customer retention


Identify and retain at-risk customers

Customer Retention

Churn is the process of customer turnover or transition to a less profitable product. With customer churn rates as high as 30 percent per year in some global markets, identifying and retaining at-risk customers remains a top priority for executives. Markets are saturated, unhappy customers leave or downsize their service usage, and a class of professional churners is beginning to emerge.

In order to combat the high cost of churn, increasingly sophisticated techniques (e.g. survival analysis) may be employed to analyse why customers churn and which customers are most likely to churn in the future. Such information can be used by customer service departments to actively monitor the customer call base and highlight customers who may, by the signature in their usage pattern, be thinking of switching to another provider.

Propensity to churn

Struggling with understanding the drivers of churn? Is your churn prediction not early enough? Having gaps in measuring customer value? Putting together the single customer view? Our churn analytics aims to retain existing customers and to prevent them from exiting to the competition (from churn). Without serious data analysis, the knowledge of customer churn behaviour is at the level of expert judgement, which turns out to be not very accurate. The churn model is designed for the maximum performance – it minimises the false positive error (predicting a customer to churn while they actually did not churn) as well as the false negative error (predicting customer not churning while they actually did churn). To boost the performance of the churn model, we use the Social Network Analysis to generate additional powerful predictive characteristics for modelling. No churn model can identify customer churn behaviour with 100% accuracy. However, experience from past projects shows that the churn rate of the customers with the lowest score is 4-10 times higher than the average churn rate in the entire customer base. Such outcome brings a solid base for a retention campaign.

Propensity to save

Imagine you have identified a customer at risk of churn. To prevent the customer from churn you consider several retention strategies, including different offers, different communication channels or different pricing options. Which one is most likely to retain the customer? Propensity to save model (PtS) is designed to answer this question. In the PtS modelling, we have successfully used classification trees,  capable of handling  small datasets usually available for the fitting. Our experience also shows that PtS models have to be updated quite frequently due to changes in the business environment (retention offers, process, customers).

Retention Campaign Optimisation

What actions should I take to retain my most profitable customers? Which customers are worth caring for and which are not? What discount can I afford to offer, so that I still make a profit on the customer if he stays? To do all those things right you should think of optimising your retention campaigns to get the highest ROI.

Deloitte’s retention campaign management offering aims to maximise the expected lifetime value of the retention marketing campaign in the complex scenario of multiple retention offers, multiple communication channels and multiple time points that we can use to treat the customer. The business value of this offering is the most efficient use of the limited resources in order to only treat the customers with the highest added value of the treatment and to avoid treating customers which are not likely to become profitable or where the treatment cost exceeds the future potential.

In essence the retention campaign optimisation is very close to the cross sell campaign optimisation and if any of those two already exist, we recommend merging the second one into the same infrastructure to save costs and to benefit from a unified approach. Retention offers and cross sell offers are then put into the same bag, the objective function is one and you can even share marketing resources between cross sell and retention in a controlled way.

Retention Campaign Optimisation


Jan Balatka

Jan Balatka


Jan Balatka has extensive experience in Data Analytics, Forensic Analytics and Electronic Data Discovery. He has international experience from projects in Switzerland, Germany, the U.S., South Africa ... More