Using CLV to Make Decisions: IBM

Let’s look at some applications of CLV. What we’re going to look at is a project I was involved with, IBM, on allocating sales force using CLV. The context was about allocating sales force for IBM’s enterprise software customers in the mid-market segment. So IBM always uses sales calls to call on its customers, check on them and see whether everything is going okay, whether they need some new software, what their needs are, how can they help them, help their customers grow better.

So during one of these sessions, they decided to evaluate how profitable their customers are and the way they did that was in collaboration with me and some other folks, how to look at the profits of IBM’s customers and we decided to use CLV. So what you’re seeing here in this chart is customers split into 10 deciles, 1-10. And then, what we did was, the customers in the decile one was people with super high CLV. They were really high in the value they were providing IBM. In decile two, we classified them as high CLV. Decile 3-7 was medium CLV and the rest of them were classified as low CLV.

One of the interesting things that came about was, when we split these customers in these 10 deciles further into whether they received a sales call or not before 2004 and whether they received a sales call. So what was really interesting to see was, if you see here there are a few customers in decile 10 who received a sales call but we’re still making a loss. But at the same time, there were these customers up here in decile one who had no sales call until 2004. So this was eye opening for IBM to see that they had so many customers who are actually very profitable that IBM had never contacted before.

So what would happen if they actually moved some of the sales calls from the customers who are loss making to the customers who were in the super high CLV segment but IBM had never called before? So this is what we did. We moved in 2005 sales calls from these low CLV segments, that IBM had been calling before and they were still making losses, onto the super high CLV segment customers that IBM had not made a call at all before. So let’s see what happened after we did this. So you see here no sales calls until 2004, that’s going to be the blue graph, and sales calls in 2005, that’s going to be the green graph.

Then in 2005, the customers who were in the super high CLV segment. The super high CLV segment. Who had not received a sales call before, that was this segment. When they were called in 2005, their sales went up 10x and the number of purchases they made went up 60%. So using CLV to identify these customers and reallocating sales force to customers that have the potential for high profits in the future paid big dividends for IBM. They were now able to increase their profits just in this group by about 19 million. So this example shows how CLV can be used to make marketing actions more effective.