|
Cellular communications firms have been losing 2.9 percent of their customers each month, or up to 35 percent per year. In addition, companies have to disconnect many customers for nonpayment of bills. Disconnecting the wrong customer can translate into significant revenue losses, while keeping delinquent customers also affects profitability.
A wireless service provider was formed by combining the customers of 11 different companies within a short space of time. Its first job was to create a unified database covering the more than 20 million customers that had been in 11 different databases. Rather than waiting years for a data warehouse, the provider created an immediate temporary marketing data mart that combined the billing data from each company. This data mart was accessed by SAS products.
The data mart allowed managers in 35 different markets to analyze the data, giving each manager the ability to handle customer trends on a local level. Local managers could immediately address churn, instituting local incentives to keep customers or to rework their existing contract.
Examples of the tactics used to reduce churn were initiatives for:
Rewarding customers by donating to the Special Olympics
Reduced rates for particular customers
A special deal on a new phone
Free movie rentals at Blockbuster Video
Price plans developed after analyzing customer behavior
Phone trade-in programs
Overall, the wireless service provider wanted to know why customers were leaving, broken down by
Amount of time with the company
Minute use
Price plan
Thirty other specific market-level and company-level trends
Armed with these objectives, the wireless service provider reorganized its IT department and hired a dozen SAS programmers to construct a decision support system. When the temporary database was built, it installed several SAS products on its Sun Enterprise servers, including Enterprise Miner and AppDev Studio.
As a result of the analysis, the company reduced its churn rate significantly compared to a number of control groups selected to match each of the the test groups receiving the promotions. Even a 1 percent reduction in churn rates produced several million dollars in additional revenue. Several valuable lessons were learned:
Old equipment turned out to be a top reason for churn. As a result, the company strengthened its trade-in program and allowed customers to get a new phone every time they signed a new contract.
The database enabled the company to determine which delinquent customers should be kept and which should be dropped based on modeling and measuring the future profitability of each.
The company was now able to make educated predictions about customer behavior, saving the company millions of dollars per year. Decision making no longer relied on educated guesses as to what would draw and hold customers, but was based on hard facts and figures that provided a detailed map for marketing strategies to follow when creating campaigns.
In the process of creating the database, the wireless service provider achieved some valuable benchmarks:
Reports could now be produced in days instead of weeks, and reports that were previously impossible could now be created.
The company could now produce both companywide and market-level analysis of customer behavior.
New fields had been added to its database. These new fields gave the company the power to make better predictive decisions.
Through customer profiling, the company was able to tailor messages to different customer groups more effectively.
|