Behavioral Design in the Wild

7 min read

Boundless Mind Digital Behavioral Design
The following topic also appears in our Free Ebook: Digital Behavioral Design


“It’s good to learn from your mistakes. It’s better to learn from other peoples mistakes.” – Warren Buffett


Behavioral Design is all around us: from the stores we shop in to the apps we use on our phone. Our interactions, experiences, and choices are designed to increase the likelihood we’ll behave in certain, predictable ways. In this blog post, we spotlight some common behavioral design techniques that, for better or for worse, modify user behavior and keep us coming back.

Reinforcement Learning and Cues: Facebook

Facebook masterfully employed Behavioral Design techniques to build one of the most habit-forming products on anyone’s phone. While they’ve used several different techniques, we focus here on two:

  1. Reinforcement Learning
  2. Cues

“PING” you have a new notification!

This alert cues a chain of events: first, you unlock your phone, then, find Facebook, and finally, open it. As described by the CAR Model, Facebook presented with a Synthetic Cue (a cue controlled by Facebook), and you performed the desired target action.  When you open the app, notifications are waiting for you, and maybe one of them is delightful! Here, Facebook employed Reinforcement Learning by giving you variable delight for opening their app.

Reinforcement Learning leverages rewards to reinforce user behavior. Nir Eyal models of the three types of rewards:

  1. Tribe – the desire to be accepted and embraced by our peers
  2. Hunt – the desire to seek and find useful information or resources
  3. Self – the desire for personal gratification or success

Largely, Facebook’s notification system relies on the reinforcing power of rewards of the Self and the Tribe to induce delight.

Over the years, you may have noticed that the types of notifications you received have changed. They’ve broadened notification content to include more than just the direct, rewarding interactions. They now also include random, superfluous notifications. Whether a deliberate design decision or the consequence of another goal, it created the feedback, or variability, necessary in the CAR Model to counteract the diminishing effect of constant reinforcement.

Why does variability matter?

If there was a reward waiting for you every single time you opened Facebook, the effect of reward would diminish, and wouldn’t compel you to open Facebook as often. Today, you might be notified often of things you don’t care about. This serves to make the reward of a notification more variable. By increasing the variability, they’ve also increased the probability that you’ll form a habit out of opening Facebook.

Reinforcement Learning: Pinterest

Social media apps are great at using Behavioral Design techniques because their business models often rely on maximizing the time you spend on the platform. With this goal in mind, these teams successfully combined multiple techniques to form digital products we will spend hours in.  Pinterest was one of the first digital leaders to use Behavioral Sciences and Behavioral Design as driving product philosophies.

How do they use it?

When you use Pinterest, your primary action is scrolling. As you use the app and scroll through your feed, you don’t know if the next row of content you encounter will be something you like. Much the opposite, most of the content you see won’t be compelling to you. This unknown creates a form of variable reinforcement since you can’t predict if you will like the next piece of content. The effect is that your scrolling behavior is highly reinforced. This Behavioral Design technique uses the reward of the Hunt to keep us seeking new and useful information.

Stopping Rules: Instagram

How many hours have you, or someone you know, spent scrolling Instagram? Next question: how many hours did they intend to spend scrolling?  We’re willing to bet there’s a pretty big difference there. Instagram was intentionally engineered to fetch more content as you scroll such that you never need to stop scrolling. There is no “Next Page” button, no “Continue” button, etc. there is only scrolling.

Why was it built like this?

From a UX engineering perspective, infinite scroll is a highly desirable feature. It simplifies the user experience and removes a unnecessary labor and choice. In this design pattern, users don’t take any extra action to continue their current flow of scrolling and viewing. Viewed from this perspective, we could consider it a design decision that helps the user.

However, when we look at it from the Behavioral Design perspective, we uncover a different story. Infinite scroll is desirable not because it helps the user, but because it leads the users to consume the feed past their satiety point. With no end to the feed and no change in action required to continue, the user never pauses to check in nor encounters no good stopping point to exit the feed.  From a revenue perspective, this is obviously a productive design decision. However, from an ethics and user respect perspective, it is questionably ethical.

Real World Experiment: Infinite Soup & Infinite Scroll

When we perform a behavior, our brains rely on natural signals from the environment to know when the behavior is completed. These signals tell our brain to perform the next behavior we’d planned on doing and continue on with our lives. In real life, we naturally receive these start and stop signals all around: we begin a walk and end at our destination, we call a meeting and end when the topic is address, or we make ourselves a drink and finish when the glass is empty.

What happens if you never get that STOP signal?

In 2005, a team at Cornell University set to find out…

Brian Wansick, James Painter, and Jill North explored what would happen if the brain never received a signal to stop eating. They cleverly rigged-up bowls of soup that would automatically and imperceptibly refill unbeknownst to the experimental subjects. Groups of 4 sat down.  Some groups received normal bowls filled intermittently by a ladle, and the other groups received the self-refilling bowls. In the experiment they invited the subjects to eat as much soup as they wanted. The results: on average, the self-refilling bowls caused people to eat 73% more than those with normal bowls. Without the stopping cue of the empty bow, the subjects just kept eating and eating and eating.

You’d expect the subjects who ate 73% more to report feeling fuller or more satisfied, right? No such luck. They reported no difference. These subjects overate without noticing and were no more satisfied. Their brains were deprived of what Wansick and his team called “Visual Cues.” In this case, without seeing the soup in their bowl empty, there was no indication to brain that the food is gone meaning it’s time to stop. This is exactly what you experience in the Behavioral Design pattern of Stopping Rules inside Instagram. Like a bottomless soup bowl, an endless feed compels us to consume far past “full”, all with no detectable change in satisfaction.

Choice Architecture: Your Local Supermarket

Picture your local supermarket. Where do you typically find staples: eggs, milk, veggies, meat, cheese? It’s no consequence that the staple items are stocked in the periphery of your local supermarket rather than together in a central, “essentials” section. This distribution of essential items throughout the store, is a Behavioral Design decision that increases the chances that you’ll purchase more, unrelated items.

Choice Architecture is the technique used here by those who design the interiors of supermarkets to create default habits our of how we navigate a supermarket.  By placing perishable staples on the far walls of the market, the customer walks through aisles stocked with the nonperishable food they might not be interested in buying, but with repeated exposure will increase the chance they purchase.

There’s no logistical reason that the milk couldn’t be placed in a refrigerated aisle in the center of the store (as is the beer, or the frozen food.) Rather, placing it afar ensures we’ll pass food we didn’t otherwise consider buying, and be forced to make the decision to not select it.

Where do you see Choice Architecture in the digital space?

Think back to any app or subscription signup. By default, you will be opted into marketing emails, promotions, recurring charges, etc. and the app will require you to “opt-out” or unselect these. In the same way supermarkets force us to walk by unhealthy food and actively decide not to purchase it, these apps auto-subscribe us to things and require us to make the decision to deselect them.

Stimulus Devaluation: Space

Habits can be created, but they can also be destroyed. One effective way to break a habit loop is by introducing a time delay between an action and its associated Variable Reinforcement to mediate how effective the reinforcement is.

At Boundless Mind we explored this concept, called Stimulus Devaluation, with our app, Space. With Space, we want to help people regain control of their relationships with apps by reducing the impact of the CAR Model.  Stimulus Devaluation operates by increasing the time delay between an action and the reinforcement that a user receives for that interaction.

How does this delay work to break habits?

It’s critical that users receive feedback and reinforcement as-fast-as-possible after completing an action. There is a short window of time after an action is completed when your brain learns from the consequences of that action if it’s worth repeating. By delaying the reward, you provide the user with the feedback that they’ve completed the correct action while decreasing the chance that the reward will effectively reprogram their behavior.

In Space, we’ve introduced a time delay before an app loads granting users the freedom to remove the habit forming effects of variable rewards, while still enjoying the apps they love. Our delay screen encourages users to slow down their breathe to sync with a ring that slowly expands and contracts. This “breathing room” not only disables the reward circuit, but also provides users a moment of stillness to reflect on whether they actually want to continue.  With Space, we adjust the delay length with respect to usage, giving app users a powerful way to manage behavior, introducing more friction when they are binging and less when the usage is low.

Ambient Communication: Apple Watch

Good design means creative problem solving given constraints. So how do you communicate several quantitative reports, on a small display, to a distracted audience? You do exactly what the Apple Watch’s design team did in their activity counter. You use Ambient Communication.

The Apple Watch’s activity app communicates information visually via a series of concentric, colored rings. By color-coding the rings, they communicate several different data streams simultaneously. By making the rings non-overlapping, the app improves clarity about which dataset corresponds to which behavior. Most importantly, the partially completed rings effectively communicate the progress towards goal in a way the user can understand at-a-glance.

Sunk Cost: LinkedIn

The more we’ve given to a person or process, the more likely we are to overestimate how valuable it is to us. Sunk Cost encourages continued investment of time, money, or information so that we become less likely to leave, leaving all that effort to waste.

In Linkedin’s onboarding experience, users are prompted to give information to setup their profile. Yet even when profiles are set up, Linkedin prompts users to add more information and reach “All Star” status. Dangling this carrot, reward of Self, is an effective way to induce them to continue to invest in their profile. From Sunk Cost we know, the more information they’ve invested, the less likely they are to quit.


Behavioral Design is an effective way to modify user behavior. The most successful products are those that can figure out how to use it to drive adoption, engagement and retention.  As a product leader, it’s important to use these tools ethically to the benefit of the user and the business. A quick check is to ask yourself: would the user agree they want to do this action more often?  If you’d like to read more about the ethical framework we abide by at Boundless Mind, download our ebook.  If you are interested in using Behavioral Design to increase user engagement in our product drop us a line.

Matt Mayberry