Behavior Design Part 1: Habit Design
You are different because of the technology you use. We tend to think of our relationship with tools as one way. I use a tool to change the world. But in reality that’s a two way road. The world, through contact with my tools, changes me. From calloused hands to cobblers’ femur our tools have always affected us. But information technology tools are different because they primarily reshape our minds.
Especially in consumer technology, the effect that our tools have on us are much larger then the effects our tools give us on the world, and we should start designing our technology with this effect as our primary concern. How can we design tools, optimizing for how it will change the tool user? I call this approach behavior design, and we’re going to start by looking at how technology shapes one type of behaviors in particular; habits.
What does Pinterest Want?
Technologies want users to behave certain ways. Most teams don’t have a systematic plan for designing these behaviors, but the insights of Neuroscience (and especially NeuroEconomics) offer us frameworks for how behaviors work and how we can build behaviors. The teams that use these insights have a huge advantage in getting the behaviors that they want from users. And in this post I’ll start with an overview of the framework I use for thinking about how to create a very important class of behaviors: habits.
What is a habit?
Most people think of a habit as “something that I do a lot,” or “Something I can’t stop doing.” To better understand our users (and ourselves) there is a more useful way to think about habits — as fast links between what we perceive and what we do.
A habit isn’t just an action; it’s the pairing of a “trigger” (something we perceive) and an “action” (something we do) where the action automatically follows the trigger. When designing habits, we’re really strengthening the connection between a trigger and an action. Neuroscience gives us both an understanding of what happens when the brain pairs triggers to actions, and how we can make that happen. Scientists describe a habit as a “trigger”-“action” pairing.
Here’s my model for creating and strengthening that pairing:
- Present a trigger
- User performs an action
- Always provide feedback
- Sometimes provide a reward
Let’s break this down step-by-step.
Trigger
“A visual image in the hand of an artist is merely a tool to trigger a mental image.”
-Roy H. Williams
Your phone chimes as a text message arrives and you automatically reach for your phone. The same chime goes off a few feet away as your friend receives a message on his phone and you still automatically reach for your phone. What’s happening here? This is an example of a tightly coupled trigger–action habit. The noise is a great trigger because it’s ubiquitous, and it is “reinforced” (meaning you are likely to do the action when you experience the trigger) because it is only rewarding sometimes.
While many triggers (like the chime) come from the world around us, some are internal and come from inside of us. For example, many people respond to the trigger of boredom or anxiety with the action of logging on to Facebook or Reddit. Watch yourself the next time you get bored or anxious. For you, is there an action that is tightly coupled to these trigger emotions?
Nir Eyal has a framework for transitioning a habit from an external trigger (like a cellphone chime) to an internal trigger (like an emotion or thought). I’ll be talking more about this idea here and be explaining what I believe is a better framework for achieving this transition.
Action
When you provide a trigger you are asking for an action. So what action should you ask for?
As a behavior designer, you need to balance user-motivation against task-difficulty. If an action is too hard, the user won’t do it. If the action is too easy, then you’ve wasted the user’s motivational potential. They were ready to do something hard, but you didn’t ask enough from them.
Stanford psychologist BJ Fogg has a model for thinking about balancing task difficulty and user motivation. In the next few weeks I’ll be discussing this topic as well as how to measure user motivation and figure out why a user didn’t preform an action.
Feedback
Users need to know that they preformed the correct action.
When I talk with people about the habits they are designing, feedback is the part that is most often neglected. The point of feedback is to let the user know that they preformed the correct behavior.
If you’re trying to create habits, “clicked”-state “ON” buttons and “throbbers” are bad feedback. They let the user know they completed an action, but it doesn’t let them know if they completed the correct action. A score summary page or level complete page is better.
Notice that feedback is deliberately separated from reward. This is important because habits are best formed when feedback happens every time a correct action is completed, but the reward only is delivered sometimes.
CodeAcademy does a good job of separating feedback from reward. Every time you submit an answer to an exercise, you receive feedback about your performance: the program tells you if you got the question right or wrong. Sometimes, you earn a badge (reward) to recognize your achievement.
Designing great feedback can be a challenge, but my CoFounder Ramsay and I will be discussing it in more detail in the next few posts.
Reward
If some rewards are good, more rewards must be better . . . right?
While at face value this seems intuitive, our brains respond to reward in non-intuitive ways.
This framework raises some open questions: because of how our brains work, it’s not the size of a reward that reinforces a habit; it’s the size of the reward minus the expected size of the reward. How and why can that possibly make sense? Dopamine-producing neurons in our brain automatically subtract-away expected rewards, so only rewards that are unexpected (or are unexpectedly large) reinforce habits. If you deliver a reward every time, the user comes to expect it and the reward ceases to be reinforcing. Rewards are only meaningful if they are unexpected.
Only rewards that are unexpected (or are unexpectedly large) reinforce habits.
- What counts as a reward?
- When exactly should I deliver a reward?
- Am I delivering rewards too often (or too rarely)?
- How big is this reward for this user? For this task?
I’ll also write more about how to deal with this paradox: the user had to know what kind of behaviors might produce a reward, but they still need to be surprised by the reward.
As designers of habits we play a large role in deciding what behaviors our users learn. That comes with an obligation to design behaviors that are in their best interests. But there are also very good cynical reasons to design habits that are in users best interests.
Dr. B.F. Skinner PhD did more then anyone else to uncover and advance the science behind behavior design. He was vilified for this work because his contemporaries believed that he was reducing humans to programmable machines (admittedly, he contributed to this perception). But he actually increased our freewill. Because of what Skinner showed us, we have brought our habits under our conscious control. You can’t live without habits; but thanks to Skinner we can choose what those habits are. Or rather, as developers, you can choose what those habits are for your users. Choose wisely.
