PREDICTIVE ANALYTICS IN RETAIL CASE #303: IMPROVE CUSTOMER RETENTION

Predictive analytics brings science to retail marketing. Along with personalization and precision targeting to maximize campaign ROI, customer retention is a perfect use case for predictive analytics, and arguably its the most important. Customer retention is an overarching name for several key use cases designed to improve lifetime value and reduce customer churn. Here are a couple good ones:
While many marketers think the goal of retention campaigns is to resuscitate inactive customers, the response rates will be low when targeting such weak customers. However, customers who have not purchased in a while, e.g. a year or more, are not all the same in terms of their probability to buy again. Predicted spend in the next 30 days for example might be low, but each customer is different, and that value can be used to rank and separate the customers worth more effort, and ultimately indicating how much you should invest in different groups.

Often overlooked for retention are the historically loyal customers who are just starting to fade. Churning a “Loyalist” customer could have the same revenue impact as losing many ordinary customers. The problem is that slow but steady changes in risk of churn are hard to detect from subtle changes in recency, frequency, or monetary value alone. Instead, precision-modeled predictions for future loyalty and the probability of churn risk can act as early warnings that trigger counter-acting interventions, such as just-in-time offers to pull fading customers back into their former loyal buying pattern. Read the case study below to learn how we helped a major retailer generate 3x revenue by targeting the right high risk customers for retention programs.

CASE STUDY

Our client, a major B2C and B2B retailer of tools and shop supplies, had been periodically targeting their high recency customers for retention campaigns. Our solution included a two-prong approach:

  • First we predicted the churn risk and expected value for each high recency customer.
  • Then we selected customers with a churn risk above 50% that had the highest expected value to purchase in the next 30 days.

Next we tested two audience groups via email promotions: Client’s unique selection (High Recency Customers) vs Loyalty Builders’ unique selection (High Risk/Value Customers). The result: The Loyalty Builders unique selection group required about half the time and resources to capture 3x the revenue. After the campaign, churn risk was significantly reduced for more customers in the Loyalty Builders’ group, and for higher quality customers.

To learn more about Predictive Analytics in Retail, read our previous articles on Marketing ROI and Personalization.

Peter Moloney is CEO of Loyalty Builders, a cloud-based predictive analytics service enabling marketers to get revenue lift from more relevant communications to their customers.

Tags :