparent, with some local tailoring still needed. But introducing the answer too early can weaken the very ownership needed to carry it and appear as tokenism. People are more likely to engage constructively when they can see the nature of the gap, where it sits, and why it matters, rather than being asked to align themselves with a model that feels already settled. The distinction is subtle but important:

engagement around the problem tends to invite interpretation, challenge, and ownership; engagement around a fixed solution more often invites compliance or resistance.

That is not simply a communications issue. It is a function of how the problem has been framed.[2]

Poor problem definition also distorts later decisions in more technical ways. It weakens options appraisal, because options are only as good as the baseline against which they are being judged. It weakens prioritisation, because leaders are asked to allocate effort against symptoms rather than underlying constraints. It weakens implementation, because interventions are selected without a stable enough view of where the operational or financial leverage actually sits. And it weakens accountability, because later underperformance becomes harder to interpret: was the intervention poor, or was the original problem statement too vague, too narrow, or simply wrong? Those downstream effects are one reason this stage deserves more scrutiny than it usually receives.[1][4]

What stronger problem definition looks like in practice

What distinguishes stronger practice is not endless analysis. It is the ability to define the problem to the point where the intervention space becomes more credible, more bounded, and more decision relevant. In practice, that usually means combining multiple evidence forms rather than privileging one.

• Quantitative data may define the scale and distribution of the gap.

• Organisational evidence may show where the constraint sits operationally.

• Stakeholder evidence may reveal where the problem is experienced differently across the system.

• Practitioner evidence may expose the difference between what appears to be happening and what is actually happening in live delivery.

The skill lies in synthesising these into a framing that can support action without collapsing prematurely into a preferred answer. That is expert work, not a routine early-stage exercise.[1][3][4]

I have seen the practical value of this in NHS transformation work. In one project supporting the decommissioning of trauma theatres, the issue was not simply whether performance could be sustained after the move. The more difficult task was to define the true operational, financial, and capacity gap across the wider theatre footprint in a way that could support decision-making by making the full implications of the planned change visible. Our gap analysis module brought together hundreds of disparate data points to generate a much clearer view of that gap across the pan-trust model, expanding the problem beyond the services immediately affected. That created a firmer basis for scenario modelling and options appraisal, rather than forcing decisions to rely on intuition or partial views of the issue.

In practice, it helped clarify the true scale of rightsizing required. The important point is not simply that more data was used. It is that the problem was defined in a way that became operationally and financially decision-relevant by showing how the different factors interacted and how pressures in one part of the system affected another.

Conclusion

Ultimately, better transformation usually starts with a more disciplined definition of the problem than most organisations realise. The challenge is not just to describe what is going wrong, but to frame the problem in a way that is analytically credible, operationally meaningful, and useful for later decisions. That requires more than data, more than stakeholder opinion, and more than a preferred solution looking for justification. When this work is done superficially, the weakness often travels downstream into options appraisal, engagement, implementation, and performance interpretation.

When it is done well, later change is more credible because it is grounded in a problem that has been understood properly in the first place.[1][4]

References

[1] Rousseau DM. Evidence-based change management. Annual Review of Organizational Psychology and Organizational Behavior. 2022.

[2] van der Bijl-Brouwer M. Problem framing expertise in public and social innovation. She Ji: The Journal of Design, Economics, and Innovation. 2019.

[3] Chen E, Leos C, Kowitt SD, Moracco KE. Complementary approaches to problem solving in healthcare and public health: implementation science and human-centered design. Translational Behavioral Medicine. 2020.

[4] Goodrich DE, Miake-Lye I, Braganza MZ, et al. The QUERI pre-implementation roadmap: steps to improve implementation of evidence-based practices. In: The QUERI Roadmap for Implementation and Quality Improvement. 2020.