At my last company, we invited a strategy consultant to come assist the board with getting a tighter rein on our strategy. We had a board filled with conflicting objectives, and we stood to benefit both from the sound process and the outsider perspective of a good consultant.
Overall, the consultant's process was quite good, and we implemented the process nearly by the (her) book. She hadn't developed anything earthshaking or new. In fact, it was largely text-book strategy set to a rational and clear formula to assist with discipline and retention. This is a good thing. But there was one other aspect of her process that was also text-book that made me uncomfortable, and that's what we're examining today.
She took a traditional b-school marketing approach to customer lifetime value (CLV). That approach says that some customers are simply not as valuable as others, and a business should drop them and focus on the ones that are profitable. Now I'm all for dropping unprofitable customers! But the methodology for defining who those unprofitable customers are is largely incorrect, and leaves a lot of money on the table.
In the case of the consultant, she advised that customers who fit our strategic definitions should get first everything: calls should be routed to answer those customers first, they should receive the direct mail offers, they should receive all available attention. According to the consultant, we could serve the low value customers but only if we didn't spend any additional money or effort to do so.
The traditional b-school approach is called a standard CLV model. It demands that you calculate the net present value of all revenue and marketing dollars over a period of time (usually 1-3 years, but could be as much as decades, depending on the business) for each customer, and only keep the customers who show positive net present value for the marketing investment. Makes sense, right? Frequently, no.
You see, the standard CLV model is just a snapshot in time. There is a cost to acquire a customer, and a cost to maintain a customer. Once the customer is acquired, the maintenance cost can be leveraged over long or short periods of time – it just depends on when you decide to exercise your option to drop them. So at that static moment in which the measurement is made, it may make sense to drop a whole batch of customers, based on anticipation of their behavior in the upcoming time period. But if you run the same projection in a month, or even a week, you could get a different result. I always found the standard CLV model sort of frustrating, because while I felt in my gut that something was wrong with it, I had nothing more substantive to back it up.
Recently I stumbled across a study published in the Journal of Marketing in 2006 that explains the weakness of the standard CLV approach. First, the article confirms that standard CLV models present an overall biased approach. The study authors point out that the standard CLV model assumes that unprofitable customers will continue to be unprofitable. This is true for some presently unprofitable customers, and not for others. Once a customer is cut off, the firm can no longer exercise options to benefit from the opportunity present in that customer, so use of the standard CLV approach actually undermines the opportunity to exercise options in the future. Future options could include offering presently unprofitable customers offers that were not offered previously and which would benefit that group of customers, maintaining service to those unprofitable customers at low cost, and of course, cutting off those customers at a later date.
At the heart of the standard CLV model is the idea that marketing resources are scarce and must be carefully administered. Well, all resources are ultimately scarce, though there is no doubt that marketing resources can be expensive. Because executives have a difficult time measuring and proving how effective their marketing dollars are, it feels better to cut them than to worry about spending indiscriminately.
The standard CLV model was developed in response to business culture heavily influenced by activity based costing (ABC). Developed by Robert Kaplan (who also developed the Balanced Scorecard), the method is intended as a tool for understanding product and customer cost and profitability. In a business it is rather simple to directly measure the time it takes to make, QC, pack and ship a widget, the time and cost involved in answering a phone call, the time and cost related to processing invoices, or placing a purchase order, or to hold a meeting. What is difficult to directly attribute is the value of any particular step. We businesspeople don't like ambiguity, so we lean hard on the things that can be carefully measured. Ask an engineer the measurement of one of her activities, and she can provide the answer with a tolerance measured in hundredths. Ask a marketer for a measurement of an activity, and you will get an answer that has a lot of qualifiers in it. So we create models like standard CLV, which gives us the satisfaction of certainty and metrics, and which turns the blind eye of denial on the idea of human psychology and market behavior.
The aforementioned business consultant actually had a reasonable rejoinder to my hesitance about making the less profitable customers wait on hold (in my opinion, a good business doesn't let any of its customers wait on hold). Her process accommodated an eye toward development of lower-value customers.
But in most businesses, there is an ongoing battle between what can be measured in operations – and therefore, cut in costs – and what can be measured in marketing. As we head into a recession in the next 12-24 months, there are some important marketing adjustments businesses must make. Managing a business in recession means feeding the quacking ducks – not cajoling the ducks into a new diet. But that still requires marketing, and marketing will continue to be difficult to measure. For the metrics geeks out there, I strongly recommend the article referenced here. It provides some interesting approaches to offset the inherent bias in standard CLV. Finally, let's all remember that a good handle on human psychology and an intimate sensitivity to and awareness of our customers may cause us to make decisions that the metrics can not assure us will be good decisions. Going with your gut is inadvisable when determining the safety of an electronics device. But it's an essential part of the toolbox when working with motivation and relationships.
Haenlein, M., Kaplan, A.M., & Schoder, D. (2006). Valuing the real option of abandoning unprofitable customers when calculating customer lifetime value. Journal of Marketing, 70, 16.
(c) 2008. Andrea M. Hill