Attrition Prediction & Retention Program

In today’s environment, it’s getting harder and harder to differentiate your institution from competitors and bring on engaged and profitable customers. When you spend so much of your resources to do just that, doesn’t it make sense for you to allocate some resources to keep them?

It can cost up to five times more to bring in a new customer than to retain an existing one.

Many institutions fall into the trap of focusing purely on bringing in new customers to generate revenue… mostly because of mandates from above in the form of acquisition goals. But it’s a lot less costly to retain revenue by keeping customers – and it has a similar impact on the bottom line. So how do you positively impact your bottom line? WordCom’s program is built on 4 key components:

• Attrition Analysis: Who is leaving and what do they look like?
• Customer Analysis: How do the customers that stay differ from those that are leaving? Who do we try to retain?
• Attrition Prediction: Create a customized model to identify customers likely to leave in the near future.
• Retention Execution: Comprehensive plan to efficiently use resources and communicate to the customers who could be saved.

Attrition Analysis

The first step is to do an analysis and determine whether your institution has an attrition “problem.” While there is opportunity in stemming even the smallest of attrition, an institution that has an attrition rate of 5% is in a much better position than an institution who is losing 30% to 40% of their customers. However, the institution with the higher rate can make a much bigger impact on the bottom line with a good retention program.

Many institutions focus on “net new” for the attrition measurement, but if the institution is really good at bringing people in, that may mask the problem of many people leaving. A comprehensive attrition analysis will look at all the people leaving on a monthly basis and profile them to look for trends. A thorough profile includes account-specific variables such as single-service vs multi-service and balance levels, as well as demographic and life stage profiles. If you do not currently have updated demographics appended to your data file, it’s important to append this information to people who have left as well as those still with you. A complete profile should minimally include:

• Single-service/multi-service
• Product ownership
• Loan/deposit balance levels
• Tenure
• State/market/branch
• Profitability segment
• Number of services
• Age
• Income
• Net worth
• Life Stage segmentation

Once the profiles are complete, they are monitored for changes on an ongoing basis. Significant changes in attrition by specific groups may signal issues with products in the competitive marketplace. Creating dashboards is a fantastic way to easily review the metrics at a glance.

Customer Analysis

In addition to the Attrition Analysis, doing an extensive analysis of your existing customers is equally as important for two reasons. First, to make the findings of the attrition analysis actionable, you have to identify what is different about the people who are leaving. If the analysis identifies that 60% of the customers who leave are single-service checking and 60% of all your customers are single-service checking, the metric is not meaningful.

The second reason for the analysis is to help create actionable sub-segments for the execution of the retention segments. A lot of the same categories used for the Attrition Analysis are also used for the Customer Analysis (product ownership, Net worth, life stage, etc.) to identify customers who have the potential to be profitable in the future. In addition, segments are created based on current profitability to ensure the appropriate resources are allocated to those customers who could significantly impact the bottom line if they left.

A valuable addition to the analysis is Life Stage triggers. By having your customers monitored for certain life events (pre-movers, new movers, newly married, new baby, etc.), you can know early on if a customer has an increased attrition risk. Also important for the execution phase is utilizing Next Most Likely Product (NMLP) models to identify if customers are likely to have a specific product need.

Attrition Prediction

During the attrition analysis stage, if you have identified certain groups of customers that are more likely to leave, then you have a basic attrition prediction system. However, it’s much more predictive to build a model that uses those variables in combination with other variables to come up with a ranking system to score customers from most likely to least likely to leave. The model needs to include as much data as can be provided… from readily available variables like product ownership, tenure and balance to transaction changes and demographic factors. The more accurate the model is in predicting attrition, the more efficient we can be with resources in the execution phase.

WordCom’s process uses a combination of product ownership variables, transaction variables, and outside demographics to create separate models based on the type of products a customer has. These models place customers into deciles based on the likelihood of attrition. For one client, the model accurately predicted 49% of the attrition within the first three deciles over a 12-month period.

Retention Execution

The first step in building a communication plan to retain customers doesn’t even rely on all the analysis steps above! Having a robust onboarding program for new customers and consistent communication to existing customers with needs (NMLP is excellent for this) can go a long way in making sure attrition is minimized from the start.

However, when customers have been identified as likely to leave, the best defense is a comprehensive communications plan. Some customers will leave due to life events, which is harder to mitigate but the majority of customers will leave due to product needs or poor customer service (perceived or otherwise). A consistent communication stream involving NMLP and feel-good messaging can be instrumental in reversing potential attrition.

If you had an unlimited budget, you would communicate to all customers through all channels with a variety of messages that are relevant to the customer. Since unlimited budgets don’t exist, strategies need to be employed to efficiently communicate to those that are likely to leave who are also those that you want to keep… all the while making sure the message is relevant to customer.

WordCom’s approach is to develop a matrix to communicate the relevant message to the proper customer at the appropriate time. The variables utilized in the matrix are:

• Likelihood of attrition
• Current profitability
• Potential profitability
• Next Most Likely Product
• Communication channels

For the communication channels, it’s recommended to use as many as are available. Ideally, for high attrition/high profit customers, you will utilize calls from branch staff, which are supplemented with email, direct mail, and digital ads. As you work down the matrix, fewer channels can be used to balance the cost of the communication with the potential value of the “save.”

Putting all of these elements together will create a comprehensive program to help minimize the number of customers leaving and increase revenue to have a significant impact on the bottom line of your institution!