Jim answers questions from fellow Drillers
(More questions with answers here, Work Overview here, Index of concepts here)

Topic Overview

Hi again folks, Jim Novo here.

Often we spend a lot of time talking about analyzing “customer data”, and the implication is we are looking at marketing or sales related information. That may be true for companies just beginning to use customer data; this data often is the easiest to understand and access. But true data-driven organizations have analysts who reach across the silos for data, looking for customer service or operational customer data that can impact the current and potential value of the customer. We have one such example from a Driller today.

Sound good? Then let’s do some Drillin’!

Q:  I work as a management consultant, currently working in a project where my client (Oil & Gas company) is trying to calculate and implement Lifetime Value into one of their businesses.  One of their business units (Industrial Lubricants) sells different kinds of lubricants and services to corporate customers such as Ford, Toyota, BMW, etc.  They have already done some customer profitability analysis and they are currently trying to calculate Lifetime Value.

A:  That’s a pretty interesting place to find a concern for analyzing LTV…

Q:  My questions:

1. What’s the best way to forecast future cash flows in a B2B scenario where models such as RFM are not relevant (Recency and Frequency do not really apply given that their customers have been with them for ages and are often in long-term contracts).  How can I project customer profit over time and how can I estimate the “lifetime” of individual customers?

A:  Well, it’s not that Recency and Frequency don’t apply, they probably apply in a different way.  In most businesses driven by contracts, service is the issue.  So you need to look for Recency and Frequency of “problems”, whatever that might mean in the industry.  I imagine “logistics” is an issue for these businesses – on time delivery, quality, “ease of use” (which could cover many factory / service issues), packaging, and so forth. This can take a lot of research, particularly if there are no “systems” capturing this kind of data. But usually, even in very old line companies, there is some place where this data resides. You just have to find it and get access to it.

Often in an environment like this it is easier to work backwards – first, identify defectors, then look for service issues or changes in behavior that imply service issues – declining order size / Frequency, expanding order Latency (weeks between orders) and so forth.