Using Modeling For More Predictable Success

Background:

A leading transportation and logistics company was performing very well in the market, but they struggled to understand why customers increased or decreased their business with them.  Although they regularly collected massive amounts of transaction and experience information on their customers, this alone did not provide sufficient insight to guide their efforts to improve customer relationships.

 

Challenge:

A key challenge was to understand which customers were growing their business compared to reducing or completely severing their business with them.  They engaged Loyalty Research Center to:

  • Define growth, shrink, and quit customers,
  • Identify and model the factors that drive growth and quit behaviors, and
  • Use that algorithm to project the findings onto the entire customer database.

All models were designed to significantly enhance Client’s ability to identify and target customers in these segments for tailored messaging, intervention, and sales efforts.

 

Solution:

LRC conducted a Modeling Program utilizing Client’s current and lost customer databases coupled with primary information collected from customers by LRC.

The first step was to agree on a set of meaningful “outcome” variables.  Client and LRC agreed to establish these variables based on the percent change in total unit volume at the customer aggregate level from one year to the next.  This encompassed growth, stable, decline, and lost customers.

The second step was the challenging effort to integrate the multiple customer-oriented databases.  Why challenging?  The databases were organized by time period and pertained to:

  • Customer-level information on product/service purchases, timing of purchases, interactions with various branches, interactions with customer service and other entities around the organization, payments over time, etc.
  • Branch-level performance information on transactions combined across customers and a variety of branch-specific measures, like turnover.
  • Other databases pertaining to products (e.g. product breakdown and failure rates), individuals (e.g. sales reps/teams), and other aspects of overall corporate performance affecting the customer’s experience.

Patterns over time were useful to pick up trends for some types of customer experience.  For example, an individual branch can be trending up in terms of units associated with the branch.  This can convey a significant amount of information.

LRC determined that Lost and Shrink customers tended to be similar in terms of:

  • Small customer size
  • Low current volume
  • No participation in Client’s Customer Satisfaction program
  • Transporting hazardous material

This makes sense:  smaller customers and those that conduct a lower volume of business in general have a lower cost of switching than larger customers.  Participation in the Customer Satisfaction program was the biggest point of interest that came out of the discussion, and makes sense.  Communicating with customers through a satisfaction program opens up more opportunities to identify and resolve problems that would contribute to quitting.

Fig. 1: Example summary chart of increased and decreased opportunities

Likewise, customers that were more likely to increase their business tended to be larger in size and currently doing a higher volume than shrink or quit customers.  An interesting trend was that if a customer’s Account Manager managed more units, the customer was more likely to grow.  This was attributed to better Account Managers being more likely to have a higher volume of customers.

 

Next Steps:

These results were highly valuable for Client.  They knew that to decrease risk of a customer shrinking or quitting, better participation in a Customer Satisfaction program was critical.  Client was able to proactively communicate with customers that were more at risk in order to prevent defection.  Key customers that fit the growth model were more effectively targeted in their inside sales strategy, putting them in position for more predictable success.

 

Posted in Case Studies, Insights, LRC Blog, Professional Services.