Review

You should know what human behaviour you’re trying to change before you do any meaningful product development work. It creates clarity for the team and can give the team a leading tractable metric to get after. This book outlines a simple and seemingly bulletproof approach to behaviour change.

You Might Also Like

Curious about how are these recommendations generated? Learn more.

Designing for Behaviour Change

I made product management my profession, I’ve long been a fan of Behavioural Economics and Atomic Habits is one of my favourite books, so I was always going to like this book. It builds on prior theory, but the authors contribution is to present a few easy to follow frameworks to apply behaviour change principles to product management. Given most product outcomes are reliant on changing people’s behaviour, this is a worthwhile read for anyone building products.

andrewclark.co.uk

Hooked

The book details how habit-forming products are created using the Hooked Model. It covers the four phases of the model - Trigger, Action, Variable Reward, and Investment - and provides insights on designing products that capture users' attention and become habitual. It emphasizes the importance of understanding internal and external triggers, making actions easy to take, providing variable rewards, and encouraging user investment.

andrewclark.co.uk

Nudge

What is it about? Was it worth reading?

andrewclark.co.uk

Key Takeaways

The 20% that gave me 80% of the value.

• Nearly everything we create is designed to shape behaviour, yet we rarely acknowledge this; start with a clear behavioural goal for your creation.
• Behaviour is influenced by promoting pressures (motivations making behaviour more likely) and inhibiting pressures (factors making behaviour less likely); designing interventions involves altering these pressures.
• The Intervention Design Process (IDP) involves:
• Observing and validating a behavioural insight—a gap between current and desired behaviours.
• Crafting a behavioural statement to define the desired outcome.
• Mapping pressures influencing the behaviour.
• Designing interventions to modify pressures.
• Evaluating interventions through pilots, tests, and scaling.
• A behavioural statement outlines the desired behaviour change: When [population] wants to [motivation], and they [limitations], they will [behaviour] (as measured by [data]).
• Pressure mapping identifies promoting and inhibiting pressures affecting behaviour, serving as levers for intervention design.
• Generate numerous intervention ideas without constraints, then select promising ones based on effectiveness and feasibility.
• Conduct an ethical check before piloting interventions, focusing on:
• What behaviour you're changing.
• How you're changing it.
• Ensuring alignment with the population's motivations and values.
• Differentiate between the intention-action gap (people intend but fail to act) and the intention-goal gap (people have a goal but reject the behaviour to achieve it).
• Ethical interventions align with existing motivations and do not impose costs that outweigh benefits or conflict with other motivations.
• Transparency and responsibility are crucial; openly communicate your intentions and be accountable for the outcomes of your interventions.
• Pilots are small-scale, operationally "dirty" interventions expected not to work, aimed at minimising resources and impact while testing for behaviour change.
• Pilot validation uses qualitative and quantitative data to assess if an intervention shows promise, accepting higher p-values due to small sample sizes.
• Tests are larger-scale interventions with greater operational diligence, assessing whether scaling is worthwhile based on effect size and operational cost.
• Scaling decisions are summarised as: We are [confidence] that [intervention] will [direction] [behaviour] (as measured by [data]). Scaling this requires [effort] and will result in [change].
• Documenting failures and continuous measurement are essential to avoid repeating mistakes and to ensure interventions remain effective over time.

Deep Summary

Longer form notes, typically condensed, reworded and de-duplicated.

Introduction

• Nearly everything we interact with constructed to shape behaviour - yet we rarely connect our desire to create with the goal of behaviour change. Why not start at the end? With a clearly described behaviour change that is the explicit goal of your creation.
• For any behaviour there are promoting or inhibiting pressures. Identifying and consciously influencing the strength of those pressures is the basis of designing for behaviour change.
• Promoting Pressures: Why would people want to do that in the first place? Motivations and factors that make behaviour more likely.
• Inhibiting pressures: Why aren’t they doing it already? The factors that make the behaviour less likely.
• Interventions are the things we build to change the pressures and thus the resultant behaviours. The Intervention Design Process (IDP) is the authors approach to just that.
• Building habit forming products that circumvent our intentions by exploiting the brains need to save cognitive resources is unethical.
• BUT there’s plenty of work to do in ethical behaviour change - where we create the conditions that allows us to act on our original motivations.
• Behaviour change is best done with a transparent goal that is clear and shared by both you and those whose behaviours you seek to change.
• In behaviour change - most pressures that can be used to make behaviour more likely can also be used to make it less likely when reversed.

Part 1: The Basics of Behaviour Change

Chapter 1. The Intervention Design Process

The Intervention Design Process:

1. Observe, define and validate a behavioural insight - a gap between how people behave now and how we want them to
2. Flesh out the insight into a behavioural statement - and map the pressures that create the current state of behaviour (which provide levers for change)
3. Design interventions to modify behaviour by altering the most powerful pressures - select the most promising ethical interventions.
4. Evaluate and refine interventions through progressively larger-scale experimentation phases (Pilot, Test, Scale)
5. Put in place continuous monitoring to measure impact of interventions over time

• When validating an insight - don’t rely on data alone try to to observe qualitatively what also appears to be true quantitatively. Do different forms of research and aim for convergent validity: evidence that’s greater than the sum of its parts - when disparate sources support the same conclusion.
• Progressively larger-scale experimentation means you don’t have to be right all the time - you just need to make small mistakes instead of big ones.
• Expect the competition to fast follow your work but they may not understand the pressures you’re altering, which can be a significant handicap.
• On timings:
• 1 week to understand the insight
• 2 weeks to validate and explore potential pressures
• 1 week to design and select interventions
• 2 weeks to get pilots up and running
• 2 weeks to get back some early quantitative and qualitative signals for validation

Chapter 2: Potential Insights and Insight Validation

• A potential insight expresses the distance between how people behave now and how we want them to. Once validated they allow us to start understanding the gap and how to design interventions that bridge it.
• There are four major types of potential insights: quantitative, qualitative, apocryphal (common knowledge not directly observed in your organisation), and external (e.g. research papers).
• Finding novel potential insights is about noticing something that hasn’t been noticed before. Let the data guide you.
• Once a potential insight is surfaced, it must be validated through insight validation. Look for convergent validity: evidence from diverse sources that supports the same conclusion. Think of it like building a table—you want multiple legs to hold up your conclusion. The more varied your sources of convergent validity, the harder it is to fall victim to confirmation bias. By triangulating from data and observation and structurally resisting groupthink, we can eliminate risks and increase our chances of a successful, scaled outcome.
• Focus on what you’re validating - don’t fall into the trap of using research to post hoc affirm whatever you already think you know.

Slow is smooth, smooth is fast. There is no such thing as wasted time in validation.

Better intervention design happens when you have as many potential insights as possible at the beginning of the process—a big, wide funnel of opportunities for behaviour change that slowly gets narrower as we hone in on pressures we’re able to successfully design interventions around.

• Focus those insights on behaviours, not on how we create them. You have to fall in love with the problem, not the solution.

Chapter 3: Behavioural Statement

• To put behaviour change at the centre of our creation process we need to clearly express the behavioural outcome that we want to achieve. We have to start at the end.
• A behavioural statement is an articulation of the world we are trying to create, from a behavioural perspective. It lays the foundation for our next steps (pressure mapping and intervention design).
• Most organisations are biased toward focusing on processes and immediate actions and how we’re doing them, rather than on their outcomes.
• A behavioural statement is a set of conditions that can either be satisfied or not and come together in a single sentence:
• When [population] wants to [motivation], and they [limitations], they will [behaviour] (as measured by [data]).
• Population = the group of people whose behaviour you are trying to change
• Motivation = the core motive for why people engage in a behaviour
• Limitations = the binary preconditions necessary for the behaviour to happen that are outside your control
• Behaviour = the measurable activity you want people to always do when they have the motivation and limitations above
• Data = how you quantify that they are doing the behaviour
• Example: When people want to get from Point A to Point B, and they have a smartphone with connectivity and an electronic form of payment and live in San Francisco, they will take an Uber (as measured by rides).
• Make sure you write a behavioural statement not a vision statement.
• Make sure you focus on a real behaviour
• Make it absolute and brave (e.g. use always, everyone). It increases the likelihood you’ll achieve something big.
• Evolve it as you need.
• Put it up on the wall where everyone can see it.
• Use it to make decisions making - you can explicitly compare available options against the behaviours they are likely to produce.
• You can scale behavioural statements down as needed, creating a clear link between individual accountability and the organisation's overall behavioural goal. This ensures everyone understands the importance of their role and has a defined area of responsibility. It also establishes a clear accountability structure.
• Finding the right size of behavioural statement comes back to autonomy and accountability. It should be as large as possible but small enough that the person could be held individually accountable for its success or failure.
• A behavioural statement is similar to an objective and key result (OKR). “As measured by data” is really just the KR and the rest is just the O, but phrased in such a way as to be observably descriptive of the world you’re trying to create. If you’re already doing OKRs for your planning process, you’re already ready to simply sub in behavioural statements and reorient entirely toward behaviour.

Chapter 4. Pressure Mapping and Pressure Validation

• Behaviour change is about interventions that move us from the world as it is (Point A) to the world as we want it to be (Point B). Our insights describe Point A and our behavioural statement describes point B. Next map the pressures that create the distance between the two, so you know what it is that needs to change.

When we talk about designing for behavior change, we are actually talking about changing the pressures that determine the behavior, rather than directly changing the behavior itself.

• You can’t make your child sleep - but you can control the pressures that act on them (tiring them out, blackout curtains etc. All you can do is to change the pressures on both sides of the equation to maximise the likelihood of the behaviour you want.
• Mapping the competing pressures that create behaviour is a crucial step.
• Competing pressures create a probabilistic balance - the result of each particular episode is unknowable but populations in aggregate are more predictable. Each person might not do what we expect all the time, but the population is affected by broad pressures.

• The arrows represent the balance of competing pressures that create our behaviour. What we actually do is determined by the net product of those forces.
• Promoting pressures make a behaviour more likely
• Inhibiting pressures make a behaviour less likely.
• If the promoting pressures overcome the inhibiting pressures, we act. If the inhibiting pressures are stronger, we don’t.
• Draw the up and down arrows, and start listing pressures on each side. Start adding or removing pressures until you understand, at least broadly, what exists now.
• More affects your behaviour than you realise (e.g. the colour of M&Ms influences how appealing they are to eat). That’s why we need insights and validations and why we’ll eventually run pilots; we are poor at introspecting our own motivations.
• A behavioural science trick that helps with pressure mapping is to flip the scenario on its head and taking it to the extreme - imagine if the opposite conditions were true and thinking about how that would influence the behaviour.
• Pressures exist along a spectrum (e.g. availability of M&Ms isn’t black and white) and by adjusting the strength of the inhibiting pressure we can change the resulting behaviour by corresponding degrees. We can do this by ramping up or down all sorts of inhibiting pressures.
• Pressures are context dependent; they interact with motivation, population, and one another in unique ways. The IDP features emphasises both validation and the pilot/test/scale motion: because the only way to know you’ve identified the right pressures is to use them to build interventions that actually change behaviour.
• What is more important than the specific pressures is considering both sides of the equation.
• We have a bias to focusing on promoting pressures.This predictable error means that there is also predictable untapped upside.
• Uber don’t need to increase your desire to get to your destination - instead they drive retention by reducing inhibiting pressures (cost, wait time, cashless payment etc)
• Your pressures should come from research and convergent validity.
• Focus on just inhibiting or promoting pressures and then switch. Try reversing the polarity of your behavioural statement; if you think about how to make sure no one ever takes an Uber, you’ll likely uncover some pressures you can use to make sure they do. And make sure you have a diverse room.

Chapter 5: Intervention Design and Intervention Selection

• Don’t get overconfident and rush into listing the pressures and putting an intervention against each of them. Down that path lies madness.
• A single pressure to give rise to many interventions
• One intervention can satisfy many pressures
• By formalising our understanding of pressures, we can see how to combine and sort them.
• The pressures are the levers, interventions are how we pull them, hopefully in the right order and with the right strength.
• As this point in intervention design volume is key - generate as many ideas as possible.
• Suspend disbelief and resource constraints while generating ideas.
• Use breaks, flip the statement, focus on individual pressures, or combine novel ones, create artificial limitations and what-ifs.
• Map each intervention to the pressures though.
• Once out of ideas - focus on any pressure that seems to dominate the others (in both quantitative and qualitative research) and create several interventions around it.
• Aim to get more than 20 unique interventions listed.
• Intervention selection - involves making a judgment call about which interventions you’re actually going to bring to pilot. There’s no exact science but were making smart bets:
• You can’t pilot everything BUT you also shouldn’t pilot only one thing. Make sure they’re different - remember most won’t work, so don’t cluster around a single idea.
• Shoot for optimum distinctiveness: a range of options that together cover as much of the spectrum as possible, but with relatively little overlap.
• You can combine interventions. Remember the goal is behaviour change, not knowing precisely which part of an intervention drives the behaviour.
• Scale down interventions so they’re manageable. Is there an intervention that will work for only a portion of the population? That can be created only under very specific circumstances? That is available only at certain times?
• Look for the cheapest, easiest-to-create, most broadly reaching interventions that still change the behaviour.

Chapter 6. Ethical Check

• Everything we create aims to change behaviour, yet we seldom acknowledge this, and accusing someone of intentionally changing behaviour is often seen as slanderous.
• Self-serving bias leads us to justify our own actions differently from others', affecting our ethical judgements.
• Interventions can harm others; thus, it's incumbent upon us to use processes like the IDP responsibly, regardless of our specialisation.
• Our interventions will always appear ethical through the lens of our own values, so understanding potential issues helps us avoid future regret.
• Having selected what interventions we intend to pilot - stop and do an ethical check based on two factors:
• What behaviour we are changing
• How we are changing behaviour
• These factors are intrinsically linked to two fundamental behavioral gaps in behavioral science: the intention-action gap and the intention-goal gap.
• The Intention-Action Gap occurs when individuals have both the motivation and the intention to perform a behaviour but fail to execute it due to barriers like forgetfulness, lack of time, or environmental constraints.
• By explicitly aligning the outcome behaviour with an originating motivation, we ensure there is no ethical concern that we are forcing a behaviour on an unwilling population, because they have already endorsed it.
• E.g. If people intend to exercise more but don't due to a busy schedule, an intervention could be providing quick workout routines that fit into their day. Since they already want to exercise, we're simply helping them overcome practical barriers.
• The Intention-Goal Gap is more ethically complex and occurs when individuals have a goal but lack the intention to engage in the behaviour required to achieve it. They might want the outcome but reject the necessary actions.
• In this case, we must carefully consider both what behaviour we're trying to change and how we're attempting to change it. The motivation component isn't sufficient because the person has already rejected the behavior as a means to honor that motivation.
• Example: Someone wants to stay healthy (goal) but refuses to get a flu shot (behaviour). Even if they understand that the flu shot contributes to health, they have no intention of getting it.
• Determining the ethics in an intention-goal gap situation has a very simple rule: if your outcome behaviour is not the result of any of the motivations of the population, it is unethical. To put it differently, if you can’t construct a valid behavioural statement, you’re out of bounds. You cannot ethically change behaviour in intention-goal gaps unless we find an alternative motivation that aligns with their values without significant drawbacks.
• Flu Shot Example: If individuals refuse the flu shot for personal reasons but value community health, we can ethically promote the flu shot by aligning it with their motivation to protect others.
• Ensure that promoting one motivation doesn't significantly impede another. The intervention should offer benefits that outweigh any potential costs to other valued motivations.
• Transparency and Responsibility: Being open about our interventions and willing to take responsibility for the outcomes helps navigate these ethical complexities. It's important to communicate intentions clearly and involve the population in understanding the trade-offs.
• How We Change Behavior Matters: Even if the what is ethically sound, the how can introduce ethical issues, especially if interventions are manipulative or coercive.
• Intervention Impact: The intervention itself should not impose costs that outweigh its benefits. For example, using fear-based messaging might achieve behavior change but at the cost of increased anxiety.
• Striving for Transparency: While full transparency might not always be possible due to practical constraints, it's incumbent upon us to be as open as possible about our methods and intentions.
• Accepting Responsibility:
• Accountability: By creating interventions specifically to change behaviour, we accept responsibility for the results of that behaviour change.
• Ethical Practice: This means we should be prepared to publicly describe and justify both the outcome behaviour and the intervention methods.

Chapter 7: Pilot and Pilot Validation, Test and Test Validation, Scale Decision and Continuous Measurement

• In the IDP, "pilot," "test," and "scale" are sequential stages that increase certainty about an intervention's ability to change behaviour, as well as information about its size and cost.
• Pilots are tightly scoped interventions expected not to work; they are conducted in an operationally dirty way to minimise impact on the organisation and resource investment.
• Keeping pilots small and operationally dirty reduces abrasion on both customers and employees by minimising their investment and expectations.
• Speed and resource efficiency are crucial; focus on finding the lowest-fidelity version of an intervention that still results in behaviour change. If it takes longer than two weeks to get into the field, scale it back and go smaller.
• Pilot validation involves qualitative and quantitative confirmation that you're headed in the right direction; due to small sample sizes, statistical significance isn't expected.
• The pilot is the first time we'll actually measure the behaviour change resulting from an intervention.
• If an intervention doesn't create the desired behaviour change, decide to revise and rerun it or kill it and return to your pressure map and intervention design.
• Ultimately, we focus on changing behaviour, not the fate of individual interventions; if some interventions fail but others succeed, move on without hesitation.
• Statistics help us validate interventions, recognising that people are neither perfectly predictable nor perfectly unpredictable.
• Use within-subjects and between-subjects tests to gather evidence that an intervention might have worked.
• Quantitative research reports two numbers: effect size (how much behaviour changes) and p-value (how confident you can be that the effect is due to the intervention).
• In pilots, higher p-values (e.g., p = 0.2) are acceptable because the goal is to decide whether to proceed to a larger test, not to achieve statistical significance.
• Pilots are small to reduce confirmation bias; less effort invested makes us more willing to recognise when interventions don't change behaviour.
• Tests are like pilots but with larger populations and greater operational diligence; they help decide if scaling the intervention is worthwhile by assessing operational cost and effect size.
• Scaling decisions are summarised as:
• We are [confidence] that [intervention] will [direction] [behaviour] (as measured by [data]). Scaling this requires [effort] and will result in [change]:
• Confidence = based on p-value but phrased colloquially
• Intervention = what the intervention is
• Direction = whether it increases or decreases the behaviour
• Behaviour = the measurable activity you established in your behaviour statement
• Data = how you quantify that your population is doing the behaviour
• Effort = the resources required to scale
• Change = based on p-value but phrased colloquially
• For example: We are very confident that sending personalised flu jab letters based on member health motivations will increase the rate of getting a flu jab (as measured by flu jab claims). Scaling this requires about ten hours and £4,000 and will result in about five hundred additional flu jabs.
• Documenting failures is important to avoid the "file drawer problem" and improve future interventions; the IDP helps our interventions be inherently resilient only if we implement methods to detect change.