Recently, I came across a very useful Marketing Science article on customer lifetime value calculations. “Customer Base Analysis in a Discrete-Time Non-contractual setting,” presents an excellent and accessible way of answering the difficult question of customer lifetime value when no ongoing contractual relationship exists.
The paper itself was enlightening. Unfortunately, as is often the case with academic papers, the key business insights are hidden from plain sight behind a rather dense, high order, statistics heavy paper. I am sure their findings are as clear as glass to the authors. However, for me, I needed some time to unpack the content, understand it, and work out how to apply it for my customers.
In the process of reviewing the paper however, I discovered a rather extensive array of content that the authors have been building for the last few years in several areas of customer prediction in contractual and non-contractual customer relationships. The authors have documented very valuable set of approaches that are accessible to a range of organizations.
As a starting point, in their paper: Probability Models for Customer-Base Analysis the authors Fader, and Hardie suggest a very useful way of classifying customer bases, shown below;
Instead of looking at only contractual and non-contractual customer relationships, their approach also includes an axis for Opportunities for Transaction. This two dimensional approach to the classification of customer relationship is the starting point for the authors to explore each of the quadrants. Over a period of several years and quite a number of theoretical papers, practical notes and example spreadsheets the authors systematically create methods for estimating customer lifetime value, and other key customer variables for each of the quadrants in this chart.
While each of the papers has a heavy dose of statistics, the statistics can be implemented in Excel spreadsheets making the overall approaches accessible to many organizations.
Over the next few weeks I will be taking each of the quadrants and providing a practical summary of the authors approaches and the application of those approaches.
In the next installment of this series I will look at predicting future purchasing patterns in Customer Continuous/Non-contractual customer bases (the top-left quadrant). If you are a retailer or similar organization this will show you how to predict customer actions without having to hire your own statistics department.
Other posts in this series
- Forecasting customer value when you don’t have a contract: Discrete transactions: for use in non-contract settings where the customer can only purchase at discrete intervals, e.g. annual conferences.
- Calculating Retail Sales Forecasts, Customer Life Time Value, and other customer variables: for non-contract settings with continuous purchase opportunities (retailers, etc)