Customer Retention - The Churn Approach?
It is a thoroughly documented fact that repeat customers tend to conduct more transactions and to spend more in a given period of time. For example, research in the US indicates that the longer the relationship between a bank and its customer, the greater the dollar value of total transactions that the customer will initiate.
An exhaustive understanding of customer needs and responding to those with value is a key factor in preventing the voluntary defection of customers – or ‘churn’ as it is otherwise known. Research also shows that the higher number of products a customer takes from a bank, the lower their propensity to churn. In fact, up to 50% of individuals using one product from a bank will churn within a period of one year, while less than 1% of individuals using more that four products from the same bank will churn during the same time period.
To understand churn we have developed a five factor model as shown below:
The objective for any bank wanting to tackle the churn issue is to put in place methodologies and tools that can be readily used to answer each of these questions.
However, a number of obstacles need to be addressed:
Banks therefore face a dilemma: Is it better to wait for data to be complete or to act now? If the bank chooses to wait, how long is acceptable? Alternatively, are there ways of using existing data to the maximum extent possible, while building capabilities for capturing and analyzing the missing information?
The simple fact is that data can never be perfect or complete. Waiting for more data to be available can cost a bank invaluable time and many customer defections. Instead, we strongly recommend that banks find ways to derive customer insight from existing data. This exercise will also allow a bank to identify data gaps, and to find ways to bridge them.
Banks capture both demographic and transactional data that can be used in constructing scorecards that assess growth. Various factors need to be taken into consideration such as seasonality, the bank’s past performance, and its size. In this case customers demonstrating certain decline patterns are given a score in terms of their likelihood to defect. With this level of data we can only examine the ‘Who’, but not the ‘What’. But this can be a first step towards identifying potential defectors.
We then need to establish a thorough understanding of customer behaviour in order to answer the ‘What’ question. Examples of behavioural data are some form of credit rating of a customer; or that customer’s response rate to different types of campaigns. Through statistical analysis and profiling it becomes possible to determine the intrinsic (industry, segment, size, etc.) and behavioural (credit rating, campaign responsiveness, sales volatility, etc.) characteristics of customers who have a predisposition to churn. Once behavioural patterns for defectors are identified, the bank will be able to have a form of early warning system for churn, since customers exhibiting similar behaviour to defectors are also likely to churn.
The remaining component of the early warning system is timing. To launch a successful and cost-effective churn programme a bank needs to get the timing right: How can customers who demonstrate a predisposition to churn be identified at the right time?
A set of statistical techniques that can be used to address this question is called survival analysis. In this context survival analysis allows us to examine the relationship between duration-specific churn and the determining factors of churn. A clear understanding of how changes in the level and combination of the determinants of churn affect its timing will assist in the development of prevention programmes that are both effective and affordable.
In order to design actions to prevent customers from defecting, we need to understand why they do it. The factors that compel a customer to churn may be very complex. However, in the most general sense, customers churn due to:
Very rich data is required to statistically analyze the relationship between these intricate factors and customer churn. However, most banks do not collect such data. A popular methodology, conjoint analysis, may be used to relate the complex factors that drive churn to product/service preferences. These complex factors may be thought of as preferences/attributes. In this context, conjoint analysis would involve the measurement of these preferences/attributes and the assessment of how they relate to the perceived similarities or differences between products/services.
A conjoint study typically involves the following:
The analytical insights gained from the WHAT, the WHO, the WHEN and the WHY should be used to compile a well-structured ‘Churn Prevention Action Plan’ that addresses all of the facets and issues related to churn. This is illustrated in the figure below.
If the churn prevention programme is effective, the bank can look forward to reaping significant rewards from its efforts. After all, as the Harvard Business Review reported in 2001, a company that can retain as few as 5% of its current customers can raise its profits by as much as 75%. That’s a statistic that shows churn prevention to be crucial for the profitability of any bank.