n virtually any business, acquiring new customers is one of the essentials for growth. But a key to business success is finding how to acquire new customers at maximum profitability. To gain new customers, a business typically can choose to develop and place ads, exhibit at tradeshows, do direct mailings and telemarketing, invest in field sales representatives, and now have an internet web site. Each of these ways to communicate with the marketplace can be useful in acquiring new customers. However, when tracking the costs versus the resulting customer response for any of these actions, the "cost per lead" or "cost per new sale" can be quite high, resulting in reduced profitability. This failure to achieve maximum return is often because the marketing communication or sales effort is not targeted to the most receptive customer group, so a company is forced to contact a much larger group at a much higher cost, "just to be sure".

Companies many times overlook one of the most important tools for new customer acquisition at high profitability: their own customer database. We have all heard of the advantages of "customer database management" or "customer database mining", but what does that mean? Essentially, it is finding ways to segment your customer database to identify those current customers or prospects most likely to buy more or to buy the first time if you paid more attention to them. It is focusing your marketing and sales efforts to achieve the lowest cost per sale, resulting in higher profitability.

How can customer database segmentation be done? Perhaps its most basic form is the old "80/20" rule: roughly 80% of your revenues come from 20% of your customers, so you should pay more attention to that 20% segment. It’s surprising how many businesses have not done even this most basic customer database segmentation to identify their best current customers.

Customer database analysis to increase sales and profitability has evolved far beyond the "80/20" rule. Using sophisticated statistical correlative and predictive methods, a customer database now can be modeled to provide answers to many critical business questions that previously were mostly guesswork:

• What customers are most likely to respond to this offer, so I don’t have to waste money contacting customers who will not be interested?

• What "nonbuyers" (those who have indicated an interest in your products, but have not yet purchased) are most likely to become buyers, if I focus more effort on them?