CUSTOMERS AS ASSETS Sunil Gupta SUNIL GUPTA is Meyer Feldberg Donald R. Lehmann Professor of Business and DONALD R. LEHMANN is George E. Warren Professor of Business at f Columbia Business School, Columbia University, New York, NY 10027. All correspondence should be sent to the first author at [email protected] ABSTRACT Customers are important intangible assets of a firm that should be valued and managed. Although researchers and practitioners have recently emphasized customer relationships and customer lifetime value, these concepts have had limited impact on the business and investment community for two main reasons: (a) they require extensive data and complex modeling, and (b) researchers have not shown a strong link between customer and firm value. We address these two issues in this article. First, we show how one can use publicly available information and a simple formula to estimate the lifetime value of a customer for a publicly traded firm. We illustrate this with several examples and case studies. Second, we provide a link between customer and firm value. We then show how this link provides guidelines for strategic decisions such as mergers and acquisitions as well as for assessing the value of a firm even when the traditional financial approaches (e.g., price–earnings ratio) fail. © 2003 Wiley Periodicals, Inc. and Direct Marketing Educational Foundation, Inc. f JOURNAL OF INTERACTIVE MARKETING VOLUME 17 / NUMBER 1 / WINTER 2003 Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/dir.10045 9 JOURNAL OF INTERACTIVE MARKETING Intangible assets, and in particular, brands and of customer data as well as sophisticated models customers, are critical to a firm. On May 22, and concentrate on targeting customers with 2001, the New York Times reported that “Intan- appropriate product or communication offers. gible assets are, by definition, hard to see and While this is of great value to database market- even harder to fix a precise value for. But a ing professionals, it appears to be of limited widening consensus is growing that the impor- value to senior managers who are concerned tance of such assets—from brand names and with strategic decisions, or investors who do not customer lists to trademarks and patents— have access to internal company data. Second, means that investors need to know more about few attempts have been made to link customer them.” This interest in intangibles arises from value to the value of the firm—a link that is the recognition that market value of the largest essential if investors are to view customers as 500 corporations in the United States is almost assets. six times the book value (the net value of phys- In this article we address these two shortcom- ical and financial assets as stated on the balance ings. First, we show how one can use publicly sheet). In other words, of every six dollars in the available information to estimate the lifetime market value of a firm, only one dollar is rep- value of a customer for a publicly traded firm. resented in the balance sheet (Lev, 2001). We illustrate this with examples and case studies Although brands have been widely heralded and show how it can be useful for a variety of as important assets for a firm (Aaker & Davis, managerial decisions. Second, we provide a link 2000) and organizations such as Interbrand between customer and firm value. We show how routinely evaluate them, the use of customers as this provides a useful guideline for strategic assets has been limited. On one hand, scores of decisions such as mergers and acquisitions. We books and hundreds of articles have argued also show that this approach provides useful about the importance of creating a customer- guidelines for assessing the value of a firm even centric organization (Seybold, 2001). Further- when the traditional financial approaches (e.g., more, the abundance of customer information price–earnings ratio) fail, as in the case of In- and increasingly sophisticated information ternet firms that had negative earnings. technology and statistical modeling have led to a revolution in areas such as customer relation- ship management or CRM (Winer, 2001). Yet LIFETIME VALUE OF A CUSTOMER some studies show that while investors implicitly Customer lifetime value (CLV) is the present capitalize product development and R&D ex- value of all future profits generated from a cus- penditures, they expense marketing and cus- tomer. One common approach is to assume we tomer acquisition costs (Demers & Lev, 2001). know how long a customer will be with a firm In recent years, the marketing literature has and then generate a discounted cash flow for developed and discussed the concept of cus- that time period (Berger & Nasr, 1998; Blatt- tomer lifetime value, which is the present value berg et al., 2001; Jain & Singh, 2002): of all future profits generated from a customer (e.g., Berger & Nasr, 1998; Blattberg & Deigh- n ton, 1996; Blattberg, Getz, & Thomas, 2001; Jain mt CLV ϭ ͸ (1) & Singh, 2002; Rust, Zeithaml, & Lemon, 2000). ͑1 ϩ i͒t Arguments for treating customers as assets that tϭ1 generate future profits, however, have had lim- ited impact on the business and investment where mt is the margin or contribution for each community for two main reasons. First, the con- customer in a given time period t (e.g., a year), cept and models of customer lifetime value orig- i is the discount rate, and n is the period over inated in the field of direct and database mar- which the customer is assumed to remain active. keting and continue to focus in this domain. This formulation assumes that a customer stays Many applications require an enormous amount with a firm for n periods with certainty. In gen- JOURNAL OF INTERACTIVE MARKETING ● VOLUME 17 / NUMBER 1 / WINTER 2003 10 CUSTOMERS AS ASSETS eral, a customer has a probability to switch or However, some recent studies show that prof- defect from the firm in any time period. While its for a customer may not necessarily increase it is possible to model switching among multiple over time (Reinartz & Kumar, 2000). Even if the states, for example using a Markov chain margin for a particular customer increases, the (Pfeifer & Carraway, 2000), we follow Dwyer customer mix for a firm also changes over time. (1997) to simplify the analysis for two customer In general, a firm starts by attracting customers states: active or inactive. If rj is the probability of who are most favorably disposed toward to customer retention in period j, the probability firm’s products and services. As the company that a customer is still an active member of a expands its customer base, it tends to draw ⌸t firm at the end of time period t is jϭ1 rj. more and more marginal customers who do not Therefore, equation (1) is modified as spend as much money with the company as the original customers. Consequently average reve- t nue per customer declines over time. This is n ͹ m ϭ r t j 1 j especially true if the company’s customer base CLV ϭ ͸ (2) ͑1 ϩ i͒ expands very rapidly and if the company is ei- tϭ1 t ther a single product company or a company that does not emphasize cross selling. For ex- This seemingly simple formulation is quite data ample, at CDNow revenue per customer fell intensive, requiring per period margins and re- from $23.15 to $21.16 in 1998. In the first quar- tention rates. In addition, it also leaves n, the ter of 1999, it acquired a competitor, N2K, that length of projection period, to be determined further contributed to the decline in its revenue subjectively or by industry norms. We therefore per customer from $18.15 in Q1 of 1999 to modify the above formulation by making the $14.42 in Q2 of 1999. following assumptions: (a) margins are constant Using publicly available data, such as finan- over time, (b) retention rate is constant over cial statements, we estimated quarterly margin time, and (c) the length of the projection pe- per customer for Capital One and Ebay by di- riod is infinite. As we will show shortly, these viding the total gross margin by the number of assumptions allow us to create a very simple rule current customers in that quarter. Figure 1 of thumb to assess customer lifetime value with shows that there is no systematic pattern or time minimal and generally available information. trend for margins. A regression analysis con- Before discussing this, we provide partial justi- firmed this view. fication for our assumptions. Constant Average Margins (m) Constant Retention (r) The average margin for each customer is simply It is possible for the retention probability of a annual revenue minus operating expenses di- customer to change every time period. For ex- vided by the number of customers. Over time, ample, as a customer stays longer with a firm, he there are two opposing forces that shape aver- may become more loyal and therefore have a age margins from customers. On the positive higher retention probability. At the same time, side, as customers stay longer with a company increasing competitive activity can reduce cus- and become more comfortable doing business tomer loyalty. In fact, recent research has ar- with a firm, they may buy more, generating a gued that escalating loyalty programs may cre- larger revenue stream over time. The company ate a “prisoners’ dilemma” and raise the cost of also has the potential of cross-selling its prod- competing firms without affecting customer loy- ucts to its customer base. In addition to in- alty (Shaffer & Zhang, 2001). creased revenue, in general the longer a cus- Practically, retention rate is one of the most tomer stays with a company, the lower is the cost difficult metrics to empirically estimate.