-
To improve service, companies mtist use multiple research
approaches am.ongdifferent customer groups to ensure that they are
hearing what customers are
saying and responding to their suggestions.
Listening to the Customer The Concept of a Service-Quality
Information SystemLeonard L. Berry A. Parasuraman
The quality of listening has an impact on thequality of service.
Firms intent on improvingservice need to listen continuously to
threetypes of customers: external customers who have ex-perienced
the firm's service; competitors' customerswho the firm wotild like
to make its own; and internalcustomers (employees) who depend on
internal ser-vices to provide their own services. Without the
voic-es of these groups guiding investment in service im-provement,
all companies can hope for are marginalgains.
In this paper, we discuss the concept of a service-quality
information system. We argue that compa-nies need to establish
ongoing listening systems usingmtiltiple methods among different
customer groups.A single service-quality study is a snapshot taken
at apoint in time and from a partictilar angle. Deeper in-sight and
more informed decision making come froma continuing series of
snapshots taken from variousangles and through different lenses,
which form theessence of systematic listening.
Systematic ListeningA service-quality information system uses
mtiltiple re-search approaches to systematically capture,
organize,and disseminate service-quality information to sup-port
decision making. Continuously generated data
flow into databases that decision makers can use onboth a
regtilarly scheduled and as-needed basis.
The use of multiple research approaches is neces-sary because
each approach has limitations as well asstrengths. Combining
approaches enables a firm totap the strengths of each and
compensate for weak-nesses. Continuotis data collection and
disseminationinforms and educates decision makers about the
pat-terns of change for example, customers' shifiiingservice
priorities and declining or improving perfor-mance in the company's
or the competitors' service.
An effective service-quality information system of-fers a
company's executives a larger view of service qual-ity along with a
composite of many smaller pictures. Itteaches decision makers which
service attributes are im-portant to ctistomers and prospects, what
parts of thefirm's service system are working well or breakingdown,
and which service investments are paying off! Aservice-quality
information system helps to focus ser-vice improvement planning and
resotirce allocation. Itcan help sustain managers' motivation for
service im-provement by comparing the service performance ofvarious
tinits in the organization and linking compen-sation to these
results. And it can be the basis for an ef-
Leonard L. Berry is a professor of marketing, the J.C. Penney
Chair ofRetailing Studies, and director. Center for Retailing
Studies, TexasA&M University. A. Parasuraman holds the James W.
McLamoreChair in Marketing, University of Miami.
SLOAN MANAGEMENT REVIEW/SPRING 1997 BERRY & PARASURAMAN
65
-
Figure 1 Principal Benefits of an Effective Service-Quality
Information System
Encourages and enables management/to incorporate the voice of
the customer
/ into decision making.
/ / Reveals customers' service priorities.
Service-QualityInformationSystem
I Identifies service improvement priorities'ly/ and guides
resource-allocation decisions.
_ Allows the tracking of company andcompetitor service
performance over time
, Discloses the impact of service qualityinitiatives and
investments.
Offers performance-based data to* reward excellent service and
correctpoor service.
Source: L. Berry, On Great Service: A Framework for Action {Nevi
York:FreePress. 1995|p. 34.
fective first-line employee reward system by identify-ing the
most effective service providers. (See Figure 1for the principal
benefits of a service-quality infor-mation system.)
The task of improving service in organizations iscomplex. It
involves knowing what to do on multiplefronts, such as technology,
service systems, employeeselection, training and education, and
reward systems.It involves knowing when to take these initiatives.
Itinvolves knowing how to implement these actionsand how to
transform activity into sustainable im-provement. Genuine service
improvement requires anintegrated strategy based on systematic
listening. Un-related, incomplete studies, outdated research,
andfindings about customers that are not shared
provideinstifficient support for improving service.
Approaches to Service ResearchA company can choose from many
possible researchapproaches to build a service-quality information
sys-tem (see Table 1). A firm would not use all approachesin the
table in the same system; too much informa-tion obscures the most
meaningflil insights and mayintimidate intended tisers. Conversely,
incomplete in-formation injects needless guessing into decision
mak-ing or, worse, paints a false picture. The nature of the
service, the firm's service strategy, and the needs of
theinformation users determine which service-quality re-search
approaches to use.
An industrial equipment manufacturer might wishto use service
reviews to benefit from unfiltered dia-logue with mtiltiple users,
reach consensus on servicesupport priorities, and solidify
relationships. A restau-rant, with a transaction-oriented business,
wotild findservice reviews far less efficient than other
approach-es. Because of the relationship nature of its business,
alimotisine service should consider new, declining,
andlost-customer surveys. It should identify any nega-tives that
tarnish new ctistomers' first impressions, orcause other customers
to be less loyal or to defect, soit can take corrective measures. A
taxi company prob-ably wouldn't use these surveys because of a
minimalrelationship-marketing potential. A firm whose strat-egy
emphasizes service reliability surely wotild wantto capture and
analyze customer service complaintsto identify where the service
system is breaking down.A company whose strategy depends on
point-of-saleservice excellence should consider mystery
shoppingresearch, which generates feedback on specific
serviceproviders.
Four research approaches summarized in the tableapply to
virtually all organizations and can be consid-ered essential
components of a service-quality informa-tion system: transactional
surveys; customer complaint,comment, and inquiry capttire; total
market stirveys;and employee surveys. These approaches enstire
cover-age of the three customer types (external
customers,competitors' customers, internal customers),
doctimentfailure-prone parts of the service system, and provideboth
transaction-specific and overall service feedback.
Personal Involvement in ListeningA service-quality information
system does not replacethe need for managers to interact directly
with cus-tomers. Becoming well informed about service
qualityrequires more than reading or hearing the results
ofstructured, quantitative studies. It also requires thatdecision
makers become personally involved in listen-ing to the voices of
their customers, which can includeparticipating in or observing
qualitative research, suchas service reviews and foctis groups. And
it can includeless formal interactions with customers, such as
whenairline executives query passengers on fiights and re-
66 BERRY & PARASURAMAN SLOAN MANAGEMENT REVIEW/SPRING
1997
-
Table 1 Research Approaches for Building Service-Quality
Information Systems
Type Description
Transactional surveys* Service satisfactionsurvey of
customersfollowing a serviceencounter.
Mystery shopping Researchers become"customers" to expe-rience
and evaluatethe quality of servicedelivered.
New, declining, and Surveys to determine whylost-customer
surveys customers select the firm.
reduce their buying, orleave the firm.
Focus group interviews Directed questioning ofa small group,
usuallyeight to twelve people.Questions focus on aspecific topic.
Can beused with customer.noncustomer, oremployee groups.
Customer advisory A group of customers re-panels cruited to
periodically pro-
vide the firm with feedbackand advice on service per-formance
and other issues.Data are obtained in meet-ings, over the
telephone.through mail questionnaires.or via other means. Employ-ee
panels also can beformed.
Service reviews Periodic visits with cus-tomers (or a class
ofcustomers) to discussand assess the servicerelationship. Should
bea formal process with acommon set of questions.capture of
responses in adatabase, and follow-upcommunication
withcustomers.
Purpose
Obtain customer feed-back while service expe-rience is still
fresh; actonfeedback quickly if negativepatterns develop.
Measure individual employ-ee service behaviors for usein
coaching, training, perfor-mance evaluation, recognition.and
rewards; identify systemicstrengths and weaknesses
incustomer-contact service.
Assess the role servicequality and other issuesplay in customer
patron-age and loyalty.
Provide a forum forparticipants to suggestservice
improvementideas; offer fast.informal feedback onservice
issues.
Obtain in-depth, timelyfeedback and suggestionsabout service
qualityfrom experienced customerswho cooperate becauseof
"membership" natureof the panel.
Identify customer expec-tations and perceptionsof the company's
serviceperformance and improve-ment priorities in a face-to-face
conversation.A view of the future, notjust a study of the
past.Opportunity to includemultiple decision makersand decision
influencersin the discussions.
Highlighted approaches normally would be part of any
service-quality information system.'Frequencies of use vary among
companies.
Frequency'''
Continuous
Quarterly
Continuous
As needed
Quarterly
Annually orsemiannually
Limitations
Focuses on customers'most recent experiencerather than their
overallassessment. Noncustomersare excluded.
Subjective evaluations;researchers may be more"judgmental" than
customerswould be; expense limitsrepetitions; potential tohurt
employee morale ifimproperly used.
Firm must be able to identifyand monitor service usageon a
per-customer basis.
Dynamics of group interviewmay prevent certain issuesfrom
surfacing. Focusgroups are, in effect.brainstorming sessions;the
information generatedis not projectable to thepopulation of
interest. Focusgroup research is mostvaluable when coupled
withprojectable research.
May not be projectable toentire customer base. Ex-cludes
noncustomers. Panel-ists may assume role of"expert" and become
lessrepresentative of customerbase.
Time consuming and expen-sive. Most appropriate forfirms
marketing complexservices on an ongoing,relationship basis.
(continued)
SLOAN MANAGEMENT REVIEW/SPRING 1997 BERRY & PARASURAMAN
67
-
Tahle 1 Research Approaches for Building Service-Quality
Information Systems (continued)
Type
Customer complaintcomment, and inquirycapture*
Total market surveys
Employee field reporting
Employee surveys
Service operatingdata capture
Description
System to retain.categorize, track, anddistribute
customercomplaints and othercommunications withthe company.
Surveys that measure cus-tomers' overall assessmentof a
company's service.Research includes bothexternal customers
andcompetitors' customers.i.e., the total market.
Formal process forgathering, categorizing.and distributing
fieldemployee intelligenceabout service issues.
Surveys concerning theservice employees provideand receive, and
the qualityof their work lives.
A system to retain, cate-gorize, track, and distributekey
service-performanceoperating data, such as ser-vice response times,
servicefailure rates, and service de-livery costs.
Purpose
Identify most common typesof service failure for correc-tive
action. Identify throughcustomer communicationsopportunities to
improve ser-vice or otherwise strengthencustomer relationships.
Assess company's serviceperformance comparedto competitors;
identifyservice-improvementpriorities; track serviceimprovement
over time.
Capture and share at themanagement level intelli-gence about
customers'service expectations andperceptions gathered inthe
field.
Measure internal servicequality; identify employee-perceived
obstacles toimproved service; trackemployee morale andattitudes.
Employee sur-veys help answer "why"service performance iswhat it
is.
Monitor service performanceindicators and take correc-tive
action to improve perfor-mance as necessary. Relateoperating
performance datato customer and employeefeedback.
'Highlighted approaches normally would be part of any
service-quality information system.^Frequencies of use vary among
companies.
Frequency^
Continuous
Semiannuallyor quarterly
Continuous tomonthly
Quarterly
Continuous
Limitations
Dissatisfied customersfrequently do not complaindirectly to the
company.Analysis of customercomplaints and commentsoffers only a
partial pictureof the state of service.
Measures customers' over-all service assessmentsbut does not
capture assess-ments of specific serviceencounters.
Some employees will be moreconscientious and efficientreporters
than others. Employ-ees may be unwilling to pro-vide negative
information tomanagement.
The strength of employee sur-veys is also a weakness; em-ployees
view service deliveryfrom their own vantage point.subject to their
own biases.Employees can offer valuableinsights into the root
causesof service problems but arenot always objective or correctin
their interpretations.
Operating performance datamay not be relevant tocustomers'
perceptions ofservice. Focus is on what isoccurring but not
why.
tailers accompany customers through their stores toask them what
they see, like, and dislike.
In 1993, the cash management division of First Na-tional Bank of
Chicago changed its customer satisfac-tion stirveys fi'om mail
questionnaires to telephone in-terviews. The change was prompted by
poor responserates to the mail survey and customers' su^estions
forimproving survey effectiveness: conduct the surveys byphone
because they are more efficient and have bankemployees who can act
on problems make the calls.
First Chicago recruited senior and middle man-
agers to conduct three prescheduled twenty-minutephone
interviews per month and write reports on eachcall for the
database. Managers were trained to do theinterviews and passed a
certification test before survey-ing their first customer. They
surveyed each employeeof the client firm who had significant
contact with thebank. Bank managers were responsible for
"actionitems" that surfaced in the interviews. The bank's
vicepresident of quality assurance, Aleta Holub, remarked,"We've
really seen a cultural change from getting ev-eryone a litde closer
to the customer."'
68 BERRY & PARASURAMAN SLOAN MANAGEMENT REVIEW/SPRING
1997
-
Direcdy hearing the voices of customers, noncus-tomers, and
employees adds richness, meaning, andperspective to the
interpretation of quantitative data.The First Chicago case
illtistrates the potential impactembedded in literally hearing the
customer's voice,rather than hearing only a distilled or numeric
repre-sentation of it. McQuarrie makes the point: "Every-one
believes his or her own eyes and ears first. Keyplayers hear about
problems and needs direcdy fromthe most credible source the
ctistomer. Learning isenhanced becatise of the vivid and compelling
qualityof first-hand knowledge."^
A well-designed and -implemented service-qualityinformadon
system raises the probability that a compa-ny will invest service
improvement money in ways thatactually improve service. It also
continually under-scores the need to improve service. Continually
captur-ing and disseminating data reveal not only progress,but
problems; not only strengths, but weaknesses.Quality service is a
never-ending journey. An effecdveservice-quality information system
reminds everyonethat more work needs to be done.
Developing an Effective Service-QualityInformation SystemThe
primary test of a service-quality informadon sys-tem is the extent
to which it informs and guides ser-vice-improvement decision
making. Another impor-tant test is the extent to which the system
modvatesboth managerial and nonmanagerial employees to im-prove
service. There are five gtiidelines for developinga system that can
meet these tests:1. Measure service expectations.2. Emphasize
information quality.3. Capture customers' words.4. Link service
performance to business restilts.5. Reach every employee.
The core success factors embedded are the coverageof external,
compedtors,' and internal customers; theuse of multiple measures;
and ongoing measurement.
Measure Service ExpectationsMeasuring service performance per se
is not as mean-ingflil as meastiring performance relative to
customers'expectations. Customers' service expectadons providea
frame of reference for their assessment of the service.
Figure 2 Service Quality Ratings for a ComputerManufacturer
Note: The dots indicate perceived service. The vertical boxes
indicate customers'zones of toierance bounded on the top by their
desired service expectations andthe bottom by their adequate
service expectations.
Assume, for example, that a company measures onlycustomers'
percepdons of service performance using a9-point scale. It receives
an average perception scoreof 7.3 on the service attribute
"Performs the serviceright the first time." How should managers
interpretthis score? Is it a good score? Without knowing
whatcustomers expect, this is a difficult question. There isno
basis for gauging the rating. Managers' interpreta-tion of the 7.3
perception score wotild likely be fardifferent if customers'
average expectation rating forthis attribute were 8.2 rather than
7.0. As researchersGoodman et al. ask: "How satisfied is a
satisfied cus-tomer? When is good, good enough?
Unfortunately,companies that ask their customers how sadsfied
theyare but fail to research ctistomers' expectations cannotanswer
these questions."^
We collected service quality data from a computermanufacturer's
customers (see Figure 2). We mea-sured two levels of expectations:
desired service (whatthe customer believes the service should be
and canbe) and adequate service, (the minimal level of
serviceacceptable to the customer). The top of the tolerancezone
represents customers' average desired service-expectation score,
the bottom, their average adequate
SLOAN MANAGEMENT REVIEW/SPRING 1997 BERRY & PARASURAMAN
69
-
Figure 3 Service Quality Ratings for a Retail Chain
g
Note: The dots indicate perceived service. The vertical boxes
indicate customers'zones of toierance bounded on the top by their
desired service expectations andthe bottom by their adequate
service expectations.
service-expectation score. Service performance is su-perior if
perception scores exceed the zone of toler-ance, acceptable if
perceptions are within the zone,and unacceptable if perceptions are
below the zone.
Comparing the perceptions-only data with thecombined
perceptions-expectations data demonstratesthe diagnostic value of
measuring customers' expecta-tions. Were the computer manufacturer
to measureonly customer perceptions, its management wouldhave litde
guidance for investing service improvementresources. The perception
scores are similar across theservice dimensions. However, the
inclusion of expecta-tions data clearly shows that improving
service reliabil-ity should take priority over improving tangibles.
Al-though reliability and tangibles have identical percep-tion
scores, customers' expectations for reliable serviceare much
higher. Whereas customers' perceptions bare-ly exceed
adequate-level expectations for reliability,they exceed
desired-level expectations for tangibles.
We also contrasted perceptions-only and
percep-tions-expectations data for a retail chain (see Figure
3).Without expectations data, management may concludethat the
firm's service quality is acceptable because allperception scores
are more than a full point above the
scale's midpoint of 5. However, the addition of expec-tations
scores suggests a much different conclusion,with service
performance on four of the five dimen-sions not even meeting
customers' minimum expec-tations.''
Documenting the value of measuring customer ex-pectations in
service quality research is necessary be-cause perceptions-only
research is common. Mea-suring expectations adds complexity and
possiblylength to the survey process and can be more expen-sive.
Moreover, accurately measuring expectations isnot easy. The best
way to do it .and whether it is evennecessary are the subject of
debate.' Advocates of per-ceptions-only measurement typically point
out thatservice perception scores explain more variance in
anoverall service quality measure than a combined
ex-pectations-perceptions measure. Perceptions ratingsconsistently
explain more variance, most likely be-cause pieces of the whole
(perceptions of specific ser-vice attributes) are being regressed
against the whole(an overall service perception measure). So why is
itso critical to measure customer expectations of ser-vice?
Because, as Figures 2 and 3 show, managerslearn more about
improving service when customerexpectations provide a frame of
reference for inter-preting perception ratings.
Emphasize Information QualityQuality of information not quantity
is the ob-jective in building a service-quality information
sys-tem. The test of information quality is to ask if
theinformation is: Relevant? Precise? Useful? In context? Credible?
Understandable? Timely?
Relevant service-quaiivy information focuses decisionmakers'
attention on the most important issues to meetand exceed external
customer expectations, convertprospects, and enable employees to
improve service.The more a service-quality information system
focuseson the service priorities of the three customer types,
themore likely managers will invest in the most appropri-ate
initiatives that can make a positive difference.
70 BERRY & PARASURAMAN SLOAN MANAGEMENT REVIEW/SPRING
1997
-
Measuring the importance of service attributes is notthe same as
measuring customers' service expectations,although they are closely
related. Customers' expecta-tions are the comparison standards they
use to judgethe performance of various service attributes.
How-ever, the service attributes are not uniformly impor-
nformation precision and usefulnessgo hand in hand.
Informationthat is overly broad or general
is not useful.
tant to customers, and it is necessary to specificallymeasure
their relative importance to monitor compa-ny and competitor
performance on those attributesthat drive customers' overall
perceptions of servicequality.
Information precision and usefulness go hand inhand. Information
that is overly broad or general is notuseflil. Researcher Brian
Lunde commented: "One ofthe worst criticisms that could be made by
a line man-ager about a company's . . . information is that it is
'in-teresting.' 'Interesting' is code for 'useless.' The
infor-mation simply must be specific enough that executives. . .
can take action make decisions, set priorities,launch programs,
cancel projects."'^
Information on what must be done to improveservice is useful.
Chase Manhattan Bank has deter-mined empirically that the approval
process is the pri-mary driver of customers' quality perceptions
for itsmortgage loan service. Accordingly, Chase's
serviceinformation system tracks its performance on themortgage
approval process compared to its principalcompetitors. However,
Chase does not stop with over-all perceptions of the mortgage
approval process. Italso investigates "sub-drivers" such as quick
approval,communication, the appraisal process, amount of
pa-perwork, and unchanging loan amount. The infor-mation is
sufFiciendy precise so manners know whatto do and can assign
implementation accountabili-ties. They review data patterns
regtilarly at man^e-ment meetings.^
An effective service-quality information system pre-sents
information dynamically. At any point in time.
the system's output tells what is becoming more or lessimportant
the context. Fresh data are more valuablewhen presented in the
context of past data. The studyof trend data reveals patterns,
nuances, and insightsthat one-time data cannot possibly reveal. Is
the in-vestment in new telephone technology paying off?Was it a
good idea to redesign the account-openingprocedures? Is the
company's new investment in train-ing reducing error rates? Has
competitor advertisingabout service influenced customer
expectations? Hasthe competitor's new store prototype given its
serviceratings a boost? Only trend data can answer these andmyriad
other questions. Ongoing research using com-mon measures across
study periods generates trenddata that provide context and aid
interpretation.
A service-quality information system will not mo-tivate
managerial and nonmanagerial employees un-less the information is
credible. Employees in low-rated units may be embarrassed and
financially hurtby the system's output and may question the
infor-mation's validity. Companies can improve informa-tion
credibility by seeking input from operating unitson the design of
research approaches and the devel-opment of specific questions.
Information sessions toexplain research approaches to employees,
with anopportunity for questions and answers, also can beuseful.
Clear explanations of the research method andsample size should
accompany the dissemination ofresults. Multiple measures a
fundamental tenet ofservice-quality information systems enhance
in-formation credibility when different measures pointto similar
conclusions. The use of an outside researchfirm for data collection
can help convey impartiality.
Information quality also is determined by whetherthe information
is understandable to intended users.Relevance, usefiilness, and
credibility all are enhancedwith easily imderstood research
information. Unfami-liar statistical jargon and symbols confuse,
intimidate,and discourage users, leading to feigned use of the
sys-tem and incorrect interpretations of its output. Thereshould be
a concerted effort to design a user-friendlysystem with uniform
reports and clear presentation ofdata.
The timeliness oi^mioTma.non influences its quality.All the
other attributes of information quality are ren-dered impotent if
information is not available whendecision makers need it. Companies
should collect
SLOAN MANAGEMENT REVIEW/SPRING 1997 BERRY & PARASURAMAN
71
-
data to support their natural decision making andplanning
cycles. Monthly transacdonal survey reportsshotild be ready for the
monthly man^ement meet-ing, total market survey results shotild
feed the annualplanning and budgeting process, customer
complaintanalyses shotild be ready for the twice-a-month meet-ings
of the service-improvement leadership team. Thedesign of databases
shotild accommodate trend-dataretrieval for managers as needed.
Companies shotildcontinually explore ways to accelerate data
collectionand dissemination. Firms might fax or e-mail
ques-tionnaires to respondents rather than use the postalservice.
Research results might be distributed internal-ly on a company's
intranet.
The information quality tests of relevance,
precision,usefijlness, context, credibility, understandability,
andtimeliness are not absolutes. Improving informationquality is a
journey of trial and error, experience curveefiects, user feedback,
and new knowledge. Btiildingan effective system is a never-ending
process of refine-ment. Larry Brandt, associate director of
ctistomer ser-vice at AMP, a mantifacttirer of electrical and
electron-ic connectors, points out the necessity of
continuousimprovement: "We need to constandy evaluate whatit is
we're measuring, why we're doing it, and whetherthe results are
worthwhile in the organization's big pic-ture, or we run the risk
of wasting time and effort."'^
Capture Customers' WordsThe best service-quality information
systems are builtwith qualitative and quantitative databases,
ratherthan stricdy the latter. Quantified data are
summaries;averages of customers' perceptions of a very specific
ser-vice issue are still averages. Quantitative data bringmany
benefits to the service informadon table, includ-ing easy analysis,
comparability from one period tothe next, and potential
projectability. What numbersdon't offer are the tone, infiection,
feeling, and "wordpictures" from customers' voices. A service
quality re-port showing that 4 percent of the customer base isvery
dissatisfied and another 13 percent is somewhatdissatisfied with
the company's service may not getmanagement's attention. However,
if the report in-cludes customers' verbatim comments, it may
receivea very different reaction.
GTE Supply and Lexus customers illustrate the im-portance of
capturing customers' words. GTE Supply
purchases numerous products needed for the tele-phone operations
of its customers, the local telephonecompanies. By implementing a
systematic survey ofcustomers' needs and opinions, GTE has
improvedservice quality. The survey generates both quandtadveand
qualitative data for each customer. Current nu-merical quality
ratings are compared to previous re-sults to spot problems. In
addition, the survey askstwo open-ended quesdons: "Why do you say
that?"(in response to a closed-ended overall quality ques-don) and
"What improvements, if any, cotild be made
AIthough Lexus was manufacturingcars with few
mechanicalproblems, the extra careshown in the sales and
service
process strongly influenced buyersatisfaction.
by Supply?" The company enters the customers' ownwords into a
database and presents them to its man-agers along with the
numerical data. GTE researchersJames Drew and Tye Fussell remarked:
"Tabtilationsof survey questions can highlight specific
transactioncharacteristics in need of improvement from the
cus-tomer's viewpoint. In contrast, open-ended commentsare
especially effective in motivating first-level man-agers and giving
the tabtiladons substance and a hu-man touch."'
Toyota introduced the Lexus line of luxury cars inthe late
1980s, and by the early 1990s, the cars hadvaulted to the top of
the J.D. Power & Associates rat-ings in customer satisfaction.
Soon afber, another luxu-ry carmaker retained Custom Research Inc.
(CRI), amarketing research firm, to find out why Lexus own-ers were
so satisfied. CRI conducted a series of focusgroups to hear the
Lexus story in the owners' words.Most of the Lexus drivers e^erly
volunteered storiesabout the special care and attention they had
receivedfrom their Lexus dealer. It became clear that althoughLexus
was manufacturing cars with few mechanicalproblems, the extra care
shown in the sales and serviceprocess strongly influenced buyer
satisfaction. Owners
72 BERRY & PARASURAMAN SLOAN MANAGEMENT REVIEW/SPRING
1997
-
felt pampered and respected as valued Lexus cus-tomers. For
example, one female owner mentionedseveral times during the focus
group that she hadnever had a problem with her Lexus. However,
onfiarther probing, she said, "Well, I suppose you cotildcall the
four times they had to replace the windshielda 'problem.' But
frankly, they took care of it so welland always gave me a loaner
car, so I never really con-sidered it a 'problem' until you
mentioned it now."CRI's research showed that the Lexus policy of
alwaysoffering service customers a loaner car took almost allthe
pain out of the service experience. These insightsfrom the focus
groups helped explain the reasons be-hind the high J.D. Power
satisfaction scores. Andthey gave CRI's client a view of the Lexus
ownershipexperience not evident from the scores alone.
When ctistomers express their views on videotape,the effect is
even more compelling than printed verba-tim comments. For company
personnel, nothing beatsseeing the intensity of customers'
comments. South-west Airlines shows contact employees videotapes
ofpassengers complaining about service. Colleen Barrett,executive
vice president for customers, states: "Whenwe show the tape, you
can hear a pin drop. It's fasci-nating to see the faces of
employees while they'rewatching. When they realize the ctistomer is
talkingabout them, it's pretty chilling. That has far more im-pact
than anything I can say."'"
During the past few years, Levi Strauss & Co., oneof the
world's most successfiil companies, has beencompletely transforming
its business processes, sys-tems, and facilities. Improving the
speed and reliabili-ty of distribudon has been its principal
objective. Theteam leading the transformation tised videotaped
in-terviews with customers to help convince the em-ployees in such
a successfiil company that change wasessential. One big customer
said, "We trust many ofyour competitors implicidy. We sample their
deliver-ies. We open all Levi's deliveries." Another
customerstated, "Your lead dmes are the worst. If you
weren'tLevi's, you'd be gone.""
Companies investing in service-quality informationsystems
shotild consider using what McQuarrie calls"perennial quesdons."'^
A perennial question is open-ended and allows customers to speak
direcdy aboutwhat concerns them most. Companies should ask
itconsistendy and save responses in a database to ascer-
tain data patterns. GTE Supply's question, "Whatimprovements, if
any, cotild be made by Supply?" is aperennial question. McQuarrie
offers this example:"What things do we do partictilarly well or
pardcu-larly poorly, relative to our competitors?" Examples
ofperennial questions directed to employees include: What is the
biggest problem you face every day try-ing to deliver high-quality
service to your customers? If you were president of the company and
couldmake only one change to improve service quality, whatchange
wotild you make?"
Combining customers' words with their numbershas synergy. The
combination, when well executed,produces a high level of realism
that not only informsbut educates, not only guides but
motivates.
Link Service Perfonnance to Business ResultsInttiitively, it
makes sense that delivering quality servicehelps a company at the
bottom line. Indeed, acctimu-ladng evidence suggests that excellent
service enables afirm to strengthen customer loyalty and increase
mar-ket share.''' However, companies need not rely on out-side
evidence on this issue. Firms can develop theirown evidence of the
profit impact of service quality tomake the investment more
credible and fact-based forthe planning and budgeting process.
A service-quality information system should in-clude the impact
of service performance on businessresults. An important benefit of
new, declining, andlost-customer surveys is the measurement of
marketgains and damage linked to service quality. Surveyscan reveal
the number and percentage of new cus-tomers who selected the
company for service-relatedreasons. Declining and lost-customer
surveys can de-termine why customers are buying less or
defecting,allowing estimates of revenue lost due to service.
Cal-ctilating lost revenue because of service
dissatisfaction,categorized by specific types of service
dissadsfaction,is a dependable way to focus management attendonon
service improvement. By computing the averagecosts for reperforming
botched services and mtildply-ing them by frequency of occurrence,
companies alsocan calculate the out-of-pocket costs of poor
service.Combining lost revenue and out-of-pocket costs
at-tributable to poor service generally will produce a sumfar
greater than management wotild assume withoutformal estimation.
SLOAN MANAGEMENT REVIEW/SPRING 1997 BERRY & PARASURAMAN
73
-
Table 2 Customers'
Behavioral-IntentionsDimension
Loyalty to Company
Propensity to Switch
Willingness toPay More
External Responseto Problem
Internal Responseto Problem
Statements of Intention
ItemLabel
123
45
67
8
9
10
11
12
13
Item Wording*
Say positive things about XYZ to other people.Recommend XYZ to
someone who seeks your advice.Encourage friends and relatives to do
businesswith XYZ.Consider XYZ your first choice to buy services.Do
more business with XYZ in the next few years.
Do less business with XYZ in the next few years.Take some of
your business to a competitor thatoffers better prices.
Continue to do business with XYZ if its pricesincrease
somewhat.Pay a higher price than competitors charge for thebenefits
you currently receive from XYZ.
Switch to a competitor if you experience a problemwith XYZ's
service.Complain to other customers if you experience aproblem with
XYZ's service.Complain to external agencies, such as the
BetterBusiness Bureau, if you experience a problem withXYZ's
service.
Complain to XYZ's employees if you experience aproblem with
XYZ's service.
Each item was accompanied by a 7-point likelihood scale |1 =
"Not at All Likely" and 7 = "Extremely Likely").
Firms also can directly estimate the profit impactof effective
service recovery by measuring complainingcustomers' satisfaction
with the handling of their com-plaints and their repurchase
intentions. Technical Assis-tance Research Programs (TARP) has
conducted ex-tensive studies documenting the much stronger
repur-chase intentions of complaining customers who arecompletely
satisfied with the firm's response comparedto dissatisfied
ctistomers (complainants and noncom-plainants) who remain
dissatisfied. Firms can monitorthe relationship between service
recovery and businessresults by measuring dissatisfied customers'
propensityto complain (the higher the better because of the
op-portunity to resolve the complaint), and by measuringcomplaining
customers' satisfaction with the firm's re-sponse and their
repurchase intentions. These data canbe used to estimate the return
on investment in service
recovery, i.e., profits attributed to ser-vice recovery divided
by the costs ofservice recovery.'^
Another way to gauge the marketimpact of service quality is to
meastirecustomers' repurchase and other be-havioral intentions in
transactionaland total market surveys. The surveyscan ask
respondents to rate how likelyit is that they will, for example,
rec-ommend the firm, do more businesswith the firm in the next few
years, ortake some business to a competitorwith better prices.
Respondents' in-tentions can then be regressed againsttheir
perceptions of service quality toreveal associations between
customers'service experiences and their futureintentions concerning
the firm. Wehave investigated empirically a batteryof thirteen
behavioral intention state-ments. Using factor analysis, the
thir-teen-item battery reconfigured intofive dimensions (see Table
2)."^
Our research shows strong relation-ships between service
performanceand ctistomer loyalty and propensityto switch (see
Figure 4). Customerswhose service perceptions were belowthe zone of
tolerance were less loyal
and more likely to switch to a competitor than ctis-tomers whose
perceptions exceeded the zone. Cus-tomers exhibited some
willingness to pay more for bet-ter service, particularly as
service perceptions rose frominadequate to desired. Intentions to
complain external-ly fell slighdy across the zone.'^ (The internal
responsedimension is omitted from our analysis because it isbased
on a single item fi-om the thirteen-item scale.)
Companies that measure customers' behavioral in-tentions (or
actual behaviors) and monitor their sen-sitivity to changes in
service performance gain valuableinformation on both why and how to
invest in serviceimprovement. Assessing the bottom-line impact of
ser-vice performance will motivate managerial and non-managerial
employees to implement needed changes. Itwill help a company move
from just talking about ser-vice to improving service.
74 BERRY & PARASURAMAN SLOAN MANAGEMENT REVIEW/SPRING
1997
-
Figure 4 Relationship between Service Quality andCustomers'
Intentions for Computer Manufacturer
7 r Loyalty
Pay More
Propensity to Switch
Adequate ; Desired. Service , Service
Perceived Service Quality
Reach Every EmployeeA service-quality information system can be
beneficialonly if decision makers use it. Accordingly, it must
bemore than a data collection system; it must also be
acommunications system. Determining who receiveswhat information in
what form and when is a principaldesign challenge. Chase Manhattan
Bank vice presidentJohn Gre^ commented: "I cannot stressenough the
need to systematize the useof survey information, a key
learningpoint for tis in the last couple of years. Itis not just
how actionable the data are,but also the system for regtilarly
review-ing the data and making decisions thatdetermine
effectiveness."'"
All employees are decision makers asthey regtilarly make
decisions that deter-mine the effectiveness of their
actions;therefore, a service-quality informationsystem should
disseminate relevant ser-vice information to everyone in the
or-ganization. Front-line service providers,for example, should
receive informationabout the expectations and perceptionsof the
external or internal customersthey serve. These personnel might
re-ceive information different from whatexecutives receive and in
different
forms (for example, in training classes, newsletters, andvideos)
but they should be included in the system.Companies miss an
important teaching, reinforcing,culture-building opportunity when
they don't sharerelevant service information with employees lower
inthe hierarchy.
John Deere shares customer feedback with everyemployee. Its
system is designed so that employees indifferent functions receive
the information in an appro-priate form, e.g., via e-mail, a hard
copy of customercomments posted on btilletin boards, and
specializedmonthly reports. Les Teplicky, manager of
after-marketsupport at John Deere, stated: "You need senior
man-agement buy-in, good data collection, clear analysis but all
that won't matter unless every employee seessomething in the
information for them.""
Just as in the design of any product, knowing theneeds of
information users is critical to designing aservice-quality
information system. The system shouldrevolve around what
information different kinds ofemployees need to help them make good
decisions andhow and when to communicate the information. (SeeTable
3 for types of questions to include in both pre-design and
postimplementation surveys of targeted in-formation users.)
Packaging the right information foreach audience and presenting it
efi^ ectively is key to
Table 3 Questions for Service-Quality Information System
Users
Predesign What would you like to know about
the customers you serve
What type of information wouldhelp you improve service in
ourcompany?
What type of information wouldyou like to have about your
ownservice performance? About yourwork unit? About the
company?About the competition?
If you presently receive informationon customer service, what
type of in-formation is most valuable to you?Why? What is least
valuable? Why?
What are your preferred ways ofreceiving customer
serviceinformation? How often would youlike to receive this
information?
Postimplementation Are you receiving the information you
need to help the company improve itsservice? (for managers)Are
you receiving the information youneed to best serve your
customers?(for frontline employees)
What information on customer servicewould you like to receive
that youcurrently do not receive? How wouldthis additional
information help you?
What customer service information thatyou receive is most
valuable to you? Why?What is least valuable? Why?
Do you receive customer service informa-tion on a timely basis?
Please explain.
What could the company do to improvethe usefulness of the
customer serviceinformation it provides you?
SLOAN MANAGEMENT REVIEW/SPRING 1997 BERRY & PARASURAMAN
75
-
the success of a service-quality information system. AsPeter
Drucker stated: "Knowledge is power. In post-capitalism, power
comes ftom transmitting informa-tion to make it productive, not
hiding it."^ "
When listening to customers becomes a habit in acompany, when
managers find it unthinkable to makeservice investment decisions
unaided by relevant infor-mation, when employees eagerly await next
month'sservice performance scores to gauge progress, when
vir-tually all employees tmderstand the service improve-ment
priorities then it is clear that the organization issystematically
tising information to improve service.
References1. Quoted in "First Chicago Shelves Paper Surveys,
Asks Managers toUse the Telephone for Customer Satisfaction
Research," The ServiceEdge, volume 8, March 1995, p. 4.2. E.F.
McQuarrie, "Taking a Road Trip," Marketing Management,volume 3,
Spring 1995, p. 11.3. J.A. Goodman, S.M. Broetzmann, and C.
Adamson, "Ineffective That's the Problem with Customer Satisfaction
Surveys," QualityProgress, volume 25, May 1992, p. 35.4. For a
detailed discussion of this study, see:A. Parasuraman, V.A.
Zeithaml, and L.L. Berry, "Alternative Scalesfor Measuring Service
Quality: A Comparative Assessment Based onPsychometric and
Diagnostic Criteria," Journal of Retailing, volume70, Fall 1994,
pp. 201-230.5. See A. Parasuraman, VA. Zeithaml, and L.L. Berry,
"Reassessment ofExpectations as a Comparison Standard in Measuring
Service Quality:Implications for Further Research," Journal of
Marketing, volume 58,January 1994, pp. 111-124;J.J. Cronin and S.A.
Taylor, "SERVPERF Versus SERVQUAL:
Reconciling Performance-Based and
Perceptions-Minus-ExpectationsMeasurement of Service Quality,"
Journal of Marketing, volume 58,January 1994, pp. 125-131; andK.R.
Teas, "Expectations as a Comparison Standard in MeasuringService
Qtiality: An Assessment of a Resssesstnem," Journal of
Marketing,volume 58, January 1994, pp. 132-139.6. B.S. Lunde, "When
Being Perfect Is Not Enough," MarketingResearch, volume 5, Winter
1993, p. 26.7. J.P. Gregg, "Listening to the Voice of the Customer"
(Nashville,Tennessee: Frontiers in Services Conference,
presentation, October1995).8. Quoted in "Changes in Satisfaction
Demands and TechnologyAlter the How's, What's, and Why's of
Measurement," The ServiceEdge, volume 8, January 1995, p. 2.9. J.H.
Drew and T.R. Fussell, "Becoming Partners with Internal
Cus-tomers," Quality Progress, volume 29, October 1996, p. 52.10.
Quoted in "Some Ways to Coddle Customers on a Budget," TheService
Edge, volume 6, September 1993, p. 4.I L D . Sheff, "Levi's Changes
Everything," Fast Company, volume 2,June-July 1996, p. 67.12.
McQuarrie (1995), p. 12.13. L.L. Berry, On Great Service: A
Framework for Action (New York:Free Press, 1995), pp. 51-52.14. See
V A Zeithaml, LL. Berry, and A. Parasuraman, "The
BehavioralConsequences of Service Quality," Journal of Marketing,
volume 60,April 1996, pp. 31-46; andR.D. Buzzell and B.T. Gale, The
PIMS Principles (New York: FreePress, 1987).15. See Consumer
Complaint Handling in America: An Update Study(Washington, D .C:
Technical Assistance Research Programs Institute,April 1986).16.
Zeithaml etal. (1996).17. Ibid.18. Personal correspondence.19.
Quoted in "Rallying the Troops," On Achieving Excellence,
volumell,February 1996, p. 2.20. Interview with Peter F. Drucker,
Harvard Business Review, volume71, May-June 1993, p. 120.
Reprint 3835
76 BERRY & PARASURAMAN SLOAN MANAGEMENT REVIEW/SPRING
1997