A_service

A service positioning

matrix

1223

A service positioning matrixDavid A. Collier

The Ohio State University, Columbus, Ohio, USA andSusan M. Meyer

University of Minnesota, Minneapolis, Minnesota, USA

IntroductionA product-process matrix proposed by Hayes and Wheelwright (1979a; 1979b;1984) has gained wide acceptance in describing goods-producing decisionsconcerning strategy, process choice, and product mix. This product-processmatrix is an important management model used in almost all introductoryoperations management textbooks (Chase and Aquilano, 1992; Krajewski andRitzman, 1993; Schmenner, 1990).

However, the Hayes and Wheelwright product-process matrix does nottransfer well to service businesses and processes. Their premise is that productvolume, the number of products, and the degree of standardization/customization determine the manufacturing process that should be used. Thisrelationship between volume and process is not found in some servicebusinesses. For example, as volume increases service businesses such as retailoutlets, banks, and hotels have historically added stores, branch banks, andnew hotels (i.e. bricks and mortar) to meet demand but do not change theirprocesses. As Silvestro et al. conclude, “… in service operations, significantvolume increases can be made, and frequently are made, without any change inthe service process, as would be expected in manufacturing” (1992, p. 66).However, with the current capabilities of information technology, volume maybe more correlated with process design decisions, especially for information-intensive businesses such as banks.

Service organizations need business models that more accurately account forthe effects of service system designs and the roles of customer and service-provider choices in creating and delivering service encounters. Thedevelopment of a meaningful positioning matrix for services is an importantcontribution to the management literature. Many authors (Fitzsimmons andFitzsimmons, 1994; Kellogg and Nie, 1995; Lovelock, 1991; Schmenner, 1990;Silvestro et al., 1992; Tinnila and Vepsalainen, 1995) have tried to developclassification schemes or positioning matrices for services. To date, none ofthese schemes or matrices is truly satisfactory. For example, most matrices failto establish a clear direction of influence between matrix axes (e.g. the directionof influence in Hayes and Wheelwright’s product-process matrix is from theproduct axis to the process axis). Other problems and areas for improvement forpositioning matrices applicable to services are discussed later in this article.

In this article, we present the service positioning matrix (SPM), a model forservice organizations to follow in determining the appropriate service system

International Journal of Operations& Production Management,

Vol. 18 No. 12, 1998, pp. 1223-1244,© MCB University Press, 0144-3577

IJOPM18,12

1224

design based on the nature of the service that customers want. This matrixserves as a guide for helping managers make decisions on appropriate processdesign given the nature of service encounters that customers want. Theconceptualization and explanation of the SPM is a theory building initiativefollowed by a preliminary empirical test for theory verification.

Before the service positioning matrix is presented, a basic understanding ofthe relevant literature is established. The SPM is then presented and examplesare given to help develop a full understanding of this new matrix. Next theresults of an empirical test of the basic logic of the service positioning matrixare reported. The article concludes by discussing contributions and directionsof future research.

Literature reviewHayes and Wheelwright’s product-process matrix does not transfer well toservices and does not aid service managers in determining the most appropriateprocesses to use in the delivery of their services. Hayes and Wheelwright’smatrix and its applicability to services is discussed here, followed by discussionof other authors’ efforts to develop a matrix for services. Service encounters andrelated topics which are relevant to this research are also discussed in thissection.

Goods-producing Hayes-Wheelwright product-process matrixHayes and Wheelwright (1979a; 1979b; 1984) proposed a product-processmatrix that has gained wide acceptance in describing product-process choicesin goods-producing businesses. The vertical axis describes the process typeranging from jumbled flow (one-of-a-kind) to job shops (jobbing), todisconnected or connected line flows (batch and assembly lines), and tocontinuous flow processes. The horizontal axis defines the product mix in termsof low volume and one-of-a-kind (customized) products, to low volumes andmultiple products, to high volumes and several standardized products, to veryhigh volumes and no more than a few standardized products.

The basic theory of Hayes and Wheelwright’s model is that product mix(volume, number of products, degree of standardization) determines the choiceof process type. Here, the direction of causation is clearly from the product tothe process. The general hypothesis is that if an organization stays on thediagonal of the matrix, the product and process characteristics are wellmatched, and therefore, organizational performance is enhanced. Safizadeh etal. (1996) attempted to empirically validate this premise but provided onlylimited confirmation that performance of firms who position themselves on thediagonal outperform those who are positioned off the diagonal. Continuous flowshops were the only process choice to show statistically significant differencesin performance on and off the diagonal.

Blois (1983) applies the Hayes-Wheelwright product-process matrix to fourservice business (restaurant, hospital, fast food, and letters of credit). Thediscussion focuses on the relationship between marketing versus production

A service positioning

matrix

1225

policies but no numerical analyses is done. He concludes that many similaritiesexist when applying the Hayes and Wheelwright product-process matrix togoods-producing versus service-providing organizations but identifies a fewpuzzling differences. For example, the simultaneity of production andconsumption in services creates difficulties for the Hayes-Wheelwright matrix.

Other authors such as Noori (1989, pp. 130-59) extend the ideas of Hayes andWheelwright’s product-process matrix by suggesting that advancedmanufacturing technologies allow firms to be successful when they positionthemselves off the diagonal. These new technologies provide manufacturerswith the capability to be highly flexible and produce lower volumes of productsin greater varieties at lower costs (competitive prices). Therefore, these off-diagonal positioning strategies become viable for organizations.

Problems of trying to classify services using the Hayes and Wheelwrightmodel are discussed by several authors who have introduced alternative modelsfor service organizations. For example, Kellogg and Nie (1995) and Silvestro etal. (1992) recognize the problem of combining volume and customization on theproduct (service) axis. Measuring volume is not straightforward in services (e.g.Is one customer equal to one service, or is each transaction one service?) and thevolume of customers can fluctuate widely without a change in the servicedelivery system.

Thomas (1978) notes that “the use of product-based models and language todescribe and manage service businesses restricts thinking in a way that limitsinnovative management approaches”. Kellogg and Nie (1995, p. 333) supportThomas’ argument that new terminology is needed to adequately describeservice situations and use the term “service package” to describe whatcustomers are buying when they purchase a service. Collier (1994, pp. 63-84) hasalso sought new terminology to adequately describe the “set of tangible (goods-content) and intangible (service-content) attributes (features) the customerrecognizes, pays for, uses, or experiences”. He calls this bundle of goods andservices a “consumer benefit package”.

Service classification schemes and positioning matricesSeveral authors have introduced matrices for classifying services, but none ofthem satisfactorily defines the relationship between the service and the servicedelivery process. A matrix of service processes was defined by Maister andLovelock (1982) and refined by Schmenner (1986; 1990). One axis is defined as“the degree of labor intensity of the process”. The second axis is “the degree ofinteraction with and customization of the service for the consumer”.(Schmenner, 1990, p. 256). Problems arise with both of the dimensions of thismatrix. Kellogg and Nie (1995) argue that labor intensity is no longer applicablefor many services because information and automation have replaced labor insome industries. Schmenner’s second dimension (axis) is also difficult tointerpret (see Tinnila and Vepsalainen, 1995, pp. 70-3 for a full discussion)because it combines the degree of customer contact.

IJOPM18,12

1226

The result of this service classification scheme is four service types(quadrants): the service factory (low labor intensity and low consumer contact),the mass service (high labor intensity and low customer contact), the serviceshop (low labor intensity and high customer contact), and the professionalservice (high labor intensity and high customer contact). Please note that theMaister and Lovelock two-by-two matrix is a service classification schemewhereas Schmenner's matrix of service processes is used to facilitate theanalysis of process choice.

A framework for designing services for the banking industry is presented byApte and Vepsalainen (1993), but they do not generalize their service channelstrategies framework to other services. Their basic premise is that if thebanking services offered range from customized services to mass transactionservices, the delivery mechanism should range from a human-centeredapproach to a technology-centered approach. This theory assumes thatcustomization must be provider-driven, an idea that is contradicted in otherindustries where customers self-customize their service experiences such asestate planning, golf lessons, and using the Internet.

Another positioning matrix applicable to services is defined by Silvestro etal. (1992). They defined six dimensions on the vertical axis:

(1) equipment/people focus;

(2) customer contact time per transaction;

(3) degree of customization;

(4) degree of employee discretion;

(5) value added back office/front office; and

(6) product/process focus.

The number of customers processed by a typical service unit per day is thehorizontal axis. They conclude that there are three types of service processes:professional services are high in contact time, customization, and employeediscretion, and focus on people, front office, and process flows; service shopsexhibit intermediate characteristics on these six dimensions. Mass services arelow on contact time, customization, and employee discretion, and focus onequipment, back office, and product characteristics. Silvestro et al. used ten casestudies to show how services are located roughly on the diagonal of theirmatrix.

One of the problems with this positioning matrix is the definition of thehorizontal axis. While the number of customers processed by a service unit perday provides a surrogate measure to Hayes-Wheelwright’s product volume, it isvoid of information about the nature of the service. Knowing this one piece ofinformation (volume/service unit/day) about a service does not give the serviceprovider enough information to make decisions on the six dimensions of thevertical axis. Likewise, six dimensions on the vertical axis may be overlycomplex given that they are likely correlated. How these six dimensions define

A service positioning

matrix

1227

one construct or one axis is unclear. Additional empirical testing (beyond theten case studies discussed in the paper) may provide evidence that some ofthese six dimensions can be dropped without detriment to the conceptualizationof the matrix.

A service process analysis matrix was proposed by Tinnila and Vepsalainen(1995). The vertical axis is the type of channel that delivers the service and fourtypes of channels are defined as follows.

The market network channel provides direct customer access to market resources withminimal intermediation making it the shortest of channels, such as using an ATM. Servicepersonnel such as store clerks represent a short channel based on personal interaction. Theagent/alliance channel includes experts, third party representatives and independent agentsthat are fairly close to the customer acting as middlemen or mediators in the service channel.Internal hierarchy means that the service is procured within the organization needing it.Internal hierarchy is considered a long channel from the point of view of the potential serviceproviders since there is no real customer but instead an employment relationship."

The horizontal axis is the type of services and is defined, “on the basis ofcomplexity and contingencies involved …” Four types of services are defined(Tinnila and Vepsalainen, 1995, pp. 61-4): mass transactions are services withfew options and little customization. Standard contracts have complexspecifications but are not extensively adapted to an individual customer.Customized delivery are services tailored to individual customers involvingsome uncertainty and contingencies. Contingent relationships involve complexproblems, several interrelated activities, risk sharing and intensivecommunication between service provider and customer.

Both of Tinnila and Vepsalainen’s axes are complicated and difficult tounderstand. It is difficult to see the continuum in the definition of either axisthat is necessary for a positioning matrix. Firms would be expected to move upand down the diagonal of a positioning matrix, but at least one of theirdefinitions does not allow this. Internal hierarchy channels of service deliveryapply only to services obtained from within an organization. By definition, thatdoes not allow these services to move down the diagonal into matrix locationsof more standardized services, nor can business-to-business or customer-to-business transactions ever be located in the portion of the matrix occupied bythe internal hierarchy channels.

A service process/service package (SP/SP) positioning matrix was proposedby Kellogg and Nie (1995) that “uses customer influence to define the serviceprocess dimension and customization to define the service package dimension”.The vertical axis of the SP/SP matrix is the service process, and is organizedinto three categories of customer influence. For expert services, there is a highdegree of customer influence, and the service-provider and customer worktogether to define, produce and deliver the service package. In service shopprocesses, the service-provider and customer also jointly determine the process,but with some limitations due to standardization. In a service factory, customerinfluence is low. The service package, and the degree to which it is customized,is the horizontal axis of the SP/SP matrix, and is defined by four categories. A

IJOPM18,12

1228

unique service package is fully customized. A selective service packageincludes both customized and standardized components where the customerhas considerable discretion in selecting from a wide range of options. Arestricted service package is standardized and customer selection is limited. Ageneric service package has little or no customization.

Kellogg and Nie’s SP/SP matrix is difficult to interpret because thecustomer’s influence is present on both dimensions. For example, the authorsstate that an expert service (vertical axis), “involves a high degree of customerinfluence. An expert, working in conjunction with the customer, defines andcreates the service.” And they define a unique service package (horizontal axis)as, “The customer has considerable discretion in defining the hows, whats andwheres of the service.” The two matrix axes are confounded by includingcustomer influence on both of them.

Both Tinnila and Vepsalainen and Kellogg and Nie hypothesize that superiorperformance occurs for most organizations that stay on the diagonal of theirmatrices. For example, Tinnila and Vepsalainen (1995, p. 65) state that,“Companies close to the ‘efficient diagonal’ provide more efficient service thanthose far from it. The diagonal matches form the best service process bycombining a service with the type of channel that best suits the needs of thecustomers. The non-diagonal corner areas of the matrix are inefficient and oftenvoid of services.” Kellogg and Nie (1995, p. 329) state that, “While main diagonalpositions seem to offer the fit between the service process and service package,it is recognized that off-diagonal positions represent viable strategic choices.With technological advances, it is possible for some service firms to position inthe off-diagonal and still remain competitive.” This latter statement supportsNoori’s discussion (1989) on the impact of advanced technologies on thepositioning matrix.

A summary of the axes of the major classification schemes and matricesdiscussed here is shown in Table I. A general similarity of all matrices in TableI is an attempt to include two pervading concepts:

(1) customer and employee involvement; and(2) service system design.

Sometimes these two ideas are found on the same axis making them difficult tointerpret. As Table I documents, the differences between these servicepositioning matrices are great. The authors use many different criteria tomeasure these two ideas (constructs) which makes concise comparisonsdifficult. Empirical data analysis is necessary to determine which criteria arebest for each axis of a service positioning matrix. This paper begins thatempirical testing.

Classifications schemes are typically clusters or categories of service entitiesthat exhibit similar characteristics. Within each classification category certainmanagement issues dominate and strategies are developed to move from onecategory to another. The early work of Maister and Lovelock (1982), Lovelock(1991), Schmenner (1986; 1990) tend to fit into this definition of classification

A service positioning

matrix

1229

schemes. The two distinguishing characteristics of positioning matricescompared to classification schemes are that

(1) superior performance is assumed if the service entity follows a certainpathway (usually up and down a diagonal) on the matrix; and

(2) a direction of causation is implied by the positioning matrix, such asfrom the product to the process.

The work of Silvestro et al. (1992), Tinnila and Vepsalainen (1995), Kellogg andNie (1995), and the service positioning matrix defined here are positioningmatrices. These positioning matrices generally address the first requirement(i.e. superior performance along the diagonal) but do not have empiricalevidence to prove this superior performance. Some of these matrices are alsovague in defining the direction of causation between their two axes.

Service encountersThe concept of a service encounter is critical to the understanding of the servicepositioning matrix presented here, and the literature in this area will be brieflyreviewed. Service encounters are “face-to-face interactions between a buyer anda seller in a service setting” (Solomon et al., 1985, pp. 99-100). Solomon et al.continue, “Any encounter is assumed to contain learned and consistentbehavior patterns; each participant should enact certain behaviors in order forthe transaction to proceed smoothly”. Shostack’s (1985, p. 244) view of theservice encounter expanded the definition to include “a period of time duringwhich the consumer directly interacts with a service”. This broader definitionrecognizes that human interaction, either face-to-face or decoupled by a contact

Customer/employee Service system designMatrix developed by involvement dimensions dimensions

Schmenner (1986; 1990) Degree of customer contact Labor intensity of the processand customization

Silvestro et al. (1992) Number of customers Equipment/people, contactprocessed by a typical time, customization, employeeservice unit per day discretion, value added,

product/process focus

Tinnila and Vepsalainen Type of services (mass Type of channel of access to(1995) transactions, standard the service (market network,

contracts, customized delivery, service personnel, agent and contingent relationships) alliance, internal hierarchy)

Kellogg and Nie (1995) Service process structure Service package structure(degree of customer influence) (degree of customization or

customer discretion)

Note: Authors use different criteria and combinations of criteria per axis, so comparisons onthese two aggregate (customer/employee involvement and service system design) dimensionsare not always well aligned

Table I.A comparison of

positioning matrices for services criteria

IJOPM18,12

1230

technology such as a telephone line, is not required to complete a serviceencounter. This view includes interaction the customer has with buildings,equipment, employees’ uniforms, advertisements, etc.

Service encounters are important for several reasons. The service encounterlevel is where the customer judges the value of the service. Service providers atthe service encounter level require service management skills such as technicalexpertise, cross-selling other products and services, and good humaninteraction skills. Considerable resources are required to support serviceencounter execution. These resources include facility, process, job, andequipment design, network architecture and the location of the service facilities.

The physical setting, or infrastructure, where service encounters take placeis termed the servicescape (Bitner, 1992; 1993). The servicescape accounts forboth who is performing actions during service delivery and the physicalcomplexity of the service setting. Service delivery actions can be performed bythe customer alone, the employees alone, or some combination of the two. Bitnerdefines physical complexity as follows. “Some service environments are verysimple, with few elements, few spaces, and few forms. They are termed leanenvironments. Ticketron outlets and Federal Express drop-off kiosks wouldqualify as lean environments, as both provide service from one simplestructure … Other servicescapes are very complicated, with many elements andmany forms. They are termed elaborate environments. An example is a hospitalwith its many floors, rooms, sophisticated equipment, and complex variabilityin functions performed within the physical facility.”

Servicescapes do more than set the physical operating environment foracting out service encounters. Servicescapes also help define the customer’sroute through the service delivery system. Bitner (1993, p. 361) further describesa servicescape as follows. “The servicescape plays a variety of roles withrespect to both the customers and the employees of the service firm. Forexample, the servicescape serves as a package for the service offering in a waynot unlike that of a product’s package. Product packages are designed toportray a particular image to the consumer as well as evoke a particularsensory or emotional reaction. In a service setting, the servicescape does thesame thing through the interaction of many complex stimuli.”

The service positioning matrixThe SPM is shown in Figure 1. The vertical axis of the SPM is the number ofpathways (routes) built into the service system design by management. Thehorizontal axis is the customer’s service encounter activity sequence. These twoaxes are described in detail below. Two criteria characterize each of the axes ofthe service positioning matrix for a total of four criteria. The two criteria oneach SPM axis are assumed to be correlated and are candidates for factoranalysis (Hair et al., 1995; McDonald, 1985). Examples are provided later in thisarticle.

The theory behind the SPM is that:

A service positioning

matrix

1231

• the direction of causation is from the desired nature of the customer’sservice encounter activity sequence to the recommended service systemdesign;

• superior performance results from staying roughly on the diagonal of thematrix; and

• the two axes are conceptually independent.

This is roughly analogous to Hayes and Wheelwright’s product-process matrix(1979a; 1979b; 1984) which assumes that superior performance results frommatching product mix characteristics to process choice, and that the directionof causation is from product mix to process choice.

Figure 1.The service delivery

system matrix

Customer wants a highdegree of freedom anddecision making powerto select a serviceencounter activitysequence

Customer wants amoderate degree offreedom and decisionmaking power to selecta service encounteractivity sequence

Customer wants a lowdegree of freedom anddecision making powerto select a serviceencounter activitysequence

Fulfillment ofCustomer Wantsand Needs

ManagementDesignedService SystemCharacteristics

Unique never to berepeated serviceencounter activitysequence

Low to moderatelyrepeatable serviceencounter activitysequence

Highly repeatableservice encounteractivity sequence

CustomerRouted

Co-Routed

ProviderRouted

Many customerpathways

Managementdesigns a lowdegree of controlinto the servicesystem

Moderate numberof customerpathways

Managementdesigns a moderatedegree of controlinto the servicesystem

Limited number ofcustomer pathways

Managementdesigns a highdegree of controlinto the servicesystem

-

-

-

-

-

-

UniqueNot Repeatable

HighlyRepeatable

Customer’s ServiceEncounter Activity Sequence

Few

Num

ber

of P

athw

ays

Bui

lt in

to th

eS

ervi

ce S

yste

m D

esig

n by

Man

agem

ent

Man

y

IJOPM18,12

1232

Two criteria characterize the “horizontal axis” of the SPM which pertains tothe nature of the customer’s service encounter activity sequence. The serviceencounter activity sequence consists of all the process steps and associatedservice encounters necessary to complete a service transaction and fulfillcustomers wants and needs.

• The degree of customer discretion, freedom and decision making powerin selecting their service encounter activity sequence(s). Customers aregiven the opportunity, by management and the service system designitself, to design their own unique service encounter activity sequence(s),in any order (sequence) the customer chooses. Customer freedom toselect unique or standard service encounter activity sequences is thefocus of this criterion.

• The degree of repeatability of the service encounter activity sequence(s).Service encounter repeatability refers to the frequency that a specificservice encounter activity sequence is duplicated from one customer toanother. The degree of repeatability is limited by the service systemdesign and how customers select and configure their activity sequences.Service encounter repeatability provides a surrogate measure to productvolume for goods-producing firms. Repeatability can be counted for eachunique service encounter activity sequence (i.e. the number of cashwithdraws at a bank’s ATM). A high degree of repeatability encouragesstandardized process and equipment design, dedicated service channels,lower costs and improved efficiency. A low degree of service encounterrepeatability encourages more customization, more flexible equipmentand process designs, higher relative cost per transaction, and moreemphasis on effectiveness.

The “vertical axis” of the SPM is the number of pathways (routes) built into theservice system designed by management. This axis is defined as follows:

• The number of unique pathways (routes) that customers can take as theymove through the service system during delivery of the service. A largenumber of unique and predefined pathways (routes) indicates thatcustomers have freedom to develop their own unique service experienceby selecting their route through the service delivery system. A smallnumber of unique pathways (routes) available to the customer tocomplete the service indicates that customers have little freedom to selecttheir route through the service delivery system. Management mustdesign an infrastructure and service system that allows the customer toselect from one or many pathways. A pathway is a unique route throughthe service system that is used to fulfill certain customer wants andneeds. Pathways can be defined and counted especially by people whoare experts in operations and process management.

• Management’s degree of control designed into the service delivery system.Control of the service system results from the set of decisions that

A service positioning

matrix

1233

management makes regarding facility, job, and process design, includingthe servicescape and all of its accompanying dimensions. A low degreeof management control designed into the service delivery system resultsin more customer freedom while designs with a high degree ofmanagement control reduce the freedom of customers.

The term “service (delivery) system” used here includes:

• job design, process design, and facility design, where each of thesecomplements the others;

• technology and service innovations used during service delivery;

• customer and service provider interaction and employee training, and

• the servicescape (Bitner, 1992).

For a full discussion of service delivery systems see Collier (1994, pp. 46-50, 109-62, 223-60, 264-9; 26).

The SPM presented here is designed to be applicable at the “serviceencounter level”. For the horizontal axis, the SPM takes the “customer’sviewpoint”. Service encounters can occur through technology (e.g. computer,telephone) or in person (e.g. hospital, restaurant). Service encounters are criticalto management because this is where customers judge the value of the service.Back office processes, which support service encounters, are removed fromcustomer view and can be designed for efficiency. The SPM helps managementdesign a service system that best meets the technical and behavioral needs ofcustomers (i.e. the customer’s desired service encounter activity sequence(s)).

The service positioning matrix functions in the marketplace in two ways.First, it can be used when planning a new or revised service. Management usesmarketing research to determine the nature of the service encounter activitysequence(s) that meets or exceeds customer wants and needs for a particularmarket segment. Knowing the customer’s desired service encounter activitysequences (horizontal axis), the SPM suggests a certain service system designwhich will allow the appropriate number of unique pathways through theservice system (vertical axis). Second, once the service system design is in place,management uses it to control customer discretion and freedom during servicedelivery.

As we move down the diagonal of the SPM in Figure 1, it is hypothesized thatwe change the nature of the service system, reducing the number of pathwaysthat customers may take through the service system, so that the relative role ofthe customer becomes less prevalent while the role of the service-providerand/or the service system itself becomes more prevalent. At the same time, thenature of the service encounter activity sequence changes from completefreedom and discretion for the customer (top left corner of Figure 1) to little orno customer discretion (bottom right corner of Figure 1).

We define three states within the continuum on the diagonal of the matrixthat we hypothesize will describe most services:

IJOPM18,12

1234

(1) Customer routed services are those that offer the customer broadfreedom to select from many possible routes through the service deliverysystem;

(2) Co-routed services offer customers a moderate number of routes throughthe service delivery system;

(3) Provider routed services constrain customers to follow a very smallnumber of possible routes through the service delivery system.

Like the Hayes and Wheelwright product-process positioning matrix, it ishypothesized that an organization that ventures too far off the diagonal has atheoretical mismatch (discontinuity) between service system characteristicsand customer service encounter activity sequence characteristics.

The service positioning matrix in Figure 1 recommends that if anorganization wants to compete in the marketplace using a provider routedservice (with high repeatability and low customer discretion), it should adopt aservice system designed with few pathways for the customer to follow (bottomright corner of Figure 1). Likewise, if an organization wants to addentertainment and freedom of customer movement to the service encounteractivity sequence as a way to gain competitive advantage, the servicepositioning matrix recommends they adopt a service system design with manyto an unlimited number of unique pathways from which the customer maychoose (top left corner of Figure 1).

To understand best the SPM and its notions of service encounter activitysequences and pathways, it is informative and necessary to give examples ofhow and why certain entities fit on the service positioning matrix. We begin atthe top left corner of Figure 1 with examples of customer routed services forwhich service systems are designed with many pathways from whichcustomers may choose. We then discuss examples of co-routed services in themiddle of Figure 1. We conclude this discussion by examining the bottom rightcorner of Figure 1 where provider routed services reside, where service systemshave very few pathways (routes) from which customers may choose.

Customer routed services“Nike Town is a theatrical presentation, a glittering production number starringthe customer. People love to shop here. It’s kind of entertainment, a social thing”said Mary Burns, store manager (The Columbus Dispatch, 1993). At Nike Town,freedom, entertainment, color, fantasy, technical shoe performance information,videos, and music are part of the goods and services attributes which are bundledwith the core good – shoes. The store has been so successful, Nike has openedseveral larger stores. Here, service encounter activity sequences are left to thediscretion of the shoe buying customer. Customers set their own pace and definetheir routes through the store, processing times at each stage, when they wantself-service versus assistance from a sales representative, and how to bundlemusic, entertainment and fantasy with buying shoes. Customers may spend threeminutes or three hours in the store depending on their wants and needs at that

A service positioning

matrix

1235

time. A three-minute service experience would likely include relatively simpleservice encounter activity sequences such as checking to see if a particular shoeis in stock and immediately leaving if it was not. A three-hour service experiencemight include listening to music, playing basketball in five different shoes,watching videos, talking to store employees and other customers, etc.

The innovative service system design at Nike Town encourages customers todesign their own unique service experience and sequence of service encounters.This design-your-own service encounter activity sequence is a vision ofeconomic competition in the Service and Information Ages (Collier, 1983; 1994;1996). The store servicescape, layout, procedures, employee job design, etc.provide the opportunity for the customer to “design their own unique set ofservice encounter activity sequences”.

Surfing the Internet exhibits an endless variety of service encounter activitysequences, electronic service encounters, and nearly unlimited number of routesfor customers to move through the service system. The “customer’s mind”(rather than the customer’s body) is moving through cyberspace and self-designing and selecting real-time service experiences and pathways. Weanalyze the service system by following the customer’s mind through thesystem, not the physical hardware and software that make this serviceencounter activity sequence possible. For this service system, the customer hasalmost complete freedom to self-design the service experience. A customerrouted sequence of service encounters is not likely to be repeated in the futureby either the same customer or other customers.

Club Med provides the customer with many opportunities to have fun andinteract with other guests. The customer has nearly complete freedom anddiscretion to design his or her vacation, and the service encounter activitysequences, whenever and whatever way they want. The Club Med servicesystem design simply defines the boundaries and activities available, the rest isup to the individual. The service system also fulfills many basic needs such asconvenient and good food, safety and security, a sense of excitement, optionalactivities, freedom, and the opportunity for privacy.

Parks, museums and health clubs are examples of customer routed services.Here, customers move through the service system and its processes at their ownpace and self-select bundles of service encounters best suited for theirimmediate needs and wants. The sequence of service encounters can bereconfigured, reversed, or stopped by the customer at almost any time. But thisbuilt in degree of freedom for customers does not mean that management hasno control over the service encounter activity sequences. For example, amuseum may define a dominate pathway, such as customers entering andexiting at certain places and times, but not restrict the pathways within a multi-floor museum.

Co-routed servicesIn the mid-range of the service positioning matrix the degree of serviceencounter repeatability is moderate. A moderate number of customer pathways

IJOPM18,12

1236

dominate these service processes. Another distinguishing criterion is that thecustomer and service provider jointly decide the service encounter activitysequence. Co-routed services include consulting, investment portfolio, and legaland medical services. A golf course is an example of a co-routed service in thatmanagement defines the dominant sequence of playing from hole No. 1 to No. 18but the customer has many options within this pre-designed service system inhow to play the course. In this mid-range, customers lose some (golf course) ormost (medical services) of their control in terms of moving through the servicesystem.

The midrange portion of the service positioning matrix, where we wouldexpect to find co-routed services, contains a complex set of management andcustomer choices. At the upper left corner of the co-routed area we would expectmore unique service encounter activity sequence options for the customer suchas legal and medical customized services or a customized golf lesson. At thelower right corner of the co-routed area we would expect more standardizedservice encounter activity sequence options for the customer such as CharlesSchwab & Company’s direct on-line stock trading capabilities or H&R Block’selectronic filing of federal tax forms. There is a broad range of shared decisionmaking power between the customer and provider in this area of the matrixbecause the customer is participating in the creation and delivery of the service.

In summary, the mid-range operating environment of the SPM exhibits amoderate degree of customer discretion, shared decision making in terms ofselecting the service encounter activity sequence(s), a complex set ofintermediate pathway and process flow structures, and access to and frequentoversight from a service-provider somewhere in the activity sequence(s). Asnoted previously, due to the very nature of these mid-range services,organizations that reside here may define their service strategy in terms ofmoving to one of the extreme corners (provider or customer routed) of theservice positioning matrix. It may be difficult for management to differentiateits service system from competitors in this mid-range arena. A future researchchallenge is to break this aggregate midrange (co-routed service) down into amore specific dichotomy.

Provider routed servicesA newspaper dispenser is an extreme example of a service system design withonly one pathway, thus allowing a single service encounter activity sequence. Anewspaper dispenser is also a fixed sequence robot (Collier, 1983). Here, there isonly one way (pathway) to obtain a newspaper based on the design of thenewspaper dispenser and distribution system. The customer must have thecorrect coins, put the coins in the proper slot, pull the handle and door open, andget the newspaper out. There is no customer discretion or freedom in how toobtain a newspaper. There is no professional service-provider there to guide thecustomer through the service system. Here, the service encounter activitysequence is between a human and the simplest of machines, and highlyrepeatable.

A service positioning

matrix

1237

An automatic teller machine (ATM) is an example of a simple service systemwith a very limited number of pathways from which the customer may choosewhile moving through the electronic servicescape. The customer must have thecorrect card, put the card in the proper slot, enter information in an electronicpad to make a deposit or withdrawal. There is little customer discretion orfreedom in how to use an ATM and no professional service-provider to guidecustomers through the servicescape and its processes. Although we term thistype of service as provider routed, it is actually technology that is defining thetight service system design and limiting customer discretion. For a typical ATMthere are only a few pathways allowed by the system design. These predefinedpathways include getting cash, making a deposit, checking account balance,and moving money from one account to another.

McDonald’s (Levitt, 1972; Sasser and Rikert, 1980) fast food restaurant isanother example of a service system with few pathways from which thecustomer may choose. The standardization of McDonald's consumer benefitpackage, facility, process, layout, jobs, uniform, grooming, communication, andequipment design are well known. This tightly designed service system iscombined with tight servicescripts as documented by their extensive videotraining program and McDonald’s Hamburger University. To ensure constancyof outcome (what) and process (how) quality, each store is audited frequentlyusing standardized visitation reports. Here, it is management that is tightlycontrolling customer routes through the service delivery system.

CNN Headline News and credit cards are characterized as provider routedservices with little customer discretion in how to use or experience the servicedelivery system(s). Their processes are best described as continuous flow wherevariable sequence and playback robots dominate the technology. Here, there isa high degree of repeatability in how these service systems are used. Thecustomer has very limited options in how to use the credit card and must watchthe content programming of CNN with no option to change it. Managementdefines a very limited set of customer pathways with a high degree of servicesystem control. Of course, the customer could substitute other services forthese, such as electronic shopping and switch television channels, but withinthe context of the SPM these are totally different services.

The SPM and technologyTechnology is used in different ways at various locations in the SPM. We wouldexpect the impact of information technology to be present all along the diagonalof the SPM in Figure 1. This contrasts with Hayes and Wheelwright’s product-process matrix which generally finds technology more prevalent at the lowerend of the diagonal (as you move down the diagonal) in high volume continuousflow manufacturing processes with very high levels of standardization. In theseprocesses, technology makes the organization more efficient by replacing labor.In service organizations, technology is useful for improving efficiency andeffectiveness, and in both increasing (provider routed) and decreasing (customerrouted) standardization (Collier, 1983).

IJOPM18,12

1238

Information technology (IT) is a unique and different resource thantraditional resources such as labor, materials, and capital equipment. As Nolanand Davenport (1995, p. 5) note, IT gets much cheaper every year and can makeprevious investments in IT completely obsolete (i.e. creative destruction). IT ischanging the economics of business (i.e. economies of scale and scope) andredefining what is possible with regard to service system design. Hence,examples can be found everywhere in the SPM solution space where ITincreases or decreases the number of service encounter activity sequencesavailable to customers and the number of pathways built into the servicesystem.

For a customer routed service like the Internet, technology providescustomers with maximum freedom and unlimited unique pathways.Management designed the infrastructure but when and where the customergoes in cyberspace is the customer’s choice. For co-routed services in the mid-range of the SPM, customers have a moderate range of pathways from which toselect, and therefore, create their own service encounter activity sequence(s).For example, Charles Schwab & Company’s direct online stock tradingcapabilities provide a moderate number of pathways for the customer to checkon, monitor, or trade stocks. Although the customer has complete freedom totrade any stock, Charles Schwab’s management has built in control mechanismsto ensure the trade is accurately completed.

Technology is used in provider routed services to replace service deliveryactivities that were performed by workers in the past such as ATMs. Here,customer freedom and the number of pathways available to the customer arevery limited, as previously discussed.

Technology will likely continue to become more pervasive in services, andcan be used to open (Internet) or tighten control (ATM) of the service encounteractivity sequence(s) available to customers. These service encounter activitysequences and associated pathways range from millions of pathways movingthrough cyberspace to only one pathway to get a newspaper out of a newspaperdispenser.

Empirical test of the service positioning matrixA preliminary empirical test of the service positioning matrix was performedby administering a survey to 64 evening MBA students. Evening MBA studentswere selected for this study because we sought participants who had areasonable understanding of processes, as they are presented in mostintroductory operations management courses, in addition to currentprofessional experience with service delivery processes. The sample chosen forthis study included only students who had previously completed an MBA-leveloperations management course, and these participants averaged 5.5 years ofprofessional work experience. These qualifications ensure that participantshave been exposed to the design and functioning of processes in themarketplace, as well as some of the concepts and terminology of operationsmanagement.

A service positioning

matrix

1239

The objective of this empirical test was to test the logic of the two matrixaxes. The survey was three pages in length with the first page containingsections similar to this article’s description of service encounters and the servicepositioning matrix, defining the matrix criteria. The second page listed 26service entities (company, machine, organization, or system) to be placed on theSPM by the survey respondent, as shown in Table II. The third page of thesurvey was the service positioning matrix shown in Figure 1 but without thethree labeled circles. The 64 survey participants had previously taken anintroductory operations management course but had not been exposed to theSPM prior to receiving the survey. Respondents were instructed to place eachservice entity on the service positioning matrix in the location that wouldindicate the entity's relationship between the two axes. Each axis of the matrixconsisted of a three point scale.

A total of 1,584 entities (26 service entities times 64 survey participant minus80 missing observations) were placed on the service positioning matrix by the

Entity survey participant positioned Average scores on:on service positioning matrix Horizontal axis Vertical axis

Surfing the Internet 1.41 1.25Charles Schwab & Co. stock buying services 1.89 1.97Tax and estate planning for a millionaire 1.48 1.67McDonald’s 2.71 2.53Major museums 1.89 1.95

Club Med 1.56 1.83Newspaper dispenser 2.84 2.63CNN Headline News 2.37 2.63Automatic teller machine (ATM) 2.61 2.42Chicago’s Nike Town store 1.88 2.05

Kidney transplant surgery 2.15 2.53Federal Express 2.34 2.16Basic bank checkbook service 2.45 2.17Ritz-Carlton Hotels 1.80 1.85Health clubs 1.79 1.71

Yellowstone National Park 1.57 1.75Legal services to set up a simple will 2.11 2.21Merrill Lynch Investment Portfolio Analysis 1.64 1.93

for a specific customer accountBlood test in a hospital laboratory 2.75 2.84H&R Block Tax Services for 1040EZ form 2.59 2.64

Flying on a major airline 2.23 2.15L.L. Bean mail orders 2.05 2.09Dental check-up and cleaning 2.43 2.45Motel 6 2.41 2.41Personal lessons from a golf pro 1.61 1.79Barnes and Noble Bookstore 1.67 1.87

Note: Company names are in italic; see Figure 2 for graph of average scores listed here

Table II.Service entities and

average scores on the service

positioning matrix

IJOPM18,12

1240

64 participants. The responses show statistically significant linear anddiagonal trends, as shown in Tables III and IV and Figure 2. Table III shows thefrequency counts of participants’ placement of service entities in each of thenine cells of the SPM. Analysis of these data indicates that the Pearson chi-square test for linear association is 413.4, highly significant at the p < 0.001level, as shown in Table IV. Tables III and IV and Figure 2 also reveal that adiagonal trend can be observed in the data. Table IV shows that Pearson’scorrelation coefficient (r) is 0.423, also highly significant at the p < 0.001. Thesestatistics suggest that the survey respondents consistently interpreted the twoaxes inherent in the SPM. The Pearson chi-square and Pearson r tests are usedfor this analysis because they are the most appropriate tests for analyzing theinterval data which were obtained from the survey.

A majority of service entity placements were located on the diagonal of theservice positioning matrix, indicating that these service-driven entities useprocesses that are designed by management to match the level of customerinvolvement in the service delivery system. The matrix was analyzed as a three-by-three matrix (indicated by the boxes in Figure 1), and the three diagonalboxes captured 57 percent of all responses, while the six off-diagonal boxescontained the remaining 43 percent (see Table III). If responses were random, wewould expect to observe approximately 11 percent (176 responses) in each box.

To further analyze the data obtained from the survey at a more aggregatelevel, we averaged the responses of the 64 participants for each of the 26 serviceentities based on the three point scales used to define the axes of the SPM. Theaverage values obtained for the 26 service entities are given in Table II. Theseaverage scores for each service entity are plotted in Figure 2. For example, theaverage score for the Internet on the horizontal axis was 1.41 and on the verticalaxis was 1.25. Similarly, the average scores for a blood test in a hospital

Management designed Customer’s service encounter activity sequenceservice system Customer Provider Rowcharacteristics routed Co-routed routed totals

Many customer routes 241 81 82 404Moderate number of customer routes 144 279 144 567Limited number of customer routes 90 138 385 613Column totals 475 498 611 1,584

Note: (26 service entities × 64 students) – 80 missing observations = 1,584 total observations

Table III.Frequency counts from empirical tests of the service positioning matrix

Degrees Statistical1,584 total observations value of freedom Significance

Pearson chi-square 413.4 4 p < 0.001Pearson correlation coefficient (r) 0.423 p < 0.001

Table IV.Statistical measures of association

A service positioning

matrix

1241

laboratory were 2.75 on the horizontal axis and 2.84 on the vertical axis. Theseaverage scores have a Pearson correlation coefficient (r) of 0.915, as shown inFigure 2, highly significant at p < 0.001.

The results plotted in Figure 2 show the similarities between variousservices in both management’s design of the number of pathways through theservice system and customer discretion during the service encounter activitysequence. For example, McDonald’s and ATMs have few pathways in theirservice system designs while allowing customers little freedom in designingtheir service encounter activity sequence. In contrast, both YellowstoneNational Park and estate planning for a millionaire have numerous pathwaysbuilt into their service system, allowing customers great freedom and flexibilityto design their service encounter activity sequence.

The results provide preliminary statistical support for the theoreticalframework used in the Service Positioning Matrix. First, the linear correlationbetween the two axes of the SPM was 0.423 (p < 0.001) indicating a strong linearand diagonal relationship between these constructs. Second, the average valuesfor each of the service entities included on the survey resulted in a very highcorrelation (r = 0.915) and Figure 2 shows that survey participants place theseservice entities at or near the locations that the authors theorize they are located.

Figure 2.Average responses of

service entity placementon service positioning

matrix

UniqueNot Repeatable

HighlyRepeatable

Customer’s ServiceEncounter Activity Sequence

Few

Num

ber

of P

athw

ays

Bui

lt in

to th

eS

ervi

ce S

yste

m D

esig

n by

Man

agem

ent

Man

y

Internet

Estate planning

Golf lesson

Club Med RitzHealth club

YellowstoneM.Lynch

Bookstore MuseumC.Schwab

LLBeanNike

TownSimple will

Airline

FedEx

Checkbookservices

Kidneytransplant Dentist

Motel 6 ATM

CNN

McDonald’s

H&R Block1040EZ

Newspaperdispenser

Blood testPearson’s r = 0.915

Note: See Table II for the average scores of entities graphed here.

IJOPM18,12

1242

ConclusionsThe SPM provides a new way to think about service businesses. It is not aperfect paradigm but it is a new way of thinking. It matches the customer’sdesired service encounter activity sequence to the design and structure of theservice system and its number of pathways. Although conceptual in nature, theSPM gives managers a new way to think about their business, the customer’sservice experience, and service system design. Please remember the focus of theSPM is at the service encounter level taking either the customer’s perspective(i.e. the horizontal axis of SPM) or management’s perspective (i.e. the verticalaxis of SPM). This article contributes to the continued development of apositioning matrix for services in the following ways:

(1) The service positioning matrix (SPM) introduced in this paper is a twodimensional tool for identifying an appropriate match between thenumber of pathways (routes) allowed by the service system design andthe degree of customer discretion in selecting their service encounteractivity sequence. Management must design a service system that meetsthe wants and needs of the customers in their targeted market segments.Once the design is set, there are a certain number of pathways forcustomers to be processed through the service system. Surfing theInternet, for example, provides an infinite set of possible serviceencounter activity sequences while an ATM allows very few.

(2) The typology of the service encounter activity sequence is describedusing two classification criteria:• the degree of customer discretion, freedom and decision making

power in selecting their service encounter activity sequence; and• the degree of repeatability of the service encounter activity

sequences. The foundation for these ideas is in Collier (1994; 1996).(3) The statistical evidence indicates that the empirical data support the

theoretical framework of the SPM. We can find no other research onpositioning matrices for services with similar statistical support.

(4) Preliminary empirical support for the service positioning matrixrevealed that the matrix axes were meaningful to the survey participantsand most service entities are located on the diagonal of the matrix, astheorized. The analysis of these data also indicates a linear and diagonalrelationship between management's service system design andcustomers' discretion to self-design a set of service encounters.

(5) The SPM represents a theory building initiative for services. Thestatistical testing and results represent a preliminary effort at theoryverification.

All positioning matrices for services need more empirical testing. Researchdesigns should also compare alternative positioning matrices using a commonset of respondents (Collier and Meyer, 1998). Other areas for future research

A service positioning

matrix

1243

include determining which service positioning matrix is best or what criteriaare best per axis. Can the SPM help companies define their service strategy?Also, the midrange of the service positioning matrix (co-routed) needs moreresearch. Can it be desegregated into a more detail dichotomy? Research onthese issues is on-going.

Both goods-producing and service-providing organizations are addingservice-, information- and entertainment-content to their consumer benefitpackages. As world economies move toward services, we must build newmanagement models that reflect the role services, service systems, and serviceencounters play in gaining competitive advantage.

ReferencesApte, U.M. and Vepsalainen, A.P.J. (1993), “High tech or high touch? Efficient channel strategies

for delivering financial services”, Journal of Strategic Information Systems, Vol. 2 No. 1, pp. 39-54.

Bitner, M.J. (1992), “Servicescapes: the impact of physical surroundings on customers andemployees”, Journal of Marketing, Vol. 56 No. 2, pp. 57-71.

Bitner, M.J. (1993), “Managing the evidence of service”, in Scheuing, E.E. and Christopher, W.F.(Eds), The Service Quality Handbook, American Management Association (AMACOM), NewYork, NY, pp. 358-70.

Blois, K.J. (1983), “The structure of service firms and their marketing policies”, StrategicManagement Journal, Vol. 4 No. 3, pp. 251-61.

Chase, R.B. and Aquilano, N.J. (1992), Production & Operations Management, 6th ed., Irwin,Homewood, IL.

Collier, D.A. (1983), “The service sector revolution: the automation of services”, Long RangePlanning, Vol. 16 No. 6, pp. 10-20.

Collier, D.A. (1994), The Service/Qual ity Solution – Using Service Management to GainCompetitive Advantage, Irwin Professional Publishing, Burr Ridge, IL and Quality Press,Milwaukee, WI.

Collier, D.A. (1996), “Tight and loose comprehensive customer contact (3C) plans”, Advances inService Marketing and Management, JAI Press, Greenwich, CT, Vol. 5, pp. 153-68.

Collier, D.A. and Meyer, S.M. (1998), “An empirical comparison of service matrices”, under review.Fitzsimmons, J.A. and Fitzsimmons, M.J. (1994), Service Management for Competitive Advantage,

McGraw-Hill, New York, NY.Hair, J.F. Jr, Anderson, R.E., Tatham, R.L. and Black, W.C. (1995), Multivariate Data Analysis, 4th

ed., Prentice-Hall, Englewood Cliffs, NJ. Hayes, R.H. and Wheelwright, S.C. (1979a), “Linking manufacturing process and product life

cycles”, Harvard Business Review, Vol. 57 No. 1, pp. 133-40. Hayes, R.H. and Wheelwright, S.C. (1979b), “The dynamics of process-product life cycles”,

Harvard Business Review, Vol. 57 No. 2, pp. 127-36. Hayes, R.H. and Wheelwright, S.C. (1984), Restoring Our Competitive Edge, John Wiley & Son,

New York, NY. Kellogg, D.L. and Nie, W. (1995), “A framework for strategic service management”, Journal of

Operations Management, Vol. 13 No. 4, pp. 323-37. Krajewski, L.J. and Ritzman, L.P. (1993), Operations Management: Strategy and Analysis, 3rd ed.,

Addison-Wesley, Reading, MA.Levitt, T. (1972), “Production-line approach to service”, Harvard Business Review, Vol. 50 No. 4,

pp. 41-52.

IJOPM18,12

1244

Lovelock, C.H. (1991), Services Marketing, 2nd ed., Prentice-Hall, Englewood Cliffs, NJ.McDonald, R.P. (1985), Factor Analysis and Related Methods, Lawrence Erlbaum Associates,

Hillsdale, NJ. Maister, D. and Lovelock, C.H. (1982), “Managing facilitator services”, Sloan Management Review,

Vol. 23 No. 4, pp. 19-31. Nolan, R.L. and Davenport, T. (1995), “Reengineering: competitive advantage and strategic

jeopardy”, Note No. 9-196-016, Harvard Business School, Boston, MA. Noori, H. (1989), Managing the Dynamics of New Technology: Issues in Manufacturing

Management, Prentice-Hall, Englewood Cliffs, NJ.Safizadeh, M.H., Ritzman, L.P., Sharma, D. and Wood, C. (1996), “An empirical analysis of the

product-process matrix”, Management Science, Vol. 42 No. 11, pp. 1576-91.Sasser, W.E. Jr and Rikert, D.C. (1980), “McDonald’s Corporation” (Condensed), Case No. 681-044,

Harvard Business School, Boston, MA. Schmenner, R.W. (1986), “How can service businesses survive and prosper?”, Sloan Management

Review, Vol. 27 No. 3, pp. 21-32.Schmenner, R.W. (1990), Production/Operations Management, 4th ed., Macmillan, New York, NY.Shostack, G.L. (1985), “Planning the service encounter”, in Czepiel, J.A., Solomon, M.R. and

Surprenant, C.F. (Eds), The Service Encounter, Lexington Books, New York, NY, p. 244. Silvestro, R., Fitzgerald, L., Johnston, R. and Voss, C. (1992), “Towards a classification of service

processes”, International Journal of Service Industry Management, Vol. 3 No. 3, pp. 62-75. Solomon, M.R., Surprenant, C., Czepiel, J.A. and Gutman, E. G. (1985), “A role theory perspective

on dyadic interactions: the service encounter”, Journal of Marketing, Vol. 49 No. 1, pp. 99-111. The Columbus Dispatch (1993), “Shoe buyers move to Nike Town”, 1 August, p. 2B. Thomas, D.R.E. (1978), “Strategy is different in service businesses”, Harvard Business Review,

Vol. 56 No. 4, pp. 158-65. Tinnila, M. and Vepsalainen, A.P.J. (1995), “A model for strategic repositioning of service

processes”, International Journal of Service Industry Management, Vol. 6 No. 4, pp. 57-80.