Throttling the Customer

In January 2005, responding to a class action suit, Netflix Inc., the world’s leading delivery-based DVD-rental company, publicly acknowledged a policy it had practiced since it launched in 1999. The company made it clear that it did not always process new rentals in the order in which they were received, but instead would give priority to the least active customers, thereby occasionally slowing service to the most regular customers. The policy, dubbed by the media as “throttling,” was widely criticized by both online and mainstream media. Nevertheless, Netflix has continued the policy and managed to increase its customer base and profits by maintaining close relationships with its customers and extensive research into their needs. According to P.K. Kannan, the Harvey Sanders Associate Professor of Marketing at the University of Maryland, Netflix is simply at the forefront of something that will become more common in the retail sector — firing the customer.

Long a strategy in business-to-business settings and the insurance sector, firing the customer consists of identifying and purging your portfolio of the least profitable customers in order to increase margins or concentrate on your best clients. In the case of retail, most companies don’t want to actually fire a customer so much as change the customer’s behavior. At Netflix, all customers pay a flat fee for “unlimited” DVD rentals, with the only limiting factor being the number of DVDs a customer can have out at one time, which is based on the membership level the customer chooses. Because revenue per customer is fixed but expenses are variable due to mailing costs, the least profitable customers are the ones who rent the most frequently. This puts Netflix in the unusual situation that its most active customers are not its best customers. Steve Swasey, Netflix director of corporate communications, says the company’s allocation policy was not intended to fire the customer. “The only customer we want to fire,” he said, “is the one that we can never satisfy.” Regardless of what they are called, it is clear that such policies will become more prevalent and advantageous as more companies gain the kind of customer knowledge and intimacy that Netflix has garnered. In addition to its traditional membership information , Netflix has access to its customers’ future needs (in the form of queues for future rental), tastes (movie rankings) and peers (using the “Netflix Friends” system, which allows friends to share queues and reviews), and it has created a tightly knit community using a combination of e-mails, “favorites lists” and member comments.

As other retailers capture more information at the point of sale and gain access to better research tools, they are more capable of weeding out customers who are hurting their margins. For instance, major electronics retailers such as Best Buy, Wal-Mart and Circuit City have recently strengthened their return policies by instituting or increasing re-stocking fees. In the case of some products, particularly software, they have stopped accepting returns altogether. The point of such policies, says Kannan, is “apportioning the real costs of providing service/product to individual customers. Before this, everyone generally knew who the big revenue generators were, but getting costs into the picture gives retailers a better idea of who the big profit customers are.”

Rohit Verma, associate professor of management in the University of Utah’s David Eccles School of Business, thinks these kinds of policies will be increasingly adopted in industries in which “the frequent use of a service is not related to generating additional revenue, but [requires] additional cost and efforts for the firm to support the activity.” That includes many e-commerce businesses, the mobile phone industry, and the financial services industry, where charging fees for expensive behavior such as using an ATM out of network is already common.

There can be some risks to employing such policies, however. A company that chooses to implement a policy aimed at eliminating or altering the behavior of a customer faces two problems. The first is negative customer reaction, even from those who are not affected by the policy. Despite the relatively small number of customers affected by the Netflix policy, Web sites and blogs quickly began posting highly negative reviews of the service. What’s more, overly complex policies may lead to customers’ feeling as though they’ve been affected when they have not been. For example, customers might confuse slow delivery on the part of the postal service, over which an e-tailer has no control, with a deliberate attempt by the company to throttle.

Another problem is obviously that fired or alienated customers become potential market share for competitors. Kannan is quick to point out the dangers of trying to build market share by attracting your competitors’ least profitable customers, but there are businesses, such as online DVD rentals, where the strategy could pay off as a way to build brand awareness and reputation that will eventually attract more profitable customers. A series of new DVD rental competitors have opened in the last year, and several, such as QwikFliks Entertainment Rentals Inc., have prominently displayed promises of “no throttling” on their Web sites. Another, MovieKlub LLC, is relying on new, “disposable” DVDs as a potential advantage over sites that need their DVDs returned. Because the DVDs destroy themselves, they are quick to point out, there is no need to throttle as the DVDs don’t need to be returned and resent to other customers.

Despite the risks, we are likely to see a growing number of retailers attempt to fire their customers or at least alter their behavior. The Netflix story offers several lessons. Any company embarking on such a strategy should be sure of the costs and benefits of both the behavior it is attempting to discourage and the behavior it is attempting to encourage. It should also attempt to offset the damage of negative publicity by putting such policies in place before it begins selling its product or service. When it became clear, for example, that Netflix could not truly sustain with its most active customers the unlimited rentals it had promised in promotions, the changing policy was perceived as the company going back on its word to customers. If a policy has to be changed in midstream, however, the new policy and the reason for the change should be explained clearly and fully to customers in order to minimize negative reactions. When the class action suit (which remains unsettled) was filed against Netflix, for example, the company immediately began e-mailing customers to poll them on their service and explain the policy. “We began tracking from day one,” says Swasey, “and we knew we didn’t have a problem.” Finally, such policies should not be confusing or overly broad, lest companies find themselves altering customer behavior in unintended ways.