Dynamism of a Dog on a Leash, Giacomo Balla, 1912 (Courtesy: Buffalo Fine Arts Academy, New York)
Ideas for Leaders #538

Tracking Customers Who Leave Without Saying Goodbye

This is one of our free-to-access content pieces. To gain access to all Ideas for Leaders content please Log In Here or if you are not already a Subscriber then Subscribe Here.

Key Concept

Customers don’t always inform a business that they are no longer customers. A new model, developed by researchers from Columbia Business School and London Business School and based on customer behaviour, allows companies to disentangle customers who are still active customers from customers who have ‘silently left’. The model also identifies customers who are in danger of leaving. 

Idea Summary

Customers can stop interacting with a company in one of two ways: they inform the company that they are leaving (such as when they cancel a subscription to a magazine) or they simply stop interacting with the company. Unsatisfied patrons of a theatre, for example, may decide to seek their entertainment elsewhere. The company has no means of knowing about that decision.

From theatres to the general store, non-contractual relationships with customers are nothing new. Such relationships have always allowed customers to take their business elsewhere without informing the business (as opposed to contractual relationships in which customers have a written agreement with the business and formally inform the business that they are leaving).

In the information age, hybrid businesses with both contractual and non-contractual relationships with their customers are increasing. For example, when you open a Facebook page, you are informing Facebook that you are a customer. When you close down your Facebook page, you are informing Facebook that you are no longer a customer. However, rather than engaging in this clear contractual process, many Facebook customers take a more hybrid approach. They ‘contract’ with Facebook by opening a page, but then disengage with Facebook by letting the page go dormant — no longer visiting the page or adding posts. In marketing terms, Facebook customers can either actively churn (inform Facebook when they are leaving) or passively churn (leave without informing Facebook). The dilemma for Facebook and other businesses with passive churners is clear: how do you tell the difference between an engaged customer of your business from those who might seem engaged but have, in truth, passively left?

In the past, businesses monitored and managed their relationships with customers through RFM (recency, frequency, monetary value) models. They looked at the recency, frequency and value of a customer’s purchases to determine the state of the relationship. According to professors Eva Ascarza and Oded Netzer of Columbia and Bruce Hardie of the London Business School, RFM models are no longer adequate to manage the increasingly hybrid relationships with customers. Instead, using data from two distinct types of businesses — a daily deals website and a theatre — the team developed a model that is able to disentangle passive churners from engaged customers.

The model emerged from their research into the behaviours of customers interacting with the two businesses over a significant amount of time (drawing their data from a sample of 1,000 customers for each business). For example, the online deals site sends its customers nearly daily emails that contain a multitude of deals. If a customer clicks on one of the deals, the site earns a fee. The researchers studied whether, over a period of 2.5 months, customers opened the email, and if so, whether they clicked on at least one deal or unsubscribed from the email.

For the theatre, the researchers focused on communication efforts with patrons. The theatre sends monthly emails to its patrons describing upcoming performances, with links to the theatre’s websites where tickets can be purchased. The researchers studied whether, over a period of 20 months, email recipients opened the email and if so, whether they clicked through to the theatre’s website or unsubscribed from the service.

In analysing the data they collected, the researchers were able to identify three different sets of behaviours corresponding to three states of customer relationships:

  • The first state was the active customer. These customers opened the emails and clicked through to the deals or the theatre’s website.
  • The second state was the passive churner — customers who had silently left the business. These customers no longer read the emails.
  • The third state identified by the researchers was the customer in imminent danger of leaving the business. These customers opened the emails but never clicked on the deals or, in the case of the theatre, clicked through to the theatre’s website. Many of these customers, during the periods studied, would eventually unsubscribe from the service.

Business Application

By delineating the behaviours of actively engaged, passively departed and imminently departing customers, the model has clear implications for customer relationship management. The model reveals some of the errors that businesses might make using traditional RFM models. Looking at revenue data only, for example (that is, focusing on the customers who had clicked) would not reveal the all-important metric of customers who had opened the email and did not click — customers, the research showed, in imminent danger of leaving. Monitoring clicking behaviour alone also did not separate the actively engaged customers from those who had passively churned.

As a result, companies may have an erroneous idea of the number of customers who are engaged with them. Companies may also be missing their targets when trying to reengage customers. If the ‘silently gone’ customers don’t open emails, efforts to develop better products or email content will, adapting the familiar phrase to the information age context, fall on deaf ears.

This new model helps companies understand which customers are actively engaged, which are silently gone and which are about to leave. This information is the first step to develop targeted marketing to keep customers engaged, bring them back into the fold, or keep them from leaving in the first place.

Contact Us




Idea conceived

  • May 2015

Idea posted

  • August 2015

DOI number



Real Time Analytics