Combating Churn-CRM in Subscription-Based Businesses

The telecommunications industry faces a particularly daunting challenge in its efforts to prevent customers from defecting to the competition. With new providers, services, rates and incentives seemingly multiplying faster than rabbits, carriers are engaged in a perennial struggle against the revolving door and the financial complications caused by this subscriber "churn." New tools are now combating the problem with sophisticated technologies for predicting customer behavior. These tools are capable of reducing chum by as much as 40 percent, holding great promise for subscription-based businesses and e-businesses of all varieties.

Customer relationship management is important for every business, but in the telecom world it is truly a matter of survival. In the wireless sector of the industry, for example, as many as 7 competing carriers operate in each domestic market, and 2 to 3 percent of each provider's customer base churns every month. With an average cost of $400 to acquire a subscriber, chum cost the industry nearly $6.3 billion in 1998, or $9.6 billion including lost subscription revenue. A 1 percent reduction in subscriber defections is believed to translate into a 5 percent increase in the bottom line: By one research team's calculation, that 1 percent drop in chum can mean a $54 million increase in annual earnings for a carrier with 1.5 million subscribers.

Internet service providers grapple with a comparable set of scenarios. The typical ISP loses 4 to 8 percent of its subscriber base each month, or up to 96 percent every year. Many ISPs therefore need to grow by nearly 100 percent annually to gain any market share at all. Since industry analysts estimate that it costs 4 times as much to acquire a new customer as it does to retain an existing one, ISPs are typically in the position of having to spend enormous sums of money just to stay even.

Traditionally, telecom companies and other chum-plagued industries have responded to these business realities by luring new customers with deals. Unfortunately, this has resulted in squeezed profit margins, a perpetual cycle of competitive discounts, and more of the chum that companies were trying to combat in the first place. In response, the industry is turning to CRM.

Unfortunately, conventional CRM database analysis techniques that group customers into clusters or segments do not adequately address the needs of telecom companies and other subscription-based enterprises. Traditional CRM applications produce detailed profiles on groups of atrisk accounts, but they fail to provide the answers to two critical questions: why these customers are likely to leave, and what can be done to make them stay?

Now a new breed of CRM providers is applying some of the world's most advanced predictive technologies to answer these questions. Utilizing sophisticated computational models with names like neural networks, logic regression and decision tree analysis, these applications make it possible to predict behavior on a customer-- by-customer basis and use that data to determine which incentive offers are likely to keep individual subscribers on the company roster. The best of these CRM technologies learn over time, so that a company's knowledge of an individual customer grows with every conversation or interaction. Combine that with the ability to automatically customize individual or batch emails with the best offer for each subscriber, and the result is the ability to treat customers as individually as the store clerks of yesteryear.

At the heart of these new-generation CRM solutions are artificial intelligence systems that track and map subscriber attributes such as purchasing behavior, the types of calls made to customer support, the length of time the subscriber has been with the company, and so on. The actual variables are customized to the company's product offerings and unique circumstances. Every time the subscriber interacts with the company, the system adds information to the mapping.