Designing Behavior Change
Can you remember the last time you successfully changed a behavior or adopted a new habit? Maybe it was something like going for a run every morning, meditating more, eating smaller portions, or just remembering to take your vitamins. As you might have found, achieving long-term behavior change is an incredibly difficult task, despite our knowledge of the benefits. Fortunately, there are a growing number of companies set on helping us take this on, particularly in healthcare. Companies like FitBit, My Fitness Pal, and Runkeeper use sensors on our phones and wearable devices to track our behavior and display trending. Unfortunately, we’re finding that many users tend to disengage with these kinds of tracking systems after the initial novelty wears off.
In this article I’ll discuss how we might keep people more engaged long term by creating behavioral change systems that are more proactive, personalized, and contextually aware. I’ll also discuss how we’ve attempted to apply some of this learning to designing Healthy Bytes, a Google Glass app that encourages a healthy lifestyle, and how this experience has gotten us to think differently about what really changes our behavior for the long term.
Current Tracking Systems
So, like many others, maybe you decided to start tracking your diet or activity level in order to make healthier decisions. You could start by browsing the 20,000 apps in the iTunes store under “Health and Fitness” or you could go with one of the more popular health tracking systems like Fitbit. Let’s use Fitbit’s Aria smart scale as an example, because it’s one of the more simple tracking systems. Every time you step on the smart scale it records your weight and body fat percentage and automatically syncs this data with a smartphone app. You can then view your progress with beautiful charts and graphs and see if you’re losing or gaining weight. The assumption is that you will make healthier decisions because you can see trending, however, the reality is that these tracking systems tend to become more of a diary of failures than a tool to help you achieve long term wellness. This is because long term behavior change requires more than the knowledge of the impact of your past behaviors. It’s like an open-loop problem, where the connection to future behavior is the missing link.
This is a very real problem for companies selling behavioral tracking systems. The quantified self movement won’t graduate into the mainstream until people know what real-life impact they can achieve with the data. Our guess is that this is why people who start using this kind of tracking software tend to slowly disengage after the initial novelty wears off. At a Startup Weekend in Raleigh, NC, we interviewed 24 people on the street who had used some sort of tracking system and found that 80% of them had discontinued long term use. Participants who had discontinued use were not really able to say why, they either learned what they wanted to know from the data, or that they just stopped using it.
What we really want to know is how far we should walk, or what we should eat. This is the leap I call moving from tracking behavior to guiding behavior.
This isn’t to say that tracking systems aren’t useful. If a person already has a healthy behavior established, like running every day, then tracking systems can optimize that behavior. For example, Nike+ uses GPS and motion sensing from your phone to give information about your run. You can see how many calories you burned, the distance you traveled, and even share map of your run on social networks for re-encouragement. This feedback might make you run a little more, or at least remind you how awesome you are, which is great. I’m just saying that tracking systems are most valuable for people that already have a behavior established, not necessarily for persons interested in adopting a new habit. In order to achieve long term behavior change, we need to think much deeper about what factors influence behavior change in the real world. Current health tracking applications are all about showing you how many steps you just walked or how many calories you just ate, not about suggesting what you should do in the future. What we really want to know is how far we should walk, or what we should eat. This is the leap I call moving from tracking behavior to guiding behavior.
From Tracking to Guiding Behavior
Guiding behavior is a much more complex task because it requires us to be proactive about how we communicate to the end user. First, we need extremely low-effort way to collect data. This is a problem that many tracking systems have solved — collecting data through mobile devices, wearables, or quick self-logging. The difficult part is then synthesizing this data and building an algorithm for recommending the next most desirable action. Next, we need to be able to communicate this recommendation in the most compelling way by having a much deeper understanding of context, personal preferences, and timing. Finally, we need a way to measure the effectiveness of this communication so that we can make smarter communications in the future, effectively closing the behavioral guidance loop. Luckily, clinical psychologist, cognitive psychologist, and behavioral economists have already made great progress building frameworks for behavior change. We can now use these frameworks and apply them within the mobile context with the goal of automating behavior change.
Dr. BJ Fogg, a professor at Stanford’s Persuasive Technology Lab, has established 3 principle factors that need to occur in order for a target behavior to happen. The user must have sufficient motivation, ability, and an effective trigger all at the same instant. The third factor, an effective trigger, is a big chunk of what’s missing from tracking systems. A trigger could be something simple like a sound, a vibration, a text message, a color change, or something more elaborate like an email to your wife notifying her that you forgot to take your medication. Whatever the case, the user needs to notice the trigger, associate that trigger with the target behavior, and have sufficient motivation and ability to complete the target behavior.
The timing of a trigger is critical for success. It must occur at the most opportune moment to push the user past their activation threshold. This is what gets us so excited about using Google Glass as a platform for guiding behavior. Wearables such as the Nike Fuelband, Fitbit, or Jawbone UP require the user to pull out their phone, or engage with the hardware whereas Google Glass is omnipresent, basically blending into a user’s life. We don’t necessarily think Google Glass is the perfect platform, its success with widespread user adoption remains to be proven, but its game-changing ability to get in people’s faces at the most opportune moment can get us to think differently about how we can better humanize trigger communications. We now have the opportunity to move from charts and graphs to communications that are much more proactive, personalized, and contextually aware. This is what Amy Roberts, a statistical epidemiologist at UNC, seeks to do with Healthy Bytes.
The mission at Healthy Bytes is to encourage healthier behavior through targeted interventions on Google Glass. Currently, we’re exploring a variety of different user experiences and building prototypes to test our ideas. Right now we’re focusing on how we might personalize behavioral guidance systems. It all boils down to the fact that what motivates people to change varies greatly. Some people are more competitive, some people need positive encouragement, some people like games, and some people just want to be told what to do. We’re also focusing on the timing of trigger communications. For example, we might want to help users make in-the-moment decisions with regards to their diet. Instead of managing the complexity of communicating calories in and out, we can make a more humanized communication of what the user might eat next. In the following example we can use the user’s health goals, historic calorie information, food preferences, and GPS location to understand what restaurant they’re at and proactively recommend what is most healthy to order.
Our goal is to focus on designing communications that will be most effective in the moment. It’s less about communicating data, and more about reverse engineering what’s effective in the real world and replicating that on Google Glass. In the above example, perhaps it’s a friend who’s recommending something healthy to eat. It could also be their dietitian who’s tracking their behaviors and designing interventions at the right moment. We can also think differently about how we humanize the whole calories in and calories out loop. In the example below we’re communicating how much more activity is required to burn off an item that they previously ate.
Of course, this is really only scratching the surface on what can be achieved on this platform. There’s a massive amount of data that can be used to help personalized behavior guidance systems. Your electronic health records, personal preferences, habits on other applications, and data collected from mobile devices can all be utilized. The important part is to start turning our heads from gathering, analyzing, and synthesizing data to figuring how we can communicate that data in the most human way. If we can figure out how to more effectively change people’s behavior, even by just a small percent, the impact it can have on our lives could be massive.
Think about how much detrimental behaviors costs our nation — poor pill compliance, obesity, our happiness levels; everything boils down to the little behaviors we do every day. According to USA Today, the US spends $258 billion a year in ER and doctor visits because 50% of patients forget to take their medications. The U.S. also spends $50 billion trying to lose weight every year. Building an effective behavior change platform isn’t just an interesting business opportunity for brands to interact with people’s lives; it affects us as a society. Our daily behaviors define who we are, our level of happiness, health, and fulfillment. We don’t really think about it, but we are a culmination of habits that slowly build up over time. If we can figure out how to create platforms that allow us to be more intentional about the habits we introduce into our lives, then we can build products that are truly life changing. This is what we seek to achieve at Healthy Bytes.
What do you think would help create healthier habits using Google Glass?
Originally published at dxlabdesign.com on September 15, 2013.
