Customer Metrics and Their Impact on Financial PerformanceSunil GuptaandValarie Zeithaml1August 19, 2004Revised November 30, 20051Sunil Gupta is Meyer Feldberg Professor of Business at Columbia University, New York, NY 10027(email: [email protected]) and Valarie Zeithaml is Roy and Alice H. Richards Bicentennial Professor ofMarketing and Associate Dean of the MBA program at Kenan-Flagler School of Business, University ofNorth Carolina at Chapel Hill, NC 27599 (email: [email protected]). The authors are grateful for thecomments by Eugene Anderson, Asim Ansari, Robert Blattberg, Shirish Dant, Melinda Denton, Donald R.Lehmann, Leigh McAlister, J.B. Steenkamp, Earl Taylor, Susan Toner and participants of the MSIResearch Generation Conference in Atlanta. They thank Ricardo Montoya for his research assistance.
Customer Metrics and Their Impact on Financial PerformanceAbstractThe need to understand the relationships among marketing metrics andprofitability has never been more critical. Firms experience ever-increasingpressure to justify their marketing expenditures. The objective of this paper is tointegrate existing knowledge about the impact of customer metrics on firms’financial performance. We investigate both unobserved or perceptual customermetrics (e.g., customer satisfaction) and observed or behavioral metrics (e.g.,customer retention and lifetime value). We begin with an overview ofunobservable and observable metrics, showing how they have been measured andmodeled in research. We next offer nine empirical generalizations about thelinkages between perceptual and behavioral metrics and their impact on financialperformance. We conclude the paper with future research challenges.1
Customer Metrics and Their Impact on Financial Performance1. IntroductionCustomers are the lifeblood of any organization. Without customers, a firm has norevenues, no profits and therefore no market value. This simple fact is not lost on mostsenior executives. In a worldwide survey of 681 senior executives conducted by TheEconomist during October-December 2002, 65% of the respondents reported customersas their main focus over the next three years compared to only 18% who reportedshareholders as their main focus (The Economist 2003). Oddly enough, while seniorexecutives recognize the importance of customers, they still rely heavily on financialmeasures because customer metrics are not clearly defined (Ittner and Larcker 1996).In this paper we review and integrate existing knowledge on customer metrics(e.g., customer satisfaction, retention) and provide several generalizations about theirimpact on the financial performance of firms. As marketing strives for greateraccountability, it is critical that we understand how customer metrics link to profitabilityand firm value. This paper has three objectives: (a) provide a review of key customermetrics and the measurement and modeling issues related to them, (b) highlightgeneralizable findings about the links between customer metrics and financialperformance of a firm, and (c) suggest areas for future research.Customer metrics include a variety of constructs. We categorize them intoobservable/behavioral and unobservable/perceptual measures. Observable measuresinvolve behaviors of customers that typically relate to purchase or consumption of aproduct or service. From a customer’s perspective, these include decisions of when,what, how much, and where to buy a product. From a firm’s perspective, this translatesinto decisions about customer acquisition, retention, and lifetime value. Unobservableconstructs include customer perceptions (e.g., service quality), attitudes (e.g., customersatisfaction) or behavioral intentions (e.g., intention to purchase). In economists’terminology, unobserved constructs are stated preferences while observed constructs arerevealed preferences.Intuitively, unobserved constructs are related to observed behavior which leads tofinancial gains. Satisfaction, for example, is expected to lead to repurchase behaviorwhich translates into increased sales and profits. In Figure 1, we suggest a simple2
framework to link what companies do (i.e. their marketing actions), what customers think(i.e., unobservable constructs), what customers do (i.e., behavioral outcomes) and howcustomers’ behavior affects firm’s financial performance (i.e., profits and firm value).2Most research studies on these topics either investigate relationships in one of the boxes,or at best link relationships among constructs in two of the boxes. For example, somestudies have established a link between unobservable constructs (e.g., satisfaction) andfirm value but do not consider intervening behavioral outcomes. Several researchers havealso established a direct link between marketing actions and firm’s financial performance(e.g., Joshi and Hanssens 2005) without examining antecedents in the “black box,” theterm used by many researchers for the unobserved constructs. Given the vast literature inthis field, we will focus on three links: (1) impact of unobservable constructs on financialperformance (e.g., link between satisfaction and profitability), (2) impact of unobservableconstructs on observable constructs (e.g., link between satisfaction and retention), and (3)impact of observable constructs on financial performance (e.g., link between retentionand profitability).The paper is organized to reflect relationships indicated in Figure l. In section 2,we begin by describing key unobservable customer metrics. For each construct we brieflydiscuss how it has been defined and measured. In section 3, we describe key observedcustomer metrics and the modeling issues surrounding them. Section 4 describes mainfindings from research that links unobservable metrics to financial performance.Research results about the link between unobserved and observed metrics are discussedin section 5. Section 6 discusses findings that focus on linking observed metrics tofinancial performance. In section 7, we identify unresolved issues and suggest directionsfor future research. We conclude in section 8.2. Unobserved/Perceptual Customer MetricsResearch on the concepts in the “black box” is more extensive and has a longertradition than research on the metrics outside the black box. These unobservableconcepts have been studied extensively for many reasons. First, because they arecollected almost exclusively through surveys, they have been relatively easy to obtain2Market and competitor factors are implicit in Figure 1.3
and share. Methodologies and best practices were developed both in companies and inmarketing research organizations. During the 1990s, for example, all of the majormarketing research suppliers had units or practices in customer satisfaction and theAmerican Marketing Association sponsored an annual Customer Satisfaction Congressthat often drew close to a thousand registrants from companies. Second, using thesemetrics as dependent variables allowed companies to diagnose key attribute drivers thatcould then be addressed by specific marketing and operational strategies within acompany. Third, the measures helped companies track performance over time,benchmark against competitors’ offerings, and compare performance across differentparts of an organization (e.g., branches, units, territories, countries).Of all the unobservable metrics, customer satisfaction has been the most widelystudied by researchers and used by firms because the construct is generic and can beuniversally gauged for all products and services (including nonprofit and public services).Even without a precise definition of the term, customer satisfaction is clearly understoodby respondents and its meaning is easy to communicate to managers. Other unobservablemeasures, such as service quality, loyalty, and intentions to purchase have also hadwidespread use in companies and been examined extensively in academic research.Service quality has been widely measured since the mid-80s but is not as prevalent ascustomer satisfaction because it is limited to examining the intangible aspects of anoffering. To a far lesser extent, constructs like commitment, perceived value, and trusthave made their way into company measurement systems and academic research. Otherpossible measures, such as product quality, have not been measured consistently enoughto be linked to behaviors or financial performance in studies. We focus on the metrics ofcustomer satisfaction, service quality, loyalty, and intentions to purchase in this paperbecause of their prevalence in use and maturity in measurement. For variety of reasons,we chose to eliminate perceived value, trust, and commitment from this discussion.Perceived value was excluded because it is the most ambiguous and idiosyncraticcustomer metric. While it can be defined in a general sense, operationalizing andmeasuring the construct has proven difficult. Most definitions state that perceived valueis the consumer’s objective assessment of the utility of a brand based on perceptions ofwhat is given up for what is received (e.g., Zeithaml 1988). However, this definition itself4
is so broad and vague that the construct is virtually impossible to measure with validity,reliability and consistency. In many academic and company studies, perceived value hasbeen measured with a single item or a small number of items (Bolton and Drew 1991) butthese measures leave to the customer the precise meaning of the term. Researchers havedeveloped complex conceptualizations and measures (Sirdeshmukh, Singh and Sabol2002) but these measures are not used in any consistent manner across studies and incompanies.We also eliminated commitment as a metric in this paper. Commitment is aconstruct that has been proposed as an alternative to customer satisfaction because itsignifies a stronger attachment to a product or company. Moorman, Zaltman andDeshpande (1992, p. 316) define commitment as “an enduring desire to maintain a valuedrelationship.” A small number of studies have measured commitment in business-tobusiness contexts (Gruen, Summers and Acito 2000, Morgan and Hunt 1994), consumercontexts (Verhoef, Franses, and Hoekstra 2002), and in the context of relational tiesamong channel members (Kim and Frazier 1997, Kumar, Sheer and Steenkamp 1995).Many researchers in marketing have viewed commitment as a unidimensional conceptand measured it simply, but others have elaborated dimensions and attributes (Garbarinoand Johnson 1999, MacKenzie, Podsakoff and Ahearne 1998; Morgan and Hunt 1994).The inconsistent conceptualizations, particularly among components of commitment,have led to myriad ways to measure the concept. Because the research on commitmenthas rarely been linked to the behavioral or financial variables we emphasize in this paper,we eliminated commitment from our study.2.1. Customer SatisfactionCustomer satisfaction has been defined in many different words but essentially asthe consumer’s judgment that a product or service meets or falls short of expectations.Research has typically portrayed the evaluation of customer satisfaction asdisconfirmation of expectations (see Oliver 1997 or Yi 1990 for a full review). This viewholds that a consumer compares what is received with a pre-consumption standard orexpectation.5
One of the pivotal definitional issues in the literature is whether satisfaction is bestconceived as a transaction-based evaluation or as an overall, cumulative evaluationsimilar to attitude. Traditionally, satisfaction was viewed as transaction specific, animmediate post-purchase evaluative judgment or affective reaction (Oliver 1993).Reflecting the more global perspective, studies such as Anderson, Fornell and Lehmann(1994) consider satisfaction to be an “overall evaluation based on the total purchase andconsumption experience with a good or service over time,” (page 54).Both in practice and in academic research, customer satisfaction has beenmeasured at the transaction level (as in trailer or event-triggered surveys) and at theoverall level (as in the American Customer Satisfaction Index). In early studies,academics often focused on measuring confirmation/disconfirmation and expectations,and the nature and type of expectations varied considerably from predictive expectations(Oliver 1997, Tse and Wilton 1988), to desires and experience-based norms (Cadotte,Woodruff and Jenkins 1987). Applied marketing research tends to measure satisfactionat the transaction level but more recently as an overall evaluation, a cumulative constructthat is developed over all the experiences a customer has with a firm.2.2 Service QualityPerceived service quality is the degree and direction of discrepancy betweencustomers’ service perceptions and expectations (Sasser, Olsen, and Wyckoff 1978,Zeithaml and Parasuraman 2004). While multiple interpretations of expectations haveemerged in service quality research as they have in customer satisfaction research, thenotion that service quality is a comparative process is one of the most basic building blocksin the field.The dominant measurement approach for quantitative assessment of servicequality is SERVQUAL, a multiple-item measure first developed in the 1980s, then testedand refined throughout the 1990s (see a review in Zeithaml and Parasuraman 2004).Researchers first operationalized the service quality gap as the difference between twoscores – customer expectations and perceptions of actual service performance for theperceptual attributes that respondents indicated were pivotal. Through this early researchfive dimensions of service quality were derived as factors: reliability, responsiveness,6
assurance, empathy and tangibles (Zeithaml and Parasuraman 2004). Refinement andassessment of SERVQUAL over two decades indicate that it is a robust measure ofperceived service quality. However, concerns about SERVQUAL have been raised anddebated, including the interpretation of and need to measure expectations, theappropriatene