How offering random rewards can shape customer behavior

Alice G. Walton | Oct 05, 2018

One reason some people find slot machines so addictive is that they never know when dropping a quarter will win a jackpot. The uncertainty helps keep some people playing. 

There’s a lesson here for marketers who want people to continually repeat a behavior, such as buying a product or recycling bottles, according to Chinese University of Hong Kong’s Luxi Shen (a recent graduate of Chicago Booth’s PhD Program), Booth’s Christopher K. Hsee, and Carnegie Mellon PhD candidate Joachim H. Talloen. The researchers conducted a series of experiments to study the role of uncertainty in people’s decisions to repeat an activity. Their findings demonstrate that uncertain incentives can be more effective than certain incentives, even if the uncertain incentives are objectively less.

In one experiment, the researchers collaborated with a running club in Hong Kong, which recruited 82 new members to participate in an event that promoted active lifestyles. As part of the event, the club encouraged these new members to complete as many track laps (400 m, or about a quarter of a mile) as they wanted. Some were offered five points at the end of each lap, convertible into cash. Others were offered either three points orfive points, generated randomly, also at the end of each lap. Shen, Hsee, and Talloen find that those facing an uncertain reward completed an average of 14 laps, almost twice as many as the 7.5 laps by the runners who knew they’d get five points each time around.

“People literally ran ‘the extra mile’ (precisely, 1.61 more miles) for the uncertain incentive,” the researchers write.

In follow-up experiments, the researchers examined what makes uncertainty so motivating. Participants in one study carried out a calculation game where they could practice as much as they wanted and, in fact, were rewarded for practicing more. Those who were offered uncertain rewards with immediate resolution chose to practice twice as much as people in two other groups—one where the rewards were certain, and one where the rewards were uncertain but the uncertainty wasn’t immediately resolved. This suggests to the researchers that timely resolution is key to the effectiveness of the element of uncertainty.

And they find another nuance: uncertain incentives reduced the chances that workers chose to start doing something, in this case, complete a survey for money. But once workers finished this initial task, uncertain incentives boosted the likelihood they would repeat the task—and improved the ultimate work performance, that is, the number of surveys completed. Thus, uncertainty may not be useful in getting people to start a task but can be an effective tool to get them to repeat it.

The findings have implications for marketing, management, and other areas where a particular behavior is desired. For example, to reward customers who bring reusable bags, retailers might offer variable discounts on groceries. Variable refunds for returning recyclable bottles to the store might transform a dull errand into a satisfying game. 

“We recommend changing the certain refund to an uncertain amount,” the researchers write, “essentially transforming the bottle-recycling machine into a bottle-recycling ‘slot machine’ that we predict will gamify the otherwise tedious recycling activity and encourage people to recycle more. Then, recycling could be rewarding not only for the monetary return but also for the pleasure of uncertainty resolution.”