An exploratory study of employee turnoverindicators as predictors of customer

satisfactionRobert F. Hurley and Hooman Estelami

Fordham University, New York, New York, USA

AbstractPurpose – The service profit chain postulates that higher employee satisfaction levels lead to high customer satisfaction, and ultimately affectconsumer loyalty and profitability. One construct that has largely been ignored in most of this research has been the role of employee turnover. Thispaper proposes that employee turnover can also be a powerful predictor of employee sentiment and resulting customer satisfaction levels.Design/methodology/approach – The relationship between employee satisfaction, employee turnover and customer satisfaction ratings is exploredusing an extensive data set from a chain of convenience stores. Employee perceptions were obtained from a survey which developed and administeredto all store personnel. Turnover data were obtained from archival data. The data are analyzed using path analysis.Findings – The test of various turnover indicators suggests that certain employee turnover indicators can perform as effectively as single-itememployee satisfaction ratings do in predicting customer satisfaction.Originality/value – The finding that turnover predicts customer satisfaction as effectively as employee satisfaction is new and has importantimplications. More attention should be paid to managing customer satisfaction through managing turnover. Also, the use of turnover as an indicator ofcustomer satisfaction should be explored in light of the fact that employee turnover is a naturally collected managerial measure, and does not requirethe costly administration of employee satisfaction surveys.

Keywords Customer satisfaction, Customer services quality, Human resource management, Employee turnover

Paper type Research paper

An executive summary for managers can be found at

the end of this article.

Introduction

The causal relationship between employee satisfaction,

customer satisfaction, and profitability is a topic of growing

academic and managerial interest (e.g. Oliver, 1997;

Reichheld, 1996; Rust et al., 1995; Estelami, 2000; Heskett

et al., 1997). This stream of research has helped conceptualize

the notion of a “service profit chain” (Heskett et al., 1994,

1997), in which firm profitability is hypothesized to be

dependent on the satisfaction levels of employees and

customers of a service organization. The service profit chain

postulates that higher employee satisfaction levels lead to high

customer satisfaction, and ultimately affects consumer loyalty

and profitability. This line of thinking not only has an intuitive

appeal, but it also highlights the critical role of customer and

employee satisfaction in the profit generation process, and

provides a vision for how service organizations should re-

engineer themselves in order to improve long-term

profitability.

One construct that has largely been ignored in most of this

research stream has been the role of employee turnover.

Research on organizational learning and knowledge

management provides a strong theoretical basis connecting

knowledge residing within employees and organizational

performance (Hurley, 2002; Kim, 1993). There is also

anecdotal evidence that higher levels of employee turnover

can lead to lower levels of customer satisfaction in retail

stores. For example, Schneider and Bowen (1993) report that

Sears has experienced that stores with lower rates of employee

turnover have higher levels of customer satisfaction. High

employee turnover may not only be indicative of a poor work

environment, but it may also be reflected in the loss of

experienced employees and established customer

relationships, resulting in negative effects on the customer.

However, currently there is little empirical research to help

validate this view, and to better understand the capabilities of

employee turnover measures as predictors of customer

satisfaction. Such an inquiry would be especially appealing,

since, unlike employee and customer satisfaction surveys,

which are time consuming and costly to collect, employee

turnover is a naturally collected managerial measure in almost

all organizations. The accessibility of this measure may

therefore help service organizations gain a clearer picture of

the dynamics of the service profit chain.

The current issue and full text archive of this journal is available at

www.emeraldinsight.com/0887-6045.htm

Journal of Services Marketing

21/3 (2007) 186–199

q Emerald Group Publishing Limited [ISSN 0887-6045]

[DOI 10.1108/08876040710746543]

Received: February 2005Revised: August 2005Accepted: November 2005

186

In an earlier study, Loveman (1998) empirically tested the

components of the service profit chain and found that while

parts of the model are supported, “work is needed to refine

and simplify several critical measures” (p. 18). Loveman’s

results suggest weak, and in some cases insignificant,

relationships, between employee satisfaction and customer

satisfaction measures. This paper builds on Loveman’s (1998)

work by systematically exploring the role of employee

turnover on customer satisfaction in a service organization.

Using data from a chain of 275 convenience stores, the

relationship between employee turnover and customer

satisfaction is explored. Results show that certain employee

turnover indicators predict customer satisfaction levels well,

and their predictive ability is equivalent to that of single-item

employee satisfaction measures gained through employee

surveys. The paper concludes with the managerial and

research implications of the findings.

Employee satisfaction and customer satisfaction

It is well established that customer satisfaction is largely a

function of the discrepancy between a customer’s expectations

and the actual experience received from the act of

consumption (Oliver, 1980). In contrast to the consumption

of goods, in service encounters, the satisfaction process is

complicated by the fact that service expectations and

experiences are often multi-dimensional in nature.

Numerous facets of a service, such as the speed of delivery,

the physical atmosphere, and employee behavior can

significantly influence consumer judgments of a service (e.g.

Parasuraman et al., 1988; Zeithaml et al., 1996). The latter

factor – employee behavior – is known to play an especially

pivotal role in consumer satisfaction with service encounters,

and is therefore of focal interest in many quality improvement

initiatives in service organizations (Bitner et al., 1990;

Estelami, 2000; Estelami and DeMaeyer, 2002; Keaveney,

1995).

The quality of the encounter between the service provider

and the customer should help improve consumer satisfaction

levels. While some empirical research has shown a positive

correlation between employee satisfaction and customer

satisfaction ratings (Schlesinger and Zornitsky, 1991), much

of the existing research has focused on the drivers of employee

satisfaction, and the consequences of customer satisfaction.

For example, Schlesinger and Heskett (1991) suggest that

antecedents such as improved employee incentives, training,

and choice of front-line employees in service organizations

can help increase employee satisfaction levels. In addition, the

works of Anderson and Sullivan (1993) and Fornell (1992) on

the Swedish customer satisfaction barometer suggest a

positive causal relationship between customer satisfaction

and customer loyalty.

While the above relationships seem intuitive, practical, and

pragmatically appealing, the strength of the system of

relationships seems to be weak. For example, in their

models of the economic returns expected from quality

improvement initiatives, Rust et al. (1995) suggest that such

initiatives may prove to be unprofitable. In estimating the

return on investments made in quality improvement

programs, the authors caution that “it is possible to spend

too much on quality” and that “not all quality expenditures

are equally valid” (p. 68). As a result, while some studies have

documented a positive correlation between customer

satisfaction and profitability (e.g. Anderson et al., 1994),

others have found the relationship to be very weak or

insignificant. For example, in a retail context, Bernhardt et al.

(2000) have found no significant relationship between

customer satisfaction and profitability.

Employee turnover as a predictor of customersatisfaction

Of interest to this paper are the findings of Loveman (1998) in

a retail banking environment. While Loveman (1998) found a

positive link between employee satisfaction and customer

satisfaction ratings, he concluded that “the link between these

two components of the service profit chain is among the

weakest” (p. 260). The weakness of this relationship, also

observed in prior works (e.g. Rucci et al., 1998; Schlesinger

and Zornitsky, 1991) may be attributed to a variety of factors.

One possibility is that the majority of existing studies utilize

simple correlations as a basis of analysis, and few take into

account the structural interdependencies that may exist

among the various components of the service profit chain.

Use of structural equations modeling methods (e.g. Oliver,

1980) may help clarify the direct and indirect effects in the

service profit chain (Duncan, 1975). The use of correlation-

based approaches that do not utilize lagged measures also

poses problems in making causal inferences, whereby the

employee satisfaction-to-customer satisfaction link may

potentially be reverse (i.e. unhappy customers creating a

poor work environment, leading to employee dissatisfaction,

and to increased turnover).

A second potential explanation for a lack of significant

relationships observed between employee satisfaction and

customer satisfaction may be the fact that employee

satisfaction surveys provide measures that lack basic

reliability and validity criteria needed for empirical analysis.

Employee satisfaction measures that are often obtained

through self-administered surveys conducted at the place of

employment may suffer from a series of non-sampling biases

(e.g. Churchill, 1996) that may limit the validity of these

measures as indicators of employee sentiment. It is not

difficult to imagine employees who, due to factors such as fear

of job loss, lack of interest, or the presence of co-workers or

supervisors at the time of completing the survey, fail to

objectively respond to questions on an employee satisfaction

questionnaire. Moreover, employee satisfaction surveys, by

definition, include only current employees, and typically

exclude ex-employees, whose beliefs and attitudes about the

work environment may be just as valuable.

One remedy is for one to seek employee satisfaction

indicators that do not rely on self-reports obtained through

employee surveys. Such measures are less influenced by

measurement error, and perhaps more reflective of the

profitability associated with employee satisfaction. Day

(1994) makes the theoretical link between employee

turnover and firm performance by suggesting that recall

capabilities of a firm are eroded when people leave. When

people depart, their intelligence regarding processes,

methods, and customers also leaves. In a service

organization, a logical hypothesis is that higher levels of

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

187

employee turnover will be associated with lower levels of

customer satisfaction. In retail, audit, tax, legal or other

service settings, when a knowledgeable service provider leaves

and is replaced by someone who is learning methods and

customer interactions for the first time, some customer

dissatisfaction is expected to result. Employee turnover may

therefore provide a basis for gauging not only employee

satisfaction, but also customer satisfaction levels.

Full-time versus part-time employee turnover

Research on the relative influence of part-time and full-time

employee turnover on service quality is limited. Most of the

research on employee turnover has focused on personal selling

or industrial sales situations, rather than the retail

environment (e.g. Futrell and Parasuraman, 1984; Lucas

et al., 1987). When research on turnover has been conducted

in retail settings, the focus has been on turnover among

management personnel, and not front-line salespeople (Lucas

et al., 1990; Keaveney, 1992). Several studies have also

examined the various causes of employee turnover (Good

et al., 1988; Darden et al., 1987). However, no research has

attempted to link employee turnover indicators to service

provider performance.

It is important to note, as Keaveney (1992) did, that some

turnover may be functional (poor service provider replaced by

a good one) and some is dysfunctional (good service provider

leaving). Day’s (1994) argument surrounding the loss of

organizational learning associated with employee turnover

may be true to different degrees, depending on whether or not

the employee leaving the company has a rich knowledge of the

organization. Therefore, significant variation can be expected

between the implications of turnover among part-time and

full-time employees. Full-time employee turnover may imply

the loss of years of training and development of the employee,

and a high cost of replacement. In contrast to part-time

employees, replacing a full-time employee may require

significant investments and a long break-in period for the

incoming employee, and may be associated with poor service

delivery.

Therefore, while Day’s (1994) argument regarding the role

of employee turnover may be true for service organizations

with complex service encounters, it may not be as critical a

factor in many retail contexts where the service encounter is

not associated with a high level of technical complexity. In

such contexts, full-time employee turnover may not

necessarily imply a loss of valued skills, and part-time

turnover may actually be a more accurate representation of

the quality of the work environment. This is because part-

time employees typically experience lower levels of

organizational commitment (Hunt and Morgan, 1994;

Meyer and Allen, 1991; Sightler and Adams, 1999). Full-

time employees may be bound to the organization as their sole

source of income (Fenton-O’Creevy, 1995; Jackofsky et al.,

1986). Part-time employees are therefore more likely than

full-time employees to leave a low-quality work environment.

This discussion has established that there is support in the

literature that:. the effects of turnover on various performance measures

may differ, depending on whether those leaving are full-

time or part-time employees; and

. since part-time employees have a more marginal

relationship with the firm, it is reasonable to hypothesize

that turnover among this population may be a more

sensitive indicator of organizational performance with

regard to customer satisfaction.

Interaction between part-time and full-time turnover

In addition to their individual effects, part-time and full-time

employee turnover may behave in an interactive fashion.

Research in organizational behavior suggests that the

existence of multiple indicators of poor performance in an

organization often serves as a signal of a deteriorating work

environment (Adizes, 1999; Wexley and Yuki, 1984). A

weakly run organization produces poor results, which tend to

manifest themselves not in just one indicator, but rather in a

variety of outcome variables, such as profitability, market

share, and various employee retention measures. Existence of

high turnover at various levels of an organization is a strong

indicator of a poorly run work environment (Griffeth and

Hom, 2001; O’Malley, 2000). Simultaneously high part-time

and full-time employee turnover may be a strong indicator of

a poor work environment, with subsequent effects on

customer satisfaction. As such, we expect that the

interaction between full-time and part-time employee

turnover may help predict customer satisfaction levels.

Relationship between turnover and customer

satisfaction

While Day (1994) suggests that the relationship between

employee turnover and customer satisfaction must be

negative, the shape of this relationship has not been

investigated in existing research. The shape of this

relationship, which may be linear, or may significantly

deviate from linearity due to threshold effects, is an

important issue. This may play a critical role in determining

the ability of employee turnover indicators in predicting

customer satisfaction. For example, emerging research in

personnel management suggests that the relationship between

an employee’s turnover likelihood and his/her job

performance may be non-linear (e.g. Jackofsky, 1984;

Williams and Livingstone, 1994). This research stream

would suggest that employees who are poor performers have

a disproportionately high likelihood of being asked to leave an

organization. As a result, job performance and turnover rates

are related in a non-linear fashion.

Steers and Modway (1981) also suggest that low-

performing employees have an increased interest in

voluntarily leaving an organization, since they gain little job

satisfaction from operating within their work environment.

Support for the non-linear relationship between turnover and

job performance has also been found by Williams and

Livingstone’s (1994) meta-analysis of the empirical results of

four previous studies. However, evidence also exists to refute

non-linearity. For example, Birnbaum and Somers (1993)

found no such evidence for a non-linear relationship in their

empirical work. Since an employee’s job performance is

expected to affect the quality of service delivered to the end

customer, the existing literature is largely unclear as to how

employee turnover and customer satisfaction will relate. As a

result, one purpose of this paper is to explore a variety of

functional forms that may best capture this relationship.

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

188

Since the role of part-time and full-time turnover on

customer satisfaction ratings is unexamined in current

research, we will explore the predictive ability of these

turnover indicators, and their interaction in relation to

customer satisfaction. Moreover, the predictive performance

of these indicators and their non-linear transformations will

be contrasted with those of employee satisfaction measures

obtained through questionnaires. The turnover indicators to

be examined are:. Full-time employee turnover, defined as the total number of

full-time employees who left during a 12-month period,

divided by the average number of full-time employees in

the store during that period.. Part-time employee turnover, defined as the total number of

part-time employees who left during a 12-month period,

divided by the average number of part-time employees in

the store during that period.. Total turnover, the sum of part-time and full-time turnover.

Note that this is a simple sum of the two turnover

measures. While a weighted total turnover measure could

be produced by dividing the total number of employees

leaving and organization by the average number of people

employed during the year, the measure used here is an un-

weighted measure, which reflects the proportion of

turnover equally among the two employee populations,

regardless of their respective sizes.. Turnover product, the product of part-time and full-time

turnover.

While the first three indicators capture the raw effects of

employee turnover, while the fourth attempts to capture

interactions that may exist between part-time and full-time

turnover. This variable is designed to explore whether

turnover among these different groups interacts and creates

differential impacts on customer satisfaction. The fact that

these two types of employees interact in the store and have

different relationships with the company suggest that this is a

reasonable area for exploration. For each of the above

indicators, two non-linear transformations are also explored.

The first is the natural log of the corresponding turnover

indicator, and the second is the inversion of the corresponding

indicator. In addition, exponential transformations of the raw

turnover measures (squared) are examined. These are

common transformations utilized for capturing non-linear

relationships (Hair et al., 1998; Lehmann et al., 1998). The

relative performance of each indicator is explored below using

a survey of both employee and customer satisfaction ratings,

coupled with employee turnover data.

Empirical investigtion

Three convenience store chains involved in quality

improvement initiatives were approached to gain

cooperation for a field study. Two chains agreed to

cooperate. One chain operated 175 convenience stores in

four areas of the USA. The other chain operated 100 stores in

one state in the Northeast USA. Each store sold grocery

items, beverages, and fast food. Compared to most

convenience stores, these stores were large (2,000-5,000

square feet), and from two to ten people would be working in

the store on a shift depending on the traffic volume in the

store. Research conducted by these companies in the past

highlighted the importance of customer service and customer

satisfaction in protecting market share.

Employee and customer satisfaction surveys

Employee perceptions were obtained from a survey that we

developed, and administered to all store personnel at both

companies. The dimensions of employee perceptions to be

measured were determined based on the literature in retailing

and services that had related various aspects of organization

functioning to performance (Birnbaum and Somers, 1993;

Day, 1994; Keaveney, 1992; Schneider and Bowen, 1993). In

addition, focus groups were held in both chains with

knowledgeable employees to identify critical aspects of

employee and customer satisfaction. Aspects of employee

and customer satisfaction were converted into survey items. A

pilot instrument was reviewed by two focus groups in each

chain to assure that the items were clear, and measured each

aspect of employee and customer satisfaction. Due to space

limitations, the support in the literature for the organizational

dimensions selected is not reviewed here.

To validate the measurement dimensions, implicit theories

were elicited of the effects of store organization and operation

on employee performance. These implicit theories were

identified via open-ended interviews and focus groups with

store personnel. Survey items were written to measure these

aspects of store effectiveness, and the items were reviewed for

face validity and clarity via another series of focus groups with

a different sample of district managers, store managers, and

salespeople. The responses were confidential, and sent to the

authors without being seen by any company personnel. The

questionnaire utilized five-point Likert scales. Employees

rated each statement according to the degree to which they

agreed or disagreed with it (5 ¼ strongly agree to 1 ¼ strongly

disagree). Examples of two statements in the Standards and

Goals scale were: “In our store we emphasize quality in

everything we do,” and “In our store we have clearly defined

standards for quality”. In addition to the detailed battery of

employee satisfaction scales, a single-item employee

satisfaction measure was obtained by asking employees to

rate their job satisfaction level on a 1-5 scale.

In total, over 3,500 employees filled out questionnaires,

with the response rate per store ranging from 40 percent to

100 percent. Scale development had two purposes. First, we

wanted to develop empirically valid scales. Second, we wanted

to develop scales that provided useful diagnostic information

to the sponsoring organizations. Factor analysis was

conducted using principal components analysis, along with a

Varimax rotation, to determine the factor structure. Scales

were purified by eliminating items that had loadings of less

than 0.6 on multiple factors, resulting in a total of nine

employee satisfaction scales. In addition, item-to-total

correlations and coefficient alphas were computed. All scales

had coefficient alphas great than 0.7 and exhibited a high level

of face validity as reflected in discussions with the two

sponsoring organizations regarding their ability to tap into

distinct aspects of store management. Appendix 1 reports the

descriptive statistics, definitions and, where appropriate,

coefficient alphas for all measures. It is important to note

that the unit of analysis for this study is the store, and that the

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

189

employee survey measures represent an aggregation of

individual responses to arrive at a mean score for each store.

Customer satisfaction data were obtained from a

standardized survey that the companies had already

collected from store intercept interviews with shoppers.

These interviews, conducted by trained interviewers

employed by an outside research company, were part of the

companies’ ongoing customer satisfaction measurement

program. Customers were asked their overall satisfaction

with the store, along with their rating of specific facets of

quality, such as friendliness and speed of service. Ratings were

done on a five-point scale (5 ¼ excellent to 1 ¼ poor).

Employee turnover

Turnover data for each store were obtained from company

records. Turnover was computed separately for full-time and

part-time employees. As is typical in turnover studies, the

turnover percentage was computed as the number of

employees leaving over the last 12 months, divided by the

average number of employees in the store during that period

(Ditz, 1971). In order to ensure a true causal relationship

between the measures of interest, the customer satisfaction

survey was conducted within approximately a six-month lag of

the employee satisfaction survey, and the turnover measures

corresponded to the prior fiscal year. The customer and

employee satisfaction measures were obtained in May-July of

1995, and the turnover measures reflected the full 1995

figures, which were obtained at the end of the calendar year.

Access to this data for research publication purposes was

made possible several years following data collection in order

to protect the sensitive nature of the data and the

confidentiality of the sponsoring organizations.

This time plan was adopted in order to ensure precedence

in the cause-and-effect relationship being examined

(Churchill, 1996), and to avoid reverse-causality

explanations whereby, for example, customer dissatisfaction

may be considered as the cause of employee turnover (e.g.

dissatisfied customers making the work experience less

pleasing, and hence resulting in employee turnover), rather

than the other way around. In addition, since the frequency of

shopping varied widely among customers surveyed between

multiple visits per day to monthly, it was expected that it

would take time for changes in store conditions to be

perceived by customers. Use of this scale of lagged effects is

also consistent with prior studies in customer satisfaction

research (e.g. Anderson and Sullivan, 1993; Anderson et al.,

1994).

Results

In order to examine the dynamics of the service profit chain as

it relates to various turnover indicators, path analysis is

utilized. Path analysis enables the use of regression methods

in order to examine causal relationships between constructs of

interest, and its use in this context is consistent with prior

works on customer satisfaction research (e.g. Oliver, 1980;

Anderson and Sullivan, 1993; Loveman, 1998). Alternative

structural equations modeling approaches such as LISREL

cannot be utilized, due to the stringent sample size

requirements of their maximum likelihood estimation

procedures.

The structure of the model being examined is outlined in

Figure 1. This model represents the causal relationships

postulated in the service profit chain (Heskett et al., 1997;

Loveman, 1998). Employee satisfaction is considered to be

driven by perceptions of various store management variables

such as training, communications, and empowerment. The

resulting employee satisfaction is expected to influence

employee loyalty levels, as reflected in the turnover

indicators. Moreover, employee satisfaction and turnover

may both have direct and indirect effects on customer

satisfaction levels. The objective of our analysis is to explore

the relative predictive performance of the 16 different

turnover indicators in the above framework.

Table I presents the regression results for the drivers of

employee satisfaction. In order to facilitate interpretation, all

variables have been standardized, resulting in parameter

estimates that can be interpreted as the path coefficients

(Duncan, 1975; Hair et al., 1998). As can be seen from Table

I, the regression model achieves a high degree of fit, as

reflected by an R2 of 0.68 (F9;264 ¼ 63:3; p , 0:001). In

addition, the effects of universally accepted drivers of

employee satisfaction are in their expected directions.

Positive effects can be found for the presence of standards

and goals (p , 0:01), training (p , 0:01), scheduling

(p , 0:01), empowerment (p , 0:05), and conflict

resolution (p , 0:05).However, there are factors that show no significant effects

on employee satisfaction. These relate to the presence of tools

and methods, communications, and performance

management. These are generally expected results, since in

a front-line retail environment, workplace quality is much

more likely to be influenced by the operational aspects of the

business, such as scheduling and training, than by subjective

components such as the quality of store meetings, which may

be difficult to both manage and measure. However, as will be

discussed shortly, the results may also be attributed to the

correlation among some of the predictor variables.

In order to examine the causal relationship between

employee satisfaction and the various turnover indicators

under study, and to determine the most effective turnover

indicator, a series of regression analyses were run. In these

analyses, the independent variables used were the single-item

employee satisfaction measure, and the multi-item employee

satisfaction measures (i.e. standards and goals, tools and

methods, etc.). The dependent variables used were the

various turnover indicators discussed earlier (i.e. part-time

turnover, full-time turnover, total turnover, product turnover,

log of part-time turnover, log of full-time turnover, etc.).

These variables varied from one regression analysis to the

next, in order to identify the turnover indicator with the

strongest link to the predictor variables.

Overall, several patterns emerged from this analysis. The

first relates to the significantly higher R2 gained for part-time

turnover than for full-time turnover. The corresponding R2

for part-time turnover was found to be about four times that

of full-time turnover. This observation is consistent with

earlier works that have suggested part-time turnover to be a

more responsive measure of work-place quality. With the

lower level of organizational commitment exhibited by part-

time employees (e.g. Fenton-O’Creevy, 1995; Sightler and

Adams, 1999), their likelihood of leaving a poor quality

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

190

organization is considerably higher, resulting in a stronger link

between employee satisfaction measures and turnover rates.

This is especially true in retail environments, where the part-

time workforce is heavily involved in customer interactions.

In addition, the results indicated that the single-item

employee satisfaction measure, and most of the multi-item

employee satisfaction measures, generally have a weak

relationship with most turnover indicators, as reflected in

the low R2s. Generally, the exponential transformations

(quadratic and square root) of the turnover indicators seem to

perform poorly, and the log transformations seem to provide

the best fit. These results identified the best-fitting turnover

indicator – as reflected by the corresponding R2 level – to be

the log of total turnover. The regression results related to this

specific analysis are shown in Table II.

The regression analysis shows that factors such as labor

scheduling, conflict resolution, and performance management

help affect turnover. This is not surprising, since the

appropriate management of these factors is expected to

influence an employee’s propensity to leave the organization.

It is important to note that the insignificant or weak effect of

the various employee satisfaction variables on employee

turnover, observed in Table II, is consistent with the findings

of Loveman (1998) in a retail-banking environment. These

results, and those of Loveman (1998), suggest that due to

measurement error, lack of an attitude-behavior link,

multicollinearity or other unmeasured factors, employee

satisfaction measures are not necessarily ideal predictors of

employee behavior, especially as it relates to employee loyalty.

To further test the relative performance of employee

satisfaction measures against those of the turnover indicators,

an additional regression analysis was conducted, utilizing both

employee turnover and employee satisfaction measures

(single-item, as well as multi-item) as predictors of customer

satisfaction. The results of this regression are shown in Table

III. As can be seen from Table III, the regression is statistically

significant. Although the R2 level of 0.24 is somewhat low, it is

consistent with model fit results of prior studies involving

lagged measures (e.g. Anderson et al., 1994, 1997). Both the

log of total turnover and the single-item employee satisfaction

measure have a significant relationship with customer

satisfaction. In addition, some of the employee satisfaction

component measures (e.g. standards and goals, labor

scheduling, and empowerment) exhibit a pattern of

insignificance. The variables that seem to exhibit

significance are tools and methods, training, manager

communications, store meetings, and performance

management.

In order to explore the influence of multicollinearity among

the employee satisfaction measures, we turned to the Pearson

correlation estimates for the predictor variables (Appendix 2).

As can be seen in Appendix 2, most of the multi-item

employee satisfaction measures exhibit high levels of inter-

correlations. A principal components analysis of the employee

satisfaction measures indicates a condition number (this is the

Table I Determinants of overall employee satisfaction

Coefficient estimate Standard error Significance

Intercept 0.005 0.035 p ¼ 0:98

Standards and goals 0.219 0.072 p , 0:01

Tools and methodsa 20.036 0.074 p ¼ 0:63

Training 0.248 0.064 p , 0:01

Labor scheduling 0.207 0.052 p , 0:01

Manager communicationsa 0.020 0.102 p ¼ 0:84

Empowerment 0.213 0.093 p , 0:05

Conflict resolution 0.111 0.044 p , 0:05

Store meetings 0.074 0.044 p , 0:1

Store manager manages performancea 20.032 0.063 p ¼ 0:61

Notes: R2 ¼ 0:683; F9;264 ¼ 63:33; p , 0:001; anot significant at the p , 0:1 level

Figure 1 Components of the service profit chain under study

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

191

ratio of the largest eigenvalue to the smallest eigenvalue) of

80.1, which is considered high by conventional standards of

regression analysis in marketing applications (Belsley et al.,

1980; Ofir and Khuri, 1986). Therefore, a factor analysis with

an orthogonal transformation was conducted on the employee

satisfaction scales (see Table IV). The number of factors

retained by the factor analysis was determined based on the

eigenvalue . 1 principle (Hair et al., 1998, pp. 103-4). The

analysis resulted in two factors based on each variable’s factor

loading being greater than 0.6 (Hair et al., 1998, pp. 111-2).

The first factor loads high on the following measures:

standards and goals, training, manager communications,

empowerment, and performance management, and accounts

for 63 percent of variation in the data. It can best be

summarized as “process management”, and will be

accordingly referred to in the balance of the paper. The

second factor, which will be referred to as “people

management”, loads on the following employee satisfaction

measures: tools and methods, labor scheduling, conflict

resolution, and store meetings, and accounts for 12 percent of

variation in the data.

The factor analysis results enable a reanalysis of previous

regression results, using a reduced model, whereby the

various employee satisfaction measures are grouped according

to the uncovered factor structure. Table V summarizes these

results. Similar to Table III, in all the regressions presented,

the dependent variable is customer satisfaction. The

independent variables are the employee turnover indicator

(log of total turnover), the single-item employee satisfaction

measure, and the employee satisfaction multi-item scales (in

reduced form). As expected, the R2s from the regressions are

generally slightly lower than those observed in Table IV, due

to data reduction. However, the results seem to be quite

Table III 2 Regression analysis for customer satisfaction

Coefficient estimate Standard error Significance

Intercept 20.05 0.06 p ¼ 0:87

Log of total turnover 20.17 0.08 p , 0:05

Employee satisfaction (single-item) 0.26 0.10 p , 0:05

Standards and goalsa 20.08 0.12 p ¼ 0:68

Tools and methods 0.24 0.12 p , 0:05

Training 20.21 0.11 p , 0:1

Labor schedulinga 0.02 0.09 p ¼ 0:95

Manager communications 20.39 0.17 p , 0:05

Empowermenta 20.03 0.15 p ¼ 0:82

Conflict resolution 0.04 0.08 p , 0:01

Store meetings 0.25 0.07 p , 0:01

Store manager manages performancea 0.18 0.11 p ¼ 0:15

Notes: R2 ¼ 0:24; F11;230 ¼ 6:58; p , 0:001; anot significant at the p , 0:1 level

Table II Regression analysis for log of total turnover

Coefficient estimate Standard error Significance

Intercept 0.11 0.07 p ¼ 0:96

Employee satisfaction (single-item)a 0.13 0.08 p ¼ 0:12

Standards and goalsa 0.07 0.07 p ¼ 0:70

Tools and methodsa 20.09 0.10 p ¼ 0:84

Traininga 0.09 0.09 p ¼ 0:96

Labor scheduling 20.39 0.07 p , 0:01

Manager communicationsa 20.07 0.14 p ¼ 0:82

Empowermenta 20.06 0.12 p ¼ 0:38

Conflict resolution 20.17 0.06 p , 0:01

Store meetings 20.16 0.06 p , 0:01

Store manager manages performance 0.31 0.09 p , 0:01

Notes: R2 ¼ 0:32; F10;240 ¼ 11:1; p , 0:001; anot significant at the p , 0:1 level

Table IV Factor analysis of employee multi-item scales

Factor 1 Factor 2

Eigenvalue 5.66 1.04

Standards and goals 0.76 0.43

Tools and methods 0.59 0.64

Training 0.67 0.53

Labor scheduling 0.38 0.73

Manager communications 0.93 0.21

Empowerment 0.88 0.29

Conflict resolution 0.06 0.84

Store meetings 0.32 0.63

Performance management 0.87 0.19

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

192

stable, and the multi-item employee satisfaction factors seem

to overwhelm the regression, as indicated by their

corresponding beta coefficients.

Overall, the people management factor seems to positively

influence customer satisfaction ratings. This suggests that

activities focusing on reduction of conflict among employees,

improvement of labor scheduling, and providing the

employees with tools and methods help improve customer

satisfaction levels. On the other hand, process management

has a negative influence on customer satisfaction. This

suggests that store initiatives such as performance

management, employee training, and empowerment do not

necessarily result in higher levels of customer satisfaction.

Since two separate store chains were used in the data

collection process, an additional regression of customer

satisfaction was conducted with an added dummy variable

referring to the store chain to which the store is associated, as

well as interaction terms between the dummy variable and

predictors. This was done in order to test if the pooling of the

data between the two chains is appropriate. The dummy

variable and the interaction terms were not found to be

statistically significant, indicating that the data could be

pooled between the two store chains. However, while the

reduced pooled model reported in Table V has resulted in a

lowering of multicollinearity (condition number ¼ 10:6), thecorrelations among the employee satisfaction factors and

single-item employee satisfaction measure are still high (0.7

and higher). Therefore, to uncover the predictive ability of the

various independent variables in determining customer

satisfaction levels, a series of regressions using each

independent variable separately and relating it to customer

satisfaction (as the dependent variable) were conducted.

These are shown in Table VI.

A distinct hierarchy emerges from the results. In general,

use of the multi-item employee perception scales results in the

highest R2 level. This is followed by the reduced two-factor

model, whereby the employee perception measures are

grouped into the two individual factors described earlier. A

significant drop in model fit occurs when the single-item

employee satisfaction measure is used. These results are

expected, as the multi-item measures, by nature, contain

more relevant and detailed information that might relate to

customer satisfaction levels. However, what is interesting in

the results is that some of the turnover indicators perform

equally well to the single-item employee satisfaction measure.

In particular, this is evident in the log of total turnover, which

results in the highest R2 level among all turnover indicators.

Interestingly, this pattern is consistent with the analyses

preceding Table III, where the best model fit was obtained

through the log of total turnover. In addition, similar to the

results discussed earlier, the exponential forms of the turnover

indicators generally are poor predictors of customer

satisfaction. While the overall R2 levels are low, these results

provide some evidence that the log-total turnover indicator

can predict customer satisfaction at levels equivalent to those

obtained from the single-item employee satisfaction measure.

Discussion

The results of this paper suggest that certain employee

turnover measures can be equally useful in predicting

customer satisfaction levels as the single-item employee

satisfaction measure. A study of 16 turnover indicators

suggests that raw turnover measures in general do not have a

significant effect on customer satisfaction levels. However,

non-linear transformations of these measures have been found

to be better predictors of customer satisfaction ratings. This

suggests that the relationship between raw employee turnover

measures and customer satisfaction is non-linear.

The non-linear nature of this relationship may help explain

the weak link observed in past studies of the service profit

chain between employee turnover and customer satisfaction

(e.g. Heskett et al., 1994). From a research perspective, this

observation suggests that instead of using raw turnover

figures, future researchers may consider using the log-total

turnover indicator presented here in examining the dynamics

of the service profit chain. Moreover, from a practical,

predictive perspective, such a non-linear transformation may

provide better estimates of consumer satisfaction levels for

managerial decision-making.

It is important to note, however, that the multi-item (rather

than the single-item) employee perception ratings have

outperformed all predictors in determining customer

satisfaction levels. This observation validates the use of

employee surveys, especially in situations where the collected

data also focuses on the predictors of employee turnover, and

not simply on employee satisfaction or other perceptual

measures. However, as also observed here, the high level of

correlation that by nature exists between multi-item employee

satisfaction measures reduces the prescriptive value of such

measures, as it relates to customer satisfaction. Figure 2

provides a visual presentation of the log-total turnover

indicator that was found to be the optimal turnover

predictor of customer satisfaction. The x-axis corresponds

to the raw full-time turnover. Like any turnover measure, this

figure has a lower bound of 0.0, in which case there is no full-

time employee turnover experienced by the business. A

turnover of 1.0 indicates that the number of employees who

Table V Regression analysis for customer satisfaction – reduced model

Coefficient estimate Standard error Significance

Intercept 20.04 0.05 p . 0:1

Log of total turnover 20.10 0.06 p , 0:1

Employee satisfaction (single-item) 0.19 0.09 p , 0:1

Factor 1: process management 20.28 0.09 p , 0:01

Factor 2: people management 0.40 0.10 p , 0:01

Notes: R2 ¼ 0:196; F4;238 ¼ 14:56; p , 0:001

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

193

left the business during the year is equivalent to the average

number of employees hired by the business throughout the

year. Higher turnover figures indicated a larger proportion of

employee flight. As can be seen, generally, the higher the

turnover rates, the higher the value of the log-total turnover

indicator. However, the non-linear nature of the indicator is

evident in situations where turnover levels are low (e.g. below

25 percent). To further utilize the survey measures in

specifically determining the predictors of employee turnover,

an additional regression, reported in Table VII, was carried

out.

The dependent variable here is the log-total turnover, and

the independent variables are the reduced employee

perception factors, which were identified earlier and utilized

in Table V. Consistent with the results of Table V, it is evident

that process management initiatives result in an increase in

turnover, as evident by the positive coefficient. This is an

interesting result, as it suggests that in the retail environment,

process management efforts may provide some employees

with a stronger incentive to leave the organization. On the

other hand, the people management factor seems to

negatively influence the log-total turnover indicator. This

phenomenon may be reflective of the fact that in service

organizations that are often highly people-oriented,

improvements in how people are managed and guided may

result in a decline in employee desire to leave the

organization.

Managerial implications and future research

From a services marketing and management perspective, the

results presented here are significant in that they question the

merits of costly employee surveys, which may fail to reflect

subsequent employee loyalty levels, or the resulting customer

satisfaction. The predictive ability of the log-total turnover

indicator in determining customer satisfaction was found to be

equivalent to the single-item employee satisfaction measure.

This is especially important, since unlike employee satisfaction

surveys, which require considerable data collection effort,

employee turnover is a commonly collected measure in most

organizations. Therefore, no incremental costs are associated

with obtaining turnover data, and therefore the use of an

appropriate turnover indicator may enable the management to

gain better foresight into customer satisfaction levels. It is

important to acknowledge, however, that the intention of this

research is not to prescribe the abandoning of employee surveys,

as they often provide the unique diagnostic ability for the

management to determine the optimal path for quality

improvement.

It is important to note that the results reported here are

specific to one particular industry. As a result, managers must

examine the various indicators outlined here within the

context of their own marketplace. We cannot generalize across

all industries, and specific turnover measures may be better

predictors of customer satisfaction in some markets than in

others. As a result, an industry-specific exploratory analysis of

Figure 2 Graphic presentation of log-total turnover index

Table VI Relative performance (R2) of employee satisfaction and turnover indicators in predicting customer satisfaction

Predictor(s) R2 F value Model significance

Single-item employee satisfaction rating (base model) 0.08 F1;240 ¼ 21:2 p , 0:01

Two factor model (process management1 employee management) 0.18 F2;239 ¼ 25:9 p , 0:01

Full employee satisfaction questionnaire 0.21 F9;232 ¼ 6:8 p , 0:01

Full-time turnover 0.02 F1;240 ¼ 6:1 p , 0:05

Part-time turnover 0.01 F1;240 ¼ 2:9 p , 0:1

Total turnover 0.04 F1;240 ¼ 8:7 p , 0:01

Turnover product 0.03 F1;240 ¼ 6:3 p , 0:05

Log of full-time turnover 0.05 F1;240 ¼ 11:6 p , 0:01

Log of part-time turnover 0.05 F1;240 ¼ 12:4 p , 0:01

Log of total turnover 0.07 F1;240 ¼ 18:4 p , 0:01

Log of turnover product 0.06 F1;240 ¼ 16:0 p , 0:01

Inverted full-time turnover 0.05 F1;240 ¼ 11:7 p , 0:01

Inverted part-time turnover 0.04 F1;240 ¼ 9:7 p , 0:01

Inverted total turnover 0.06 F1;240 ¼ 16:3 p , 0:01

Inverted turnover product 0.06 F1;240 ¼ 15:4 p , 0:01

Full-time2 0.02 F1;240 ¼ 5:3 p , 0:05

Part-time2 0.00 F1;240 ¼ 0:3 p ¼ 0:56

Full-time1/2 0.04 F1;240 ¼ 10:3 p , 0:01

Part-time1/2 0.03 F1;240 ¼ 7:5 p , 0:01

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

194

the predictive ability of turnover indicators may be required

prior to their use. Moreover, it is crucial that managerial

judgments of turnover rates take into account the proportion

of turnover attributed to be voluntary versus non-voluntary.

Voluntary turnover is expected to be more indicative of poor

working environments and more closely associated with poor

customer satisfaction levels. In such contexts, root-cause

analysis may be an invaluable resource for management in

identifying the underlying causes of employee and customer

discontent.

Future research in this area can further examine the

turnover construct and its drivers in several ways. For

example, using data collection methods other than employee

satisfaction surveys (e.g. focus groups, open-ended

interviews, etc.), the drivers of employee turnover can be

more candidly studied. Moreover, the effects of turnover

intentions on employee performance may provide new

insights on the topic. In addition, while many of the existing

studies of the service profit chain have utilized correlational

approaches with little reliance on time-series data, the study

reported here has utilized lagged measures of the key

constructs, reducing the possibility of reverse-causality

explanations for the results (e.g. customer dissatisfaction

resulting in higher employee turnover rates). Future research

in this area should capitalize more on careful timing of the

measurement of the constructs of interest, and take into

account the causal and temporal relationships in the service

profit chain. Moreover, an examination of the entire service

profit chain, including business performance measures (such

as store profitability and market share), and utilizing the

improved turnover indicators presented here, may also

provide for a better perspective on the dynamics of the

profit generation process in service organizations. It is hoped

that this paper will inspire additional research and motivate a

better understanding of the drivers and implications of

employee turnover in service markets.

References

Adizes, I. (1999), Managing Corporate Lifecycles, Prentice-

Hall, Englewood Cliffs, NJ.

Anderson, E.W. and Sullivan, M. (1993), “The antecedents

and consequences of customer satisfaction for firms”,

Marketing Science, Vol. 12, Spring, pp. 125-43.

Anderson, E.W., Fornell, C. and Lehmann, D.R. (1994),

“Customer satisfaction, market share and profitability:

findings from Sweden”, Journal of Marketing, Vol. 58, July,

pp. 53-66.

Anderson, E.W., Fornell, C. and Rust, R.T. (1997),

“Customer satisfaction, productivity and profitability:

differences between goods and services”, Marketing

Science, Vol. 16 No. 2, pp. 129-45.

Belsley, D., Kuh, E. and Walsh, R.E. (1980), Regression

Diagnostics, Wiley, New York, NY.

Bernhardt, K., Donthu, N. and Kennet, P. (2000),

“The relationship among customer satisfaction, employee

satisfaction, and profitability: a longitudinal analysis of

satisfaction and profitability”, Journal of Business Research,

Vol. 47, February, pp. 161-71.

Birnbaum, D. and Somers, M.J. (1993), “Fitting job

performance into turnover models: an examination of the

form of the job performance-turnover relationship and a

path model”, Journal of Management, Vol. 19, Spring,

pp. 1-11.

Bitner, M.J., Booms, B.H. and Tetreault, M.S. (1990),

“The service encounter: diagnosing favorable and

unfavorable incidents”, Journal of Marketing, Vol. 54,

January, pp. 71-84.

Churchill, G.A. Jr (1996), Basic Marketing Research, Dryden

Press, New York, NY.

Darden, W., Hampton, R. and Boatwright, E. (1987),

“Investigating retail employee turnover: an application of

survival analysis”, Journal of Retailing, Vol. 63, Spring,

pp. 69-88.

Day, G. (1994), “The capabilities of market-driven

organizations”, Journal of Marketing, Vol. 58, October,

pp. 37-52.

Ditz, G. (1971), “Status problems of the salesman”,

in Wotruba, T.R. and Olsen, R.M. (Eds), Readings in

Sales Management, Holt & Winston, New York, NY.

Duncan, O.D. (1975), Introduction to Structural Equation

Models, Academic Press, New York, NY.

Estelami, H. (2000), “Competitive and procedural

determinants of delight and disappointment in consumer

complaint outcomes”, Journal of Service Research, Vol. 2

No. 3, pp. 285-300.

Estelami, H. and DeMaeyer, P. (2002), “An exploratory

study of customer reactions to service provider over-

generosity”, Journal of Service Research, Vol. 4 No. 3,

pp. 205-17.

Fenton-O’Creevy, M. (1995), “Moderators of differences in

job satisfaction between full-time and part-time female

employees: a research note”, Human Resource Management

Journal, Vol. 5 No. 5, pp. 75-81.

Fornell, C. (1992), “A national customer satisfaction

barometer: the Swedish experience”, Journal of Marketing,

Vol. 56, January, pp. 6-21.

Futrell, C.M. and Parasuraman, A. (1984), “The relationship

of satisfaction and performance to salesforce turnover”,

Journal of Marketing, Vol. 48, Fall, pp. 33-40.

Table VII Determinants of log-total turnover index

Coefficient estimate Standard error Significance

Intercept 0.00 0.06 p . 0:1

Factor 1: process management 0.28 0.08 p , 0:01

Factor 2: people management 20.53 0.08 p , 0:01

Notes: R2 ¼ 0:16; F2;267 ¼ 25:4; p , 0:001

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

195

Good, L.K., Sisler, G.F. and Gentry, J. (1988), “Antecedents

of turnover intentions among retail management

personnel”, Journal of Retailing, Vol. 64, Fall, pp. 295-314.

Griffeth, R.W. and Hom, P.W. (2001), Retaining Valued

Employees, Sage Publications, Thousand Oaks, CA.

Hair, J.A. Jr, Anderson, R.E., Tatham, R.L. and Black, W.C.

(1998), Multivariate Data Analysis, Prentice-Hall, Upper

Saddle River, NJ.

Heskett, J.L., Sasser, W.E. Jr and Schlesinger, L.A. (1997),

The Service Profit Chain, The Free Press, New York, NY.

Heskett, J.L., Jones, T.O., Loveman, G.W., Sasser, W.E. Jr

and Schlesinger, L.A. (1994), “Putting the service-profit

chain to work”, Harvard Business Review, Vol. 72, March/

April, pp. 164-70.

Hunt, S.D. and Morgan, R.M. (1994), “Organizational

commitment: one of many commitments or key mediating

construct?”, Academy of Management Journal, Vol. 37 No. 6,

pp. 1568-87.

Hurley, R.F. (2002), “Putting people back into organizational

learning”, Journal of Business & Industrial Marketing, Vol. 17

No. 4, pp. 270-81.

Jackofsky, E. (1984), “Turnover and job performance:

an integrated process model”, Academy of Management

Review, Vol. 9, pp. 74-83.

Jackofsky, E.F., Salter, J. and Peters, L.H. (1986), “Reducing

turnover among part-time employees”, Personnel, Vol. 63

No. 5, pp. 41-3.

Keaveney, S.M. (1992), “An empirical investigation of

dysfunctional organizational turnover among chain and

non-chain retail store buyers”, Journal of Retailing, Vol. 68,

Summer, pp. 145-73.

Keaveney, S.M. (1995), “Customer switching behavior in

service industries: an exploratory study”, Journal of

Marketing, Vol. 59, April, pp. 71-82.

Kim, D.H. (1993), “The link between individual and

organizational learning”, Sloan Management Review,

Vol. 35, Fall, pp. 37-49.

Lehmann, D.R., Gupta, S. and Steckel, J.H. (1998),

Marketing Research, Addison-Wesley, Reading, MA.

Loveman, G.W. (1998), “Employee satisfaction, customer

loyalty, and financial performance: an empirical

examination of the service profit chain in retail banking”,

Journal of Service Research, Vol. 1 No. 1, pp. 18-31.

Lucas, G.H. Jr, Babakus, E. and Ingram, T. (1990),

“An empirical test of the job satisfaction-turnover

relationship: assessing the role of job performance for

retail managers”, Journal of the Academy of Marketing

Science, Vol. 18, Summer, pp. 199-208.

Lucas, G.H. Jr, Parasuraman, A., Davis, R. and Enis, B.M.

(1987), “An empirical study of sales force turnover”,

Journal of Marketing, Vol. 51, July, pp. 34-59.

Meyer, J.P. and Allen, N.J. (1991), “A three-component

conceptualization of organizational commitment”, Human

Resource Management Review, Vol. 1, Spring, pp. 61-89.

Ofir, C. and Khuri, A. (1986), “Multicollinearity in

marketing models: diagnostics and remedial measures”,

International Journal of Research in Marketing, Vol. 12 No. 3,

pp. 181-205.

Oliver, R.L. (1980), “A cognitive model of the antecedents

and consequences of satisfaction decisions”, Journal of

Marketing Research, Vol. 17, November, pp. 460-9.

Oliver, R.L. (1997), Satisfaction: A Behavioral Perspective on

the Consumer, McGraw-Hill, New York, NY.

O’Malley, M. (2000), Creating Commitment: How to Attract

and Retain Employees, Wiley, New York, NY.

Parasuraman, A., Zeithaml, V.A. and Berry, L.L. (1988),

“SERVQUAL: a multiple item scale for measuring

consumer perceptions of service quality”, Journal of

Retailing, Vol. 64, Spring, pp. 12-40.

Reichheld, F.F. (1996), The Loyalty Effect, Harvard Business

School Press, Boston, MA.

Rucci, A.J., Kirn, S.P. and Quinn, R.T. (1998),

“The employee-customer-profit chain at Sears”, Harvard

Business Review, Vol. 76, January/February, pp. 82-98.

Rust, R.T., Zahorik, A.J. and Keiningham, T.L. (1995),

“Return on quality (ROQ): making service quality

financially accountable”, Journal of Marketing, Vol. 59,

April, pp. 58-70.

Schlesinger, L.A. and Heskett, J.L. (1991), “Breaking the

cycle of failure in services”, Sloan Management Review,

Vol. 32, Spring, pp. 17-28.

Schlesinger, L.A. and Zornitsky, J. (1991), “Job satisfaction,

service capability, and customer satisfaction: an examination

of linkages and management implications”, Human Resource

Planning, Vol. 14 No. 2, pp. 141-9.

Schneider, B. and Bowen, D. (1993), “The service

organization: human resources management is crucial”,

Organizational Dynamics, Vol. 21, Spring, pp. 39-55.

Sightler, K.W. and Adams, J.S. (1999), “Differences between

stayers and leavers among part-time workers”, Journal of

Managerial Issues, Vol. 11 No. 1, pp. 110-25.

Steers, R.M. and Modway, R.T. (1981), “Employee turnover

and the post decision accommodation process”,

in Staw, B.M. and Cummings, L.L. (Eds), Research in

Organizational Behavior, JAI Press, Greenwich, CT,

pp. 235-81.

Wexley, K.N. and Yuki, G.A. (1984), Organizational Behavior

and Personnel Psychology, Irwin, Homewood, IL.

Williams, C.R. and Livingstone, L.P. (1994), “Another look

at the relationship between performance and voluntary

turnover”, Academy of Management Journal, Vol. 37, April,

pp. 269-98.

Zeithaml, V.A., Berry, L.L. and Parasuraman, A. (1996),

“The behavioral consequences of service quality”, Journal

of Marketing, Vol. 60, April, pp. 31-46.

Further reading

Dabholkar, P.A., Thorpe, D.I. and Rentz, J.O. (1996),

“A measure of service quality for retail stores: scale

development and validation”, Journal of the Academy of

Marketing Science, Vol. 24, Winter, pp. 3-16.

James, L.R. (1982), “Aggregation bias in estimates of

perceptual agreement”, Journal of Applied Psychology,

Vol. 219 -29.

Rust, R.T., Inman, J.J., Jia, J. and Zahorik, A. (1999), “What

you don’t know about customer-perceived quality: the role

of customer expectation distributions”, Marketing Science,

Vol. 18 No. 1, pp. 77-92.

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

196

Appendix 1

Table

AI

Mea

sure

san

dde

scrip

tive

stat

istic

s

Scale

Description

Scalealpha

No.ofitem

sMea

n

Overallsatisfaction

Cus

tom

ers’

over

all

satis

fact

ion

with

the

stor

e,fr

iend

lines

sof

serv

ice,

and

spee

dof

serv

ice

0.80

34.

31

Full-tim

eem

ployeeturnover

Turn

over

offu

ll-tim

eem

ploy

ees

over

the

12-m

onth

perio

dN

/A1

97.0

1

Part-tim

eem

ployeeturnover

Turn

over

ofpa

rt-t

ime

empl

oyee

sov

erth

e12

-mon

thpe

riod

N/A

115

3.11

Overallem

ployeesatisfaction

Empl

oyee

perc

eptio

nsab

out

thei

rsa

tisfa

ctio

nw

ithth

eir

job,

pay

and

the

com

pany

0.80

113.

73

Stan

dardsan

dgoals

Empl

oyee

perc

eptio

nsab

out

how

high

qual

ityst

anda

rds

are

inth

est

ore

0.87

124.

00

Team

work

Empl

oyee

perc

eptio

nsab

out

the

degr

eeof

coop

erat

ion

and

supp

ort

amon

gst

ore

empl

oyee

s0.

8811

3.72

Toolsan

dmethods

Empl

oyee

perc

eptio

nsab

out

the

adeq

uacy

ofm

ater

ials

,to

ols

and

met

hods

used

tose

rve

cust

omer

s0.

8515

3.93

Training

Empl

oyee

perc

eptio

nsab

out

the

qual

ityan

dam

ount

oftr

aini

ng0.

795

3.45

Store

mee

ting

Empl

oyee

perc

eptio

nsab

out

the

qual

ityof

stor

em

eetin

gs0.

742

3.58

Store

man

ager

communicates

Empl

oyee

perc

eptio

nsab

out

how

wel

lth

em

anag

erco

mm

unic

ates

info

rmat

ion

0.79

53.

97

Store

man

ager

empowers

Empl

oyee

perc

eptio

nsab

out

the

degr

eeto

whi

chth

em

anag

eral

low

sem

ploy

ees

tom

ake

deci

sion

sw

hen

itis

poss

ible

todo

so0.

865

4.00

Store

man

ager

man

ages

perform

ance

Empl

oyee

perc

eptio

nsab

out

whe

ther

the

man

ager

hold

spe

ople

acco

unta

ble

and

prov

ides

corr

ectiv

efe

edba

ck

whe

nne

eded

0.79

34.

03

Conflictresolution

Empl

oyee

perc

eptio

nsab

out

whe

ther

confl

icts

can

bere

solv

edpr

oduc

tivel

yin

the

stor

e0.

702

3.93

Laborsched

uling

Empl

oyee

perc

eptio

nsab

out

whe

ther

labo

ris

sche

dule

dpr

oper

lyto

allo

wfo

rtr

aini

ng,

time

off

and

serv

ing

cust

omer

s0.

752

3.28

Note:

n¼

275

stor

es

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

197

Appendix 2

Table

AII

Cor

rela

tions

amon

gpr

edic

tor

varia

bles

Employee

satisfaction

Stan

dardsan

d

goals

Toolsan

d

methods

Training

Labor

sched

ulingCommunications

Empowermen

t

Conflict

resolution

Store

mee

tings

Perform

ance

man

agem

ent

Logoftotal

turnover

Employee

satisfaction

1.00

0.71

0.68

0.73

0.66

0.67

0.69

0.48

0.51

0.59

20.

16

Stan

dardsan

d

goals

1.00

0.81

0.72

0.51

0.74

0.74

0.41

0.45

0.67

20.

12

Toolsan

dmethods

1.00

0.78

0.59

0.62

0.66

0.55

0.44

0.57

20.

20

Training

1.00

0.62

0.68

0.67

0.40

0.47

0.62

20.

16

Laborsched

uling

1.00

0.52

0.55

0.49

0.56

0.49

20.

36

Man

ager

communications

1.00

0.91

0.30

0.44

0.83

20.

05

Empowermen

t1.

000.

370.

430.

782

0.10

Conflictresolution

1.00

0.36

0.27

20.

32

Store

mee

tings

1.00

0.45

20.

20

Perform

ance

man

agem

ent

1.00

0.03

Logoftotal

turnover

1.00

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

198

Corresponding author

Robert F. Hurley can be contacted at: bob@drbobhurley.com

Executive summary and implications formanagers and executives

This summary has been provided to allow managers and executivesa rapid appreciation of the content of the article. Those with aparticular interest in the topic covered may then read the articlein toto to take advantage of the more comprehensive description ofthe research undertaken and its results to get the full benefit of thematerial present.

The general acceptance that there is a significant relationshipbetween the job satisfaction of employees and the satisfactionof customers – a relationship which then extends to theorganization’s profitability – is one which many managementsseem comfortable with. It’s got an intuitive appeal and seemslike a common sense way of improving the bottom line.

One factor which has been somewhat ignored in studyingthis phenomenon has been the role played by employeeturnover and what effect it has on customer satisfaction andimprovements to long-term profitability. Higher levels ofemployee turnover, which can lead to lower customersatisfaction, may not only be indicative of a poor workenvironment, but may also be reflected in the loss ofexperienced employees and established customerrelationships, resulting in negative effects on the customer.

Robert F. Hurley and Hooman Estelami seek to validatethis view in the context of retail stores and to betterunderstand the capabilities of employee turnover measures aspredictors of customer satisfaction. Their results suggest thatcertain employee turnover indicators can perform aseffectively as single-item employee satisfaction ratings do inpredicting customer satisfaction.

One appeal of using employee turnover as a measure is that,unlike employee and customer satisfaction surveys, which canbe time-consuming and costly, employee turnover statisticsare something that almost all organizations collect as a matterof course. The accessibility of this measure might offer a cost-effective way of helping companies get a clearer picture of thedynamics of the service profit chain.

Apart from the cost in money and time, employeesatisfaction measures that are obtained through self-administered surveys conducted at the workplace may havelimited value. It is not difficult to imagine employees who, dueto factors such as fear of job loss, lack of interest, or thepresence of co-workers or supervisors, fail to respondobjectively to a questionnaire. And, by definition, suchsurveys only include current employees. The beliefs andattitudes of former workers about the work environmentmight be just as valuable.

Other studies have also found that, in estimating the returnon investments made in quality improvement programs, it ispossible to “spend too much on quality” and that not allexpenditure on quality is equally valid. While some studieshave documented a positive correlation between customer

satisfaction and profitability, others have found the

relationship to be very weak or insignificant.

It should also be borne in mind that the link between

customer satisfaction and employee satisfaction might, in

some instances, be in reverse. Unhappy customers creating a

poor work environment, leading to employee dissatisfaction,

and to increased turnover.

With regard to the link between employee turnover and firm

performance, Hurley and Estelami say: “When people depart,

their intelligence regarding processes, methods and customers

also leaves. In a service organization, a logical hypothesis is

that higher levels of employee turnover will be associated with

lower levels of customer satisfaction. In retail, audit, tax, legal

or other service settings, when a knowledgeable service

provider leaves and is replaced by someone who is learning

methods and customer interactions for the first time, some

customer dissatisfaction is expected to result. Employee

turnover may therefore provide a basis for gauging not only

employee satisfaction, but also customer satisfaction levels.”

When using employee turnover as a measure in this way,

managers should be aware that a significant variation can be

expected between the implications of turnover among part-

time and full-time employees. Full-time employees may imply

the loss of years of training and development of the employee,

and a high cost of replacement. In contrast to part-time

employees, replacing a full-time employee might need

significant investments and a long break-in period for the

incoming employee may be associated with poor service

delivery.

In many retail service encounters, however, a high level of

technical ability may not be needed. In such contexts, full-

time employee turnover may not necessarily imply a loss of

valued skills, and part-time turnover might actually be a more

accurate representation of the quality of the work

environment. This is because part-time employees typically

experience lower levels of organizational commitment,

whereas full-time employees might be bound to the

company as their only source of income.

It is also crucial that managerial judgments of turnover rates

take into account the proportion of turnover attributed to be

voluntary versus non-voluntary. Voluntary turnover is

expected to be more indicative of poor working

environments and more closely associated with poor

customer satisfaction levels. In such contexts, root-cause

analysis may be an invaluable resource for management in

identifying the underlying causes of employee and customer

discontent.

In questioning the merits of costly employee surveys,

Hurley and Estelami say their intention is not to prescribe

abandoning them as they often provide the unique diagnostic

ability for the management to determine the optimal path for

quality improvement.

(A precis of the article “An exploratory study of employee turnover

indicators as predictors of customer satisfaction”. Supplied by

Marketing Consultants for Emerald.)

Employee turnover indicators

Robert F. Hurley and Hooman Estelami

Journal of Services Marketing

Volume 21 · Number 3 · 2007 · 186–199

199

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

Or visit our web site for further details: www.emeraldinsight.com/reprints