Designing for Behavior Change, 2nd Edition

Preface

Eight years ago, when I started writing the first edition of this book, behavioral science consisted of a few prominent researchers and PhD students scattered across business schools and psych departments, along with some humorous books about the mistakes people make. Very few companies were trying to bridge the gap between academic theory and real products that helped real people in their lives. Personally, I had to scrounge to find companies where anyone was even familiar with the research, let alone formally trained in it. The idea of intentionally designing products to change user behavior was strange or even alarming to many people. There were pockets of innovative work—from the UK government’s Nudge Unit to BJ Fogg’s Persuasive Technology Lab to the behavioral consulting firm ideas42—but they were not well known outside of their communities.

That’s all changed now. In partnership with the Action Design Network and the Behavioral Science Policy Association—two nonprofit organizations dedicated to fostering the practical application of behavioral science, which didn’t exist eight years ago—we recently surveyed the landscape of applied behavioral science. Over two hundred teams, representing behavioral science teams with over 1,500 members, responded—and we know there are many more out there.

There are now dedicated teams applying behavioral science to develop new products, communications, and policies that serve their users across the world: from Qatar to Spokane, Washington. While Silicon Valley firms like Google and Uber are well represented, so are stalwart mainline companies like Walmart and Aetna, with dozens of small consulting shops scattered around the US, Europe, and beyond.

What are they doing? While each effort is unique, each of these groups is trying to develop products, communications, or policies that cause their users to do something different in their lives. In other words, they are designing for behavior change. Whereas traditional design is fundamentally about solving a user need, behavioral design is about solving a user need where doing so entails, in a sense, “solving the user” too: changing the person in order to solve the problem.

And that’s where this book comes in. This is a guidebook on how to do it yourself: how to identify behavioral problems your users face, develop clever solutions to help them overcome those obstacles, iteratively learn from the process, build a team that does this, and make it successful in an organization. And, underneath it all, how to think about this work in an ethical, thoughtful way and avoid the serious mistakes and abuses that some practitioners are facing now—and that threaten our field as a whole.

Along the way, I’ll try to teach you the fundamental understanding of the mind that many behavioralists like me have: of a quirky, elegant, but necessarily imperfect decision-making process that guides your users’ decisions and actions. I’ll give you the core lessons and a framework to understand the research literature, but this book isn’t fundamentally about behavioral science theory (for me, Nobel Laureate Daniel Kahneman’s Thinking, Fast and Slow is still the best introductory book on behavioral science out there—I highly recommend it). Instead, our focus will be on action: what you can do, in your work, right now.

How This Book (and This New Edition) Came About

In 2019, O’Reilly asked if I could do a second edition of the book, updating it for tremendous growth in the field since it first came out. I was happy to do so—if only to celebrate that growth and help support its further advancement in my small way.

The ideas and process I discuss here come, foremost, from my own experience working in this field over the last 11 years: first at the personal finance company HelloWallet, and now at the investment research company Morningstar. It is enriched by countless conversations with other practitioners in the field: those with existing teams and those just seeking to enter the field. I’m particularly grateful to the Action Design Network, the nonprofit organization that I helped start in 2013, which has grown far beyond my dreams or certainly my intentions, with events all across North America on applied behavioral science. I will also be weaving in lessons from the survey I mentioned before, which we believe is the largest and most comprehensive survey of teams designing for behavior change ever conducted (though the way the field is growing, it won’t be for long!).

Personally, I started doing applied behavioral science while completing my PhD and working at HelloWallet. HelloWallet began as many start-ups do: with a great deal of energy, a clear problem to solve, and a fundamentally misguided notion of how to solve it. We wanted to help everyday Americans improve their finances—and wrote an application that encouraged them to set up budgets and save for the future. There was just one little problem: everyone already knew how to do that. The problem people faced wasn’t a lack of understanding; it was that they didn’t act on it. There was a tremendous gap between people’s intentions and their actions.

Thankfully, in addition to having a great (but misguided) idea, we had the ability to measure whether we were being successful. We looked at whether people used our app. They didn’t. We also looked at whether it was helping the few people who used it to improve their finances. It wasn’t. I’m thankful for those hard lessons because it’s better to know early when something isn’t working and to be able to learn from that and do better—which will be a recurring theme throughout this book.

At HelloWallet, we set up an engine for behavioral experimentation: building scientific experiments into the underlying platform of the application. My motto was to make it “easier to test than to argue”—i.e., where product managers or designers disagreed about which version of a feature would more effectively change spending behavior, it was better to test both approaches in the field than to argue about which might be right. And by and large we succeeded—both in making it easy to run tests and in having the impact we sought.

We found that we could help people change their bank habits, to cut the money they lost to ATM and other banks fees by 25%; for lower-income families that equated to nearly a day’s wages per month. We found that we could successfully nudge people to put aside money for savings by giving them a simple and easy-to-understand display of their finances, compared to that of their peers. And we also found that some of our efforts—like congratulating people on their budgeting success—actually backfired: they spent more! We quickly discontinued that feature. The first edition of Designing for Behavior Change was, in many ways, a description of what we were learning as we learned it. It detailed the process we were using in practice. In fact, the book arose out of an internal guidebook that I’d written for my teammates on applied behavioral science.

Since then, I’ve gained a great deal more experience with the challenges of designing products that help people change behavior. But, I’ve also verified that the fundamental approach was sound; the four-step process laid out in the book is what I still use today—though I’ve started using different names for the steps. I’ve modified the process around the edges, making it more efficient and flexible, but it is still tremendously useful in my own work. After the first edition, I also subsequently learned that many other teams had independently developed a similar iterative approach.

About five years ago, I moved to Morningstar (along with some of my teammates from HelloWallet) and was given the opportunity to set up a much larger behavioral science team. We work across the company, on issues ranging from effective investing to retirement saving and spending to challenges of internal decision making. We conduct applied research both with our in-house team and with leading behavioral scientists from academia.

This edition draws upon my more recent work at Morningstar, formal and informal consulting I’ve done with other organizations in our space, and especially work through the Action Design Network. The Action Design Network has grown from our initial monthly meetup in DC to a broad base of volunteers organizing events across 15 North American cities. Through it, I was able to learn from the challenges many other teams face, the approach they use, and their solutions.

In this new edition of the book, I’ve sought to bring together these diverse streams of insight to a coherent guidebook for practitioners to help you learn from the best of our knowledge in the field so you can design your own products to more effectively help users change their own behavior.

Who This Book Is For

As you can probably tell by now, this book is aimed at practitioners—the people who design and build products or communications with specific behavioral goals. Teams that design for behavior change should generally include the following roles, and individuals in each of these roles will find practical, how-to instructions in this book:

• Interaction designers, information architects, user researchers, human factors experts, human–computer interaction (HCI) practitioners, and other UX folks

• Product managers, product owners, and project managers

• Marketing and communications professionals

• Behavioral scientists (including behavioral economists, psychologists, and judgment and decision-making experts) interested in products and communications that apply the research literature

The person doing the work of designing for behavior change could be any one of these people. At Morningstar, we have each of these roles, but most of my team is composed of behavioral researchers. This work can be, and often is, done wonderfully by UX folks. They are closest to the look and feel of the product and have its success directly in their hands. This approach enriches their current practice by adding an extra theoretical layer to design hypotheses and tests.

Product owners and managers are also well positioned to seamlessly integrate the skills of designing for behavior change to make their products effective. Finally, there are other behavioral scientists (like me) working in applied product development and consulting at organizations like ideas42 and the Center for Advanced Hindsight. So, the people designing for behavior change probably wear other hats as well.

In addition, this book is for entrepreneurs and managers. If you’ve ever read Nudge, Blink, or Predictably Irrational,¹ and wondered how you could apply them to your own product and users, read on. While the book is about helping users take action in their lives, that doesn’t mean that designing for behavior change is incompatible with a for-profit business model. Businesses make a profit; that’s how they exist. So, you’ll find suggestions for building a successful business model on voluntary behavior change. If in addition to making a profit, you are helping your users take action and change their behavior, this book can help you do it.

Nonprofit entities and some government agencies often explicitly focus on helping users change their behavior; Designing for Behavior Change can help. For example, the UK’s Behavioural Insights Team is widely applying behavioral research to public policy and services around the world. Where relevant, I’ll note parts that are particularly important for nongovernmental organizations (NGOs) and government agencies. Because it’s more compact to write, I’ll refer primarily to “companies” here. In almost all cases, I really mean companies, organizations, and relevant government agencies.

Finally, my expertise is software development, so I’ll use the terminology that I use in my day-to-day life—applications, software, and programs. You don’t need to be in software development to find this relevant to you. In fact, some of the most innovative work in persuasive design, one of the fields that this book draws inspiration from, is in the design of everyday objects.² As you apply Designing for Behavior Change to your work, whether in software or beyond, I’d love to talk with you and share notes! You can reach me at [e-mail].

Combining Research, Data, and Product Expertise

One of the book’s recurring themes is that understanding how the mind works is not enough to build behaviorally effective products.

In addition to behavioral science research, we need two sets of skills to support the process. First, we need to plan for data analysis (both qualitative and quantitative) and for refinement and iteration based on that data. That means adding metrics to the application and conducting user research to understand individual behavior, analyzing the data, and making improvements over time based on it.

Second, we need to build products that people actually enjoy using. I know it sounds obvious, but it’s something that’s often forgotten as we build products designed to educate, motivate, or otherwise help our users. We tend to focus on the behavior change (and how important it is) and forget the fact that people still have to choose to use our products. Users avoid boring, frustrating, ugly applications, so we should remember the lessons of good product design, from identifying user needs and frustrations to designing an intuitive, beautiful user interface.

When you bring these raw ingredients together—behavioral research, product design or marketing expertise, and data analysis—you have what’s needed to design for behavior change:

What You Need to Know to Benefit from This Book

This book gives you enough knowledge in each of these three areas to get oriented and to start working on concrete products and communications. It covers most of the behavioral research you’ll need to finish the products as well, but at some point along the way, you’ll need people who are experts in qualitative or quantitative data and in product design. Chapter 17 provides a detailed look at the skills required for a team that designs for behavior change, including where you can find (or develop) them.

If you are an expert in one of these areas, all the better. The book will show you how designing for behavior change builds upon and complements your existing expertise. You’ll find out how to leverage your existing skills to play a leading role in the development of behaviorally effective products and communications within your organization.

What Types of Behaviors This Can Help With

The techniques I’ll talk about here assume that the product will support an action that people aspire to but have had difficulty undertaking. Learning a language. Sticking to a diet. Meeting new people. This may seem like it applies only to a narrow set of products, but I’ve found that there are two big groups of behaviors that fit these criteria:

• Behaviors that users want to change within their daily lives

• Behaviors within the product itself that are part of using the product effectively

Or, from the perspective of the company making the product, behavior change is either:

• The core value of the product for users

• Required for users to extract the value of the product

In the first case, users have some behavioral problem in their daily lives and buy the product to help them with it. In the second case, users have some other need that the product solves, but they must adapt and change their behavior in order for the product to deliver on its promise.

The first case includes:

• Controlling diabetes

• Paying off credit card debts

• Getting back in shape

• Getting involved in their communities

Often these behaviors relate to big-picture social issues like health and wellness. When we design products that support these behaviors, we help the individual and impact our society at the same time. Oracle Utilities’s Opower and Google’s Nest, for example, are products that help individuals decrease individual energy usage: saving people money and helping the environment at the same time. Other products that change behavior in this way are Fitbit (exercise) and Weight Watchers (diet).

As I send this book off for final production, the coronavirus COVID-19 is spreading across the globe. We’ve seen a rapid mobilization in the healthcare community, including work by behavioral scientists to help promote social distancing and handwashing behavior.³ Researchers are mounting large-scale, rapid turnaround studies to test techniques to keep people safe—by applying behavioral science to the design of communications and products.⁴ It’s an impressive display of the power of behavioral science to help people take action that’s in the best interest of the individual and society as a whole.⁵

The second type of behavioral change is far more mundane. Individuals often seek to change behavior as a means to an end. Let’s say a user wants to learn a new language and gets a software package to help do it. Simply learning how to effectively use the software can take some substantial changes in behavior within the product—building new habits to log in daily and practice the language, overcoming fears about looking foolish while doing it, and so on. The user wants to take an action (learning the language) but struggles. A well-designed product can help the user make those personal adjustments.

This second type, and the products that require it, is much broader than the first. It covers the sweep of voluntary changes in behavior that people might make to benefit from products they’ve already chosen to use. It touches upon a huge swath of the consumer product space, from Yelp to Square to Rosetta Stone. Some examples of actions that occur within software products that one might try to improve include:

• Organizing email contacts

• Drawing decent flowcharts

• Formatting documents

As with many other behavioral scientists and designers, I believe that no design is neutral.⁶ Anything we design that interacts with other people—communications, products, services, etc.—has an impact on their behavior and ultimately their lives and outcomes. Here, we talk about how to make that process intentional and, hopefully, beneficial.

For both types of behavior change, the goal is to develop products that help users take action and to deliver the value that the company offers. This voluntary, transparent support for behavior change helps companies be successful as well.

What This Book Is Not About

If you’re looking for a book on how to make people do something you want them to do (even if they don’t want to do it), you’re in the wrong place. I’m not necessarily judging your motives or aims; I just can’t help you much.

In particular, this book isn’t designed to help with persuasion. There are many ethical and thoughtful uses of persuasion—and we all seek to persuade each other in our daily lives. While there are many similarities with designing for voluntary behavior change, other issues arise such as educating users about a product, building a convincing argument, rhetorical delivery, building rapport over time, and more. We won’t cover them here. The topic of voluntary (i.e., pre-persuasion) behavior change is big enough as it is!

In addition, this book isn’t intended to help with trickery or coercion (for both practical and ethical reasons). Sadly, there are many companies trying to do just that, though—and it’s dangerous for our field. I’ll delve into the details of what’s happening across our industry and how we’re starting to face well-deserved backlash from both government regulators and thoughtful technologists in Chapter 4.

The Chapters Ahead

In the following pages, I walk you through each of the skills you need to design for behavior change, starting with a firm foundation in how the mind makes decisions. Then, I show each step that’s required to develop a new product: moving from discovery to design to implementation and refinement. I introduce each concept where it is first needed. In Part III, I step back and give some additional information on the scope of the industry (and where the jobs are), how to build a behavioral team at your organization, and likely problems you’ll face along the way. That’s it.

However, if you’re looking for a more formal chapter outline, here you go:

Part I: How the Mind Works

Chapter 1 arms you with your first skill: an understanding of how the mind makes decisions. You’ll get an overview of the current literature on decision making, as well as a dozen key lessons and their implications for the design process.

Chapter 2 then describes the six high-level factors that must come together at the same time for a person to take action. They form the CREATE Action Funnel—Cue, Reaction, Evaluation, Ability, Timing, and Experience—which shows you what needs to be addressed in your product and where users usually drop off.

Chapter 3 looks at the reverse problem: how to help your users stop unwanted habits or improve poor decisions.

Chapter 4 delves into detail about the ethical challenge facing the field: how too many practitioners have used these techniques to manipulate users into buying (or overusing) their products, how we’re fooling ourselves if we think we wouldn’t do the same under the right (or wrong) conditions, and how we can tackle these problems head on.

Part II: A Blueprint for Behavior Change

Chapter 5 introduces you to the larger process of designing for behavior change—how your new knowledge about the mind can be used in practice—using the acronym DECIDE: Define the problem, Explore the context, Craft the intervention, Implement within the product, Determine its impact, and Evaluate next steps.

Chapter 6 starts charting your course by defining the problem: the overall outcomes you hope the product will deliver and who it seeks to help. From there, it demonstrates how to elicit a potential idea for behavior change—how, specifically, your users will get fit or take control of their finances, for example.

Chapter 7 explores the context in which your users will act with a behavioral map: a narrative of how the product team envisions user interaction with the product and how users will change their behavior. It then evaluates the various behaviors they could change in light of their needs, interests, and prior experience. It refines your initial plan to finalize the specific behavior the product will support.

Chapter 8 introduces you to the meat of the DECIDE process: designing the intervention itself. We look, at a high level, at the main strategies you can use to help users change behavior, using the story of a fish stranded on the beach.

In Chapters 9 and 10, we look at how to craft specific interventions, based on behavioral research, to support your users to take action. Chapter 9 presents interventions that are appropriate when users face problems of Cue, Reaction, or Evaluation; next, Chapter 10 looks at interventions for Ability, Timing, and Experience.

Chapter 11 wraps up the discussion of crafting an intervention with two extensions: how you can handle multi-step complex interventions and how to help the user hinder negative unwanted actions.

Chapter 12 presents some tips for implementing the intervention within the product itself. Applied behavioral science doesn’t require a particular development approach or technology; it does, however, require high-quality data on outcomes. We discuss how to build those metrics, and the ability to measure changes in them, in the initial product development—rather than as a hacked-up afterthought.

Chapter 13 focuses on determining the product’s impact, its success or failure. It starts with the most powerful tool there is for impact measurement: the humble A/B test. Since most readers are likely familiar with A/B tests already, this chapter dives into the details of how to use it effectively to gather rigorous data and the common pitfalls that practitioners face.

Chapter 14 looks at other ways to determine impact, when A/B tests or other randomized control trials aren’t available. I cover how to utilize statistical models to gain insight. I talk about the challenges of gauging the causal impact of software on real-world behavior and how to overcome them.

Chapter 15 concludes the DECIDE blueprint for behavioral design by helping you Evaluate next steps after your implementation. It looks at how to find problems that limit impact, including how qualitative and quantitative analyses are both needed, and work hand in hand.

Part III: Build Your Team and Make It Successful

Chapter 16 provides the results from the Behavioral Teams survey in detail. Case studies are scattered throughout the book, but here we dig into the numbers: how large the field is, where the jobs are, and what challenges and successes other teams are having.

Chapter 17 looks at what it takes to start applying behavioral science at your organization with a team of 1 or a team of 20. We look at how to make the case to stakeholders in-house and the skillset you or other team members will need.

Finally, Chapter 18 wraps things up with a quick review of the designing for behavior change process and key takeaways on how to make it happen in your organization. It also covers many of the questions that can arise when putting these lessons into practice.

At the end of the book, there’s information for those who are looking to dive even deeper:

• Glossary of Terms: a glossary of key terms, like behavioral map and data bridge

• Bibliography: a comprehensive list of the works cited in this book

In the first edition of this book, there was a list of online resources (Appendix B). We’ve moved that online. Each of the core chapters in Part I ends with “A Short Summary of the Ideas” if you only have a few minutes. They are a useful wrap-up that also give advice if you’re just looking for an informal process to sketch things out before you jump in head first.

The core chapters of Part II wrap up with “Putting It into Practice”—sections with concrete examples and exercises to help you put each chapter’s ideas to work.

Let’s Talk

My hope with this book is to further the conversation about voluntary behavior change and help build up the tools needed to develop behaviorally effective products. However, even after writing two editions (and likely even after 20) I have no illusions about the completeness of this work—there’s still a tremendous amount to figure out. We’re all going to learn as we go along.

Personally, I’m always looking to learn, share, and collaborate, so don’t hesitate to reach out if you have a cool story to share, an idea for a research project that would further develop the field, or an idea for a behavior-changing project that you’d like to bounce off someone. You can find me under sawendel on Twitter, LinkedIn, and AngelList; my contact information is on my website, http://about.me/sawendel.

If you think there’s something that can be improved in this book or find something that is inaccurate, please reach out to me and tell me about it. One of the many benefits of working with O’Reilly Media as a publisher is that a lot of you will be reading this book in an electronic format and can quickly get an updated version of the book if corrections need to be made. For those who are reading this in paper form, I’ll keep a list of corrections, additions, and other updates online at behavioraltechnology.co.

O’Reilly Online Learning

Note

For more than 40 years, O’Reilly Media has provided technology and business training, knowledge, and insight to help companies succeed.

Our unique network of experts and innovators share their knowledge and expertise through books, articles, and our online learning platform. O’Reilly’s online learning platform gives you on-demand access to live training courses, in-depth learning paths, interactive coding environments, and a vast collection of text and video from O’Reilly and 200+ other publishers. For more information, visit http://oreilly.com.

How to Contact Us

Please address comments and questions concerning this book to the publisher:

• O’Reilly Media, Inc.
• 1005 Gravenstein Highway North
• Sebastopol, CA 95472
• 800-998-9938 (in the United States or Canada)
• 707-829-0515 (international or local)
• 707-829-0104 (fax)

We have a web page for this book, where we list errata, examples, and any additional information. You can access this page at https://oreil.ly/designing-for-behavior-change.

Email [e-mail] to comment or ask technical questions about this book.

For news and information about our books and courses, visit http://oreilly.com.

Find us on Facebook: http://facebook.com/oreilly

Follow us on Twitter: http://twitter.com/oreillymedia

Watch us on YouTube: http://www.youtube.com/oreillymedia

Permissions

Since the first publication of Designing for Behavior Change, I’ve continued to build upon and refine the theoretical and practical tools presented in the book. The tools that I use with my own team, that I’ve trained other companies and organizations with, are extensions of those initial ideas (and indeed, the first edition itself was a description of our practice and approach at HelloWallet at the time). In this edition, I’ve brought together a number of those artifacts and lessons from applying this approach over that last six years. In particular:

• The exercises in the “Putting It in Practice” sections of Chapters 6–14 incorporate parts of a workbook I developed for my team at HelloWallet to apply these concepts.

• Chapter 13 incorporates sections from a guide to experiments I wrote for my team at Morningstar.

• Chapters 1, 3, and 4 incorporate part of an introduction to behavioral science that I wrote for a book exploring the applications of behavioral science to one’s personal and spiritual life.

• Chapter 16 describes a survey I conducted in 2019 with the nonprofit organizations the Behavioral Science Policy Association and Action Design Network, and have published separately on our websites.

In each case, the source material was used as a starting point, often with significant adaptations and changes. All materials are used with permission.

Acknowledgments

With the first edition of this book, I had many people to thank for their help in developing the ideas and process presented within. The list has only grown.

I continue to talk with and gain inspiration from people I started working with on the first edition—from Rob Pinkerton and Katy Milkman to the Action Design Team, especially Zarak Khan, Matthew Ray and Erik Johnson. Here, I’d like to call attention to additional people I’ve learned from since that first edition. Top of the list would be my intellectual sparring partner, Ryan Murphy, and my team at Morningstar: Heymika Bhatia, Sarwari Das, Jatin Jain, Sam Lamas, Sagneet Kaur, Alistair Murray, Sarah Newcomb, Shwetabh Sameer, Stan Treger, and Leon Zeng. Many thanks to Ray Sin as well, for your thoughtful research. Outside of Morningstar, I’m particularly grateful to Paul Adams, Julián Arango, Florent Buisson, May C., Jesse Dashefsky, Clay Delk, Barbara Doulas, Darrin Henein, Fumi Honda, Peter Hovard, Anne-Marie Léger, Jens Oliver Meiert, Brian Merlob, Shafi Rehman, Neela Saldanha, Nelson Taruc, and Mark Wyner for their comments on the draft.

In terms of the rich stream of behavioral research that I draw upon here, there are simply too many people than can be thanked. Even though it is not the current fashion to do so in popular-press books, I make a point to cite authors of the specific studies I reference throughout the body of the book, though I don’t have space to offer a full literature review. For some readers, this may seem overly academic or, frankly, boring. That’s certainly not my goal! Rather, it is the simple acknowledgment of the many thoughtful researchers who have done great work and that both I and you, as readers of this book, benefit from. I want never to overstate my ingenuity or to dim the glow of their bright light. My primary error, I fear, is not giving enough credit to the amazing work of other researchers. Please accept my apologies for any citations that are missing.

And as always, I’d like to thank Alexia, Luke, and Mark—for putting up with me and my obsession with writing and for loving me nevertheless.

¹ Thaler and Sunstein (2008), Gladwell (2005), Ariely (2008).

² Dan Lockton has a set of papers that provide an extensive review of the various domains in which intentional behavior change has been applied.

³ See recommendations to the Irish government, for example.

⁴ Kwon (2020); Jachimowicz (2020)

⁵ We’ve also seen what happens when incomplete lessons from behavioral science are divorced from the process of science (including rigorous testing and empirical validation)—in the furor over the UK government’s initial resistance to strict social distancing based on the concept of behavioral fatigue. See Yates (2020) for the political debate, and UK Behavioural Scientists (2020) for a thoughtful response calling for proper scientific testing of the concept before using it in government policy with lives at stake.

⁶ In Nudge, Thaler and Sunstein make this case, for example, and a discussion of the unintentional and intentional impact of design has long been a part of the design community (e.g., Nusca 2019, Vinh 2018).