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

Part III. Developing the Conceptual Design

Chapter 8. Preparing the User

One of the best studies I’ve ever seen on behavior change was run by Tim Wilson, a social psychologist at the University of Virginia. He took a group of first-year college students who were struggling—they weren’t doing well in school and were worried about their future—and randomly assigned the students into one of two groups: one group received a short, 30-minute intervention; the other received nothing special.^[97^]

Wilson was concerned that the students saw themselves as failures. His intervention entailed giving the students information about potential interpretations of their bad performance in school:

We gave them some facts and some testimonials from other students that suggested that their problems might have a different cause…namely, that it’s hard to learn the ropes in college at first, but that people do better as the college years go on, when they learn to adjust and to study differently than they did in high school… (Gilbert and Wilson 2011)

The randomly selected group that reinterpreted their bad grades got better grades in the future. They got better grades all the way to their final year in college; they also were less likely to drop out of college. While the study did not track their full academic performance over time, we can posit that the effects were not immediate. Rather, it appears that students would have slowly changed how they saw themselves and gradually changed the amount of effort they put into their studying, after this initial push.

A 30-minute intervention that changed performance for years? Impressive.

Wilson is a leading proponent of the idea of “story-editing”: like the students in his experiment, we can reinterpret what’s happened to us in the past by changing the story we tell ourselves about it—our self-narratives ([ref188]). That reinterpretation then affects our future behavior. When we change our behavior, we also change the experiences we’ll have in the future, making them marginally more likely to support our self-narratives. And with each new experience, our internal story of who we are changes a bit more, spurring a new cycle of behavior change.

For the students, it would have worked like this: Wilson helped half of them interpret their performance differently. Those who saw themselves as going through a temporary tough spot (and not as failures) would be slightly more likely to work harder and perform better on the next test. They would then look back at that (improved) performance and reinforce their understanding of themselves as students who could study and could overcome the challenges of first-year life. They would then work even harder on the next test, perform better, and so forth. With time, the internal stories, or self-narratives, of the two groups diverged, thanks to a small push from the initial intervention.

We interpret and reinterpret our experiences every day of our lives, and thus shape our self-narratives and our future behavior. These cycles of interpretation and behavior can clearly support beneficial changes, like studying more. They can also lead to negative ones, like when someone feels like a failure and doesn’t put in effort to try to change that. It depends on how we use our past experiences and whether we see ourselves in control of the outcomes of our lives.

Tactics You Can Use

Tim Wilson’s story-editing technique is an example of preparing the user to take action in the future. By changing the students’ perceptions of the past, they were more likely to succeed in the future. This chapter is about techniques to prepare users for action. In each case, the intervention helps the user think about things differently, and that shapes how they act. Here are the three tactics:

Narrate

Change how users see themselves.

Associate

Change how users see the action.

Educate

Change how users see the world.

I’ll focus on the preparation that occurs before the action itself; in Chapter 7, we covered how the user can be “prepared” to act via increasing motivation at the time of action. This approach is different—it’s about changing some part of the individual that carries over into the future, often for a lasting and sustainable impact. Let’s start with the self-narrative.

Narrate the Past to Support Future Action

Our self-narrative, as you saw in Wilson’s experiment, is how we label ourselves, and how we describe our behavior in the past. Products can help people see themselves differently. The goal, from a behavior change perspective, is to help people see themselves as someone for whom the action is a natural, normal extension of who they are.

In other words, if you want to help people begin exercising (like the Moves app does),^[98^] help them see themselves as people who have already been exercising in small ways,^[99^] and just need to do more (e.g., first-time Moves users may be surprised to find out how far they normally walk each day). An easy way to support this process is by merely asking people about things they’ve done in the past that are related. And congratulate them for the work they’ve already accomplished. Where their past experiences have been negative, as they were with Wilson’s first-year college students, help them see those experiences as surmountable challenges and not inherent character defects.

Essential to a supportive self-narrative is the belief that one can actually succeed in the action (i.e., users need to feel that the action is under their control and that they have the skills and resources to—potentially—make it happen). That’s the sense of self-efficacy ([ref10]) discussed inChapter 2 under the ability to act. Reminding people about their prior successes at related tasks can help build that sense of self-efficacy; so can the “small wins” and positive feedback described in the last two chapters.^[100^]

Stepping back a bit, there is another very subtle influence that one’s self-narrative, or overall mindset, has on their bodies apparently burned more calories due to how the mind interpreted the exercise in the maid’s normal routines ([ref39]). The effect was a placebo, as NPR’s Alix Spiegel put it: “If you believe you are exercising, your body may respond as if it is” (2008). Our beliefs not only affect our behavior directly; they also affect our bodies.

Associate with the Positive and the Familiar

Chapter 1 and Chapter 2 talked about how, for many of our choices in life, we intuitively know whether taking an action feels right or not to us. A big part of that is our prior associations—our learned experience that buying a fancy pair of shoes is going to make us feel great when we walk out of the store with them, at least for a few days.

Products can build these associations to help a person change behavior. In Chapter 6, we talked about changing the action itself so that it leverages prior experiences. Here, the product can be changed so that it helps users make the mental connection between the action they want to take and their prior experiences. I call this a behavioral bridge, because it helps the user cross from one type of behavior to another by making it less “new” and difficult. The bridge connects past experience with future actions.

Here’s an example: Speek is a conference-calling application that allows people to switch from using a dial-in number and long PIN code to using a simple link in a URL (Figure 8-1). Dial-in numbers and PINs are frequently misplaced and annoying to enter. When a user clicks on the URL, Speek calls them and patches them into the conference line. The challenge is that using a URL is new and strange. In the product, the company highlights the new and unique aspects (ease of use, etc.) but is also careful to leave a behavioral bridge in place—a comforting bit of information about how users can treat it like a normal conference call if they need to, since the underlying technology is a conference line with a dial-in number and access code.

Figure 8-1. Speek does something new: URL-based conference calls. It also provides a behavioral bridge to associate it with a more familiar action—using a standard dial-in conference call (in gray under the main box).

Educate Your Users

Education is all about giving people the information they need, and then hoping that when the time comes to act, they will make an informed choice to act.

Focusing on information is a noncontroversial and common approach to behavior change. We think that everyone would believe like we do if only they had access to the same information and training. This has been an approach taken by many of my otherwise favorite NGOs and government agencies.

Unfortunately, information doesn’t equalize us and doesn’t make us behave the same. As we saw in Chapter 1, conscious information may have nothing to do with action at all—the action may be habitual or based on intuitive reactions. Or, when it does have an impact, it is filtered through all of the rest of our experiences and information.

Providing users with information can be immensely powerful. But one should be thoughtful about how and when it’s applied. Education efforts falter when:

§ The action that people will take isn’t consciously thought through at all—it’s habitual or otherwise automatic.

§ People are overwhelmed with too much information.

§ The information comes too long before (or after) the decision needs to be made. We rapidly forget unconnected, unused facts.

An example of an education-only approach that clearly didn’t work is mortgage disclosures in the United States. Mortgage lenders are required to provide reams of documentation on everything from how the loan works to the fact that most older homes have lead paint in them. This is all important information—and rarely read. It’s too much to take in, not structured to demand attention, and isn’t clearly actionable. The only “action” the mortgagee can take by the time he or she gets all of the disclosure documents is to walk away, with no house, and no certainty about where the escrow money goes.

To better understand when education can be effective, let’s use a common example: educating people about the importance of saving for retirement with financial literacy seminars. In the 1990s, there was a significant increase in the use of retirement planning seminars, as employers have shifted from pension plans to 401(k) plans that individuals directly contributed to and managed ([ref15]). These seminars, and other financial literacy programs in high schools and beyond ([ref125]), have been the subject of quite a lot of controversy, with numerous researchers questioning their impact ([ref123]).

Consider three different approaches that a retirement education product could take. It could educate people about why an action is important, how the action works and the raw data needed for a good decision, or what to do to take the action:

Why

By and large, we already know that saving for retirement is important. No one wants to die in poverty. However, for other beneficial actions, users of the product may not honestly know the importance of the action; for example, I didn’t know that skin cancer screening was important for (relatively) young people until recently.

How

We don’t understanding the inner workings of 401(k) plans and score poorly on financial literacy questions about basic topics like compound interest ([ref121]). And, there’s evidence that this knowledge helps us make good financial choices ([ref91]). But if the information is delivered too far in advance from a moment of decision, as many financial literacy programs are, we simply forget it.

What to do

We don’t know what to do when confronted with dozens of 401(k) options—in fact, we often take a naïve strategy of putting an equal amount into each fund we’re faced with. A simple heuristic for easy diversification—use a stock market index fund or a target date fund—can simplify and shape that decision to users’ benefit.

Which type(s) of education is best depends on the particular situation. With voluntary behavior change, we assume that the user already has some motivation to act. Information about how a system works can be fascinating for those already in the know, but overwhelming and too removed from the actual decision for those who aren’t. Logistical information (what to do) can provide clear actionable guidance and increase the user’s ability to act immediately, a key component in the Create Action Funnel.

How Training Your Users Fits In

You can effectively provide users with information—educate them—at the moment they start interacting with your product, and thus prepare them for the tasks ahead. But you can’t suddenly change their skills once they start; that takes time. You can, however, use their early interactions with the product to boost skills that are important for later interactions.

In Chapter 6, we already broke down the task into small steps that provided small wins and built toward the final goal. The early steps should already provide the structure you need to train the user for later interactions. For example, in Chapter 6, we talked about an application that encouraged people to run regularly. Some of the early steps included buying running shoes, planning out a path for the first run, and placing the running shoes by the door to remind the person. Here’s a quick way to check that the necessary skills are being developed, using that running app as an example:

1. Look over the sequence of actions the user is taking. For example, getting running shoes, planning out a course, and so on.

2. Starting from the end, identify the most challenging parts where the user is most likely to fail. What skills are required for the person to take that action? For example, your users might need the skill of adapting to bad weather and maintaining their routine, or they’ll get out of the groove.

3. Given what you know about the user, and the previous actions in the list, would the user have those skills already?

4. If not, are there ways to build in similar, but smaller, challenges earlier in the process?

Building in a series of incrementally harder challenges is, of course, related to a standard technique in game design: game designers look at ways to build a succession of challenges that grow progressively harder as the gamer gains experience with the system. Similarly, it’s a standard technique in education. We learn addition before algebra, and algebra before calculus. Sometimes we forget these lessons in designing our product, because we ourselves already have the skills required to act. This is just a brief reminder to look at the process from the perspective of the skills your users have, now that you’re almost fully done with your behavioral plan.

In some cases, you can add in skill-building exercises within the app; in others, it may require a previous training program. For example, I remember that I had to take an intensive math training over the summer (aka a “catch-up class”) before I was allowed to even apply for a particular program in school. If your company has a suite of related products, they can be explicitly designed to build on one another: like Rosetta Stone’s language-learning programs.

Update the Behavioral Plan

Use the tactics from this chapter to review and update the behavioral plan. Specifically:

§ With respect to the action, how do users think of themselves? Do they see themselves as people who can (generally) be successful? Yes, it’s kitschy, but it’s true—if people don’t believe in themselves, good luck getting them to take action. Story editing can help turn a negative self-perception around.

§ What positive, similar experiences can the product hook into? How can it build a behavioral bridge between those experiences and the action the user needs to take?

§ Is there any information that the user absolutely needs to know before embarking on the journey? I say absolutely, because overloading users with less-than-necessary information can distract and slow them down.

§ Are there skill-building exercises or experiences that should be added early in the process to prepare users for harder, later steps?

Once you’ve updated the behavioral plan, let’s take a look back at how these techniques relate to what we learned about the mind’s decision-making process in Chapter 1.

How Behavior Change Techniques Relate to the Thought the Behavior Requires

Remember in Chapter 1, where there was a spectrum showing the different ways that the mind makes decisions, from those that required very little thought to those that required lots of conscious thought and attention? Now that we’ve gone over the major techniques you can use to create a supportive context for action, let’s revisit that spectrum.

Each of the techniques discussed thus far maps onto the spectrum—each is appropriate for particular types of behaviors and their level of thought:

§ Building habits is important for often-repeated, familiar situations (go to refrigerator, get snack), where we’re not really making any “choice” to act at all. Techniques to change habits are similarly relevant where there’s an existing habit in place; and most conscious efforts simply won’t do much.

§ Automation and “cheating” are even more extreme, requiring no thought at all except for the initial step of getting consent. They are appropriate in cases where the user’s involvement isn’t required to adapt and personalize the solution; and it’s acceptable (and possible) just for the person to sign off on the action occurring on their behalf.

§ Building associations with prior experiences and changing one’s self-narrative are important for semi-familiar situations where we make a conscious choice, but still use our intuitions.

§ Structuring the action into a sequence of small, easy steps decreases the amount of work a user needs to do at a time. That’s most appropriate when there is a lot of work to do—when there’s some conscious thought required.

§ Educating your users, as part of a broader array of tactics, is especially important for unfamiliar situations where the most work is done—and you consciously weigh the pros and cons of a behavior.

Revisiting the graphic from Chapter 1, now with ways to change behavior, we have a set of inventions (on the bottom) and the scenarios for which they are appropriate (on top); see Figure 8-2.

Figure 8-2. In familiar situations or where little thinking is required, products can employ habits or automation to shift the burden away from conscious user attention

This graphic can help you choose the right intervention for a given behavior. Misunderstanding the behavior and using the wrong intervention can have serious consequences. For example, there are countless well-meaning programs that try to educate people about the dangers of smoking to get them to stop. But by the time someone is a smoker, smoking is already a habit—people don’t think about it much. In fact, some great research with brain scans shows that attempts to provide information to smokers about the dangers of smoking just triggers the habit—they want to smoke more, right then ([ref117])!

So far, this book has been designed to help you carefully plan out the application to make it effective and avoid such errors. By combining knowledge of how the mind works with specific strategies and tactics to change behavior, you can understand the type of behavior change problem you face, and develop a conceptual plan for the product. Next up: how to apply the same understanding, strategies, and tactics to developing wireframes and mockups of the application.

On a Napkin

Here’s what you need to do

§ Help users see themselves as people who would naturally take the action.

§ Help users build strong associations between things they are already familiar with and enjoy and the new action.

§ Provide users with clear instruction on what they need to do and any other essential information to act.

How you’ll know there’s trouble

§ You’re providing users with reams and reams of information or trying to get them to think exactly like you do about the action.

Deliverables

§ The updated behavioral plan, which describes how the user moves from inaction to action, including how the product prepares them for that action