The Behaviour Change Wheel: Making Sure Change Sticks

You can build a well-funded, well-intentioned intervention and still get poor results if you diagnose the wrong problem.

In this book review, I explore how The Behaviour Change Wheel helps organisations understand what is really preventing a behaviour before deciding whether training, communication or something entirely different is needed.

When the obvious solution solves the wrong problem

For years, road safety bodies assumed young drivers crashed more often because they lacked the necessary driving skills. Advanced driver training seemed like the obvious solution.

It didn’t work.

The problem was not capability. Many novice drivers already knew how to drive safely. Their motivation encouraged different behaviour: speeding, taking risks and paying less attention.

More training could not fix the real barrier.

This is the central value of The Behaviour Change Wheel by Susan Michie, Lou Atkins and Robert West. It slows down the jump from “this is the behaviour we want” to “let’s train people” or “let’s communicate it more clearly.”

Start by defining the behaviour

The framework begins with a deceptively simple question: what exactly do you need people to do differently?

A behaviour needs to be observable. Who needs to do what, when, where and how often?

Instead of saying, “We need people to adopt AI,” we might ask:

Which people, in which roles, should use which tool, for which part of the workflow, to make which decision or produce which output?

That level of specificity changes the conversation. Once the behaviour is clear, we can start to understand why people may be ignoring a useful tool.

What is COM-B?

At the centre of the Behaviour Change Wheel is COM-B. It proposes that behaviour depends on three conditions:

Capability: Does the person have the physical and psychological ability, knowledge and skills?
Opportunity: Does the physical and social environment support the behaviour?
Motivation: Do conscious beliefs, habits and emotional responses encourage it?

A weakness in any one of these areas can prevent the desired behaviour.

Once you identify the real barrier, the framework helps you select an appropriate intervention. Training may help with a capability gap. Environmental restructuring might be more useful when systems or workflows get in the way. Modelling may help when the behaviour is not visible or normal among colleagues.

This diagnostic approach also sits behind behaviour-led learning design: understanding the cause before building the solution.

What this looked like in practice

A client once asked us to run a workshop to help subject matter experts use a new authoring tool.

When we investigated, the main barrier was not skill. The experts were worried about producing poor-quality learning because there were no clear standards, templates or quality controls.

We created those conditions instead of relying on a workshop alone. Around 50 subject matter experts gained the ability to build digital learning independently, with early projections suggesting savings of approximately 2,000 hours and £100,000 in internal costs, for only one business unit.

The presenting request assumed a capability gap. The main barrier was a mix of opportunity and capability.

Choosing an intervention you can defend

The book goes beyond diagnosis. It maps COM-B barriers to nine intervention functions, seven policy categories and 93 defined behaviour change techniques.

It also introduces APEASE: six criteria for assessing options based on affordability, practicability, effectiveness, acceptability, spillover effects and equity.

That gives leaders a structured way to explain why an intervention was selected rather than defending a familiar solution on instinct.

The examples in the book mainly come from health, but the method applies anywhere people need to adopt a tool, follow a process, manage risk or work differently.

If your organisation is planning a change, these three questions to ask before an AI rollout offer a practical place to start.