Predictive Churn Analytics: A new approach to Churn Management Over the last decade, Indian telecom industry has managed to grow at a very fast pace with subscriber CAGR of 50%. However, with teledensity reaching close to 75% and urban teledensity surpassing 150%, the scope for future acquisitions is decreasing. Industry growth has entered in mature phase and the marginal cost of each acquisition is bound to increase. Marketing axiom “It is less costly to retain an old customer than acquiring a new one” is going to be even more evident. Under such circumstances, strategic focus of telcos should shift to customer retention. Traditionally, telcos have been relying on excel based analysis & dashboards to measure and control customer churn. The common metrics used are AoN, ARPU, MoU, Service Plan & Service Area etc. The major constraint with the current approach is that the retention efforts are reactive i.e. they are triggered only after customer has already expressed his willingness to abandon current service operator either voluntarily or involuntarily. In some cases customer might have already switched to other operator. In Indian scenario, the problem is more compounded by the usage of dual SIMs. This approach not only leads to dolling out of costly retention products to customers by telcos but also exact root cause of churn is often missed. Some customers are even smart enough to circumvent the long retention calls by simply saying that they are leaving service area. In lot of cases it was found that customer had already switched to competition with better service plan. The question which often baffles retention managers is how to predict which customer can churn. The question can be answered with the help of predictive analytics. With the help of predictive models based on customer’s demographic and usage patterns one can predict churn with certain degree of certainty. However, there are certain issues which need to be addressed by telcos before they can reap the benefits of predictive analytics. Every analytics require large amount of historic & demographic data to predict the future trend based on historic patterns. While there is huge amount of customer’s usage data available in telco’s database, there is very less demographic data available. Even if the data would be available, it would not be integrated with CRM rendering it unusable. Therefore, first & foremost telcos should collect this data from its customers and integrate it with CRM & other legacy systems. In future, the best way to capture this data is at the time when sale is made. The information should be captured along with CAF & properly fed into the system. If required, CRMs should also be modified appropriately to store this data. The other area where telcos need to focus is the entire way of measuring customer experience i.e. his lifecycle journey through whole system. Currently telcos measure performance of various touch points like contact centre, service centre etc independently. There is nothing wrong in such approach and efficiency & effectiveness of each touch point must be measured & improved at both macro & micro level. However, it misses out on measuring a customer’s experience while he uses various touch points simultaneously for various queries & complaints. Therefore, the approach needs to be turned upside down & everything must also be viewed through customer’s lens. It is critical to examine that at various stages of lifecycle where, with whom and for what customer has interacted with the system and what was his experience during these interactions. By plotting customer’s journey in this way we can determine his residual feeling after each interaction. Since it is not practically feasible to plot such journeys for subscriber base of 30-40 lakhs, we need to develop such techniques & metrics which can capture such picture. The metrics which can help envisage such picture could be customer’s complaint per month, no. of times he contacted contact centre, no. of times he visited service centers, no. of first time resolutions for this customer, resolutions within TAT etc. These inputs can then be fed into predictive models to determine customer’s level of satisfaction and propensity to churn in near future. Such type of analysis will not only help telcos to reduce churn but will also help them target product offerings in better & effective way. Whether or how sooner telcos adapt to these new methods to create service differential is the question that will be answered in due course of time. The sooner, the better. By Simarjeet Kaur Worked with Idea Cellular as Retention & Loyalty Manager
Posted on: Fri, 23 Aug 2013 06:49:37 +0000