McKinsey Quarterly

Using behavioral science to improve the customer experience

| Article

Service operations seem a natural setting for the ideas of behavioral science. Every year, companies have thousands, even millions, of interactions with human beings—also known as customers. Their perceptions of an interaction, behavioral scientists tell us, are influenced powerfully by considerations such as its sequence of painful and pleasurable experiences. Companies care deeply about the quality of those interactions and invest heavily in effective Web sites and in responsive, simplified call centers.

Yet the application of behavioral science to service operations seems spotty at best. Its principles have been implemented by relatively few companies, such as the telecommunications business, which found that giving customers some control over their service interactions by allowing them to schedule field service visits at specific times could make them more satisfied, even when they had to wait a week or longer. Many more companies ignore what makes people tick. Banks, for example, often disturb the customer experience by altering the menus on ATMs or the interactive-voice-response (IVR) systems in call centers. They fail to recognize the psychological discomfort customers experience when faced with unexpected changes.

Likewise, for every restaurant that surrounds a bill’s arrival with a succession of complementary desserts—thereby capitalizing on the customer’s preference for service encounters that end positively—there are a lot of call centers that ignore the importance of a strong finish. Indeed, many companies actively work against one by placing so much emphasis on average handling times that they inadvertently encourage agents to end a call once its main business is complete, leaving customers with memories of brusque treatment.

It doesn’t have to be this way. Academics such as Professor Richard Chase at the University of Southern California’s Marshall School of Business have used research on how people form opinions about their experiences to design actual services. In a 2001 Harvard Business Review article,1 Chase and his team even laid out principles for managers to consider when designing any customer interaction. Get bad experiences over early, so that customers focus on the more positive subsequent elements of the interaction. Break up pleasure but combine pain for your customers, so that the pleasant parts of the interaction form a stronger part of their recollections. Finish strong, as the final elements of the interaction will stick in the customers’ memory. Give them choice, so they feel more in control of the interaction. And let them stick to their habits rather than force them to endure the discomfort and disorientation of unexpected change.

Here we review the experience of an insurance company that used those principles to improve its customers’ satisfaction significantly, with no incremental costs or fundamental changes in people or infrastructure. A systematic approach like this one is needed to counteract the natural tendency of service operations to focus on the needs of IT systems and work flows, not to mention the preferences of employees, managers, and service providers, largely ignoring the way customers perceive their service interactions. If companies in a broad range of service industries—including banking, telecommunications, and retailing—applied a rigorous approach, they would reap significant economic benefits, ranging from reduced churn to greater cross-selling to additional customer referrals.

Setting the stage

Executives at a leading North American health insurer sought to help patients manage their treatment programs for serious long-term illnesses, such as diabetes or congestive heart failure. Conditions like these are difficult to manage because treatment is often protracted and outcomes can depend on the patients’ willingness to make significant lifestyle changes.

Patients participating in an experimental health-management program received regular, scheduled calls from a team of nurses over a period of several months. The calls aimed to deliver additional support to patients undergoing long-term treatment, by helping them understand the available options and stick to their treatment regimes, as well as reinforcing lifestyle changes recommended by their doctors. Improved compliance helps insurers too, as better outcomes reduce the overall cost of treatment.

In the past, the clinical-treatment program for each patient had determined the content of such calls, and the company used what it considered to be a tried-and-true method for managing them. Team members had received guidelines on the objectives of the calls and used a checklist to sequence discussions with customers.

Behavioral science in action

To see if this approach could be improved, the company divided the nurses into two groups—approximately 20 in a pilot group and another 20 in a control one—and began applying a behavioral-science lens to the interactions of the former to test different versions of the call structure. Postcall surveys measured the customers’ satisfaction with each call and with the company. Key customer and operational metrics (including sign-up rates) helped estimate the financial impact. The pilot team used behavioral-science principles throughout the interactions.

1. Get bad experiences over with early

The team identified difficult issues—for example, the forthcoming lapse of certain insurance benefits or the need to transfer from one facility to another—and moved them to the start of the call. It also set up a later phase built around constructive coaching from the nurses on how to deal with the issues raised earlier. In addition, general questions that were likely to make patients uncomfortable (about current pain levels, smoking habits, eating patterns, and alcohol consumption, for instance) were moved from the end of the call to the beginning.

2. Break up pleasure and combine pain

By combining the most challenging elements of a call in its first phase, the health-management team could focus on positive aspects during the rest of it. The team found that patients responded very positively to coaching by nurses, so there was an effort to ensure that coaching on multiple topics was an explicit part of every phase of the call. A nurse might, for example, discuss the next treatment steps, how the patient could take advantage of all covered benefits, and ways of minimizing out-of-pocket expenses. There was also an effort to resolve all possible issues within a call and to transfer it to other groups only as a last resort.

3. Finish strongly

The conclusion of the health-management calls was scripted to finish on a positive note by emphasizing the tangible insurance benefits available to patients and, where medically appropriate, the likelihood of a successful outcome to the agreed-upon action plan. At the end of a program lasting several months, with calls taking place every month or so, patients received a final call from their health-management nurse. This call ended by celebrating their progress, reviewing the goals they had met, and summarizing the positive steps they had taken to achieve those goals.

4. Give customers choice

The company made an effort to give customers explicit choice on three critical elements: the type of treatment plan, which facilities to visit and which doctors to see, and the timing of future calls. In each area, the nurse was guided to tell the customer, “You have a choice; let me give you some options.” Customers explicitly had the right to make the ultimate decision, though the outcome may have been limited or strongly suggested—for example, “Hospital A is closest to your home, but B is only 15 minutes further away, and it has a specialist unit with a great track record at treating your condition.”

5. Let customers stick to their habits

In many situations, it was important for patients to change their lifestyles—say, by eating different foods, consuming less alcohol, or exercising. To encourage patients to make these changes while minimizing the discomfort they generated, nurses introduced them gradually over a series of calls. Dietary changes might be discussed initially, for instance, followed by encouragement to begin exercise. The nurses also tried to reframe the patients’ perceptions of the severity of the changes by comparing them with more unfavorable alternatives: for example, “instead of eliminating your favorite foods altogether, why not just try picking low-fat varieties next time you are in the store.”

The team also worked to ensure that the calls themselves became a positive habit for the patients. This approach gave them the option of having the same nurse on follow-up and promoted a consistent approach for every call, so that they became used to the interactions.


The effect of the changes was significant. Patients in the test group reported an average satisfaction level seven percentage points higher than that of patients in the control group—for calls with the same basic content. These patients’ satisfaction levels with the company was on average eight percentage points higher than that of the control group. More important, patients in the test group were on average five percentage points more likely to say that the calls had motivated them to make positive changes in their behavior.

Notably, the program didn’t significantly affect the company’s costs or change key operational metrics, such as the length of a call or the number of calls a day. Moreover, test group nurses reported an average level of job satisfaction higher than that of the control group nurses. Finally, the impact was rapid. Most of the increase in the satisfaction levels of the test group patients happened within two weeks.

Many other service industries could benefit from a similar approach. By breaking down frontline transactions and rebuilding them with behavioral and experiential principles, companies could systematically achieve rapid, measurable improvements in customer satisfaction.

Explore a career with us