Have you heard the joke about the two statisticians on the front line of a battle? Crouched in their foxholes, they spied an enemy soldier coming down a hill. Taking careful aim, one fired 10 feet below and the other fired 10 feet above the target. Smiling, they turned to each other and said, "We got him!"

Well, on average, they did. However, averages are convenient fictions that often mislead us more than they give us information. In spite of this, customers are often evaluated relative to averages.

For review, let's think about what an average is. We all know how to get an average; add up a column of numbers and divide by how many numbers are in the column. What most people forget is that an average makes sense only when it comes from a particular kind of distribution of numbers. If the things that are averaged don't come from a normal distribution ­ one that follows the bell-shaped curve ­ the average makes no sense. For example, average income is usually a worthless number. Anyone looking at the distribution of income will see that, in most cases, there are a lot more poor people than rich people.

I don't suggest one abandon using averages. Use them where they are appropriate. However, in this and subsequent columns, I will explain why segmentation is a better alternative to averaging and how it can add a whole new dimension to understanding your customers. Segmentation is the key to successful customer relationship management (CRM) because it provides the most efficient mechanism for target marketing and customer care.

Essentially, segmentation is carving up your customer files in ways that group the people into categories. What kind of categories? You decide, as long as the categories make sense to you and help you run your business profitably.

One simple and useful way to segment your customer files is with a Pareto analysis. It provides a way of dividing customers according to their value to the organization.

To construct a Pareto diagram, select a criterion scale, such as customer sales. Next, rank and order the customers from the one who spends the most with the company to the one who spends the least. Finally, divide the total list into deciles, those 10 percent of customers spending the most, second most, and so on. Plotting the data will produce a distribution as illustrated in Figure 1.

Figure 1: Sample Pareto Distribution

In this Pareto chart, 70 percent of sales come from the top two deciles. This group of customers is a segment. This analysis has many different implications. For example, applied to CRM, these 20 percent of customers providing 70 percent of revenues obviously must be given the best treatment. They are the best customers. Manage-ment should know who they are and what they are buying.

This segment gives management insight into what it is doing best. By comparing its characteristics to the attributes of the three deciles below it, a segment constituting 21 percent of sales in the chart, management may learn what it has to do to increase sales to the middle-tier segment.

Additionally, the remaining deciles may be costing the company money. A company may use this type of analysis to control calls to its service center. Upon initiating contact, the customer is asked to type in his customer number using his telephone keypad. Depending on the customer's decile, he is either moved to the head of the call queue or forced to wait until it is his turn. In other words, good customers get the best treatment. If low revenue customers hang up and leave to do business with the competition, so much the better.

Most, if not all, businesses have similar distributions of customer sales. The common term is the 80/20 rule. Eighty percent of sales come from 20 percent of customers. These "heavy users" constitute a segment that is vital to the business.

Segmentation represents a more accurate and informative way of examining your customer database. The Pareto method is just one example of segmenting your data into groups that give you insight. Future columns will examine other segmentation schemes, analytic methods and ways to use segments to increase your marketing effectiveness.

The important thing to remember is that there is no average customer. Each customer represents a different set of challenges and opportunities. Segmentation is the only viable way to cope with today's dynamically fragmenting consumer marketplace. By using segmentation, marketers are more effective in channeling resources and discovering opportunities.