Using Behavioral Science to Power Marketing

Behavioral science is an essential area of research that marketers rely on to effectively serve consumers. The work behind behavioral science helps marketers better understand consumer behavior and drive desired outcomes.

Recently, Chicago Booth’s Kilts Center hosted a panel of alumni to discuss the power of behavioral science in marketing. The moderator, Oleg Urminsky, professor of marketing, studies consumer decision-making and its implications for the field.

“Increasingly, when we’re trying to interact with consumers, we’re designing a choice architecture,” Urminsky said, referencing research by Nobel Prize–winner Richard H. Thaler, the Charles R. Walgreen Distinguished Service Professor of Behavioral Science and Economics. “We’re creating an environment in which they’re going to make that decision.”

The panelists shared how they use behavioral science to create choices, win over key stakeholders, and improve consumer motivation.

Finding the Sweet Spot

When working at Match Group, a Dallas-based company that owns multiple online dating platforms, Sushil Sharma, ’11, found that forcing users to make choices can lead to richer results.

Sharma—who now serves as chief growth officer for online mortgage company Better in New York—worked for years as chief product and revenue officer at Match. When users were presented with profiles of people they may be interested in dating, they could answer with yes, no, or maybe.

Sharma and his team observed that about 60 percent of responses fell into the maybe category.

“People did not want to make a decision,” he said. “And that was the bucket that we were trying to solve for, because maybe is not actionable. If we know it’s a no, then we can build on your preferences. If we know it’s a yes, we can find other people you like.”

The team came up with an idea: show two profiles at once and force users to choose one. If they skip both, that means that they’re saying no to both—no maybes.

“All the tests we run normally take weeks to be statistically significant,” Sharma said. “This one we launched at 11 a.m. Within a couple of hours, there was a signal. You could see double the number of yeses.”

The number of connections between users grew so rapidly that some reported feeling overwhelmedby having too many new matches. “But that’s a nice problem to have,” Sharma said.

This experiment highlights an interesting aspect of behavioral science, Urminsky commented, which is that there are often trade-offs. In this case, Match gave users more connections by forcing a decision, but the additional connections also made them feel a bit more worn out.

“There’s often an optimal level,” Urminsky said. “If you take it too far, users may drop out and say, ‘This is too pushy.’ Finding that sweet spot requires a deep understanding of the consumer and some experimentation.”

For this reason, Sharma said that anyone looking to run a similar experiment should start small and scale slowly. Forcing consumers to choose is powerful; taken to extremes, it can act as a deterrent.

The Power of Experiments

Gayatri Narayan, ’08, is working to digitize processes within the Chicago–based food and beverage company PepsiCo, where she serves as senior vice president of digital products and services. In many cases, she’s found that behavioral science is key to convincing stakeholders to accept new innovative products.

In one example, Narayan’s team  worked with PepsiCo retail partners to develop an app that would make stocking their stores easier and more efficient. The app used an algorithm to understand demographic information, pricing behavior, and other data from each store, predicting how much product the retailer should carry.

“We designed the software to inform retailers on exactly what products and quantities they need in stock, but many felt they had more knowledge than the black box algorithm,” Narayan said. As a result, many retailers ignored the app and its advice.

To win them over, Narayan used the common behavioral science concept of reframing. Rather than telling retailers about the power of the app, her team found a way to show them. “Within the region that the salesperson was covering, we would say, ‘Let’s run these experiments at a few stores and you can observe how they perform,’” Narayan said. “We had to peel back and build the experiment infrastructure.”

By creating experiments, Narayan said that they were able to demonstrate to retailers how well the algorithm could predict orders compared to their typical processes. Following these experiments, PepsiCo offered them an option for whether to adopt the new app. Taking these steps, Narayan’s team helped empower the retailers to make decisions as well as be involved in offering feedback for how to improve it along the way.

Urminsky said that a process like this builds a feedback loop between a company’s technology and its real-world employees. By bringing both together, the technology and the people can improve each other.

“It’s easy to fall into the trap of saying, ‘I knew it all along,’” Urminsky said. “Experimentation is one way to break that trap and counter some of these behavioral biases that we all have. It helps us be more calibrated in making our decisions.”

Making Data Easy and Motivational

When trying to inspire action, Carolyn Sleeth, ’05, saw that data had to be presented to consumers in a way that was both easy and motivational.

Sleeth, vice president and general manager of aortic at the medical device company Medtronic in Minneapolis, said that many of the company’s products link into smart devices, such as smartphones, and there is an ability to display a great deal of data for patients and their doctors.

But this mountain of information, taken alone, can overwhelm their customers rather than help them. Without guidance as to when to take an action and what they should do, doctors and patients sometimes ignored the data.

“Our goal is to use that data to help a patient or physician make a decision,” she said.

The Medtronic team saw that they had to present the patients’ data differently in order to make them useful, Sleeth added. They reduced the quantity of information and provided guides on how to use it. These steps helped motivate physicians and patients to take actions that were in their best interests.  

For instance, when patients needed to get their blood pressure down to a specific target, Medtronic used patient data to define targets for the doctors to use in guiding their patients.  

This is a perfect example of how paying attention to their consumers’ behavior was key to Medtronic’s ultimate goal to improve health outcomes.

“That’s a theme we see in research on behavioral science all the time,” Urminsky said. “Simplifying something can have a huge impact.”