Would you rather have 3,500 or 5,000 new customers? The answer may seem obvious, and in the context of designing a marketing strategy, you are likely most excited about the program that will generate more customers at the lower cost per acquisition (CPA).
However, I’ll now remind you of something you probably already know … not all customers are equal. The 80/20 rule typically proves to be pretty accurate — 20% of your customers generate 80% of bottom-line value. 80/20 White Paper — Do You Really Know Who Your Best Customers Are?
This is where smart direct marketing comes in. Although you will always have a mix of more and less profitable customers, with the right predictive model you can aim your direct marketing toward the prospects that are most likely to turn into highly valuable customers.
Typically, direct marketers think of building targeting models designed to maximize response rates. A response model will predict the likelihood of each prospect to become a customer. By targeting your marketing to folks who are most likely to respond, you can boost volumes of new customers and minimize CPA. However, the customers who are most likely to respond are not necessarily likely to be the most valuable. In some cases, the most responsive prospects could actually be particularly unprofitable, especially if they’re the type to jump at short-term incentives.
A more powerful modeling approach will focus on finding the type of customers you want. This could generally mean focusing on prospects who are likely to be the most valuable, or at times this could mean focusing on optimizing more specific metrics.
The specific targeting approach to use in a particular situation will depend on your business priorities and which metrics need to be managed most aggressively. Regardless of the specific metric, the conceptual approach of the targeting model is to predict the likelihood of each prospect to do business with you AND meet a specific performance threshold — e.g., sales > X; gross margin > Y; lifetime value > Z.
The modeling approach you choose can have a huge impact on the volume and performance of customers that you acquire. As the hypothetical example in the table below shows, choosing a response model vs. value model will lead to very different business results. The response model will generate the largest volume of accounts, but the customer value model will generate 23% more profitability.
The key point is that you can build targeting models that are aimed at different objectives and make deliberate choices about the trade-offs.
And back to the original question: While there may be times that you care most about account and customer volume, from a long-term perspective it’s likely more important to maximize your bottom line, which should mean choosing fewer more profitable customers over more less profitable ones.
William Cao | Chief Analytics Officer
William has earned an MBA from The University of Chicago Booth School of Business, an MS in statistics from Kansas State University, and an MS in applied mathematics from Southeast University. William and his team provide data and analytic leadership to Catalyst’s clients.