We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.


Customer Categorization Tips for Identifying “Good” Customers

Originally published April 25, 2013

How do you characterize what makes a customer a “good” customer? In one of my previous articles, we looked at some example variables contributing to calculating the value of a customer, some of which described the revenue potential associated with the customer, some described costs, and others evaluated levels of different kinds of risks. But aside from using these variables to calculate some quantifiable valuation, we can also look at how the values of the collection of these variables contribute to segmentation of the customers into different “qualification” strata. There are two aspects to this segmentation: defining the customer categories and their characteristics, and developing methods for classifying customers into those categories.

Let’s focus on the first of these aspects: specifying the criteria for evaluating customer “goodness.” The first step is to suggest the levels for categorization. The simplest approach is to define two levels of customers: those who are good customers and those who are not good customers. We can do this by providing greater precision for the conceptual variables and measures used for evaluating customer value and setting thresholds for the measures of each variable.


 Variable  Measure  Definition  Threshold for "Good"
Revenue Stream
Net annual number of products purchased
The count of products purchased - the number of items returned during the calendar year
8 or more total net products purchased
Revenue Stream
Total annual net revenue in dollars The sum total of all sales transactions - the sum total of refunds for returned items during the calendar year   $1,000.00 or more in total net sales
Customer Lifetime
Total number of months The number of months that the customer has been a customer 24 or more months
Maintenance Cost
Total annual count of customer support calls The number of calls the customer has initiated during the calendar year Fewer than 3 calls
Maintenance Cost
Total annual minutes of customer support calls The total number of minutes the customer has spent on support calls during the calendar year Fewer than 20 minutes
 Risk Cost
 Credit rating
FICO score
FICO score > 700


Table 1: Example criteria to qualify "good" customers

For example, consider the variables and measures shown in Table 1. In this approach, any customer whose measures meet or exceed all the threshold values is considered a good customer, while all others are not considered to be good customers. There are some drawbacks to this simple model that limit its usability, such as:
Granularity – This example only showed two types of customers, but in reality there may be multiple levels of qualifications and classifications.

Weighting – Not all of the measures contribute equally to the qualification of a good customer. For retail businesses, the revenue-based measures may be more important, while in financial services the risk-based measures may take precedence.

Precision – The thresholds may be defined too strictly and some good customers might not qualify when their measures are close to the thresholds, but do not meet them.

Completeness – In this example we provide one set of measures and thresholds, but do we have all the right measures?
Ultimately, one clear objective is to be able to not just classify the customers, but also determine the percentages of value that are attributed to each customer segment. There are some ideas for refinement, including:
Additional Levels – Suggesting additional customer categories and corresponding thresholds. Instead of dividing the customer community into good and not good customers, provide some tiered categories such as “Best,” “Good,” “Mediocre,” or “Bad” customers, all determined based on threshold scores for the defined set of measures.

Improved Weighting – Test different weighting factors associated with the measures to determine sensitivity as a way to improve the accuracy of the model.

Improved Thresholds – By allowing some flexibility in the overall model for defining the categories, customers may be appropriately categorized even when they barely miss meeting a few of the criteria.

Continued Refinement – Continually review the full set of value measures to identify new candidates for inclusion.
All of these ideas must be framed within the corporate vision and provide measures that are mapped to the customer categories in ways that are optimized to maximize the company’s key performance metrics such as customer profitability (how much money you make by customer), accounts receivables turnover (how quickly your customers pay their invoices), cost to serve (how much is spent in maintaining the customer), or customer satisfaction. For example, you may designate your best customers as those whose profitability is at least 150% of the average customer’s profitability, who have the longest 5% customer lifetimes, and 20% of the average cost to serve. The next tier (“good”) would be defined using lower levels for the same sets of measures, and so on for each subsequent level.

The end result of this process is a set of measures used for customer categorization, but it is only useful in the context of the existing customer base. The next step would be to identify the demographic variable values the customers in each customer category share that can be used in a number of ways for improving value, as we will see in my next article.

Recent articles by David Loshin

 

Comments

Want to post a comment? Login or become a member today!

Be the first to comment!