Are Your Customers About to Defect?

How to Tell If They Are Planning to Take Their Business Elsewhere

By Published on .

Chris Dickey
Chris Dickey

In today's tough economic environment, the most precious asset an organization has is its current customer base. And maximizing returns on that asset begins with a solid retention strategy. But in many organizations today, retention strategies are simply not proactive enough to pick up on hints that customers are about to defect -- hurting customer profitability.

But it is possible to curb customer defection before it's too late. Here are three ways to maximize customer value by decreasing defection.

1. Determine the warning signs. This begins by developing a predictive model to identify the behavioral changes that are precursors to defection. The variables that predict defection, churn and other negative behaviors are often, but not always, intuitive: usage (recency and frequency), average ticket, satisfaction, visits, engagement. A model can identify the most significant variables -- those that are able to predict the negative behavior -- so the marketer knows which to monitor.

A model also enables a marketer to identify the threshold of concern for each behavior, that point below which the likelihood of defection is high enough to indicate a true warning sign.

For example, the model may indicate that usage is the most meaningful variable in predicting defection, and that a 20% change in a given time period exceeds the threshold of concern and is, indeed, a warning sign to be concerned about. (See sidebar below for more examples of common warning signs and thresholds of concern.)

The good news is that predictive models are viable based on the types of data you should be capturing today.

2. Track and monitor changes to warning signs. Once marketers know which customer behaviors to focus on and what levels of change should be cause for concern, they can isolate and track behavior changes over time. They can then monitor those changes in real time by setting up automated tracking processes within a data warehouse.

Marketers can monitor behaviors with database-mining software and view them at a macro level using dashboards.

3. Develop a trigger-based engagement strategy to address behaviors in real or near real time. Marketers can actively intervene before a warning sign becomes a real defection. A trigger strategy must be developed over time, testing and optimizing to see which interventions can most quickly reverse negative behaviors. Interventions should be built around a highly relevant and dynamic message, an offer of relevant and significant value communicated in a channel that takes advantage of real-time data (such as e-mail, text or voice messaging).

If usage is down, for example, at a fitness club, a triggered response may be a message of encouragement and an offer for a friend or family member to join at a reduced rate, or an offer for a free personal training session. If a customer's average order has waned, a retailer could use a new product or service trial offer to reacquaint a customer with a category. Customers might also view that type of offer as a value and soft benefit if it's messaged as a "sneak peek."

For a customer who has neglected to reload a prepaid card, a marketer could send a reminder message and an offer to reload and receive a free trial of a product that would move the customer into another daypart. For a decline in website visits, a marketer could provide a new reason to view the site, including special "best-customer-only" experiences, new content or members-only tools.

Chris Dickey is senior VP-director of customer relationship management at Barkley. He oversees agency analytics and relationship marketing for 24 Hour Fitness, Sonic Drive-In and Build-A-Bear Workshop, among others.

Trigger-based programs are generally easy to set up and automate, and they are highly measurable. They can be geared to a wide variety of customer behaviors, and business rules can be developed to change the message, offer or channel based on other factors beyond just the behavior (for example, the value of the customer, the geography, customer demographics, product mix or severity of behavior change).

By understanding the key warning signs within your customer-behavior data and developing proactive trigger-based programs to intervene, you can more precisely hone your marketing effort to those communications that drive the highest incremental return on investment. Trigger-based programs are successful because the communication is highly relevant (you definitively know that the behavior has changed) and timely (you can react to the behavior almost instantly through automation).

Triggers are becoming the core communication platform for all customer-relationship-management-based programs. If you haven't developed and deployed one in your organization, do it now. Pilot a program and begin making significant steps toward understanding the warning signs of defection, actively intervening and maximizing your organization's most valuable asset.

Examples of Warning Signs and Thresholds of Concern

  • A member routinely visits three times a week for six months. In the past two weeks, usage is down 30%.
  • A customer buys across four dayparts consistently for nine months. In the past month, she has dropped the evening and breakfast dayparts from her purchase pattern.
  • A registered high-frequency stored-value cardholder continuously reloads below a $5 threshold. In the past two weeks, he has maintained a balance of zero.
  • A retail customer's average ticket is in the top 20% range. Her average ticket drops 10% to 20% in a month.
  • A top customer rates you at five consistently. Overall satisfaction drops from five to four in two consecutive quarters.
  • A "best" customer visits your website weekly, with an average engagement time of 20 minutes. Average time spent on the site falls 30% in three weeks.
Most Popular
In this article: