Skip to main content

Triple yes for retention

Bob Rietveld

Managing churn is one of those challenges that requires foresight. Customers generally leave without saying goodbye (Ascarza, Netzer, and Hardie 2018), so predicting which customers are likely to leave has become common practice.

It’s tempting to focus efforts on customers who have the highest probability of leaving, but this is a suboptimal strategy (Ascarza 2018). Why spend time on customers who are going to leave anyway? Ask the following three questions to design a retention strategy that may keep your customers in. But before spending resources, remember that churn is a natural part of business. Sometimes it can’t be prevented, and not all churn is bad. Low-margin or high-demanding customers can drain energy and revenues. But if you decide to try and keep more customers onboard, answering the following three questions might help.

graph TD
A[1. Is this customer likely to churn?] -->|Yes| B[2. Is this customer valuable?]
A -->|No| Exclude[Exclude]
B -->|Yes| C[3. Will this customerrespond positively to my retention action? ]
B -->|No| Exclude
C -->|Yes| Include
C -->|No| Exclude

Is this customer likely to churn?

Using historic retention/churn data, companies can make predictions about which customers are most likely to leave. Counter-intuitively, it’s customers with a medium to high probability where a retention effort might be sensible. If churn probability is low, customers are going to stay anyway, and if the churn probability is high, they are going to leave anyway. Understanding the churn distribution is a critical starting point.

Is this customer valuable to my company?

Understanding the value of every customer relationship is best expressed in the customer lifetime value (CLV). CLV provides a clear framework for assessing the potential lost value of a client AND the upside of the budget you’re willing to spend. Differentiating the retention budget based on value leads to increased profitability on a campaign-by-campaign basis (Lemmens and Gupta 2020).

Will this customer respond positively to my retention action?

A final question remains: how sensitive is a customer to the proposed retention offer? It’s useless to give a discount, if a customer is leaving because of poor customer service. Understanding how customers are going to respond requires data which probably does not exists today. An experiment is needed experiment to find out what works and for whom. Doing the experiment will allow you to understand what works AND apply these leanings to the rest of the customer base. Using modeling techniques, like uplift models, based on the experiment data allows firms to be very specific in what retention offer to bring to which customer. This practice is far more effective than focusing on high churn probability, as it accounts for the differences in customer responses to your offer (Ascarza 2018)

In the pursuit of effective customer retention, it’s essential to look beyond churn probability and conduct a more comprehensive analysis. However, even the most sophisticated analysis cannot always reveal the underlying reasons why customers are considering leaving. To gain a deeper understanding, it’s crucial to engage with your customers directly and ask for their feedback. Their insights will not only help you formulate better hypotheses for retention experiments but also enable you to tailor your retention offers to address their specific pain points, whether it’s pricing, service quality, or product features. By combining data-driven analysis with customer feedback, you’ll be well-equipped to develop a powerful retention strategy that keeps your valuable customers loyal and engaged. Remember, the key to successful customer retention lies in understanding your customers’ needs and responding to them effectively.


Ascarza, Eva. 2018. “Retention Futility: Targeting High-Risk Customers Might be Ineffective.” JMR, Journal of Marketing Research 55 (1): 80–98., Eva, Oded Netzer, and Bruce G S Hardie. 2018. “Some Customers Would Rather Leave Without Saying Goodbye.” Marketing Science 37 (1): 54–77., Aurélie, and Sunil Gupta. 2020. “Managing Churn to Maximize Profits.” Marketing Science 39 (5): 956–73.