Applying ACTMAN: Profiling

Profiling works by analyzing the characteristics of a firm's current customers and then identifying those characteristics that distinguish between the best customers, the worst customers and non-responders. The firm then targets potential customers whose characteristics match those of the firm's best customers. Success in this endeavor depends on access to external databases. It is possible to learn about existing customers using surveys or other methods, but without external databases that contain comparable information about non-customers, it is impossible to target prospects for acquisition. The profiling process involves three steps:

	Step 1:	Identify available data on customers and prospects.
	Step 2:	Determine which variables distinguish between the best customers, the worst customers and non-responders.
	Step 3:	Use these variables to target prospects that have the highest potential.

Given ample data, profiling is a simple undertaking. Its weakness stems from its lack of statistical sophistication: It does not include the statistical analysis necessary to assess the relative strength of various characteristics as predictors. In other words, profiling identifies the characteristics of a firm's best customers, but it does not tell analysts which of these characteristics are the best identifiers of the best customers.

For example, suppose a financial services firm has analyzed its customer base to determine its best and worst customers. The firm also has data on prospects that were solicited but did not respond. Table ACTMAN-1 displays the demographic factors that discriminate between its high-value customers, low-value customers and non-responders. It shows that the best customers have higher incomes and larger families. The firm now can target prospects who fit this profile.

Examining the profile of the non-responders can make this targeting method more efficient. Note that the non-responders are similar to the high-value customers except in household income. The non-responders' income is slightly lower than that of the high-value customers but higher than that of the low-value customers. This suggests that the firm could be failing to reach prospects with moderate to high potential value. In light of this, the firm may want to reevaluate its modes of communication with prospects whose profiles resemble that of the non-responders. It is possible that the traditional mode of communication does not reach them.

To get a better idea of why profiling has such great value, consider this example about a firm in the personal computer industry. This firm has a direct sales force that calls on large corporate accounts. The sales force was told to obtain the following information about each prospect:

	- The number of personal computers currently in operation;
	- Key criteria for selection of a personal computer vendor;
	- Sales potential, assessed by looking at the number of new personal computers and the number of those more than two years old;
	- Current vendor(s); and
	- Satisfaction with current vendor(s).

On the basis of these data, the sales force developed a prospect targeting strategy. First, it analyzed the importance each factor had in predicting whether a prospect becomes a customer. It then developed a scoring model, which weighted each factor to create an overall score. The firm gave each profiled prospect a score, and then divided the prospects into accounts A, B and C according to the scores. Prospects in Account A received four follow-up calls, those in Account B received two, and those in Account C received only one. In this way, profiling significantly improved the efficiency and effectiveness of this sales organization. It also increased Customer Equity™, because the cost of acquiring each customer decreased significantly, which allowed the firm to acquire more customers for the same budget.