There is no question that churn is the plague of the telecom industry. In the mobile business, annual churn rate of 20%-30% are standard. In developing countries, churn can be as high as 60% per annum. Do you know any other industry where companies lose 20%-60% of their customers between the 1st of January and the 31st of December?
High levels of churn are the results of both the supply-side and demand-side peculiarities of telecom service providers. On the supply side, operators engage in intense marketing activities, launching new tariffs, handsets and promotions on a continuous basis to lure end-users to their own service offering. On the demand side, barriers to switch are very low: SIM cards are mostly free, prepaid customers usually do not have to provide much information, and in countries where Mobile Number Portability has been implemented, customers can take their phone number with them to their new service provider.
But if churn is the plague, you don’t have to be fatalistic. Tomorrow’s winners will be service providers that better understand what customers want and better anticipate how customers will behave. This applies to churn as well and raises questions such as: can we forecast which customers are likely to churn in the near future? Can we explain churn? Can we retain customers who are about to churn?
Yes, we can. Churn can be reduced, and should.
INVESTAURA has developed a five-step Churn Prediction and Customer Retention process as follows:
- Step 1: in the initial step we use data mining techniques and statistical analysis to develop churn prediction models that accurately predict which customers are likely to churn in the next 1-2 months. Typically, these models will identify 50%-70% of the near-term churners within 10% of the total customer base. Note that these models can be very good at predicting churn, but they are not always so good at explaining churn. Yes, they do provide typical characteristics of churners broken down by customer segments; but say, if men are more likely to churn than women, the models do not really explain why. So hypotheses can be formulated, but these need to be crosschecked with customer feedback.
Also, some of the identified customers are much more valuable than others, which leads to the following step.
- Step 2: Customer Lifetime Value (CLV) is estimated for each customer, on a monthly or bimonthly basis. The calculation of the CLV will draw on the results of the previous step, using in particular the survival function to calculate the probability that a current customer will still be customer and generate revenues in the next months and years. Also the calculation of the CLV is greatly facilitated when the telco has undertaken a service costing exercise in the past, as profit margins are then known on a service (voice, data), call types (on-net, off-net) as well as user basis. At the end of step 2, you will know your target for retention: those customers who are both likely churners and are particularly valuable. Let us say that these represent 2%-4% of your total customer base, in a given month.
- Step 3: in this step, we plan and design the retention campaign(s). Customer retention can be expensive, so even if we know from Step 1 and Step 2 which customers should be addressed, the retention campaign has to be designed in such a manner that it is not only low cost but also bring maximum impact.
There are two main cost components in a retention campaign: the cost of contacting customers and the cost of the incentive provided to customers so that they do not churn. SMS as communication channel is used extensively in marketing, in particular for cross-selling and up-selling. Although it is a very cheap way to approach customers, it is not always appropriate for customer retention for two reasons: first it does not provide any feedback or insight on the reason why customers might be thinking of churning; secondly it might provide costly incentives to customers who are not planning to churn in the first place (remember that the churn prediction is not perfect!).
More often that not, customers churn because they are not satisfied with the quality of their interaction with their service provider, not because of price. Think about it: did you ever get a call from your service provider asking whether everything was allright?
This is the reason why we believe that customer retention based on a dialogue strategy should also be implemented: telcos can collect customer feedback that helps understand the true nature of churn, and minimise the cost of incentives / compensation. While talking to customers, front-office staff can react flexibly to customer needs and show that they do care, which does not necessarily mean giving something away for free.
- Step 4: in this step, the retention campaign is executed over a 2-6 weeks period. Various customer segments can be approached in parallel, using various communication channels and various short-term retention measures and incentives. In this phase, customer feedback needs to be channelled back to the back-office team for further analysis and suggestions. This feedback provides useful input data for the identification of medium-term churn reduction measures such as:
- launching new tariff options and products that better meet the needs of individual customers segments
- improving the quality of service (sales, CRM, network)
- improving brand strength
- creating barriers to exit
- Step 5: this phase is about quantifying the overall performance of the churn prediction and customer retention process (Step 1-4): did we address the right customers? How many customers could be retained? How much did this cost? How much additional revenues and profit were generated from the customers retained? What did we learn from customers? Do we understand the drivers of churn better than before? What can we improve in the next cycle?
The impact of a professionally orchestrated churn reduction programme can be massive. One of our customers could reduce its churn rate by more than 20% and is saving millions of dollars in the process, while increasing its market share. A number of mobile operators have communicated that they could decrease churn between 15% and 50% depending on customer segments. According to INVESTAURA’s estimation, reducing churn by 10% points leads to an improvement in EBITDA margin of 2%-4% points.
What about you? Are you doing all what it takes to reduce churn and unlock value in your business?