Compulsive Behavior Sells
A middle-aged woman sits before a computer screen on the 11th floor of Expedia’s glass-clad headquarters in Seattle. Two electrodes are taped to her brow just above her left eye, two more on her left cheek. A one-way mirror reflects her face as she responds to requests issuing from a speaker mounted in the ceiling.
Behind the glass, a researcher directs the test subject as a half-dozen designers, engineers, and executives look on in rapt silence. “Okay, Shannon,” the researcher says. “Go to Expedia and start shopping for your trip to Hawaii.” The audience gazes intently at a large video display. A running graph of the electrodes’ output trails across the screen. The electrodes on the brow measure contraction of the muscles that activate frowning—a sign, according to the theory of facial electromyography, of concentration, tension, or irritation. Those on the cheek track the play of muscles involved in smiling, evidence of the warm glow of delight that occurs when the brain’s reward circuitry is activated.
Though she has not been told, Shannon has been brought in to test a new Expedia feature, known as Scratchpad, that the online travel broker hopes will bring travelers back to the site daily between the time they start planning a trip and the day they make a purchase. Scratchpad automatically records the hotels and flights a customer has viewed, allowing users to pick up on previous searches without having to re-create them.
The first part of Shannon’s test is simply finding the Scratchpad button on the Expedia.com page. She hasn’t found it yet, but when she looks at photos of the Westin Maui, the smile sensors spike: a jolt of joy—and potential paydirt for Expedia. The company aims to make the experience of shopping so pleasurable that using the site becomes a habit.
Forging new habits has become an obsession among technology companies. In an age when commercial competition is only a click away, the new mandate is to make products and services that generate compulsive behavior: in essence, to get users hooked on a squirt of dopamine to the brain’s reward center to ensure that they’ll come back.
The rise of mobile computing has intensified that imperative. The small screen crowds out alternatives, focusing a person’s attention on a limited number of go-to apps. The ones that get used are the ones people click on impulse while they’re drinking their morning coffee, waiting for the bus, or standing in the checkout line.
For a long time, the methodology for designing habit-forming products was haphazard: build it, put it before the public, and watch it go viral or fade into oblivion. In recent years, though, product teams have become more deliberate. Principles derived from behavioral science play an increasing role in software design, creating a demand for experts who can guide developers in the art—and science—of behavior engineering.
Among the most influential is Nir Eyal, an entrepreneur turned user-experience guru who has become Silicon Valley’s most visible advocate of habit-forming technology. His blog, Nir and Far, has attracted more than 25,000 subscribers hungry for insights into shaping user behavior, and his writing has appeared in both the mass-market pages of Psychology Today and the insider club of TechCrunch.
He has worked for some of the biggest names in technology (most of whom don’t want to talk about it) and has presented workshops from Norway to Thailand. His inaugural Habit Summit, held last March on the Stanford campus, drew hundreds of participants from startups and blue-chip firms alike. Eyal promotes a scheme he calls the hook, a simple set of steps derived from his observation of numerous online products and services and undergirded by a wide range of psychological and neurological research. The hook, he says, is the magic behind Facebook, Google, Instagram, Pinterest, Snapchat, Twitter, and just about every other icon of the consumer Internet. It leads users into a repetitive cycle that transforms tentative actions into irresistible urges.
John Kim, Expedia’s chief product officer, brought Eyal in last year to help the company develop compulsive experiences, and now Shannon, the volunteer in front of the computer screen, is testing the fruit of their labor.
After several minutes, she still hasn’t discovered the Scratchpad feature. Finally the researcher guides her to its unobtrusive button in the menu bar at the top of the Expedia homepage. Low-level tension registers on the graph at the front of the room—concentration? frustration?—but then she recognizes a picture of the Westin Maui, which Scratchpad captured automatically. A shimmer of delight ripples across the graph. “I like the fact that it saves me time,” she comments. “I’d go back and use this again for sure!”
Eyal’s workshops offer a four-hour immersion in the mechanics of the hook. On a warm spring day several weeks before the Expedia research exercise, he’s getting ready to run a session at the office of Zurb, an airy design studio not far from Apple’s headquarters in Cupertino, California. Clad in a plaid shirt that hangs untucked over jeans, he’s coat-hanger thin with eyes that glitter beneath a clean-shaven crown.
He opens with a disclaimer. “I’m not an advocate for creating addiction,” he says. “Addiction has a specific definition: it always hurts the user. I talk about the pathways for addiction because the same things that occur in the brain help us do something that can be good.”
Thus he initiates 67 attendees from companies including Hewlett-Packard, the New York Times, and Samsung into the mysteries of the hook.
It starts with a trigger, a prod that propels users into a four-step loop. Think of the e-mail notification you get when a friend tags you in a photo on Facebook. The trigger prompts you to take an action—say, to log in to Facebook. That leads to a reward: viewing the photo and reading the comments left by others. In the fourth step, you inject a personal stake by making an investment: say, leaving your own comment in the thread. This pattern, Eyal says, kicks off a cycle that lodges behaviors in the basal ganglia, the part of the brain where automatic behaviors are stored and where, according to neuroscientists, they last a lifetime.
The psychology behind the hook dates back at least to the 1930s, when the American psychologist B. F. Skinner showed that he could induce desired behaviors in animals. Skinner is famous for training pigeons to do seemingly intelligent things, like reading signs and following instructions by manipulating the equivalent of Eyal’s trigger-action-reward sequence.
Other researchers have refined Skinner’s theories in the decades since. On the screen behind him, Eyal flashes a slide drawn from the seminal work of Stanford behavior theorist B. J. Fogg. It’s an x-y plane with axes labeled “motivation” and “ability,” a curve tracing a diagonal smile from upper left to lower right. According to Fogg, a behavior happens when a trigger coincides with both motivation and ability—but only in the right proportion. If a trigger consistently fails to initiate the desired action, the theory goes, habit designers should aim to enhance the user’s ability. Motivation is hard to influence, because you can’t make people do what they don’t want to do. Ability is more malleable: simply make the behavior easier to execute.
Still, the reward must promise enough pleasure to drive people to take the intended action. In training animals to execute complex behaviors, Skinner discovered that varying the payoff—from highly desirable to nothing at all—both increases a behavior’s frequency and helps keep it from fading once the rewards stop.
A classic example is slot-machine gambling. The player never knows whether the next pull might bring a $5 win or a $50,000 jackpot. The unpredictability of the reward—and the randomness of its arrival—is a powerful motivator to pull the lever again and again.
Eyal draws a parallel between Skinner’s variable rewards and the endless variety to be found on, say, Pinterest: the user can scroll endlessly, scanning for distinctive items amid a sea of banality. “This,” he says, swiping his finger downward as though scrolling a touch screen, “becomes this”—he moves his arm up and down as though he were cranking a slot-machine lever.
The hook’s final stage, investment, closes the loop by “loading the next trigger,” Eyal says, an idea inspired in part by work on game psychology by Jesse Schell, a Disney Imagineer turned Carnegie Mellon professor. Take Twitter. When you make an investment by posting a tweet, a follower’s reply to your contribution triggers an e-mail notification to your in-box, inciting you to take yet another spin through the cycle.
The workshop hums with activity as the students form small groups to work on their own projects. One of the participants is an Expedia executive named Pooja Vithlani, who is part of the team developing Scratchpad. (Her title is senior product manager of compulsion.) John Kim sent her to learn more about how to apply the hook.
She has a clear sense of the Scratchpad user’s action (shopping for airline tickets), reward (a handy list of possible travel arrangements), and investment (curating the list by eliminating options that prove obsolete or impractical). However, the trigger proves elusive. Vithlani muses on the anxiety occasioned by oncoming holidays and the attendant travel plans. Can she nudge potential customers to check Scratchpad at the very moment they’re feeling pressure to lock in airfares? Relief from the stress of holiday travel might be the bait on Scratchpad’s hook.
Eyal readily admits that many of the ideas he promotes aren’t his own. “I don’t do original research and don’t intend to,” he says over his habitual lunch, a build-it-yourself burger at an eatery near his home in Palo Alto, California. “There’s more research than we know what to do with.”
To his clients, he offers a simple, practical scheme that keeps the arduous process of product design focused squarely on a user’s impulses, desires, and motivations. And there’s a chance he’ll come cheap. He charges some clients a day rate determined by rolling a pair of dice and multiplying the result by 100. “It’s a variable reward,” he says with a sly grin.
Eyal developed his interest in habits early. Born in Israel, he moved with his family to Florida at age three. Eating provided the surest relief from the alienation of being a foreigner with a funny name. “My parents took me to a fat camp when I was 12,” he says. He racked up Cs and Ds despite being pegged as a gifted student.
In adolescence he began to shed the pounds and focus on his studies after reading The T-Factor Diet, which emphasized methodology over willpower. The experience showed him how much unconscious impulses influenced his own life, and the power to be gained by working with rather than against them.
Eyal graduated from Emory University with a degree in journalism and eventually landed at Stanford’s Graduate School of Business. After Mark Zuckerberg spoke to the class, “overnight everyone was making an app,” he says.
He realized that all those apps would need a way to generate revenue. In 2008, he persuaded the prominent venture capitalists Kleiner Perkins and Mike Maples to fund a company, AdNectar, that would broker the coming wave of socially driven brand messaging. He sold the business three years later to social-commerce website Lokerz for a sum that neither party will disclose.
The sale set Eyal adrift. The time, effort, and anxiety of running a startup had weakened his closest relationships and added 15 pounds to his lanky frame. Slowly it dawned on him that as mobile devices got smaller, screen real estate shrank and habits became more important. He realized that habits could be at the heart of his next business—and his own rejuvenation.
Throwing himself into studying consumer psychology, he devoured research on how products influence behavior. The successful apps he had encountered through AdNectar, he noticed, had in common a cyclical feedback loop of user behavior.
Eyal’s fascination with the mechanics of habit-forming technology coincided with a dramatic rise in the Internet’s potential to influence behavior, enabling software developers to manipulate many of the behavioral dynamics Skinner and other researchers had identified. Smartphones became a ubiquitous channel for delivering triggers, while apps reduced complex actions to the simple press of a button. The social Web delivered a panoply of interpersonal rewards. Game designers began to speak of forming a “compulsion loop.” Entrepreneurs went from tracking monthly active users to a new measure called “compulsion rate,” the percentage of users who returned from day to day.
Riding a cresting wave, Eyal developed his ideas on his blog. Eventually he decided to collect his writing into a coherent form. “I set out to write a 30-page document,” he says, “and I ended up with a 256-page book.”
As the manuscript neared completion, Eyal employed the hook to galvanize his audience. He sent an e-mail (the trigger) to his blog subscribers inviting them to read the book in progress and critique his work. Nine hundred people responded. They received the reward of reading the advance copy, then made the investment of adding their comments on Google Docs. Eyal promised to include their names in the final product, loading the trigger to buy a printed copy and post a review on Amazon.
The scheme worked. Eyal released the self-published edition of Hooked in early January 2014. Within a week and a half, he had racked up 125 reviews, and the book was lodged at the top of Amazon’s ranking of product design books.
It’s the afternoon following Shannon’s session, and the Expedia team has gathered in a conference room to decide on the next steps. The test subject understood intuitively what actions to take with Scratchpad. The facial sensors indicated that she experienced an appropriate psychic reward. She made her investment in a provisional form; Scratchpad added hotels she looked at to the pinboard automatically. (The ability to curate the list is on the drawing board for a future update.)
But the trigger remains a glaring issue. Shannon had to be told to click the Scratchpad button. Without a cue, she didn’t even know it was there.
Data has shown that using Scratchpad doubles Expedia.com’s compulsion rate—in other words, visitors who register for Scratchpad and use it are far more likely to return to the site within 24 hours than those who don’t. But only a small percentage of visitors actually use it. More effective triggering could go some distance toward fixing that.
Vithlani has an idea that she thinks might do the trick. She calls it continuous shopping. Travel shoppers are often paralyzed because prices shift constantly, she notes. Expedia loses sales because people forget what they’ve found in past searches. “Continuous shopping will give them perfect memory and fresh prices,” she says.
The Scratchpad window would open automatically when a visitor arrived at Expedia.com and offer to track price changes in return for registering. Then prospective customers would receive a daily e-mail telling them whether prices had risen or fallen; they would need to click through to Scratchpad to see the details. “This could develop a compulsion loop because we’re getting people used to the fact that we remember what they looked at,” she explains.
It would take only a few weeks to implement and test. (A year later, variations of Vithlani’s idea are live on the website, and Expedia reports that using Scratchpad now triples the compulsion rate and doubles the repeat-visit rate.)
Eyal himself is not immune to the siren call of behavior engineers. In an article he wrote for Forbes entitled “Strange Sex Habits of Silicon Valley,” he candidly laments the impact of mobile devices on intimacy between himself and his wife. Not long ago, upon slipping into bed at night, he often found himself reaching for his tablet rather than his spouse. Drawing on B. J. Fogg’s behavior model, he broke the spell by pushing his late-night browsing down the ability scale. He installed a timer that turned off his Wi-Fi router at bedtime, forcing him to switch it back on before he could satisfy his craving for a late-night online fix.
The gambit improved his sex life, but the larger issue remained, he wrote: “The confluence of increased access and greater sharing of personal information, and at higher transmission speeds, has created the perfect storm of addictive technology.”
Eyal’s worry isn’t idle. Slot-machine designers are renowned for inducing behaviors that resemble addiction, and in 2011 the American Society of Addiction Medicine began defining addiction in terms of behaviors that activate the brain’s reward circuitry, whether or not substance abuse is involved.
“The ethics of this have still to be worked out,” says Chris Nodder, author of the archly titled user-experience manual Evil by Design.
Hooked concludes with a chapter on ethics that directs behavior engineers to focus on applications that improve users’ lives and that the engineers themselves find helpful. On the whole, though, Eyal views behavior engineering as a grand opportunity. “Wouldn’t you like to want to exercise without thinking about it?” he asks. “Or save money every day by being more frugal? That’s what this technology makes possible.”
In any case, neuroscience suggests that eliminating the potential for addiction would require eliminating pleasure itself. “In the brain, our pleasure center and memory center are in close proximity, as though nature wanted us to reproduce and remember how,” explains Howard Shaffer, a Harvard psychiatrist.
A future of smart watches and biometrics may make engineering new habits even easier. “Now the interface disappears, which provides all kinds of new triggering opportunities,” says Eyal. “I think we’ll see a golden era—I hope—of habit formation and interesting ways to help people live better lives.”
The inside story of how ChatGPT was built from the people who made it
Exclusive conversations that take us behind the scenes of a cultural phenomenon.
How Rust went from a side project to the world’s most-loved programming language
For decades, coders wrote critical systems in C and C++. Now they turn to Rust.
ChatGPT is about to revolutionize the economy. We need to decide what that looks like.
New large language models will transform many jobs. Whether they will lead to widespread prosperity or not is up to us.
Design thinking was supposed to fix the world. Where did it go wrong?
An approach that promised to democratize design may have done the opposite.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.