Customer retention is one of the biggest challenges facing businesses in a wide range of industries, from subscription media and mobile phone operators to fitness clubs and insurance providers. Our client, a national UK business with a customer base of over 800,000, wanted to understand what drives customer retention and attrition. Which types of customer stay longest? What drives customers to leave? And what can be done to increase the lifetime value of customers?
Select were asked to analyse a large dataset generated by customer activity over the last several years in order to identify drivers affecting the loss of customers (otherwise known as churn). We considered a wide range of factors: indicators of customer behaviour and engagement, customer demographics, indicators of the customer experience, and external factors such as competitor activity.
We used survival analysis to model how the hazard (risk of leaving) of an individual evolves as the factors associated with them change over time. The fitted model can be used in several ways to gain valuable insight:
- The model quantifies the strength of each driver, showing where the client should focus their energy in improving the customer experience in order to improve retention.
- It predicts the lifespan of an individual with a given set of characteristics, allowing the client to identify the types of customers with the highest expected lifetime value.
- When used with live data, the model predicts the chance of an individual customer leaving imminently. This allows pre-emptive action to be taken to save high risk customers.
We also fitted a model to examine the factors that affect the chances of a lost customer returning. This allows a longer term view to be taken, since the loss of a customer who is likely to return has less business impact than the loss of someone who is most likely gone forever.
The fitted models were encapsulated in a user friendly Excel tool that allows the client to extract the relevant information and predictions of customer retention under different scenarios.
Prior to our analysis, the client had a number of hypotheses based on anecdotal evidence about which factors are important in retaining their customers. Our analysis was able to confirm some of their hypotheses while refuting others. This allowed the client to move to an evidence based strategy for improving retention in which they focus resources on the areas that have been shown to have high impact.