Customer Lifetime Value & Churn
All Customers Are Not Created Equal – Calculating CLV the Right Way...

Many companies have realized that they should not treat all customers the same. In the banking world, customers who frequently use teller services have been discouraged from doing so and have been prompted to use the ATM more often. On the other hand, gaming companies and airlines have long provided extra perks to their highest rollers and their most frequent travelers. This is not to say that companies should mistreat customers, but only that by identifying and investing in their best customers can companies increase loyalty, reduce the likelihood of defection and, most important of all, increase profits.

Additionally, in some cases, companies will "give away" one product with the knowledge that long-term purchasing of an add-on or complementary product will yield positive cash flow. In this case, the identification of the best customers becomes critical to sustaining a viable business. Amazon's Kindle and Barnes and Noble's Nook are both examples of a base product sold at a discount with the hope that customers who purchase the hardware will yield a long revenue stream from electronic book purchases. The best customers at both of these companies often receive free or discounted shipping rates.


Customer Lifetime Value

The lifetime value of a customer refers to the total revenues generated by an individual customer over time from all purchases of the company's products and services, discounted by the time value of money. CLV is critical both from the standpoint of marketing investment and for corporate valuation. Since resources are always limited, understanding how much it costs to acquire a customer and how much revenue they will generate allows companies to allocate resources toward those customers who are likely to be the most profitable and to employ the tactics that best suit them. Extending that analysis to the full customer base, companies can generate a measure of the company's worth that can then be used by analysts as one metric to help quantify the overall value of a company.

Complicating the calculation of CLV are three key issues:

  • Are customers contractually bound to the company for a period of time, as telecommunication customers might be, or can they come and go as they please, like supermarket shoppers?
  • Are purchases discrete and limited, as might be found in the purchasing of tickets in a concert series, or can purchases occur at any time, like movie rentals from Netflix?
  • Are the company's financial and IT systems up to the task of collecting, analyzing, and reporting information at the level of the individual customer?

The combination of contractual commitment and usage occasions yields four different situations:

  1. Contract with continuous usage (e.g., mobile phones);
  2. Contract with discrete usage (e.g., baseball season tickets);
  3. No contract with continuous usage (e.g., credit card use); and,
  4. No contract with discrete usage (e.g., attending a play).

Peter Fader of the Wharton School at the University of Pennsylvania and David Schweidel of the Goizueta Business School at Emory University, along with their co-authors, have demonstrated that each of these four situations requires different data and a unique calculation of CLV.

Moreover, many companies' IT systems may not be set up to collect and store data at a fine level of granularity. We know of one fast food chain, for example, that collects all purchase information and then aggregates the information to the weekly level for each location, simply because its systems are not engineered to store purchases at the most elemental level. Other companies store the most disaggregate information, but only for a short period of time, after which they are aggregated and archived. Fortunately, as the cost of disk storage has decreased, it is now becoming possible for companies to collect, store, and analyze this fine-grained information.


Individual-Level Customer Base Analysis

Fader (on our Board of Science Advisors), in his book Customer Centricity Essentials: What It Is, What It Isn't, and Why It Matters (2011), details why companies should pay more attention to CLV. Moreover, while an advocate of aggregate or segment-level calculations of CLV, Fader makes a strong case for customer-level calculations of CLV, which, by definition, will explicitly incorporate the important notion of variability (heterogeneity) across customers.

By explicitly recognizing that not all customers are created equal, and that they have different lifetime values, CLV must be calculated at the level of the individual customer.

in4mation insights has worked with Fader, Schweidel, and others to operationalize all of their CLV models within the hierarchical Bayesian context so as to provide robust, accurate, state-of-the-art metrics for every customer.

In doing so, we provide, among others:

  • An estimate of the lifetime value of every customer;
  • An estimate of the Residual Value of every customer (given the CLV, how much of that amount is left to be gained in the future?);
  • Prediction of the likelihood and timing of future churn/defection;
  • Quantification of the effects of marketing programs and tactics on CLV and churn; and,
  • The ability to understand all of the above by any characteristic or combination of characteristics of the customer at the deepest, drill-down level.