Speakers’ Corner: Data mining for gold-standard retention


Intense competition, number portability and deregulation have given customers more choice than ever in the mobile market.

According to figures from Mediacells, more than 5,000 individual deals (different contract offerings with different phones and incentives) were available in the UK last month. Customer choice is eye watering.

We all know it costs less to keep customers than it does to acquire new ones, yet mobile remains an industry where churn rates can be as high as 25 per cent.  When you could lose one in four customers each year, the pressure on gaining new subscribers increases substantially.

Churn itself is not, however, the most critical issue; churn of high value customers is. The real challenge is knowing which customers a network wants to retain and which it can afford to lose by creating a process through which intelligent choices can be made.

What is not always considered is the fact that the industry already has the information available to it to retain the best customers. Every call that someone makes and every text message they send, or receive, tells a story about the customer’s experience. 

Consider, as an example, an area of poor mobile coverage around the M25. If customers receive a poor experience in the area, they blame the network. If it happens often enough, they may churn. All this information and more is readily available to the networks through call detail records. 

The problem of course, is the vast quantity of call detail records that are created. Millions are created every hour, billions every week. Capturing, storing and analysing that volume of information to gain meaningful data requires serious amounts of computer processing power.

Full article in Mobile News issue 445 (August 10, 2009).

To subscribe to Mobile News click here