Behavior Change Part 3: Building Products
This is the final post in a three part series on behavior change, covering theory, case studies, and how to build products for behavior change.
The question then becomes, how can we build better products to drive behavior change? We have many formulas and frameworks above, but it will be useful to factor out some of the key elements and think how we might build a better service from the ground up.
Overview of the proposed model for productizing behavior change.
1. Psychological Foundations
Before embarking on behavior change, it’s important for three things to be true:
You strongly desire the outcome
You believe that you have the power to change
You have mental tools to build resilience and reframe your relationship with the problem (e.g. applied CBT)
These will lay the foundations to change your life for the better and provide fuel for your life change. No matter what, it’s going to require work, and to do the work you have to have a strong “why”.
2. Better Habit Formation
It’s clear that the “New Year’s Resolution” approach of just starting a big, new behavior from scratch isn’t a good approach. Work backwards from your goal and focus on making the habit small and consistent, and ensure the cue is obvious, the action is attractive, it’s easy to perform, and that you get a satisfying reward afterwards.
An important fork to consider is whether all of these elements are intrinsic to your product or not. It’s easier and more scalable for the cue to be an app notification and the reward to likes or badges within the app, but sometimes the most powerful cues, actions, and rewards are in the real world, and there’s an opportunity to help people set these conditions up in their own lives.
Consider behavior change as a fractal set of habits. Noom helps create a habit around food logging (you eat food, you open up the app to track, you get rewarded in the app for doing a good job), which is one important component of achieving a more general goal of weight loss.
3. Structure
Should you have an opinion on habit formation? It’s a lot easier to just provide tools than to provide users with all of the elements needed for habit formation. It’s harder and messier, but clearly a more reliable path for behavior change.
3.1 Goals and Plans
Figuring out what to do next can be a lot of work. Whether it’s deciding on the best diet to become healthier or a bodyweight workout training plan, people practically need to become experts in a domain before they can confidently decide what to do next. There’s an enormous amount of value in finding the intersection between the latest science and best practices, and a given person’s goals, needs, and motivations. People want to be able to relax, trust the plan, and focus on executing.
A plan should draw the most efficient path from someone’s current state and their desired destination. It should start as small as possible to establish the habit, and use feedback to ensure you’re making as much progress as you can given your constraints. Was last week too hard? Let’s dial it back this week. Just right? Let’s keep pushing at a similar pace.
3.2 Tracking
Tracking inputs is important for three reasons:
Ensure plan adherence. You can’t progress if you’re not doing the work. If you’re having trouble sticking to your plan, try to adjust the difficulty, or revisit how you’re building new habits so they can stick.
Iterate on what is or isn’t working. If you’re doing the work and not getting results, this can help you zero in on what needs to be changed.
Provide motivation. You can gamify this with streaks and badges, and create a habit loop where logging another successful day is a reward in itself.
Ideally tracking can be done automatically, like with sleep tracking or logging runs with an Apple Watch, but for tracking food, mood, or other hard-to-track activities, manual logging is needed. Think hard about how to lower the friction of tracking each datapoint, and create a habit loop around this data collection.
3.3 Feedback
Sometimes, an action can require a lot of nuance to perform correctly. You need good form when you’re running or squatting to avoid injury. You need advice from a teacher to nail the dynamics, phrasing, and timing of a new piano piece. A coach or a more experienced partner can be very helpful in delivering this feedback, and it’s something that technology, today, isn’t very good at since it requires perceiving and understanding the human’s action, and comparing it to an understanding of what “good” looks like. However, this is actively being worked on, for example with tempo.fit.
3.4 Outcome Metrics
It’s critical to know whether what you’re doing is working, and getting you closer to achieving a goal, for two reasons:
It proves that what you’re doing is working, and you’re progressing towards your goal
Provide motivation and reward to reinforce your habit loops
The best outcome metrics have the following four properties:
Accurate and trustworthy. Do you believe the number? Some companies make up arbitrary scores without explanation to the user, and this undermines their utility.
Meaningfully approximate your goal. If I’m trying to improve my productivity, a metric on the amount of time spent on my computer will be accurate, but correlate weakly with the thing I care about.
Easy and cheap to measure. Ideally measurement is automatic and frequent. Continuous glucose monitors are better than fingerpricks, which are better than getting a datapoint at the doctor every month or two.
Sensitive to short-term changes. If a metric isn’t moving, it has limited utility to provide feedback on whether your plan is working, nor will it provide much motivation.
Running is a great example. How much faster and longer are you running vs. last week? How far are you from your target 8:00 minute pace for your upcoming half-marathon? Managing blood pressure is another good example. You can measure it yourself non-invasively using a blood pressure cuff, with medication and lifestyle changes it can change within weeks, and it predicts risk of heart disease very well.
3.5 Iteration
These ingredients combine to form a feedback loop. Make a plan, do the work, see if it works, and repeat, until you’ve reached your goal.
4. Ensuring Success
Having a perfect plan and tracking setup is often not sufficient to ensure success.
4.1 Carrot: Motivation
Motivation is fickle. It comes and goes, and you shouldn’t rely on it for the success of your plan. That’s why habits are so important; the correct behavior becomes automatic, even if you don’t feel like doing it. However, motivation is crucial to get the journey started, and get you through patches when your habits waver. This can come intrinsically, when you get fired up about how badly you want to succeed, or extrinsically, when someone hypes you up. Sometimes tough love can help you break through tough mental barriers and get to the next level.
4.2 Stick: Accountability
Accountability helps convert long-term desires (of course I want to run four times this week!) into short-term action (I don’t really feel like running right now, but I better get up and do it). If you lean too heavily on guilt or potential punishment it could build bad associations and undermine your whole program, but the prospect of incurring some personal cost can be very helpful. Examples include wagers where failure means donating money to a cause you dislike, or setting up social engagements where backing out would disappoint the other person.
5. Community
Community adds value in a few interesting ways:
Increases accountability and motivation. You might disappoint your peers or feel a loss of status if you don’t make progress, and naturally wanting to fit into a group can drive action.
Increases stickiness. If you become friends with a group of people you’ll want to keep hanging out with them, which will make it more likely you’ll keep improving over time.
There are many forms of community. Having a morning running partner can help you get up in the morning. Having a cohort of people working on similar goals can be motivating. Coordinating groups of humans can be messy, but many new health and education startups are taking on this complexity because it can be so effective in driving behavior change.
A social network is a form of community that can provide significant motivation. It’s interesting to think about how to apply the principles in Status as a Service, and set up a system in which you’re “mining” for social capital through beneficial personal transformation. Strava is probably the clearest attempt, but generally hasn’t been a breakout success. I’d bet it’s because the the content is relatively homogenous (little variable rewards) and isn’t particularly valuable/engaging for consumers, especially when compared to other consumer social apps. Peloton also has some social features that can help motivate, with leaderboards for intensity and consistency and shoutouts from instructors.
6. Coaching
Coaches are usually a bundle of three components:
Creation of a personalized plan.
Feedback on whether you’re doing activities correctly.
Motivation and accountability to ensure you’re doing the work and staying on track.
It’s becoming increasingly clear that self-serve tools aren’t sufficient to drive long-term change for most people, and there are a lot of startups popping up to address this need with coaches (e.g. Future Fit).
Coaches are expensive, which begs the question, how much could be done by technology? Automatically creating a personalized plan is relatively easy. Providing nuanced feedback about whether you’re doing activities correctly is hard and requires specialized knowledge, though advances in computer vision are helping in specific verticals. Providing motivation and accountability is an emotional task and would be difficult to automate, but perhaps could be done as effectively by a partner from the community.
Thanks for reading. I’m always curious to hear personal stories of behavior change, please let me know what techniques have worked for you and what you’ve changed.
P.S. If you’re hungry for more mental models, check out this one from Patricia Mou at Wellness Wisdom.