11 1998 Itner Lacker Nonfinancial Measure

Journal of Accounting ResearchVol. 36 Supplement 1998

Printed in U.S.A.

Are Nonfinancial MeasuresLeading Indicators of Financial

Performance? An Analysis ofCustomer Satisfaction

CHRISTOPHER D. ITTNER AND DAVID F. LARCKER*

1. Introduction

This paper examines three questions on the value relevance of cus-tomer satisfaction measures: (1) Are customer satisfaction measures lead-ing indicators of accounting performance? (2) Is the economic value ofcustomer satisfaction (fully) reflected in contemporaneous accountingbook values? And (3) Does the release of customer satisfaction mea-sures provide new or incremental information to the stock market? Manyargue that improvements in areas such as quality, customer or employeesatisfaction, and innovation represent investments in firm-specific assetsthat are not fully captured in current accounting measures. Accordingto these authors, nonfinancial indicators of investments in "intangible"assets may be better predictors of future financial (i.e., accounting or stockprice) performance than historical accounting measures, and should sup-plement financial measures in internal accounting systems (e.g., DeloitteTouche Tohmatsu International [1994] and Kaplan and Norton [1996]).

•University of Pennsylvania. The support of KPMG Peat Marwick, Ernst & Young, andthe Citibank Behavioral Science Research Council is gratefully acknowledged. We alsothank Pamela Cohen, John Core, Neil Fargher, Robert Kaplan, Richard Lambert, MichaelMaher, Richard Willis, an anonymous reviewer, and workshop participants at HarvardBusiness School, the 1996 Stanford Summer Camp, and the 1998/ourna/ of Accounting Re-search Conference for comments on earlier drafts.

Copyright ©, Institute of Professional Accounting, 1999

2 ENHANCING THE FINANCIAL REPORTING MODEL: 1998

This same discussion has produced calls for disclosure of nonfinancialinformation on the drivers of firm value (e.g., Wallman [1995], Edvinssonand Malone [1997], and Stewart [1997]). A report by the American In-stitute of Certified Public Accountants [1994, p. 143], for example, con-cluded that companies should disclose leading, nonfinancial measureson key business processes such as product quality, cycle time, innovation,and employee satisfaction.

One nonfinancial measure emphasized in these discussions is cus-tomer satisfaction. We examine the value relevance of customer satis-faction measures using customer, business-unit, and firm-level data. Thecustomer-level tests provide evidence on the fundamental assumptionthat future-period retention and revenues are higher for more satisfiedcustomers, making customer satisfaction measures leading indicators ofaccounting performance. The business-unit tests extend the analysis byexamining the cost and profit implications of customer satisfaction, aswell as spillover effects such as growth in customers due to positiveword-of-mouth advertising and enhanced firm reputation. The business-unit tests also allow us to investigate the ability of typical business-unitsatisfaction measures (which are based on aggregated responses from asmall sample of customers) to predict accounting performance and cus-tomer growth. Finally, the firm-level valuation tests and event study ex-amine whether customer satisfaction measures provide information tothe stock market beyond the information contained in current account-ing book values.

We find that the relations between customer satisfaction measures andfuture accounting performance generally are positive and statisticallysignificant. However, many of the relations are nonlinear, with someevidence of diminishing performance benefits at high satisfaction levels.Customer satisfaction measures appear to be economically relevant tothe stock market but are only partially reflected in current accountingbook values. We also find that the public release of these measures isstatistically associated with excess stock market returns over a ten-dayannouncement period, providing some evidence that the disclosure ofcustomer satisfaction measures provides information to the stock marketon expected future cash flows.

Section 2 reviews the literature on the measurement and performanceconsequences of customer satisfaction. Section 3 examines the relationbetween customer satisfaction indexes and subsequent purchase behav-ior of individual customers. Section 4 presents our business-unit tests,followed by firm-level tests in section 5. Section 6 concludes the paper.

2. Literature Review

Calls for greater emphasis on nonfinancial customer satisfaction mea-sures are motivated by the perceived absence of information on oneof the key drivers of firm value. The marketing literature contends that

NONFINANCIAL MEASURES AS INDICATORS OF PERFORMANCE 3

higher customer satisfaction improves financial performance by increas-ing the loyalty of existing customers, reducing price elasticities, lower-ing marketing costs through positive word-of-mouth advertising, reducingtransaction costs, and enhancing firm reputation (e.g., Anderson, For-nell, and Lehmann [1994], Fornell [1992], and Reichheld and Sasser[1990]). These advantages are believed to persist over time, suggestingthat the net benefits from investments in customer satisfaction may notbe fully reflected in contemporaneous accounting performance (Ander-son, Fornell, and Lehmann [1994]). However, achieving higher customersatisfaction is not without cost. Economic theories argue that customersatisfaction (i.e., customer utility) is a function of product or serviceattributes. Increasing customer utility requires higher levels of theseattributes and additional cost, particularly at higher satisfaction levels(Lancaster [1979] and Bowbrick [1992]). Likewise, traditional opera-tions management theories maintain that the investments needed to im-prove product or service quality increase exponentially at high qualitylevels (e.g., Juran and Gryna [1980]). Thus, improvements in customersatisfaction may exhibit a diminishing, or even negative, relation to cus-tomer behavior and organizational performance.

Despite lack of agreement on the specific association between customersatisfaction and financial performance, most firms track some form ofcustomer satisfaction measure (Ross and GeorgofF [1991]). These mea-sures are inputs for improvement programs, strategic decision making,and compensation schemes. Ernst & Young [1991] found that customersatisfaction measures were of major or primary importance for stra-tegic planning in 54% of the surveyed organizations in 1988 and 80% in1991, and were expected to be of major or primary importance in 96%by 1994. Ittner, Larcker, and Rajan [1997] found that 37% of firms usingnonfinancial measures in their executive bonus contracts include cus-tomer satisfaction measures, while William M. Mercer, Inc. reported that35% of firms use customer satisfaction measures in determining compen-sation and another 33% planned to do so {HR Focus [1993]).

A survey of vice presidents of quality for major U.S. firms, however,found that only 28% could relate their customer satisfaction measuresto accounting returns and only 27% to stock returns (Ittner and Larcker[1998]). Similarly, a survey by Arthur Andersen & Co. [1994] indicatedthat the top-two problems in implementing customer satisfaction initia-tives were: (1) linking customer satisfaction and profitability, and (2)understanding the point of diminishing returns for customer satisfac-tion initiatives. The accounting firm's study of the food, toys/games, air-lines, and automotive industries also found little systematic relationbetween customer satisfaction levels and profitability, leading them toconclude that "the assumption that profits flowed inevitably from cus-tomer satisfaction simply didn't hold up" (Arthur Andersen & Co. [1994,p. 1]). In contrast, Anderson, Fornell, and Lehmann's [1994] study ofthe performance consequences of customer satisfaction in 77 Swedish

4 CHRISTOPHER D. ITTNER AND DAVID F. LARCKER

firms supported the hypothesis that customer satisfaction is positively as-sociated with contemporaneous accounting return on investment, aftercontrolling for past return on investment and a time-series trend.Banker, Potter, and Srinivasan [1998] also found that customer satisfac-tion measures were positively associated with future accounting per-formance in 18 hotels managed by a hospitality firm. Foster and Gupta[1997], however, found positive, negative, or insignificant relations be-tween satisfaction measures for individual customers of a wholesale bev-erage distributor and future customer profitability, depending on thequestions included in the satisfaction measures. Anderson, Fornell, andLehmann [1997] found positive contemporaneous associations betweencustomer satisfaction and return on investment in Swedish manufactur-ing firms, but weaker or negative associations in service firms.

Mixed evidence also exists on the extent to which customer satisfac-tion measures provide value-relevant information beyond that containedin current accounting statements. Using surveys and revealed prefer-ence experiments, Mavrinac and Siesfeld [1997] found that institutionalinvestors ranked customer satisfaction indexes only eleventh most usefulamong nonfinancial measures, and that participating investors put noweight on customer satisfaction measures when valuing companies.

Related research by Aaker and Jacobson [1994] examined the associ-ation between stock returns and customers' perceptions of brand quality.Using data on 34 brands included in the EquiTrend survey by TotalResearch Corporation, the authors regressed stock returns during the14-month period prior to the survey on "unexpected" accounting returnon investment, "unexpected" quality, and "unexpected" brand awareness.^Aaker and Jacobson found a positive association between perceived brandquality and stock returns after controlling for unexpected accountingreturns. Since the stock price returns preceded the measurement of per-ceived quality, their results suggest that the market (at least partially) im-pounds customer perceptions of brand quality into stock price. However,the use of prior-period stock returns provides no evidence on whetherperceived brand quality is a forward-looking indicator of economic perfor-mance.

In summary, prior empirical studies provide mixed evidence on the re-lation between customer satisfaction indexes and financial performance,and no evidence on whether there are diminishing or negative returns tocustomer satisfaction. More importantly, prior research offers no sup-port for claims that customer satisfaction measures provide incrementalinformation to the stock market on the firm's future financial prospects.

' The "unexpected" components of these measures represented the residuals from afirst-order autoregressive model pooling 102 time-series and cross-sectional observationsfrom each series. The accounting return on investment figures related to the fiscal year-end occurring during the 14 months prior to the survey period.

NONFINANCIAL MEASURES AS INDICATORS OF PERFORIVIANCE 5

3. Customer-Level Analyses

Our initial analyses examine whether current satisfaction levels for in-dividual customers are associated with changes in their future purchasebehavior and firm revenues. One of the fundamental assumptions ofcustomer satisfaction measurement is that higher satisfaction levels im-prove future financial performance by increasing revenues from existingcustomers (due to higher purchase quantities and lower price elastici-ties) and improving customer retention. We examine the effects of cus-tomer satisfaction on the purchase behavior of existing customers usingdata from a major telecommunications firm. This analysis provides aninitial test of customer satisfaction measures' ability to predict future ac-counting performance and is similar to procedures used by firms to de-velop new marketing strategies and plans for individual customers.

The telecommunications firm has approximately 450,000 customersfor this service, which typically is sold to small businesses competing inlocal markets. In 1995, the average customer had sales of $230,000 (me-dian = $175,000) and had been in business 8 years (median = 10 years).The mean (median) customer purchased approximately $3,000 ($1,150)in services during 1995. The firm faces a number of national and re-gional competitors for this service, which is not regulated. To increaserevenues from existing customers and attract new customers, new orenhanced services are introduced each year. The firm considers themeasurement of customer satisfaction and the identification of its de-terminants to be key inputs into their quality and customer service initi-atives and overall corporate strategy. Given the characteristics of thisservice and the firm's emphasis on customer satisfaction, we expect theirsatisfaction measures should predict future-period customer behaviorand revenue.

The firm measured customer satisfaction for a random sample of2,491 business customers buying a specific service in 1995. The cus-tomer satisfaction index (CSI) is based on three questions assessing: (1)overall satisfaction with the service (from 1 = not satisfied at all to 10 =extremely satisfied), (2) the extent to which the service had fallen shortor exceeded customer expectations (from 1 = has not met expectationsto 10 = exceeded expectations), and (3) how well the service comparedwith the ideal service (from 1 = not at all ideal to 10 = absolutely ideal).The index is constructed using Partial Least Squares (PLS) to weight thethree items such that the resulting index has the maximum correlationwith expected economic consequences^ (customers' self-reports of recom-mendations, repurchase intentions, and price tolerance^). The resulting

2 See Wold [1973; 1982] and Fornell and Cha [1987] for detailed econometric discus-sions of PLS.

^Recommendation ranges from 1 = not recommend to others to 10 = strongly recom-mend to others. Retention ranges from 1 = not at all likely to continue using this service


scores are rescaled to range from 0 (least satisfied customer) to 100(most satisfied customer). Mean (median) CSIin 1995 was 62.3 (66.7).^

We assess future purchase behavior using 1996 retention rates andrevenue, and percentage changes in revenues between 1995 and 1996.Retention rates allow us to test claims that more satisfied customers are lesslikely to move to a competitor or stop using the service. The revenue leveltests examine whether more satisfied customers purchase more of the ser-vice than less satisfied customers. Finally, the revenue change tests examinewhether customers at higher satisfaction levels increase purchases morethan those at lower levels. The telecommunications firm has attemptedto increase revenues from existing customers by introducing new serviceofferings and enhancing existing offerings. If it is easier to cross-sell newproducts to more satisfied customers or to upgrade them to more ex-pensive services, revenue growth should be positively associated with sat-isfaction levels. However, if highly satisfied customers tend to buy moreservices but already had filled their requirements by 1995, revenue levelsfor this set of customers may be higher but revenue growth may be zero.^

Customer retention is coded one for 1995 customers who purchasedservices again in 1996, and zero otherwise (660 customers were not re-peat purchasers). The one-year lag is typical in this business because cus-tomers sign annual contracts. Revenue was measured in 1995 and again in1996, with percentage revenue changes defined as [(1996 revenue/1995revenue) - 1]. The revenue change for lost customers is -100%. Obvi-ously, many factors other than satisfaction levels may influence customerpurchase behavior; we are limited by data availability to two additionalcontrol variables. Because larger firms are more likely to purchase moreservices, we control for size using the customers' sales in 1995 (denotedSITE). We also control for the number of years the customer has been in

to 10 = extremely likely to continue using this service. Price tolerance is based on threesimilar questions asking whether the customer is likely to continue using this service ifprices increased by 15%, 10%, and 5% (1 = not at all likely to continue using this serviceand 10 = extremely likely to continue using this service).

* One critique of studies using customer satisfaction measures such as these is that themeasures are ordinal rather than cardinal. Although a valid criticism, we are attempting toprovide evidence on whether the types of customer satisfaction measures used in practicefor decision-making, compensation, and disclosure purposes are associated with subse-quent financial performance, despite limitations in their measurement properties.

^ We assume that revenue growth primarily captures additional sales to customers whoremained at a given satisfaction level, rather than customers who increased revenues be-cause they moved to a higher satisfaction level between 1995 and 1996. This interpretationis consistent with marketing research which finds that customer satisfaction levels are fairlystable over time (Anderson, Fornell, and Lehmann [1994]). However, because we only havecustomer satisfaction measures for a single period (i.e., individual customers typically arenot surveyed in multiple years), the revenue growth measure will capture both economiceffects.

NONFINANCIAL MEASURES AS INDICATORS OF PERFORMANCE

TABLE 1OLS Regressions Examining the Association between 1995 Customer-Level

Satisfaction Scores and 1996 Customer Retention, Revenues, and Revenue Change for2,491 Business Customers of a Telecommunications Firm^

(t-Statistics in Parentheses)

Intercept

CSI

AGE

SIZE

Adjusted R^F-Statistic

Retention0.482***

(13.41)

0.002***(6.16)

0.013***(3.99)

0.000(0.39)

0.02119.04***

Revenue-535.29

(-1.38)

19.464***(4.92)

48.137(1.34)

0.003***(9.85)

0.04943.36***

Revenue Change-0.447***

(-8.89)

0.003***(5.74)

0.004(0.90)

0.000(1.44)

0.01312.07***

*** Statistically significant at the 1% level (two-tail)."A customer in 1995 is defined as retained in 1996if that customer also purchased the service in 1996.

Revenue change is defined as [(1996 revenues divided by 1995 revenues)- 1]. Customers that were notretained are given a revenue change score of -1.0. 1996 revenue from customers that were not retainedis set to zero. Customer satisfaction (CSI) scores range from 0 (least satisfied) to 100 (most satisfied).AGE is the number of years the customer has been in business. SIZE is the customer's total revenue.

business (denoted AGE) to account for the high rate of business failuresin young firms.^

Linear regressions examining the associations between 1995 CSI scoresand customer retention, revenue levels, and revenue changes in 1996 arereported in table 1. All three models are significant {p < 0.001, two-tail),with adjusted Rh ranging from 1.3% to 4.9%. This low explanatory powersuggests that customer satisfaction is only one of many factors influenc-ing customer purchase behavior in this segment of the telecommunica-tions industry. For example, the small business customers surveyed arelikely to exhibit volatile and unpredictable cash flows, making it difficultto forecast purchase behavior one year into the future. Therefore, it isimportant to benchmark our results against the inherent difficulty offorecasting customer behavior in this setting.

The point estimates for the regression coefficients, however, are eco-nomically significant. 1995 CSI was positively related to customer re-tention, revenues, and revenue changes in 1996 {p < 0.001, two-tail),supporting claims that customer satisfaction measures are predictive ofsubsequent customer purchase behavior.' The coefficients imply that a

also included measures for the customers' metropolitan statistical area (MSA) tocontrol for regional differences in economic environments, competition, etc. These mea-sures were not statistically significant and are excluded from the reported tests.

'We also estimated the retention model using logit. The results were virtually identicalto those using OLS.


ten-point increase in CS/was associated, on average, with a 2% increasein retention, a $194.64 revenue increase, and 3% higher revenue change.Revenues also increased with customer size and retention with customerage, but neither control variable was significantly associated with revenuegrowth.

So far, we have assumed a linear association between satisfaction andcustomer purchase behavior. The large sample of customers allows usto test for potential nonlinearities in these associations. The nonlinearfunctions linking retention, revenue levels, and revenue changes to sat-isfaction are developed using additive nonparametric regression with vari-ance stabilization (S-Plus [1991, chap. 18]). This method, an extensionof the transformation procedures developed by Box and Cox [1964], fitsan additive nonlinear regression model to the criterion and predictorvariables. The nonlinear transformations of the variables (selected usingthe supersmoother procedure^) are selected to maximize the correlationbetween the transformed criterion and predictor variables such that theresidual variance of the transformed criterion variable is constant.

The nonparametric functions linking 1995 CS/levels to 1996 customerretention, revenue levels, and revenue changes are presented in figures1-3, respectively. The figures plot the predicted values of the depen-dent variables (selected using the optimal nonlinear transformation ofCSI) versus actual CSI.^ Regressions of the dependent variables on theirnonlinear transformations of CSI yield adjusted i?s for the retention,revenue, and revenue change models of 1.72%, 0.90%, and 1.40%, re-spectively (p < 0.001, two-tail), and ^statistics for the associated regres-sion coefficients of 6.68, 4.85, and 6.04, respectively {p < 0.001, two-tail).

Figure 1 indicates that over much of the CS/range, average 1996 reten-tion was increasing in 1995 CSI. For example, a customer with a CSI of30 in 1995 (on a 0 [least satisfied customer] to 100 [most satisfied cus-tomer] scale) had a 64% retention rate, while a customer with a CSI of60 had a 75% retention rate. The plot shows a distinct increase in reten-tion at a CSI of about 67, while scores above 70 produced no increase inretention rates. Over 25% of customers were above this score, whichsuggests that investments to increase the satisfaction of a large propor-tion of the customer base would yield little change in retention.

The revenue levels function in figure 2 shows a nearly linear relationbetween 1995 CS/and 1996 revenue. A movement in CS/from 40 to 60,

each observation, the supersmoother procedure estimates a linear regression us-ing data on each side of that point (i.e., ft-nearest neighbors). The smoothed value is esti-mated using the regression coefficients and the actual x value for the observation. The sizeof the span (i.e., k) is selected using a complex cross-validation technique that minimizesthe mean square error between actual y and the smoothed value of y. See S-Plus [1991, 18-40-18-44] for additional details.

^The plots in figures 1—3 do not control for AGE and SIZE. Plots using residuals fromregressions of the three dependent variables on AGE and SIZE had very similar shapes.

M

U


0.8

0.75

0.7

I 0.65

0.6

0.55

y

/./

>

20 40 60 80

Customer Satisfaction Index in 1995

100

FIG. 1.—Retention analysis for business customers of a major telecommunications firm(n = 2,491). The nonlinear function linking retention to satisfaction is developed using ad-ditive nonparametric regression using variance stabilization. The nonlinear transformationsof the variables (selected using the supersmoother procedure, where the span is chosenusing local cross-validation) are selected to maximize the correlation between the trans-formed criterion and predictor variables such that the residual variance of the transformedcriterion variance is constant. Figure 1 plots the value of retention predicted using the opti-mal nonlinear transformation of customer satisfaction versus actual customer satisfaction. Acustomer in 1995 is defined as retained if that customer also purchased the service in 1996.

for example, increased predicted customer revenue by roughly $400 peryear. Like the retention results, the revenue function also shows a dis-tinct "step" at a CSI of about 70. Because this service can be purchasedin several difFerent "sizes," the revenue step suggests that this CSI thresh-old explains customer moves to a larger service offering. In contrast toretention rates, predicted revenue levels continued to increase until CSIwas maximized at 100, although the predicted revenue difference be-came progressively smaller. For example, a six-point CSI difference wasassociated with a predicted revenue difference of $74.80 per year whenmoving from a CSI score of 88 to 92, $37.52 from 92 to 96, and $25.81from 96 to 100.

Figure 3 displays the function linking revenue changes to the previousyear's CSI. Predicted revenue changes are negative throughout the rangeof CSI scores, because of the loss of existing customers (which was morethan offset by the addition of new customers) at all satisfaction levels,with the negative values driven by the -100% change in revenue assigned


3000

«, 2800

? 2600

"oO

Iu.

•a£a.

2400

2200

2000

1800

1600

1400

1200

1000

>: ^'

.z '

^^

i

20 40 60 80

Customer Satisfaction index in 1995

100

FIG. 2.—Revenue analysis for business customers of a major telecommunications firm(re = 2,491). The nonlinear function linking revenue dollars to satisfaction is developedusing additive nonparametric regression using variance stabilization. The nonlinear trans-formations of the variables (selected using the supersmoother procedure, where the span ischosen using local cross-validation) are selected to maximize the correlation between thetransformed criterion and predictor variables such that the residual variance of the trans-formed criterion variance is constant. Figure 2 plots the value of revenue predicted usingthe optimal nonlinear transformation of customer satisfaction versus actual customer satis-faction. The 1996 revenue for customers that were not retained is set equal to zero.

to lost customers.'*^ The predicted revenue change increased until CSIreached approximately 80, indicating that average revenue reductionsfor current customers declined as satisfaction increased. However, con-sistent with the retention function, revenue changes generally stopped

'" Overall, the firm experienced a 13% increase in customers and a 19% increase in totalrevenues between 1995 and 1996. Revenue growth for the retained customers in our sam-ple ranged from -97.6% to 500.0% (mean = 7.5%, median = 4.2%). We do not restrict therevenue levels and change tests to retained customers to avoid selection biases. However,when we repeated the linear and nonparametric regressions using only retained customers,CSI was positive and significant {p < 0.10, two-tail) in both the revenue level and revenuechange models. The shape of the estimated revenue function from the nonparametricregression was very similar to the plot in figure 2 and again exhibited steadily decliningdifferences in revenues at high satisfaction levels. In the revenue change plots, revenues in-creased until a satisfaction score of 45 was reached, after which there was almost no changein revenues until CSI = 80. Revenue then continued to grow almost linearly between CSIscores between 80 and 100. These results indicate that highly satisfied customers, if theywere retained, purchased more of the service in the future.


3-0.15 -

•0.2

I(u-0.25u(0

O -0.3

-0.35

» -0.4

2-0.45! r"

20 40 60 80

Customer Satisfaction Index in 1995

100

FIG. 3.—Revenue change analysis for business customers of a major telecommunicationsfirm (n = 2,491). The nonlinear function linking revenue change to satisfaction is devel-oped using additive nonparametric regression using variance stabilization. The nonlineartransformations of the variables (selected using the supersmoother procedure, where thespan is chosen using local cross-validation) are selected to maximize the correlation be-tween the transformed criterion and predictor variables such that the residual variance ofthe transformed criterion variance is constant. Figure 3 plots the value of revenue changepredicted using the optimal nonlinear transformation of customer satisfaction versus actualcustomer satisfaction. Revenue change is defined as 1996 revenues divided by 1995 reve-nues minus one. Customers that were not retained are assigned a revenue change of-1.0.

above this score, indicating that on average increasing the CSI of cur-rent customers above 80 did not produce greater revenue changes aftertaking lost customers into account.

Further evidence is provided in table 2. Rather than using nonlinearestimation techniques, we form ten portfolios based on the customers'CSI and then compare mean retention, revenue levels, and revenuechanges for each decile using general linear model {CLM) methods. Thisportfolio approach makes no assumptions about the functional form un-derlying the associations. Instead, CLM conducts an analysis-of-variancetest of differences in means across portfolios, after controlling for SIZEand ACE. Least squares means, which represent the means for each per-formance measure after controlling for the two covariates, can then becompared to assess whether mean performance was statistically differentacross the deciles.

The GLM results suggest that the relation between CS/and retention ischaracterized by several customer satisfaction "thresholds" that must be


TABLE 2Least Squares Means from General Linear Model (GLM) Estimates ofthe Association betweenPortfolios Formed on the Basis of Customer-Level Satisfaction Levels and Subsequent CustomerRetention and Revenue Change far 2,491 Business Customers of a Telecommunications Firm

Decile

123456789

10

Mean 1995Customer Satisfaction

14.2038.9447.9955.8164.4469.6376.2681.3588.2398.30

F-Statistic for CS/Effect/^-Statistic for Model

Mean 1996Retention

0.600.7010.7010.7210.7410.811-50.781-"0.771-20.78'-"O.771-"0.0244.74***5.61***

Dependent Variable ̂

Mean 1996Revenue

1393.231714.551548.022266.75'-32238.7612477.071-32250.801-32291.111-33I88.141-82776.811-3

0.0513.14***

12.18***

Mean 1996Revenue Change

-0.38-0.271-0.231-0.211-0.201-0.121-5-0.181-2-0.201-0.141-2-0.141-20.0153.84***3.44***

*** Statistically significant at the 1% level (two-tail).''The reported least squares means for each dependent variable represent the means for each cate-

gory after controlling for the customers' size (revenues) and years in business. The coefficients on thetwo control variables are not reported to simplify presentation. Customer retention and revenue changewere measured between 1995 and 1996, customer satisfaction (CS/) in 1995, and revenue levels fromcustomers in 1996. Superscripted numbers next to the least squares means indicate that the mean issignificantly larger (p < 0.15, two-tail) than the mean for the indicated decile (e.g., a superscripted 1indicates that the mean is significantly larger than the mean for decile 1).

reached before retention increases. The lowest retention rates (60%) arefound in the bottom decile of C5/scores. Deciles 2-5 have higher reten-tion {p < 0.15, two-tail) than decile 1 (70%-74%), but mean retentionrates within these four deciles are not statistically different. Mean reten-tion increases to 81% in decile 6, significantly higher than retention indeciles 1-5. Retention rates in deciles 7-10 (the highest CSI scores)range from 77%-78%. These rates are larger than those in deciles 1-4in most cases {p < 0.15, two-tail) but are not statistically different fromeach other or from the 81% retention rate in decile 6. The statisticallyequivalent results in the upper deciles contradict claims that customerretention is maximized when satisfaction scores are at their highest levels(e.g., Jones and Sasser [1995]).

Revenue levels also exhibit a series of satisfaction "thresholds." Thelowest mean revenue levels ($1393.23) are found in the bottom decileof CSI scores. Revenue is marginally higher but statistically similar indeciles 2 and 3 ($1714.55 and $1548.02, respectively). Mean customerrevenue increased to $2266.73 in decile 4 (/> < 0.15, two-tail) and re-mained close to this level through decile 8. In decile 9, mean revenuepeaked at $3188.14, greater than revenues in the lower eight deciles(p < 0.15, two tail). Mean revenue levels in decile 10 ($2776.81) werelower than (but statistically similar to) those in decile 9 but were not sta-tistically greater than mean revenues in deciles 4-8.


Similar to the retention and revenue results, revenue changes werelowest in the bottom decile of CSI scores (-38%). Revenue changes werelarger in deciles 2-5 than in decile I {p < 0.15, two tail), but revenuechanges in these deciles were not statistically different from one another.Revenue changes increased significantly in decile 6 (-12%). Results aremixed in the remaining deciles, ranging from -20% in decile 8 to -14%in deciles 9 and 10. In most cases, revenue changes in the top four decilesare significantly larger than in deciles 1 and 2, but revenue changes indeciles 7-10 are not statistically different from one another.^

While the customer-level results generally support claims that customersatisfaction measures are leading indicators of customer purchase behav-ior, the evidence also indicates that the retention and revenue growthbenefits from improved customer satisfaction diminished at higher satis-faction levels. Finally, the GLM tests provide some evidence of customersatisfaction "thresholds" that must be reached before customers changetheir purchase behavior.

One potential explanation for the customer-level results is the firm'suse of Partial Least Squares (PLS) to compute its satisfaction measure.Some marketing researchers claim that customer satisfaction measurescomputed using PLS have superior measurement properties relative tothe satisfaction measures used by most firms (Fornell [1992] and Fornellet al. [1996]). To provide some evidence on the relative ability of alter-native satisfaction measures to explain customer behavior in this firm, weexamine six additional customer satisfaction measures commonly used inpractice. "Top box" represents customers answering in the "top box" ofthe scale for a single question on their satisfaction with the service (e.g.,a response of 5 on a 1-5 scale, where 5 = very satisfied). Since this firmuses a ten-point scale in its satisfaction survey rather than the more com-mon five-point scale (Ryan, Buzas, and Ramaswamy [1995]), we code "topbox" 1 if the customer answered 9 or 10 to a question on their overall sat-isfaction with the service (from 1 = not at all satisfied to 10 = extremelysatisfied), and 0 otherwise. Similarly, "top-two box" is coded 1 if the cus-tomer answered 7 or above to the same question, and 0 otherwise. The"secure customer index" {SCI), a measure of customer loyalty, is coded 1if the customer answered 7 or above (on ten-point scales) to each ofthree questions (overall satisfaction with the service [from 1 = not atall satisfied to 10 = extremely satisfied], likelihood of recommendation

" CLM tests using only retained customers found the lowest revenue change in thebottom-two deciles (3.4% and 4.7%, respectively) and the highest change in deciles 9 and10 (11.0% in each), with the rates in the two groups significantly different (p < 0.15, twotail). However, revenue changes in deciles 1, 2, 9, and 10 were not statistically differentthan those deciles 3-8. CLM revenue tests using retained customers found somewhatlower revenue levels in decile 10 ($3544.75) than in decile 9 ($4140.37), though the figureswere not statistically different. Mean revenues in decile 9 were significantly greater thanthose in deciles 1-8, while mean revenues in decile 10 were only statistically larger thanthose in deciles 1-3.


[from 1 = not recommend to 10 = strongly recommend], and expectedretention [from 1 = not at all likely to 10 = extremely likely]), and 0 oth-erwise.̂ ^ "Equally weighted" is the equally weighted standardized re-sponse to these three questions. "First principal component" is the firstprincipal component factor score for the three questions. Finally, "singlequestion" is the customer's response to a single question on their overallsatisfaction with the service (from 1 = not at all satisfied to 10 = extremelysatisfied).

The linear regression results in table 3 indicate that the choice of cus-tomer satisfaction measures has little effect on the significance of thecustomer satisfaction coefficients or the explanatory power of the mod-els. Despite the use of PLS estimation methods, the firm's CS/measureexplains future customer purchase behavior no better than the simplermethods used by most firms. Thus, the customer-level results do notappear to be driven by the firm's customer satisfaction measurementmethodology.

4. Business-Unit Analyses

Although the customer-level tests indicate that customer satisfactionmeasures predict subsequent purchase behavior of existing customers,they provide no evidence on the costs or profits associated with highersatisfaction levels, the effects of customer satisfaction on growth in newcustomers, or the extent to which organization-level customer satisfac-tion indexes, which are typically based on aggregated survey responsesfrom a relatively small sample of customers, are leading indicators offinancial performance. We therefore extend the analyses to examine theextent to which business-unit customer satisfaction measures predict fu-ture accounting performance and number of customers.

We conduct these tests using data from 73 retail branch banks fromthe western U.S. region of a leading financial services provider.^^ Thebank is a relative newcomer to this region and faces considerable com-petition. To achieve its strategic goal of gaining substantial worldwidegrowth in customers, the firm has made customer satisfaction one of fivecorporate "imperatives" (along with achieving financial results, manag-ing costs strategically, managing risk, and having the right people in theright jobs) incorporated in its "balanced scorecard" performance mea-surement system. Customer satisfaction scores form a major componentof quarterly performance evaluations and bonuses for branch-level man-agers and above.

'2 The "secure customer index" and SCI are registered trademarks of Burke CustomerSatisfaction Associates.

•'After deleting "outliers" (i.e., observations with studentized residuals greater than anabsolute value of three in the regression analyses), the number of observations in our testsranges from 71 to 72.


o

3

CO COr - l O

o oo o

00 O) - ^ CO COO O rH rH rH

p p o o po o o o o

* * * ** * * ** * * *O OCO

oo

CO lO

O rH

l O CO

o

CO

oo

CO

oop p

o d o o o o o

p p p p p p p^ d ^ d d

* * * * * * ** * * * * * *^̂ ^ _̂̂ ̂ 00 00 ^̂ 5 ^̂ ^ ^̂ ^oi i> o 00 cd ^ O)

Ol t^ Ol

rH Ol(M Op p6

O) OlrH rH

p p

* * * * * * ** * * * * * ** * * * * * *CM »n "^ O) CN CM CNO iC O C31 O O Op P rH p p O Od d d d d d d

VT3C

cucoa.T3 O

S 2 -2 5 SP

U s x̂=2 .5 cr.bz /3 W fc

t i .

£ cfl

" 2

s B o


The firm computes quarterly customer satisfaction measures based onthe average of three monthly surveys of 25 retail customers per branch.The customer satisfaction index {CSI) is a composite of 20 items. Scoresfor each of the items equal the percentage of customers responding inthe "top-two boxes" of the question's scale (i.e., 6 or 7, where scores rangefrom 1 = not satisfied to 7 = very satisfied). The most heavily weighteditem (45%) asks customers to rate "the overall quality of [the branch's]service against your expectations." The remaining items include the qual-ity of tellers versus expectations (7.5%), six additional items concern-ing tellers (7.5%), six items concerning nonteller employees (7.5%), thequality of automated teller machines (ATMs) versus expectations (7.5%),three additional items concerning ATMs (7.5%), and one item measur-ing problem incidence (10%). In the second quarter of 1996, branchsatisfaction scores ranged from 43 to 70 (mean = 57.2, median = 58)on a scale from 0 (no top-two box responses) to 100 (all top-two boxresponses).

The firm provided customer satisfaction and accounting data for thethird quarter of 1995 through the second quarter of 1996. Although thistime period is short, the frequent repurchase cycle and relatively lowcustomer switching costs in retail banking imply short lags between acustomer's experience and observed changes in purchase laehavior andeconomic performance. We examine six performance variables: reve-nues {REV), expenses {EXP), margins {MAR), return on sales {ROS), re-tail customers {RETAIL), and business and professional customers {B&fP).Margins are defined as revenues minus expenses, and return on sales asmargins divided by revenues. We estimate three basic models:

PERFit.,1 = a + PiCS/j ( + Pg PASTPERFit +

= a P i i t ^^

+ P4%AB6'P,-j^i + 8,-

= a p%ACSI1 + £,-,+ 1

where PERF is either revenues, expenses, margins, return on sales, retailcustomers, or business and professional customers, CSI is the branch'scustomer satisfaction index, PAST PERF is the dependent variable'svalue in the prior period, t denotes the third and fourth quarters of 1995,and t + 1 denotes the first and second quarters of 1996. Averages overthese quarters are used for levels variable, and percentage changes be-tween the quarters for change variables. To control for other factors thatmay influence accounting performance (e.g., product mix, demograph-ics, regional growth rates, etc.), we include RETAIL and B&P when thefour accounting measures are dependent variables, and we include PASTPERF, both to control for time-series trends and to examine whether CSIprovided incremental information on future performance.


To assess whether branches with higher satisfaction levels have morecustomers and more revenues per customer, we examine the associationbetween CSI levels in the third and fourth quarters of 1995 and account-ing and customer levels in the following quarters (table 4, panel A). CSIhas a stadstically positive association with revenues (p < 0.05, two-tail)and the number of business and professional customers (p < 0.10, two-tail), after controlling for performance in the prior period. Since therevenue model controls for the number of customers, these resultsindicate that, on average, branches with higher satisfaction scores hadhigher revenue per customer. However, CSI is not statistically associatedwith expenses, margins, ROS, or retail customers.^* The number of B&Pcustomers in the first and second quarters of 1996, in turn, is positivelyassociated with revenues, expenses, and margins {p < 0.10, two-tail) andwith ROS {p < 0.15, two-tail). This evidence suggests that higher cus-tomer satisfaction levels have an indirect effect on accounting perfor-mance by attracting new customers, one of the bank's strategic goals.

The efFects of CSI levels on subsequent percentage changes in per-formance are reported in panel B of table 4. Like the customer-leveltests, this model examines whether branches with higher satisfactionlevels experienced greater improvement in accounting performance andcustomer growth. Customer satisfaction levels are positively related tosubsequent percentage changes in margins and ROS {p < 0.10, two-tail),after controlling for prior changes in margins and returns, implyingthat branches with higher satisfaction levels increased future profits ata greater rate than other branches. Percentage changes in retail cus-tomers also exhibit a positive association with changes in each of theaccounting measures, while business and professional customers are pos-itively associated with changes in revenues and negatively associated withchanges in expenses. However, CSI levels are not statistically related tosubsequent growth in either customer group. Finally, the coefficients onPAST PERF are negative in each of the financial performance models,suggesting mean reversion in accounting performance.

Results using percentage changes in CS/are presented in panel C of table4. Customer satisfaction changes exhibit no significant direct effect onsubsequent changes in revenues, margins, or ROS. However, CS/changesare positively associated with future changes in retail customers, andretail customer changes are positively related to changes in revenuesand margins. Similar to the levels tests in panel A of table 4, CSI changesappear to have an indirect effect on accounting performance throughgrowth in customers. Consistent with this interpretation, percentagechanges in CSI are positively associated with changes in revenues and

''' The insignificant results for these variables are not due to the inclusion of past per-formance in the models. In fact, none of the CSI coefficients was significant when pastperformance was not in the model.

TABLE 4OLS Regressions Examining the Association between Customer Satisfaction Scores and

Subsequent Accounting Performance and Customers for 7B Retail Branch Banks'(t-Statistics in Parentheses)

Panel A: Perfonnance Levels on Customer Satisfaction Levels in Prior PeriodDependent Variables''

Intercept

CSI

PASTPERF

RETAIL

B&P


REV

-73.985(-1.42)

2.089**(2.35)

1.178***(21.82)

-0.046***(-5.05)

0.114*(1.83)

0.97482.20***

Panel B: Percentage Changes

Intercept

CSI

%APASTPERF

%&RETAIL

%AB&'P

Adjusted R^f-Statistic

%AREV-0.520

(-0.37)

0.002(1.16)

-0.376***(-2.93)

0.972***(3.97)

0.272*(2.48)

0.529.11***

Panel C: Percentage Changes

Intercept

%ACSI

%&PASTPERF

%ARETAIL

%AB&P


%AREV

0.110***(7.28)

0.130(1.22)

-0.324**(-2.48)

0.849***(3.13)

0.256**(2.19)

0.298.41***

EXP-33.453(-0.92)

0.731(1.20)

0.998***(14.19)

0.009**(2.20)

0.097**(2.47)

0.87128.83***

MAR

-25.42(-0.42)

0.714(0.49)

1.058***(13.98)

-0.037***(-3.32)

0.370***(3.86)

0.91191.59***

ROS

0.074(0.67)

-0.001(-0.55)

0.874***(9.47)

-0.000(-0.91)

0.0002*(1.65)

0.6840.27***

RETAIL91.978(0.70)

0.929(0.42)

0.953***(79.38)

—

—

0.983165.07***

B&fP

-22.425(-1.40)

0.489*(1.78)

1.056***(42.46)

—

—

0.96906.26***

in Performance on Customer Satisfaction Levels in Prior PeriodDependent Variables''

%AEXP0.157»

(1.54)

-0.002(-1.32)

-0.042(-0.64)

0.975**(5.47)

-0.152*(1.83)

0.308.84***

%AMAR-0.454

(-1.22)

0.012*(1.86)

-0.245(-1.37)

1.713***(2.69)

0.414(1.37)

0.164.23***

%AROS-0.406*

(-1.70)

0.009***(2.18)

-0.187(-1.39)

0.802*(1.96)

0.201(1.03)

0.113.10***

in Performance on Percentage Changes in (Dependent Variables

%AEXP

0.027***(2.70)

-0.074(-0.99)

-0.037(-0.56)

1.062***(5.69)

-0.149*(-1.85)

0.308.56***

%AMAR0.225***

(6.13)

0.293(1.07)

-0.139(-0.78)

1.28*(1.91)

0.439(1.43)

0.133.54**

%AROS

0.109***(4.57)

0.073(0.41)

-0.130(-0.95)

0.600(1.37)

0.217(1.07)

0.051.83*

%ARETAIL0.024

(0.69)

-0.001(-1.17)

0.518***(8.29)

—

—

0.4934.39***

foAB&P-0.098

(-0.68)

0.001(0.39)

0.200*(1.48)

—

—

0.011.23

Gustomer Satisfaction

%ARETAIL

-0.021***(-5.73)

0.080***(2.74)

0.495***(6.42)

—

—

0.4529.64***

%AB&P

-0.042**(-2.64)

0.000(0.00)

0.205*(1.50)

—

—

0.001.15

***, **. *. * Statistically significant at the 1%, 5%, 10%, and 15% levels (two-tail), respectively.'Outliers are deleted from the models. The resulting sample sizes range from 71 to 72.''REV= revenues, EXP = expenses, MAR = margins (revenues- expenses), ROS = return on sales (margin/sales),

RETAIL = the number of retail customers, B&T = the number of business and professional customers, CSI = the cus-tomer satisfaction index, and PAST PERF = the level or percentage change in the dependent variable in the priorperiod. Percentage changes in CS/and past performance are measured between the third and fourth quarters of 1995. Allother percentage changes are measured between the first and second quarters of 1996. Customer satisfaction and priorperformance levels ate the averages for the third and fourth quarters of 1995. All other levels variables are the averagesfor the first and second quarters of 1996.


margins {p < 0.10, two-tail) when retail and 56^customers are excludedfrom the model (not reported).

Because the small sample size prevents us from estimating the non-parametric regressions used in the customer-level tests, we examinepotential nonlinearities by forming quartiles based on CSI levels or per-centage changes in CSI}^ GLM then is used to test for differences inmean performance changes across quartiles, after controlling for thecovariates. GLM tests of the associations between performance levels andCSI levels, provided in panel A of table 5, indicate that revenues arehigher in the top quartile of CSI scores than in the lower three quartiles{p < 0.15, two-tail). Margins are also larger in quartile 4, but signifi-candy so only relative to quartile 3. Branches in the top quartile of CSIscores also exhibit greater expenses and more retail and B(fP custom-ers, but the differences are not statistically significant at the 0.15 level(two-tail).

The GLM tests in panel B of table 5 examine the association betweenpercentage performance changes and quartiles based on CSI levels. Thesmallest accounting changes are found in branches with the lowest sat-isfaction levels (quartile 1). Mean changes in revenues, margins, andROS are not statistically different from zero in this group and all aresignificantly lower than means in the other quartiles {p < 0.15, two-tail).Although the least squares means for quartiles 2-4 are significantlylarger than those in quartile 1, the performance changes in these threegroups are not statistically different from one another. Changes in ex-penses were statistically lower in branches in the top quartile of CSI scoresthan in the bottom quartile (-1.3% in quartile 4 vs. 2.6% in quartile 1).Percentage changes in retail customers were significantly higher in thethird quartile (1.8%) than in the first quartile (0.5%) but fell slightly(though insignificantly) to 1.5% in quartile 4. Changes in B&'P custom-ers were negative in every quartile except the fourth, but the differenceswere significant {p < 0.15, two-tail) only between quartile 2 (-5.7%) andquartile 4 (0.5%).

Results using percentage changes in C5/ (table 5, panel C) indicate that ac-counting improvement rates were not significantly different until branchCSI changes ranked in the top quartile. Quartile 4 reported significantlygreater improvement in revenues than quartile 1 (12.1% vs. 8.3%), inmargins than quartiles 2 and 3 (28.2% vs. 15.9% and 11.0%, respectively),and in ROS than quartile 3 (12.7% vs. 3.8%). None of the other com-parisons of quartile means is statistically significant. The top-two quartiles

alternative methods for examining nonlinearities in a sample of this size are theintroduction of squared terms for C5/or the use of interactions between C5/levels and CSIchanges (e.g., if the changes term was statistically positive and the interaction term wasstatistically negative, the results would suggest that the effect of CS/changes is contingenton the branch's CS/level). However, when we attempted to implement these approachesusing the branch data, we encountered serious levels of multicollinearity, with correlationsbetween the independent variables exceeding 0.90.


CO i nV II

(A

t

art

& crtV

S

«5 :=

in• *

OJ inV II

1̂o

Io

.a

n4, Ol

1 SQ1 )

*40

*

CTlCOCO

t oOlo

I >OJ^î nTt"

toOJoOJ

oTt"

Tt"

t oCO

Tt"

RE

V**

25*

CO00

0000o

CTl

oOJ

O)OJ

oOJ

^^00ooOJ

i n

Ol

EX

P

**

00

00OJ-^

OJ

o

r H

0000COOI

Tt"

oi n

OJ

OJCO

oCOOl

CO

OJOI

MA

R

**

Ol

t oOJ

oo

O)Tt"

o

00Tt"o

oi no

Tt;

RO

S

*

13*

CT^t oi n

00

o

00Tt"o

OJ

Tt"

OJ

t oOJtoCO

OJ

i n

i n

O l

RE

TAIL

*

98*

OJCOTt"

t ot31O

0000t o

^

OJ

CO

*""

OJt oOJ

11.

1

a

acti

o

1

Cus

ti

0

*s(3

ea41

a0

•s

les

For

m

ia

Pan

el B

:

• ^

flj

^ ^

* J

rt3

CO

rti

Qua

CM

'S

ile

1

u

1st

itat

t«

li.

0

S

CM

^ -

o

!>.CM(OII

Co

rt

:s

toIIIS:

iu

i n

54.

I I

J'c3c

Mea

OIdinII

c3

Mea

n

*

60*

t o

00COo

CTlOO

OJ00

oo

'po

Tt"CO

oo

%A

RE

V

*

19*

t o

t oCOo

0 0

oo1

Tt"OJ

oo

i nooo

1

t o

oo

%A

EX

P

*

30*

CO

Tt"OJo

ODi nOJo

OJ

OJ

o

COOJOJo

COinoo

te;

36*

OJ

OQ^ô

o

o

CO'PI

o

_̂oo

OJooo

%A

RO

S

*

.48*

i >^-t

i no

i n

oo

1

00

oo

1

t o

oo

1

'inooo

1

1

.72*^^

o^-<o

i nooo

Tt"

oo1

inoo1

CTlOOO1

05

a0

11

mer

o

U.S

nges

Id

U

a

Per

ci

(4H0

. |

ca

on t

hle

s F

orm

0?

Pan

el C

:

• ^

3

°c

rti

Qua

OJu

irti

3

ile

l

urt

ist

Itat

'v

0

S

ft;

•T lo

CM

dCMII

<

Mea

n

o

<

crtos

Olooi

1II

<

ean

9

s

COCM

1

II

ean

%A

s

**

.00*

t o

t oCOo

o f l

OJ

O

OJinoo

ooo

CO00

od

1

*

46*

i n

COCOo

goo

1

00

ooo

CO

oo

od

1

*

.78*

OJ

^Ôlo

enCJOJ00OIo

o

o

CTli n

o

OJCTl

d

te;

inin

CO^ô

OJ

O

00CO

oo

Tt"CTlO

o

COCTlO

d

*

.95*

• *

^ ^

t^Tt"O

ooo

1

Tnoo

°

Tj"

oo

1

COCO

od1

COCTl

o

i noo

00

oo

oinoo

1

oCO

o

-0.

oco

1

<"2 iS.is i i


both experienced significantly lower retail customer losses than quartile1, but CSI changes were not statistically associated with percentagechanges in B&P customers. In general, the change tests in panel C sugrgest that relatively large increases in customer satisfaction were necessaryto improve performance.

The business-unit results indicate that satisfaction measures computedusing aggregated responses from a small sample of customers have somepredictive ability for future accounting performance. However, many ofthe accounting gains appear to come indirectly through growth in newcustomers, rather than direcdy through increased profits from existingcustomers. There also appear to be customer satisfaction "thresholds" thatmust be reached before branch performance improves. We find no evi-dence that branches with higher satisfaction scores had greater expenses,and some evidence that the highest-rated branches were able to controlcosts better than low-rated branches in subsequent quarters. Finally, port-folio tests using CSI changes indicate that fairly large improvements incustomer satisfaction were necessary to produce significant changes inthe number of customers or accounting performance.

5. Firm-Level Analyses

Although the preceding tests provide qualified support for claims thatcustomer satisfaction measures are leading indicators of financial per-formance, they provide no evidence on whether the stock market viewscustomer satisfaction as a forward-looking performance indicator. Weprovide evidence on this issue using firm-level data from the AmericanCustomer Satisfaction Index (ACSI), a national economic indicator ofcustomer satisfaction managed by the National Quality Research Centerat the University of Michigan Business School and the American Societyfor Quality. Each firm's customer satisfaction score is estimated usingtelephone survey data obtained from several hundred consumers whopurchased or used the company's product within the past six months.

ACS/scores are based on 15 questions rated on ten-point scales. Thequestions are formed into four latent variables (perceived quality, cus-tomer expectations, perceived value, and customer satisfaction) that arelinked in a causal model to two latent variables for expected outcomes(self-reported customer complaints and customer loyalty). The customersatisfaction construct (i.e., the ACSI) is a combination of three questions(overall satisfaction, confirmation of expectations, and comparison toideal) weighted using Partial Least Squares {PLS) such that their linearcombination is maximally correlated with the customer complaint andcustomer loyalty constructs. The scores for individual customers are re-scaled to range from 0 (least satisfied) to 100 (most satisfied). Firm-levelACS/scores represent average scores for customers using the firm's prod-ucts. Additional details are provided in National Quality Research Cen-ter [1995] and Fornell et al. [1996].


The ACSI scores are based on a uniform measurement methodologyand represent independent assessments by an external organization. Inaddition, the sample is not biased toward a self-selected group of com-panies that voluntarily disclose customer satisfaction measures. However,the ACSI's minimum sample size requirements and random samplingtechniques constrain the sample to very large firms with substantialmarket share (mean [median] market value of equity = $12,528 [$7,283]million). The survey also focuses on consumer products and ignores cus-tomers' satisfaction with commercial products (i.e., business-to-businessproducts). The ACSI may also provide noisy measures for diversifiedconsumer products firms, which can be represented in only one productcategory (such as Proctor 8c Gamble, included only in the "Personal Careand Cleaning" category) or which can receive scores in multiple catego-ries (PepsiCo appeared in the "Soft Drinks" category and three times inthe "Restaurants-Fast Food-Carry Out" category for KFC, Pizza Hut, andTaco Bell).^^ This measurement error will cause estimates of the coef-ficients linking ylCS/scores to firm performance to be inconsistent.

5.1 VALUE RELEVANCE OF CUSTOMER SATISFACTION MEASURES

ACSI scores for individual firms were released publicly for the firsttime in the December 11, 1995 issue of Fortune (Fierman [1995]). Thearticle provided scores from the initial 1994 ACSI survey, as well as up-dated 1995 scores computed 3, 6, 9, or 12 months after the initial ACSImeasurement (the timing varied by industry).^' The Fortune article re-ported 1994 ACSI scores for 138 firms and 1995 scores for 140 firms,with the scores ranging from 63 to 90 (mean = 78).

We examine the extent to which the ACSI scores are associated withthe market value of equity, after controlling for information containedin contemporaneous accounting book values, using a cross-sectional val-uation model of the form:

i = Po + Pi ASSETSi + Pg LIABi + P3 ACSIi + e,,

where MVEj is the market value of equity for firm i, ASSETS^ is the bookvalue of assets, LIAB^ is the book value of liabilities, ACS/j is the satis-faction score, and e, is random error (e.g.. Landsman [1986] and Barthand McNichols [1994]). Since data collection for the 1994 AC5/endedon July 22, 1994, we use Compustat data for the fiscal year-end closest toJuly 1994 to measure MVE, ASSETS, and LIAB when 1994 ACSI scoresare used in the model. Starting with the 1995 ACS/survey, different in-dustrial sectors are measured during different calendar quarters, so we

'^When a firm is represented more than once in the ACSI, we use the average of themultiple scores in our analyses.

I'The 1994 and 1995 ACS/scores have a correlation of 0.91. Consequently, we do notconduct firm-level tests using changes in satisfaction scores.

-16775.01***(-2.23)

1.73***(16.95)-1.77***

(-15.89)243.20***

(2.53)0.74

126.16***121

-16917.10**(-2.21)

2.19***(18.22)-2.25***

(-17.13)235.67**

(2.39)0.77

149.20***124


TABLE 6OLS Regressions Examining the Association between the Market Value of Equity and the American

Customer Satisfaction Index (ACSI) Scores' Published in Fortune(t-statistics in parentheses)

1994 ACSI Scores 1995 ACSI Scores

Intercept

ASSETS

LIAB

ACSI

Adjusted R'^i^-Statisticn

***, **, * Statistically significant at the 1%, 5%, and 10% levels (two-tail), respectively."The dependent variable in the models is the market value of equity for the fiscal year-end closest

to the month the ACS/scores were collected. ASSETS is the book value of assets, IJAB is the book valueof liabilities, and ACSI is the customer satisfaction score. Outliers and firms without complete IIBIEIS,CRSP, and Compustat data were deleted from the sample.

use Compustat data for the fiscal year-end closest to the month of datacollection for MVE, ASSETS, and LIABin analyses of 1995 scores. Incre-mental value relevance of the ACS/implies Pg > 0.

Our goal is to determine whether an aggregate nonfinancial measure,such as customer satisfaction, provides incremental information for ex-plaining differences in the market value of equity, after controlling forbalance sheet information. We use an aggregate customer satisfactionmeasure, rather than individual determinants or drivers of customer sat-isfaction, because firms use similar measures for decision making andperformance evaluation. Moreover, extending the analysis to individualperceptual drivers of customer satisfaction is not feasible with the ACSIbecause the necessary data are not available. We also ignore income state-ment accounts (e.g., advertising or marketing expenditures) in the cross-sectional valuation model, even though these may affect the statisticalsignificance of customer satisfaction. Our objective is not to identify thedrivers of customer satisfaction but to determine whether this measureprovides incremental explanatory power in a traditional cross-sectionalvaluation model.

Results from the valuation tests are provided in table 6.̂ ^ The co-efficients on v4CS/are positive and significant {p < 0.05, two-tail) usingeither 1994 or 1995 scores, implying that customer satisfaction measures(or information correlated with these measures) provide insight into

elimination of firms without complete data and the deletion of outliers reducethe sample sizes to 121 using 1994 AC5/scores and 125 using 1995 scores. White's [1980]test revealed no evidence of heteroscedasticity in any of the models.


firm value that is not reflected in current accounting book values.^^ Thecoefficients on ACSI imply that a one-unit difference in the index wasassociated with a difference in the market value of equity of between$236 to $243 million, after controlling for accounting book values.

We augment the association tests reported in table 6 with analysis basedon the Edwards-Bell-Ohlson (EBO) or residual income valuation model(e.g., Ohlson [1995] and Feltham and Ohlson [1995]) which maintainsthat equity market value is a function of equity book value plus discountedresidual future earnings. The EBO model can be characterized as:

= Po + Pi ASSETSi + Pg LIABi + P3 EUTUREi + e ,̂

where MVE, ASSETS, and LIAB are defined above, and EUTUREi is * ediscounted value of future earnings in excess of a capital charge basedon the opportunity cost of capital for firm i. If current customer satisfac-tion levels are incorporated into forecasts of future cash flows, ACSIshould be positively associated with the variable EUTURE. We test this as-sociation using a variant of the forecasted residual earnings measure inFrankel and Lee [1995]. The first available median consensus earningsand long-term growth forecasts for years +1 and +2 (where the year ofACSI computation is denoted year 0) are obtained from the IIBIEIS files.Earnings forecasts for years +3 to +5 are computed by multiplying theearnings forecast for the previous year by one plus the long-term growthforecast. Capital charges are computed by multiplying the equity cost ofcapital by book values at the end of the previous year. The equity costof capital is estimated using the systematic risk over year 0 (where themarket is approximated by the value-weighted daily CRSP index) andthe average risk-free rate (0.06) and market risk premium (0.074) pro-vided by Ibbotson [1996]. The actual book value of equity for year 0 isused to compute residual earnings for year +1. Book values for comput-ing residual earnings in subsequent years are computed using the bookvalue calculated for the preceding year plus the preceding year's fore-casted earnings multiplied by one minus the dividend payout rate (prox-ied by the average dividend payout over the five years ending in year 0).Forecasted residual earnings are computed as the present value of thefive annual residual earnings forecasts (year +1 to year +5), discounted

provide further evidence on the value relevance of customer satisfaction mea-sures, we repeated the analysis using two alternative measures of market value: the earn-ings-price ratio and the market-to-book ratio. Following prior studies of the determinantsof these ratios (e.g., Beaver and Morse [1978], Ohlson [1990], and Alford [1992]), we es-timated earnings-to-price and market-to-book ratios as a function of systematic risk, divi-dend payout, median IIBIEIS consensus forecasts for long-term earnings, and the ACSI.The coefficients on ACS/were positive and statistically significant (p < 0.10, two-tail), againsuggesting that customer satisfaction measures provide value-relevant information to themarket.

1994 ACS/Scores-3226.09*

(-1.96)39.23*(1.87)1.41***

(13.68)0.62

97.08***121

1995 ACS/Scores-4442.75**

(-2.49)55.44**(2.41)1.49***

(14.23)0.64

109.46***125


TABLE 7OLS Regressions Examining Long-Term Forecasted Residual Earnings

as a Function of ACSI Scores and Current Earnings ̂(t-statistics in parentheses)

Intercept

ACSI

EARN

Adjusted R^F-Statisticn

*•*, **, * Statistically significant at the 1%, 5%, and 10% levels (two-tail), respectively.' The dependent variable in the models is the discounted value of future earnings in excess of a capi-

tal charge reflecting the opportunity cost of capital for firm i, and is based on the median consensuslong-term growth and earnings forecasts in IIBIEIS. ACSI is the firm's customer satisfaction score, andEARNis annual earnings for the fiscal year-end closest to the month the ACS/scores were collected. Out-liers and firms without complete IIBIEIS, CRSP, and Compustat data were deleted from the sample.

using the equity cost of capital. Residual earnings beyond year +5 (andthe terminal value) are assumed to be zero.^^

Results in table 7 indicate that ACSI measures are positively associatedwith forecasted residual earnings. Even after controlling for current an-nual earnings, both 1994 and 1995 ACSI scores are predictors of analysts'long-term forecasts of residual earnings {p < 0.10, two-tail). This evi-dence suggests that at least some of the expected benefits from customersatisfaction are already impounded into earnings forecasts.

Similar to the approach used in the business-unit analyses, we testfor potential nonlinearities in the market's valuation of customer satis-faction by dividing the sample into quartiles based on the firms' ACSIscores. A GLM model is estimated with the ACSI quartiles as predictorvariables and the book values of assets and liabilities as covariates. Re-sults in table 8 indicate that the lowest mean market values (after con-trolling for book values) are found in quartile 1 (the lowest ACS/scores).When the valuation model is estimated using 1994 data, quartiles 2-4all have larger mean market values than quartile 1 {p < 0.15, two-tail).However, mean values within these three quartiles are statistically equiv-alent. Similar results are obtained using 1995 data. This plateau in thebenefits from higher customer satisfaction levels is similar to the thresh-olds found in the branch bank and suggests there are diminishing re-turns at higher satisfaction levels.

adjusted I? for a base case model regressing market value of equity on bookvalues of assets and liabilities is 0.73 (0.77) using 1994 (1995) data. The addition of ourresidual earnings variable increases the adjusted fC to 0.92 (0.94). Thus, our proxy for re-sidual earnings appears to have some descriptive validity.

8273.76

10939.28'

12064.36'

12153.57'

0.7575.89***121

10213.87

10367.86

12991.19''2

12153.57'

0.7890.07***125


TABLE 8Least Squares Means of Market Value from General Linear Model (GLM) Estimates of

the Firm-Level Associations between Portfolios Formed on the Basis of CustomerSatisfaction Scores Reported in Fortune and the Market Value of Equity'-

1994 Scores 1995 ScoresQuartile 1(Mean CS/= 71.7/70.1)''Quartile 2(Mean CSI= 77.2/76.2)Quartile 3(Mean C5/= 81.2/80.3)Quartile 4(Mean CS/= 85.7/84.8)Model i?2f-Statisticn

*** Statistically significant at the 1% level (two-tail).'The least squares means represent mean market values for each quartile after setting the two covar-

iates (book values of assets and liabilities) equal to their means. Outliers and firms without completedata are deleted from the models. Superscripted numbers next to the quartile means indicate that thefigure is significantly larger {p < 0.10, one-tail) than the figure for the indicated quartile (e.g., a super-scripted 1 indicates that the mean is significantly larger than the mean for quartile 1).

^The first figure is the quartile's mean 1994 ACSI score; the second figure is the mean 1995 ACSIscore.

5.2 INDUSTRY DIFFERENCES

The preceding tests assume that custotner satisfaction's effect on firmvalue is similar across industries. However, Anderson, Fornell, and Rust[1997] find significantly different contemporaneous associations betweencustomer satisfaction and accounting returns in different industries.Based on these results, they predict that high customer satisfaction is as-sociated with higher performance in manufacturing firms and in mostservice businesses. However, in retailing, where customer satisfaction ismore dependent on providing customized services, high satisfaction maybe associated with lower accounting returns because of the high cost ofincreased customization. Anderson, Fornell, and Rust [1997] also con-tend that customer satisfaction is less important in monopoly businessesor in industries with high switching barriers, such as regulated utilities.However, other authors note that utilities and telecommunications pro-viders now realize the need to increase customer satisfaction as marketsare opened to competition (Bertrand [1989], Electric World [1994], andSchlossberg [1993]).2'

We examine industry effects by estimating separate valuation modelsfor five industry categories with ten or more observations (nondurable

2' The increased emphasis on customer satisfaction in regulated industries is consistentwith research finding greater weight on nonfinancial measures such as customer satisfac-tion in utility and telecommunications executives' bonus contracts than in other indus-tries (Bushman, Indjejikian, and Smith [1996] and Ittner, Larcker, and Rajan [1997]).


TABLE 9OLS Regressions Examining the Association between the Market Value of

Equity and the 1995 ACSI Scores for Broad Industry Categories(t-statistics in parentheses)

Intercept ASSETS UAB ACSI AdjustedNondurable Manufacturing -14045.66

(-0.87)

Durable Manufacturing

Transportation, Utilities,Communication

Retail

Financial Services

14514.59(0.36)

-35712.12**(-2.83)

-52956.32*(-2.02)

-20750.92(-1.14)

0.77*(1.89)

1.13***(9.20)

0.68*(1.98)

1.93***(4.13)

0.86***(3.09)

0.55(0.82)

-1.10***(-7.95)

-0.08(-0.16)

-1.58**(-2.50)

-0.84**(-2.80)

180.77(0.90)

-128.22(-0.26)

451.53**(2.71)

-703.22*(-1.99)

353.18(1.43)

0.88**

0.87***

0.66***

0.88***

42

18

40

0.82*** 18

10

***, ••, * Statistically significant at the 1%, 5%, and 10% levels (two-tail), respectively. See table 6 for variabledefinitions.

manufacturing; durable manufacturing; transportation, utilities, andcommunications; retail stores; and financial services).^2 Based on Ander-son, Fornell, and Rust's [1997] predictions, we expect ACS/to be posi-tively associated with market values in the manufacturing and financialservices industries and negatively associated with market values in re-tailers. We make no predictions about the expected associations in thetransportation, utilities, and communications groups.

Industry-specific tests using 1995 data are provided in table 9. Al-though the sample sizes are small, the evidence suggests that the valuerelevance of customer satisfaction measures varies across industries. Thecoefficients on ACSI are positive but statistically insignificant in dura-ble and nondurable manufacturing firms, and are statistically positivein transportation, utility, and communication firms. Coefficients on ACSIin financial service firms are also positive but not statistically significant(due in part to a sample size of only ten). In contrast, the negative asso-ciation between ACSI and market value for retailers supports Ander-son, Fornell, and Rust's [1997] claim that in retailing, the benefits from

22 The desirability of estimating industry-specific models is unclear. If customer satis-faction levels are a function of the level of competition in an industry, the effect of com-petition on customer satisfaction will be removed. This will tend to produce conservativetests of the association between satisfaction levels and market value. In addition, the in-dustry samples are limited to a relatively small number of large, surviving firms. If there islittle cross-sectional variation in customer satisfaction measures within these firms, we willfind little association between satisfaction and firm value, even if customer satisfaction isimportant.


increased customer satisfaction can be exceeded by the incrementalcost.23

Differential industry efFects also emerge when we examine the two in-dustries (food processing and utilities) with ten or more observations(not reported). The coefficient on ACS/is positive for utilities and nega-tive for food-processing firms {p < 0.15, two-tail). The negative result forfood processors may be due to the already high scores in this industry.Moore [1997] argues that firms in industries with high customer satis-faction levels have less opportunity to use customer satisfaction to dif-ferentiate themselves from competitors. Consistent with this claim, thehighest-rated food-processing firms had the top satisfaction scores inthe ACSI survey, and the group as a whole was second only to the threefarms in the soft drink industry (mean ACS/score = 84 vs. 86). Utilities,on the other hand, had some of the lowest satisfaction levels in the sur-vey (mean = 74), potentially providing room to improve economic per-formance by enhancing customer satisfaction.

5.3 STOCK MARKET REACTIONS TO THE RELEASE OF THE ACSI

Results based on valuation tests indicate that customer satisfaction in-dicators can be incrementally value relevant to stock market partici-pants but provide no evidence that the public release of customer satis-faction measures provides new (or incremental) information to the stockmarket. We investigate the information content of customer satisfactionmeasures by examining the stock market response to the initial disclo-sure of individual ACSI scores. Although the cover date for the Fortunearticle was December 11, 1995, this issue was received by many Ameri-can subscribers on November 27, 1995 and was loaded into Lexis/Nexison December 1, 1995.̂ ^ Given the wide distribution period, we computecumulative abnormal stock market returns using two trading periods:the five trading days from November 27, 1995 to December 1, 1995, andthe ten trading days from November 27, 1995 to December 8, 1995. Ex-pected returns are computed using the market model:

where Rji = the return on the security of firm i in period time t, R^t -the return on the market portfolio in period t (measured using the

2* When the industry models were estimated using 1994 data, the signs on ACS/did notchange from those using 1995 data and remained positive and significant for the trans-portation, utihties, and communications group. ACSI was also positive and significant infinancial services. However, the coefficients on ACSI, though still negative, were no longerstatistically significant in the retail group.

2*1996 ACS/scores for individual firms were released in the February 3, 1997 issue ofFortune. However, the time period surrounding the distribution of this issue was con-founded by earning announcements by most of the firms. As a result, we do not examinethe market's response to the release of the 1996 scores.


value-weighted index in CRSP); a, = the rate of return for firm i whenRmt = 0; P, = the systematic risk for firm i; and 8,7 = the random errorfor firm i in period t.

Market model parameters are computed using the 100 observationspreceding day -10 (i.e., day -110 to day -11). Abnormal returns for eachfirm are estimated using the formula:

e,7 = ^it- a,-- ^iRmt for -10<t< +10.

To be included in the final sample, each firm requires complete stock re-turn data for the estimation and event periods and no significant con-founding events from November 20, 1995 to December 18, 1995. Theserequirements eliminate 28 firms.

We estimate the following cross-sectional regression of cumulative ab-normal returns (CAR) following the release of the ACSI scores on thefirm's customer satisfaction index:

CARti = Po + Pi ACSIi + Ej,

where CARt, represents cumulative abnormal returns for firm i over thefive or ten days following the ACSI disclosure and ACSI, is the customersatisfaction score published in Fortune. If the ACSI scores released in For-tune provided new, economically relevant information to the market, Pjshould be positive.^^

The results are reported in table 10.̂ 6 Coefficients on ACSIzre positivebut insignificant when five-day abnormal returns are examined (panel A);however, ten-day abnormal returns are significantly related to both 1994and 1995 ACS/scores (panel B).^' The coefficient on 1994 ACS/impliesthat a five-unit difference in satisfaction (roughly one standard deviation

25 We do not have an expectations model for customer satisfaction. Consequently, ourmodels assume the market's ex ante beliefs were that customer satisfaction was the samefor all companies. We repeated the tests using changes in ACS/scores from 1994 to 1995.AC5/changes were not statistically significant (p > 0.15, two-tail).

28 We also conducted the analysis using standardized cumulative returns, where the ab-normal return for each firm is standardized using the estimated standard deviation of theresiduals from the market model and a factor reflecting the increase in variability due toprediction outside the estimation period (Patell [1976]). The advantage of this approachis that it controls for the likely cross-sectional differences in the distributions of stockprice return processes across the firms in our sample and should provide statistically well-specified tests. The results using standardized returns were nearly identical to those in ta-ble 10.

2'One potential limitation with regressing cumulative abnormal returns on ACS/scoresis that the time period selected for computing the abnormal returns may coincide withtrends in firm-specific returns arising from factors other than ACSI disclosure. We there-fore included an additional variable in the models representing cumulative abnormal re-turns over the five or ten days prior to the ACSI disclosure. The coefficient on ACSIremained insignificant when examining five-day returns and significant when examiningten-day returns.

1994 Scores

1995 Scores

-0.050(-1.05)

-0.042(-0.98)

0.001(1.08)

0.001(1.03)

Panel B: Ten-Day Abnormal Retums1994 Scores

1995 Scores

-0.123*(-1.96)

-0.102*(-1.73)

0.002*(1.97)

0.001*(1.76)


TABLE 10OLS Regressions Examining the Association between Cumulative Abnormal Retums at the Release of

the ACSI Scores in Fortune and the Firm's Customer Satisfaction Index "(t-statistics in parentheses)

Intercept ACSI Adjusted R^ i^-StatisticPanel A: Five-Day Abnormal Retums

0.00 1.16

0.00 1.05

0.03 3.90*

0.02 3.11*

'Statistically significant at the 10% level (two-tail)."Outliers are deleted from the models. The resulting sample sizes range from 103 to 105. The

dependent variables in the models are cumulative abnormal returns following the initial public releaseof the firms' ACSI scores. Cumulative abnormal returns are computed using the market model, withthe return on the market portfoho measured using the vahie-weighted index in CRSP. Five-day (ten-day) returns are measured over the five (ten) trading days following the initial release of individualACS/scores in Fortune.

from its mean) was associated with a 1% increase in ten-day cumulativeahnormal returns.^^

We investigate potential nonlinearities in the market's response toACSI disclosure by forming quartiles based on the firms' ACSI scores andtesting for differences in mean abnormal returns across the portfoliosusing CLM. The quartiles are computed by assigning firms to portfoliosbased on their 1994 or 1995 satisfaction scores. The 28 firms with con-founding announcements during the event period or firms with missingreturns are then deleted. The resulting quartiles range in size from 24to 33 observations.

The GLM results are presented in table 11. ACS/scores are significantpredictors of cumulative abnormal five-day returns using 1995 scores,and of cumulative abnormal ten-day returns using scores from eitheryear (p < 0.15, two-tail). All four models exhibit negative CARs in quar-tile 1, with an average ten-day return in the 1994 (1995) model of -1.9%(-0.5%). Least squares means are also negative in quartile 2 using 1995scores, with an average ten-day return of -1.1%. With the exception of

*̂ We repeated the analyses in table 10 using broad industry groupings with ten or moreobservations (manufacturing nondurables; manufacturing durables; transportation, utili-ties, and communications; and retail). Transportation, utility, and communications firmshad significant positive ACS/coefficients using ten-day returns and 1994 scores or five-dayreturns and 1995 scores. In the two specific industries with ten or more observations, foodprocessors had statistically negative associations between satisfaction and market perfor-mance in all four models, while utilities had statistically positive ten-day returns using ei-her ACSI score.


3O• ^

2

Has

,

CO1*.

13•a

c

2 « ^

^ CM

o t̂eo 00

CO <rq qd d

CO lO

o o

to too oq qd d

?- ooO -Hq qd d

sgo o

q qd d

t^ Olo oq qd d

o o

ij - d*• 9 91 ? ?

o

IQ

?5 W5 («

. . 0 0

a 2

o-xi b<" u I/)^J .^ Lavi a. oj

<" ^ ;(̂

- g a-o u 2

•H.-S Si S

3 60 !- U Ol

f

in G (^ "^ '-^ "J

1

•SO 2 S i - S -«. • U 3 C 5

>^^ .> S C ^= rt -; S rt w•ii « 3 •« £ -S > p .S c .-•= o 3 •" ^ "̂'^ o c S3 .2 gj

2 S •=


ten-day returns using 1994 scores, mean CARs in the two bottom quar-tiles were statistically equivalent. The negative or insignificant returnsin these groups indicate that the market did not react favorably to thedisclosure of ACSI scores for firms with satisfaction scores below themedian.

The largest positive responses occurred in the top-two quartiles, pro-viding some evidence that the market viewed higher satisfaction scoresas positive indicators of future cash flows. Least squares means in thesequartiles were significantly larger than means in the first quartile using1994 scores, and significantly larger than means in the first and/or sec-ond quartiles using 1995 scores. The average ten-day abnormal return inthe models was 0.75% in quartile 3 and 1.075% in quartile 4. However,the least squares means in the top-two quartiles were not statisticallydifferent, providing further evidence that improving customer satisfac-tion beyond a certain point may yield little additional economic gain.

Overall, the results in tables 10 and 11 generally are consistent withthe release of ACSI scores providing new information to the stock mar-ket. While the survey and experimental results of Mavrinac and Seisfeld[1997] indicated that institutional investors place little or no weight oncustomer satisfaction measures when valuing firms, our valuation mod-els and event study results provide some support for the hypothesis thatnonfinancial measures such as customer satisfaction affected the mar-ket's assessment of future cash flows.

6. Conclusion

This study contributes to a growing body of accounting research on thepredictive ability and value relevance of nonfinancial performance mea-sures (e.g.. Amir and Lev [1996], Foster and Gupta [1997], Mavrinacand Seisfeld [1997], and Banker, Potter, and Srinivasan [1998]). Usingcustomer and business-unit data, we find modest support for claims thatcustomer satisfaction measures are leading indicators of customer pur-chase behavior (retention, revenue, and revenue growth), growth in thenumber of customers, and accounting performance (business-unit reve-nues, profit margins, and return on sales). We also find some evidencethat firm-level customer satisfaction measures can be economically rele-vant to the stock market but are not completely reflected in contempo-raneous accounting book values.^^ However, some of the tests suggestthat customer behavior and financial results are relatively constant overbroad ranges of customer satisfaction, changing only after satisfactionmoves through various "threshold" values, and diminishing at high sat-

^ Compared to prior studies, an important difference in our firm-level tests is the anal-ysis of the market's response to newly released measures (i.e., the American Customer Satis-faction Index) rather than to nonfinancial information that is not contained in financialstatements but is readily available to market participants (e.g.. Amir and Lev [1996]).


isfaction levels. Taken together, our results offer qualified support forrecent moves to include customer satisfaction indicators in internal per-formance measurement systems and compensation plans (Kaplan andNorton [1996] and Ittner, Larcker, and Rajan [1997]).

There are a number of limitations to our analyses. First, although cus-tomer satisfaction is a choice variable for firms, our tests assume that thismetric is exogenous. The issue of endogeneity is particularly problem-atic when assessing the relation between performance and customer sat-isfaction. If all firms optimally select customer satisfaction levels based onexogenous factors, there should be no statistical relation between perfor-mance and customer satisfaction, after controlling for the exogenous de-terminants. We also assume that our results are due to nonoptimizingbehavior by firms rather than to model misspecification. Second, thecustomer satisfaction measures in our tests, like all satisfaction measuresused in practice, have somewhat arbitrary measurement properties (e.g.,weighted questionnaire scores are transformed to range between 0 and100), as opposed to being measures with cardinal properties. The arbi-trary nature of these measures makes it difficult to identify the functionlinking customer satisfaction to customer behavior and organizationalperformance. Third, customer satisfaction is likely to be measured witherror (e.g., the ACSI surveys only individual consumers and not busi-ness-to-business customers), causing a variety of well-known economet-ric problems. Fourth, the appropriate functional form linking customersatisfaction to future customer behavior and financial performance isunclear. Although we used several common models, we are uncertainwhether customer satisfaction is best measured in absolute or relativeterms, whether the variables in the models should be expressed in levels,changes, or percentage changes, and what is the correct lag between cus-tomer satisfaction and performance. Finally, the low explanatory powerof our customer satisfaction measures precludes a strong substantive in-terpretation on the appropriate role of this measure in predicting ac-counting performance and explaining economic value.

Our analysis is also limited to customer satisfaction metrics; researchusing a broader set of nonfinancial metrics may provide insight into themodest ability of customer satisfaction measures to explain future ac-counting and stock market performance. In addition, future researchmight probe the reasons for the differential results across industries.For example, the competitive structure of an industry, customer switch-ing barriers, the length of customer repurchase cycles, and other relatedfactors may help explain industry differences in customer satisfaction'simpact. Perhaps the most interesting extension would be a sophisticatedanalysis of the negative relation between customer satisfaction and per-formance in selected industries. In particular, it would be useful todetermine whether these results are due to model misspecification or tosituations where firms in a given industry have "overinvested" in cus-tomer satisfaction.


REFERENCES

AAKER, D. A., AND R. JACOBSON. "The Financial Information Content of Perceived Qual-ity." Journal of Marketing Research (July 1994): 191-201.

ALFORD, A. W. "The Effect of the Set of Comparable Firms on the Accuracy of the Price-Earnings Valuation Method." Journal of Accounting Research (Spring 1992): 94-107.

AMERICAN INSTITUTE OF CERTIFIED PUBLIC ACCOUNTANTS. Improving Business Reporting—A Customer Focus. New York: AICPA, 1994.

AMIR, E., AND B. LEV "Value-Relevance of Nonfinancial Information: The Wireless Com-munications \nA\xstry." Journal of Accounting and Economics (August-December 1996): 3-30.

ANDERSON, E. W.; C. FORNELL; AND D. R. LEHMANN. "Customer Satisfaction, MarketShare, and Profitability: Findings from Sweden." Journal of Marketing Research (July1994): 53-66.

ANDERSON, E. W.; C. FORNELL; AND R. T. RUST. "Customer Satisfaction, Productivity, andProfitability: Differences between Goods and Services." Marketing Science 2 (1997): 129-45.

ARTHUR ANDERSEN & Co. Customer Satisfaction Strategies and Tactics. Chicago: ArthurAnderson, 1994.

BANKER, R.; C. POTTER; AND D. SRINIVASAN. "An Empirical Investigation of an IncentivePlan Based on Nonfinancial Performance Measures." Working paper. University of Texasat Dallas, Cornell University, and University of Pittsburgh, 1998.

BARTH, M. E., AND M. F. MCNICHOLS. "Estimation and Market Valuation of EnvironmentalLiabilities Relating to Superfund Sites." Journal of Accounting Research (Supplement1994): 177-209.

BEAVER, W., AND D. MORSE. "What Determines Price-Earnings Ratios?" Financial AnalystsJournal (July/August 1978): 65-76.

BERTRAND, K. "Putting Customers First." Business Marketing (December 1989): 30-34.BOWBRICK, P. The Economics of Quality, Crades and Brands. New York: Routledge, 1992.Box, G. E. P, AND D. R. Cox. "An Analysis of Transformations."/ourraa/ of the Royal Statisti-

cal Society, series B (1964): 211-52.BUSHMAN, R.; R. INDJEJIKIAN; AND A. SMITH. "CEO Compensation: The Role of Individual

Performance KMduation." Journal of Accounting and Economics (April 1996): 161-93.DELOITTE TOUCHE TOHMATSU INTERNATIONAL. "Performance Measurement." Working

paper, 1994.EDVINSSON, L., AND M. S. MALONE. Intellectual Capital: Realizing Your Company's True Value

by Finding Its Hidden Brainpower. New York: HarperBusiness, 1997.EuxTRic Womjx "Utility-Customer Loyalty Still Up for Sale." (May 1994): 27-28.ERNST & YOUNG. The International Quality Study. New York: Ernst & Young and the Ameri-

can Quality Foundation, 1992.FELTHAM, C , AND J. A. OHt^ON. "Valuation and Clean Surplus Accounting for Operating

and Financial Activities." Contemporary Accounting Research (Spring 1995): 689-732.FIERMAN, J. "Americans Can't Get No Satisfaction." Fortune (December 11, 1995): 186-94.FORNELL, C. "A National Customer Satisfaction Barometer: The Swedish Experience."yoMr-

nal of Marketing (January 1992): 6-21.FORNELL, C , AND J. CHA. "Partial Least Squares." In Latent Variable Path Modeling with Par-

tial Least Squares, edited by J. B. Lohmoller, pp. 52-78. New York: Springer-Verlag, 1989.FORNELL, C ; M. D. JOHNSON; E. W. ANDERSON; J. CHA; AND B. E. BRYANT. "The Ameri-

can Customer Satisfaction Index: Nature, Purpose, and Findings."/ourraa/ of Marketing(October 1996): 7-18.

FOSTER, G., AND M. GUPTA. "The Customer Profitability Implications of Customer Satis-faction." Working paper, Stanford University and Washington University, 1997.

FRANKEL, R., AND C. M. C. LEE. "Accounting Valuation, Market Expectation, and theBook-to-Market Effect." Working paper. University of Michigan, 1995.

HRFocus. "Incentive Pay Focuses on Quality." (July 1993): 15.IBBOTSON, R. Stocks, Bonds, Bills and Inflation: 1996 Yearbook. Chicago: Ibbotson, 1996.


ITTNER, C. D., AND D. F. LARCKER. "Innovations in Performance Measurement: Trends andResearch impMcations." Journal of Management Accounting Research (Fall 1998).

ITTNER, C. D.; D. F. LARCKER; AND M. V RAJAN. "The Choice of Performance Measures inAnnual Bonus Contracts." The Accounting Review (April 1997): 231-55.

JONES, T O.; AND W. E. SASSER, JR. "Why Satisfied Customers Defect." Harvard Business Re-view (November/December 1995): 88-99.

JURAN, J. W., AND F M. GRYNA, JR. Quality Planning and Analysis. New York: McCraw-Hill,1980.

KAPLAN, R. S., AND D. P. NORTON. The Balanced Scorecard. Boston: Harvard UniversityPress, 1996.

LANCASTER, K, Variety, Equity, and Efficiency. New York: Columbia University Press, 1979.LANDSMAN, W. "An Empirical Investigation of Pension Fund Property Rights." The Account-

ing Review (October 1986): 662-91.MAVRINAC, S., AND T. SEISFELD. "Measures That Matter: An Exploratory Investigation of

Investors' Information Needs and Value Priorities." Working paper. University of West-ern Ontario and Ernst & Young, 1997.

MOORE, E. G. "Adoption of Customer Satisfaction and Marketing Constructs in ExecutiveBonus Plans: The Effect on Firm Performance." Working paper. University of Michigan,1997.

NATIONAL QUALITY RESEARCH CENTER. American Customer Satisfaction Index—MethodologyReport. Milwaukee, Wise: American Society for Quality Control, December 1995.

OHLSON, J. A. "A Synthesis of Security Valuation Theory and the Role of Dividends." Con-temporary Accounting Research (Spring 1990): 648—76.

OHLSON, J. A. "Earnings, Book Values and Dividends in Security Valuation." ContemporaryAccounting Research (Spring 1995): 661—87.

PATELL, J . M. "Corporate Forecasts of Earnings per Share and Stock Price Behavior:Empirical Tests." foumal of Accounting Research (Autumn 1976): 246-76.

REICHHELD, F F , AND W. E. SASSER, JR. "Zero Defections: Quality Comes to Services." Har-vard Business Review (September/October 1990): 105-11.

Ross, J., AND D. GEORGEOFF. "A Survey of Productivity and Quality Issues in Manufactur-ing: The State of the Industry." Industrial Management (January/February 1991): 3-5,22-25.

RYAN, M. J.; T BUZAS; AND V RAMASWAMY. "Making CSM a Power Tool." Marketing Re-search: A Magazine of Management and Applications (Summer 1995): 10-16.

SCHLOSSBERG, H. "Competition Sparks Gas Company to Become Customer Friendly." Mar-keting News (April 12, 1993): 8.

S-PLUS USERS MANUAL. Vol. 2. Seattle, Wash.: Statistical Sciences Inc., 1991.STEWART, T. A. Intellectual Capital: The Wealth of Organizations. New York: Doubleday/

Currency, 1997.WALLMAN, S. M. H. "The Future of Accounting and Disclosure in an Evolving World: The

Need for Dramatic Change." Accounting Horizons (September 1995): 81-91.WHITE, H. "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test

for Heteroskedasticity." Econometrica (May 1980): 817-38.WOLD, H. "Nonlinear Iterative Partial Least Squares (NIPALS) Modeling: Some Current

Developments." In Multivariate Analysis, edited by P. R. Krishnaiah, pp. 383-407. NewYork: Academic Press, 1973.

. "Soft-Modeling—The Basic Design and Some Extensions." In Systems under IndirectObservation, edited by K. Joreskog and H. Wold, pp. 1-54. Amsterdam: North Holland,1982.

11 1998 Itner Lacker Nonfinancial Measure

Documents

Transcript of 11 1998 Itner Lacker Nonfinancial Measure