Designing for Behavior Change: Applying Psychology and Behavioral Economics (2013)

Part VI. Putting It into Practice

We’ve covered how the mind makes decisions, how to design products that leverage that knowledge, and how to test and refine those products over time. We’re almost done.

At each step of the way, this book has provided guidance on how to apply these lessons in your daily work. But there may be some places you still have questions you aren’t sure about. Part VI tries to answer some of the most common questions and reinforce how the whole process works to help orient you at each stage.

Chapter 15 covers a range of questions about the decision-making process, what that means for product development, and how products that change behavior affect a company’s bottom line. It also provides a simple, start-to-finish example of using this process for a new product.

Chapter 16 wraps things up. It summarizes the major techniques learned in the book and the common themes that underlie the book’s approach to behavior change.

Chapter 15. Common Questions and a Start-to-Finish Example

An Example of the Approach

The method described here comes from our practical experiences at HelloWallet and from the inspiration we’ve gained talking with other companies in the field. Throughout this book, there are examples of particular companies that we’ve learned from and how they, and we, have applied this approach. But there hasn’t been an opportunity to analyze a single, start-to-finish scenario. Let’s do that here.

For this scenario, let’s move past the standard examples used in this book of exercise trackers, personal finance apps, and such, and show how the approach can be used in a novel context—working with contractors on your home.[155] Not all behavior change is sexy and socially important. In fact, as I mentioned in the Preface, mundane changes in everyday life are the most numerous, but the least obvious, targets for behavior change. So without further ado, here’s the scenario:

Jake has a software company that develops apps for building contractors, helping workers with a range of tasks, from avoiding injury to making efficient use of building materials.[156] One day, some of his clients approach him and want him to develop a new application that will improve thequality of their work, decreasing the rate and severity of errors on the job site. Most errors are caused by people making careless mistakes, so Jake recognizes that the success or failure of this new application will depend on helping contractors and their workers change their behavior. It’s something they already want to do, it’s something the contractor’s customers (the homeowners) want, and there’s always room for improvement.

So, let’s walk through the four stages of designing for behavior change—understand, discover, design, and refine—for Jake’s new application (see Figure 15-1). At this point, all Jake has is an initial request from his clients and a hazy vision of what the product will do.

Figure 15-1. The four stages of designing for behavior change

Understand

As the foundation for designing for behavior change, Jake learns about how the mind makes decisions. The most important lessons can be summarized in the Create Action Funnel (Figure 15-2): in order for the proposed application to be successful at changing the behavior of the contractors, he needs to ensure that they are cued to think about the app regularly; don’t have a strong negative reaction (e.g., getting angry at being told what to do); evaluate the situation and see that the benefits of taking action will outweigh the costs; have the ability to actually make those quality-control changes; and feel that the timing is right for action (because of urgency or setting aside time to do it) to use the app and improve the quality of their work.

Figure 15-2. If you remember one thing from the “Understand” phase of designing for behavior change, remember this: the Create Action Funnel

Discover

Next, Jake will start planning out the new product (Figure 15-3). Jake already has a vision of what he wants to do: improve building quality. He’ll flesh out that vision to clearly define the real-world outcome that he wants to achieve. After talking with potential clients for this application, he realizes that what they care about most is decreasing the number of problems that are euphemistically known as “nonconformities.” That’s when the contractor doesn’t end up building what had been agreed upon, leading to arguments with customers (i.e., the homeowner) and problems with the building inspector. Often, the contractor has to tear out the nonconformity and redo the work, at his own expense. No one is happy in that situation.

Figure 15-3. The discovery phase: your goal is to clarify the outcome, actor, and action

Jake works with his team and other experts in the industry to generate a list of potential actions that the contractors can take to decrease nonconformities. For example, taking additional training courses, adding additional supervision, and using testing and monitoring devices to ensure that the weather and other conditions are right for each task (weather conditions can significantly affect the quality of newly poured concrete, for example).

As part of the brainstorming, Jake discovers that sometimes contractors don’t follow the building plans because they don’t have the most up-to-date version, don’t have the opportunity to check it regularly, or look for shortcuts to save time. Jake adds an idea to the list of potential actions: helping contractors frequently refer to the building plans.

He does user research in the field and discovers that both the foreman and the individual workers have the ability to decrease nonconformities. But the foreman is most likely to be receptive, and have the necessary skills, to use Jake’s app. The foreman is thus the actor who will take action to decrease errors.

Putting the actor, outcome, and list of possible actions together, he evaluates the costs and benefits of each option—in terms of how difficult it will be to change the behavior, how effective that change will be at ensuring contractors keep to the building plans, and the revenue that his company can make from selling the application. The action of “helping the foreman frequently refer to the most up-to-date building plans” is the best option.

Design

With a clear sense of the behavior that the product hopes to change (outcome, actor, and action), Jake and his team start designing the product (Figure 15-4).

Figure 15-4. The design process focuses attention on changing the action, environment, and user preparation, in turn, to support behavior change

The first task is the conceptual design, or behavioral plan—a narrative that describes how contractors will change their behavior over time, using Jake’s product. He breaks the action into its component steps so that he can structure the action to ensure success. In order to regularly refer to the building plans (action), the user will need to (a) bring a device (a phone or a tablet) with the app to the job site, (b) load the correct building plans, (c) make the time to review them, (d) actually review them, and then (e) build in accordance with the latest plans.

Jake looks over this list for places he can simplify the process or tailor it to the existing skills and experience of his target users, the foremen. For example, he can use GPS to automatically load the correct building plans, and simple database logic to ensure that only the most up-to-date version is shown by default. He can schedule the review process to align with consistent cues in the user’s environment—arriving at the job site, taking a break, etc. Those consistent cues also lay the foundation for building a long-term habit out of the behavior.

Next, he designs the environment in which the user acts. Part of that is the application itself, and part of it is on the jobsite. He looks for ways to cue action (reminders within the product, or papers taken to the site) and highlight the users’ existing motivation (e.g., using loss aversion to highlight the cost of fixing nonconformities after the fact).

Finally, he looks at how to prepare the user. He adds a short training video within the application for first-time users, and integrates the application into the quality-control procedures the users already follow (to create a narrative in which using the product is a natural extension of their work). Thankfully, some of those procedures already occur on the foreman’s smartphone, and so Jake plans to directly link the new application to the ones that they are already using.

The resulting behavioral plan tells Jake’s team what the product should do. Next up, the team will work on the interface designs—the look and feel of the product that delivers this functionality. The team chops the behavioral plan into specific user stories for their agile development process, like, “As a user, I want to quickly learn how to use this software, so I don’t need to waste time figuring out how it works” or, “As a user, I want to automatically see the right building plan for the job I’m on, to save time searching for it.”

The interaction designer develops wireframes based on these user stories; along the way, she will use the numerous behavioral tactics in her toolbox—priming the user to activate thoughts of quality, avoiding choice overload by restricting the complexity of decisions that need to be made at one time, and eliciting implementation intentions around how to resolve nonconformities if they are found.

Afterward, Jake reviews the wireframes using the same process that he used for the behavioral plan—focusing on the action, the environment, and the users’s preparation in turn. For each interaction the user has with the system, Jake checks whether the action could be simplified, whether the product delivers a single, clear call to action, and whether the user is properly prepared for action and knows what to do.

When the wireframes are solid, the team does qualitative testing with potential users, generates a clickable prototype for additional testing, and refines the designs accordingly. When the designs are done, the team builds the product.

Refine

No product is perfect, especially one that tries to change deeply ingrained, real-world behavior for people on the job. So, Jake and his team assess the impact of the product by doing a simple comparison between clients who were randomly selected to receive the trial version of his product, and those who weren’t (Figure 15-5). He compares nonconformities across the two groups, and finds that yes, the product works—it substantially reduces errors. But there’s room for improvement.

Figure 15-5. Refining the product starts with a clear sense of its current impact, finding ways to improve it, then testing those changes in the field

During data collection, Jake realizes that many of the mistakes that occur on the job happen in the late afternoon when workers are tired and hurrying to get home. He has some ideas on how to fix that, such as having the app recommend breaks, and making the foreman consciously aware of that problem. He evaluates these behaviorally informed changes along with additional requests from his current clients and from his marketing team.

Finally, he develops a plan to change the product that should improve its impact and increase revenue. He makes the change, measures it, and it works. Nevertheless, he continues to gather data about the performance of his app and iteratively improves it in the future.

Questions About How and Why We Act

Now, let’s turn to a set of questions that I’ve encountered about how our minds’ decision-making process works. There’s no strict ordering or relationship here; these are just some of the most common issues that people have wanted to better understand. Feel free to jump to those questions that are relevant to you.

What Happens Before People Take Action the First Time?

The Create Action Funnel provides the five preconditions for action, and the design process puts those five preconditions in place with a product. But how do people normally progress from inaction to action over time?

There’s no consensus in the literature about how individuals move from inaction to action on their own. There are many pathways by which people become active, each of which brings together the preconditions for action in their own way. The intentional design of products to support action, described here, is just one of those pathways.

Consider the example of a person who wants to clean the house. He has a positive intuitive reaction and decides that the benefits (girlfriend won’t leave him) outweigh the costs (an afternoon of TV watching lost) after a conscious evaluation. But there are always more pressing things to do (no time pressure), he forgets to do it (no cue), or he doesn’t know how to do it effectively and doesn’t have the right cleaning supplies (lacking ability).

Here’s a quick overview of different pathways by which this person could move from failing to take the action to successfully doing it:

Self-directed action

He thinks about cleaning a few times, and it doesn’t happen. When he’s at the grocery store, he remembers to pick up cleaning supplies, but doesn’t set aside time to use them. One day, he says enough is enough and plans out a specific day and time to do it, sets a reminder, and removes other distractions.

Intentional help from another person

A friend hears him complaining about his dirty house and decides to push a bit to get it done. The friend reminds him at the grocery store to get supplies and makes him commit to cleaning on a particular day.

Directed help from a product

He downloads a task-management app based on David Allen’s book Getting Things Done (Penguin Books, 2001). He records all of the things he needs to do (cleaning and otherwise), and, with the help of the app, works out the logistics and timing to finally clean up the house.

Sudden change in environment

He wants to clean the house, but just doesn’t get it together. For other reasons, he changes jobs and moves. His new house is much smaller, and he gets rid of most of his stuff. It’s also much easier to clean, and he finds it’s easy to continue a habit of keeping the house clean once he’s started.

Sudden change in life circumstances

The guy gets married. His wife won’t put up with his messiness and lays down the law for him. He shapes up.

Social drift

For various reasons, he gets to know new people at work, and some of them become friends. He visits their houses, which are pleasantly clean. When he goes home, he looks at his house differently. This happens a few times, and he’s ashamed of bringing other people over. He gets up the gumption and does it.

These are common pathways, but of course there are many more. The point is that there simply isn’t a single route by which people change their behavior over time. Sometimes there’s a single, culminating event that forces change (marriage); at other times, things move more slowly (social drift).

While these examples point to multiple possible routes, anecdotes aren’t science. Thus far, researchers haven’t established clear, generally accepted rules for how individuals move from inaction to action over the course of time. There’s a great deal of activity in the addiction and health space, with models such as the Transtheoretical Model[157] ([ref150]) and Health Belief Model ([ref97]), but their general applicability is unclear. We know a lot more about what happens at the moment of action (a cue, a conscious choice to act, etc.) than we do about what happens before that action.

LESSONS FOR BEHAVIORAL PRODUCTS

This section speaks to the transition from inaction to action, but it’s especially relevant when you’re trying to increase adoption of your product. There’s no simple answer to how people will decide to adopt, but you can improve your chances. If possible, identify life transition points for your users, and ensure your product cues them then. Beyond that, you can use open-ended campaigns to gather data about unknown entry routes, and subsequently focus efforts if major opportunities are found.

How Do the Preconditions for Action Vary From Day to Day?

Each of the five preconditions can vary from day to day and moment to moment. In Part III, we discussed how the five preconditions are characteristics of the action, the environment, and the actor (the “context” of the user’s decision to act, or not). Both the actor and the environment vary over time, and thus so do the conditions for action:

Environment

As a person moves through the day, the environment generally changes. The different environments shape whether the person will take a particular action or not. For example: someone may have Internet access while waiting for a meeting, and can use mobile apps, but not while driving. The person may see reminders to exercise at home, but not while at the bar.

Actor

Over the course of the day, people also vary in themselves. Their wakefulness varies, their tendency to be distracted varies, and of course, so do their emotions. An action that seems interesting and worthwhile to a rested, well-fed person can seem overly complex, too much work, and not urgent to someone who is hungry and tired.

This variation is a mix of happenstance and structure. Day to day, a person’s work schedule may be too irregular to identify stable moments of opportunity for engagement with the product. But the commuting schedule may be stable and clear. Where feasible, the product should align with the moments and situations of opportunity, in which some of the preconditions for action are already naturally in place, and then fill in the gaps.[158]

From the perspective of product design, look for structure within this day-to-day (and moment-to-moment) variation, and build on it to encourage action. Are there particular times and places when the user of the product is least distracted? Or most motivated to act? Where are the cues to act in the person’s daily life?

LESSONS FOR BEHAVIORAL PRODUCTS

You have two main options to harness the natural variation in people’s days. You can do user research, and try to find structured opportunities for engagement that are common across the user base. Alternatively, you can prompt users to self-identify the best times for them—when they want to receive reminder messages, when they want to be motivated with uplifting stories from other users, etc.

How Do the Dynamics of Using a Product Change as the User Gains Experience With It?

There are many ways in which people can change as they start using a product. There’s no guarantee that over time, people will become experts or master the product and behavior. For example, some people keep using a product they love indefinitely, and others stop after the first time, or after a few weeks.

Rather than a proscriptive pathway, there are different states that users can be in, relative to the product. In each of these states, the dynamics of the Create Action Funnel (cue-reaction-evaluation-ability-timing) and the context that supports it (action-environment-actor) are somewhat different. These states are:

§  Not a user of the product

§  Dynamics. For someone that has never used the product before, its very newness is a benefit and a curse. Cues are likely to be noticed. Newness can increase motivation to explore. But newness can make the user unsure about the logistics of actually using it, and that person’s ability to succeed. Also, our intuitive reaction to the idea is based on our experiences with similar activities, which could be good or bad.

§  What the company should do. Ensure the user knows about the product at all (cue). Make sure the value is clear (evaluation), and associate the product with familiar, pleasant things (reaction).

§  A one-time user of the product who had a positive experience

§  Dynamics. Once we gain experience with a product, our future intuitive reactions take that into account. If we liked the product the first time we used it, excellent. We gain knowledge of how to use it (decrease costs, increase logistical ability).

§  What the company should do. Keep cueing the user; it’s unlikely that the user has built a strong association between the product and an existing cue in his environment. Highlight the positive experience (evaluation and reaction). Use knowledge gathered during the first use to try to align the product with times and situations when the user isn’t distracted and can take action (ability and timing).

§  A one-time user of the product who had a negative experience

§  Dynamics. If the first experience was negative, we have two obstacles to overcome—the allure of newness is gone, and we have an intuitive negative reaction. It is much harder to bring these users back.

§  What the company should do. Honestly, focus attention on other users. This is a tough group to win back. And focus attention on improving the experience of first-time users in the future.

§  A user who’s returned to the product one or more times

§  Dynamics. When a user has successfully returned to the product, that’s a sign that the conditions are ripe for future use. Something in the user’s context has pulled him back. However, it’s not clear yet how stable that something is—it might be a temporary assignment at work that makes the user think about or need the product. It might be a core desire. At this point, minor disruptions to the user’s context (a different routine at work, etc.) can easily stop him from returning.

§  What the company should do. Keep cueing the user until there’s a strong association between the product and an existing environmental cue. Highlight prior positive experiences and successes. With user research, try to understand the user’s context: are there temporary factors pushing him to use it for which the product must find a substitute?

§  A user who regularly uses the product

§  Dynamics. When the user returns to the product repeatedly, then there’s a stable context that pulls him back. Only major changes in the user’s context are likely to disrupt continued use—the user changes jobs, gets divorced, takes on significant new activities that pressure his time, or the product suddenly drops the particular functionality he loves.

§  What the company should do. Don’t screw it up. Be very careful when changing functionality that’s driving usage. Look for regular users who drop off—if the context disruption is temporary, there’s a very good chance you can win them back (but don’t take it for granted; see[ref131]). If you are working with a potentially habitual behavior, make sure the cue and routine are constant over time.

§  Habitual user

§  Dynamics. If the user returns on a regular basis and responds automatically to an environment cue, you’ve successfully built a habit of use. (You can see whether it’s a habit by looking at the usage pattern in the data and through user research.) Since the behavior is largely on autopilot, it’s highly resistant to change. Only major changes in the cue or a lack of ability to act are likely to break it.

§  What the company should do. Don’t mess with the cue or change the fundamental learned routine that users enact.

In addition, Nir Eyal talks about how to leverage the initial usage of an application to build future interest and make it easier or more valuable for users to return. He refers to it as the “investment” step that occurs right after an individual has had a positive experience with the product. He focuses on investment in the formation of habits, but it’s an insight that can be applied for non-habitual behaviors as well ([ref59]).

LESSONS FOR BEHAVIORAL PRODUCTS

As users gain experience with a product, the type of support they need changes. The most critical periods for adopting a new product to change behavior is the first experience (when curiosity and newness is replaced by an evaluation of the product), and in the transition from a returning user to a regular user (when minor changes in circumstance can disrupt the new behavior).

Does the product provide an excellent first-time user experience? Does it provide a compelling, ongoing reason to return to the product—especially on a regular schedule with a stable cue to facilitate habit formation?

Questions About the Mechanics of Building Behavior Change Products

How Can a Company Get Help Testing Its Product, Especially from Outside Researchers?

Experimentally testing a product, especially when the target behavior can’t be measured within the product itself, can be an intimidating endeavor. Believe it or not, academic researchers would probably love to help test your product’s impact on behavior. Many of them can’t be “hired” in a traditional sense—because they have full-time jobs in academic institutions and for professional reasons can’t accept consulting contracts. But you can build partnerships of mutual benefit if you have enough users of your product to support a scientifically valid study.

Here’s how you can go about building research partnerships:

§  Find researchers in your field. This is a lot easier than it used to be; you can start with Google Scholar (http://scholar.google.com/) to search for the topics you work on, and see whose names are in the most commonly cited articles (the results are sorted by the number of citations). Academic conferences on your topic are another good way to start but take more investment of time and energy.

§  Contact them, asking for suggestions about academics who might be interested in studying the topic with your user base. In particular, ask who might be interested in testing the impact of interventions. Follow-up and contact the suggested researchers; the researchers you initially contacted may be interested, of course.

§  Discover what they need to aid their work. Describe your product and user base. Ask them what they need. Depending on the person, and their status in relation to their career and field, that will vary. Here are common options:

§  Access to unique data. In some disciplines, it is extremely costly and difficult to obtain detailed information about individual users, from demographic information to observed preferences, and especially about changes in behavior over time.

§  Access to a large user base. The power of scientific tests increases with more users, but it is costly for most researchers to gather a large enough test population on their own. If your product already has them, excellent.

§  Access to funding. Financial support from companies to academic social science researchers can be highly problematic; it can taint researchers’ independence and undermine their ability to publish the results of a study with the company. However, if the company is looking for expert advice, and not academic publication of the results, some researchers certainly can be hired for paid consulting arrangements. Other options include providing grants for research on the topic, and supporting grant proposals submitted by the researchers to third-party grant agencies.

§  Develop a shared research plan. With a small, trial project, try working together. Set clear expectations on access to data, funding (if any), staffing (on both sides), and timing of the study and analyses.

§  Don’t try to restrict the results or ideas. Academic research runs on innovative ideas and the data to back them up. Companies can’t lock down the ideas learned from the study—they must be shared between the company and researchers. Similarly, if there is a whiff of restriction on the publication of negative results, that will undermine the reception of the study and turn away most researchers. Paid consulting arrangements work differently, of course.

§  Respect the need for specific testing protocols. In order for a study to be scientifically valid, researchers will have to execute the study according to specific rules. For example, they may require very specific wording for questions asked of users. That’s part of the bargain; and, it will often help the company formulate much more solid conclusions than if it had informally developed slapdash surveys.

LESSONS FOR BEHAVIORAL PRODUCTS

If the testing process seems overwhelming, don’t despair. First, look to off-the-shelf tools for testing applications (discussed in Chapter 12). If those aren’t useful, look for professional research partnerships with the academic community. If your company is doing innovative things to help users change their behavior (which of course you are!), researchers may be interested in working with you.

How Can You Sustain Engagement With Your Product?

Products that hope to actively change long-term behavior can’t do that if people start using the product, then drop off. Sustained engagement with a product requires an act of behavior change just like the product’s target behavior. At a high level, continued usage of a product follows the same rules outlined throughout this book.

There are five preconditions to (re-)engaging with a product. A cue to do so, a positive intuitive reaction, a positive conscious evaluation, the ability to do so, and reason to do so at a particular time (rather than endless procrastination). These preconditions can be met with the three big strategies for behavior change:

1.    Helping the user make a conscious choice to reengage each and every time

2.    Building a consistent habit of logging into the application, with a cue, routine, and reward

3.    “Cheating”—making the product a required part of the person’s daily life (like a time-keeping app at work) or making it a default behavior (like a home page on a browser)

To support the conscious choice to reengage,[159] companies can design around the behavior of reengagement:

§  Tailor the action to the user. Make it dirt simple to log back in again (e.g., by using a single sign-on procedure with Facebook instead of a unique username and password).

§  Design the product and overall environment to support it, by making sure the motivation is there (i.e., that the product is worth using!), providing a clear cue, and removing distractions.

§  Prepare the user by ensuring that prior experiences with the product were worthwhile and enjoyable and by making use of the product seem a natural thing to do, and removing any logistical barriers to actually using the product.

Too often, companies just focus on providing value and wonder why users don’t come back. Value is clearly important—users have to want to use the product (i.e., they must have a positive conscious evaluation of the product), or they won’t voluntarily use it. But there’s much more involved, and much more that companies can do.

Of the five preconditions for reengaging with a product, it’s the cue that I think has received the least attention and thought in product development circles outside of growth hacking (where it is an essential part of the discussion). Perhaps it’s because we think that if we build the greatest product in the world, people will naturally come and use it.

To cue individuals to reengage with a product:

Provide value

Absolutely, this is essential. If users don’t think your product is worthwhile, you aren’t going to be able to grab their attention repeatedly (they’ll change their environment to avoid you). So, value is the first step. But it’s only the first step.

Uniquely become part of the person’s environment

One way to remind people to use a product is to ensure it’s seen—by placing the Nike+ FuelBand by the side of the bed or by making your application the home page on a browser. By uniquely, I mean that there aren’t other shiny things the person is seeing at the same time; that’s a real problem with tweets and emails, for example—they are extraordinarily crowded channels.

Uniquely become part of the person’s expected routine

At a particular time of day (or situation), train the user to uniquely think of the application as a way to do something or relieve boredom. “Training” isn’t happenstance: ask the user to plan out a particular time to use the product (i.e., as part of implementation intentions). If you’re providing value, then help users form a habit around getting that value.

Be so darn cool and memorable that people think of you, on their own

We all aspire to this, and we think that a beautiful product will make users dream about us. I don’t know about you, but I’ve seen a lot of beautiful artwork in museums. I don’t dream about them, and neither do I dream about more than a handful of products. So, invest in other ways to get attention.

Build strong associations with something that is part of the user’s environment or daily routine

If you can’t get in front of users’ eyeballs directly or reserve a slot on their daily calendar, build on what’s already there. For example, whenever there’s a crisp spring weekend day in D.C. (an existing aspect of my environment), I think of the bike trailer my kid loves riding in and head to the garage to get it.

Be useful each and every time they see you

Don’t train the user to ignore you. Does your carpet catch your attention? No. It’s in your line of sight, and you step on it as part of your daily schedule. But it does nothing for you most of the time. Same thing with sending lots of emails to users. Don’t be the carpet.

Be new and different each time

One way to avoid being the carpet is to make sure that each attention-grabbing piece of content contains something new (or potential for newness)—social network notifications do this beautifully with their teaser emails about friends doing stupid stuff. It’s more than a random reward for logging into Facebook or Twitter—the attention grabber itself is different each time.

LESSONS FOR BEHAVIORAL PRODUCTS

You can think about behavior change as a two-stage problem. First, there’s the behavior of engaging with your application. Second, there’s the behavior that the application seeks to support. The same tools taught throughout this book for the latter behavior are also applicable to the former.

Pay particular attention to how the application is cueing people to log in—is there a consistent reminder in the user’s environment that either you can construct uniquely for your product, or that you can associate with your product?

When You’re Seeking to Change Behavior, How Do You Avoid Going Too Far?

There is sometimes a fine line between helping people take an action they want to take, and pushing them to take the action they wouldn’t otherwise take. How do you avoid going too far into coercive tactics? I’m not your mother, and I’m not judging you. But here are some ways to think about the issue.

Personally, I approach the issue pragmatically. In the end, we are trying to change behavior. But we’re trying to change behavior in ways that (a) people want and that (b) are beneficial. At the extremes, we can all reasonably distinguish products that are enabling from products that arecoercive. For example, an arm band that automatically tracks exercise and reports the data to the wearer—that’s enabling. An ankle bracelet that shocks you and reports you to the police whenever you leave the perimeter of your house—that’s coercive. However, in the middle, things get fuzzy.

One reason it’s fuzzy is that we all want many things. And we want different things over the course of our lives and even over the course of the day. Our self-control waxes and wanes, we change our behavior in subtle ways at work and at home, and, let’s face it: we all have “moods.” No person, and certainly no product, is smart enough to judge what another person’s “true” desires are in a given moment.

Things also get fuzzy because “beneficial” behavior change is in the eye of the beholder. There are trade-offs in every action we take, especially as we build products to change our own and other people’s behavior. I might think that it’s a reasonable trade-off between privacy and beneficial behavior change to use an app that monitors my credit card for cigarette purchases and then alerts my spouse. You might not.

While there are gray areas, I don’t think that the ethics of behavior change are relative, squishy things. There are clear lines we should not cross if we claim to help users take action (and not to be forcing them to do so): involuntary actions; not telling users what we’re doing; and manipulating people’s emotions so they make poor choices. Around those lines, we need to be a bit humble and find ways to check ourselves and our motivations. Here are some approaches you can try:

§  Tell users what you’re doing. If the person wants to take the action, at least at some point in the daily ups and downs of life, great. Telling the user what you’re doing shouldn’t cause a problem and is a good, simple check on excess.

§  Make sure the action is optional. Leave it up to the user to decide whether or not to use the application, and don’t hide behind false options—for example, an app that monitors employee productivity at work isn’t optional (the job may be “optional,” but the app isn’t).

§  Ask yourself whether you’d want someone else to encourage you to use the product. Is this product really designed to help you?

§  Ask others, especially strangers, if they would trust the application.

Avoiding coercion doesn’t mean that you encourage users to do anything they want to do. The company will have, and must have, a stance on the behaviors it wants to encourage. “Dieting” and “eating everything you want” aren’t two equally valid options for weight loss. One works (sometimes) and the other doesn’t. You can talk about and be up front about that stance. If you’re helping people diet, don’t be ashamed about it. Do it, and do it proudly, but be transparent and make participation optional.

Many types of products, even those that are explicitly coercive, can be good and useful. The ankle bracelets used for home detentions probably fall into that category. On net, society is better off because of their use. But that’s a different type of product, which deserves scrutiny and thought. Here, my goal is to spur thought about products that enable voluntarily behavior change so that we are clear about what we’re doing and the means we’re using to affect user behavior.

LESSONS FOR BEHAVIORAL PRODUCTS

Talking about product ethics may be an unusual topic for a book aimed at practitioners. But we can’t outsource ethics. We should feel proud of our work. Part of that means double-checking that the product is truly voluntary, is up front about the behaviors it tries to change, and seeks to make a beneficial change for its users.

How Does Designing for Behavior Change Affect My Business Model?

Helping users voluntarily change their behavior can directly support the core business goals of a company, but these types of products do raise special business considerations.

Recall from the Preface that there are two main targets for behavior change:

§  Behaviors that users want to change within their daily lives

§  Behaviors that occur within the product itself and are required to use the product

In the latter group, the business impact is clear and straightforward. Let’s say the behavior change task is learning how to better organize email in an email client. If you help users do this more effectively, they are more likely to (a) buy later editions of the product in the future (or renew, if the product is on a subscription model) and (b) recommend it to others. In both cases, behavior change means more revenue.

For behavior that users want to change in their daily lives, there are a variety of options, depending on whether the behavior is frequently repeated or occurs once. Let’s say the behavior change task is exercising. It’s a repeated behavior that people often want to change in their daily lives. Standard business models work perfectly well for this type of product and align user success with business success:

§  Revenue increases with the time users spend using the product.

§  Subscriptions and reoccurring fees. For example, a product that helps people set, track, and compete with friends over exercise goals. If the product is successful, both exercising and revenues are sustained. A similar model exists for products that automate user tasks altogether—like target date funds in the investment space, which charge an ongoing fee for automating the process of asset allocation.

§  Advertising. For example, the same exercising tracking/competition product could be free to the user but advertise fitness products. If the product is successful, it continues to deliver an interested, engaged audience for advertising.

§  Revenue increases with new sales.

§  Market penetration. For example, an exercise tracker like the Nike+ FuelBand. People don’t really need two exercise trackers. But if the product is successful at changing behavior, then current customers refer others, and sales grow. If the market becomes saturated, that’s a nice problem to have, and the company can move on to new product lines (up-sell).

§  Cross-sell and up-sell. For example, the company has multiple exercise products (a tracker, shoes, clothing, etc.). If one product is successful at changing behavior, then the customer is more inclined to buy other products from the company.

§  Revenue increases with the success of users at changing their behavior.

§  Incentive-aligned third-party support. For example, a third party benefits from the same thing that users benefit from: behavior change. The third party would then pay for the product on the user’s behalf. Employers benefit from their employees being physically healthy and pay for products like Keas to help employees exercise. Energy companies benefit from their customers being energy efficient, and pay for products like Opower.

As you can see, products that change repeated behaviors follow the same business models as any other product—plus the somewhat unusual option of third-party support.

Products that change a one-time or infrequent behavior have a more difficult time aligning user success with business success. Consider, for example, the behavior of getting a mortgage, and products that help people shop around for a better one. Thankfully that’s not something that most people do very often. Here’s how standard business models become problematic:

§  Revenue increases with the time users spend using the product.

§  Usually not relevant. Customers just want to find a good mortgage; they won’t pay more if it takes longer to find one.

§  Revenue increases with new sales.

§  Cross-sell and up-sell. This is feasible if the company offers related services—like banks that offer mortgages and checking accounts. However, customers have a hard time assessing the quality of one-time services like mortgage-buying assistance; that uncertainty leaves room for chicanery. In the mortgage market, there’s a significant temptation (financial incentive) to shift from providing unbiased decision-making support to pushing mortgages that generate high, lead-generation fees.

§  Market penetration. Generating new sales based on high-quality decision support for mortgages is possible. However, because customers have difficulty assessing quality, the same temptations exist for the companies that seek to “help” users make that decision. Hence, businesses use gimmicks that made customers excited up front, but leave them loaded with significant fees and risk, like balloon mortgages.

§  Revenue increases with the success of users at changing their behavior.

§  Incentive-aligned third-party support. A one-time change in behavior would need to be very significant to attract (incentive-aligned) third-party support. In the case of decision-making support, Fannie Mae and Freddie Mac have appropriate incentives for end users to receive optimal mortgages, but few other entities have incentives aligned with the end user for such a product.

And, of course, there are business models in the behavior change space that clearly are not aligned with the interests of user. Gyms are the classic example of businesses that profit because their users fail to change their behavior. Many gyms rely on fees paid by users who plan to exercise but never do.

LESSONS FOR BEHAVIORAL PRODUCTS

This probably isn’t in the domain of designing for behavior change per se, but the company should think about how a focus on behavior change affects its bottom line. Does the company rely on subscriptions and would lose money if users simply solved their problem? That’s a misalignment of the company’s goals with the user’s goals and spells trouble. Thankfully, there are numerous incentive-aligned options, especially for repeated behaviors.