Design and Behaviour Design Are Not the Same Job

Design and Behaviour Design Are Not the Same Job

Both shape what people do. One is judged by the artefact, the other by the behaviour. Confusing them is why so many well-designed products, services and change programmes fail to change anything.

1. Two questions that look alike

Ask a designer to improve a hospital's hand-hygiene programme and you will get better dispensers, clearer signage, a cleaner poster and perhaps a more pleasant soap. Ask a behaviour designer and you will get a count of how often hands are actually washed, broken down by ward, shift and role; a map of where in the care pathway the behaviour is skipped; an analysis of why (no dispenser at the point of need, time pressure, no visible norm, no feedback); and a set of interventions ranked by expected effect, most of which are not artefacts at all: moving the dispensers to the bedside, making compliance rates visible on the ward, changing who is expected to prompt whom.

Both professionals care about people. Both will say they are "user-centred". Both may produce a dispenser. The difference is not in the output but in the question each is answering, the evidence each accepts as success, and the range of levers each considers legitimate.

2. Why the confusion is so common

Three forces push organisations to collapse the two.

Design has absorbed the language of behaviour. "Nudge", "friction", "habit loop" and "choice architecture" are now part of every product designer's vocabulary. The words travelled faster than the method. A designer who talks about friction is not necessarily measuring it, and a team that calls a feature a nudge has not necessarily tested whether it nudges.

Artefacts are visible and behaviour is not. A redesigned app, a new signage system or a refreshed onboarding flow can be shown to a steering committee. A two-percentage-point change in a behaviour, measured against a baseline over six months, cannot be put on a slide with the same impact. Organisations fund what they can see.

Good design does change behaviour, sometimes. This is the honest reason for the confusion. A clearer form gets completed more often. A better-placed button gets clicked more. Chandon and colleagues showed that extra shelf facings raise a brand's evaluation entirely through the attention they attract, which is design doing behavioural work (Chandon et al., 2009). The same study showed the limit: attention gained did not always carry through to evaluation. When the only barrier to the behaviour is the artefact, fixing the artefact is behaviour design, and the designer deserves the credit. The error is to generalise from these cases to all the others, where the barrier is motivation, social context, competing demands or the absence of any need for the behaviour at all, and no amount of artefact improvement will reach it.

3. What behaviour design does that design does not

It starts from a count

Before anything is sketched, behaviour design establishes who does what, how often, under which conditions. This baseline is what every later claim is measured against. Design projects usually begin with a brief, a persona and a set of pain points, which describe experience rather than behaviour. The two are related but not the same: people report pain points they have adapted to and omit the behaviours they do not notice themselves doing.

It diagnoses before it designs

The COM-B model asks whether the missing ingredient is capability, opportunity or motivation. The answer determines the intervention class. A capability gap needs training or simplification; an opportunity gap needs changes to the environment or the social context; a motivation gap needs incentives, feedback or identity work. Design, lacking this diagnostic step, defaults to the intervention it is good at, which is improving the artefact, and improving the artefact only addresses a subset of opportunity and capability gaps.

It treats the artefact as optional

The most powerful behavioural interventions often involve no new artefact. A changed default. A reallocation of who is responsible. A different time of day. The removal of a step. Making the desired behaviour unnecessary by producing the outcome another way. A designer's brief rarely permits these moves because they fall outside the thing being designed. A behaviour designer's brief requires considering them first, because they are cheaper and more reliable than anything that depends on people noticing and using a new artefact.

It measures behaviour, not satisfaction

Usability testing establishes that people can use the artefact and like it. Neither predicts whether they will use it when nobody is watching, next month, under time pressure, with their phone buzzing. Behaviour design insists on field measurement of the actual behaviour, ideally against a comparison group, because the gap between "can and likes" and "does" is where most interventions die.

It takes effect sizes seriously

Behavioural science has an unusually honest literature on how much its interventions achieve: typically a few percentage points for a light-touch nudge, an order of magnitude more for defaults and structural change. A behaviour designer uses these figures to size the intervention to the goal. Design has no equivalent calibration, which is why a redesign is so often expected to deliver a transformation and so often delivers a modest improvement that is then recorded as a disappointment.

4. What design does that behaviour design must not forget

The argument so far could be read as a case for replacing designers with behavioural scientists. It is not. Behaviour design without design produces interventions that are correct and unusable: the right default buried in an incomprehensible settings page, the right feedback delivered in a chart nobody can read, the right prompt at the right moment in language that insults the person receiving it. Three things design brings that behavioural work routinely lacks:

• Craft in execution. An intervention is experienced as an artefact. If the artefact is ugly, confusing or disrespectful, the behavioural mechanism never gets a chance to operate. Attention, the scarcest resource in any behaviour change, is won or lost on design quality.
• Synthesis across the whole experience. Behavioural interventions are often point solutions. Design thinks in journeys, systems and coherence. A behaviour change that is locally optimal and globally jarring will be worked around.
• Respect for the person. Design's user-centred tradition, at its best, insists on dignity, agency and delight. Behavioural science, at its worst, treats people as bundles of biases to be steered. The design tradition is a useful corrective, and the ethical questions behaviour design must answer are often questions design has been asking for decades.

Three of the levers named above have their own papers in this series: attention, which design is best placed to win and behaviour design most often forgets to measure; behaviour-free design, which is the lever design briefs can never reach; and the calibration of goals to effect sizes, which tells you how much any artefact can be expected to deliver.

5. Where CLEAR places the line

The CLEAR framework is a behaviour-design method in the sense this paper has defined: it starts from an outcome and measurable key results (Clarify), diagnoses the system before choosing any intervention (Leverage), tests on a small scale against the baseline (Experiment), measures behaviour rather than satisfaction (Analyse), and iterates (Refine). Design enters at two points. In Leverage, when the map shows that the artefact is a genuine constraint, design is the intervention. In Experiment, every intervention, artefact or not, must be executed well enough that its mechanism can operate, and that is design work. CLEAR does not replace design. It tells you when design is the answer and holds it to a behavioural standard of evidence when it is.

6. Consequences for practice

Brief differently. A design brief specifies the artefact and its qualities. A behaviour-design brief specifies the behaviour, the people, the baseline, the target and the time horizon, and leaves the intervention open. Most organisations write the first kind of brief and expect the second kind of result.

Hire for the question, not the output. Designers and behaviour designers both produce prototypes and run tests. The difference is what they test for. Ask a candidate how they would know whether their last project worked. A designer will describe the artefact and its reception. A behaviour designer will tell you the baseline, the measured change and the confidence they have in it.

Measure behaviour from the start. If behaviour is the goal, the baseline must exist before the design work begins, or there will be nothing to compare against and the project will be judged on the artefact by default.

Let behaviour design decide whether to design. The most valuable outcome of a behavioural diagnosis is sometimes the conclusion that no new artefact is needed, that the lever is a default, a responsibility or a removed step. A process that begins with a design brief can never reach that conclusion.

7. Conclusion

Design and behaviour design share tools, vocabulary and a concern for people. They differ in what they are for. Design makes things good. Behaviour design makes behaviour change, using good things when good things are what the situation needs and other levers when they are not. The confusion between them costs organisations twice: once when a design project is expected to change behaviour it was never equipped to reach, and again when a behavioural insight is executed so poorly that it never gets the chance. Keep the questions separate, staff each with the people who can answer it, and judge each by its own evidence.

© 2026 Erik Bohjort · clear-framework.com