European Management Journal Vol. 16, No. 2, pp. 169–189 ...€¦ · help design intangible elements. We conclude the paper by identifying gaps in the literature and sug-gesting

European Management Journal Vol. 16, No. 2, pp. 169–189, 1998 1998 Elsevier Science Ltd. All rights reservedPergamon

Printed in Great Britain0263-2373/98 $19.00 + 0.00PII: S0263-2373(97)00086-8

A Structured ProductDevelopment Perspectivefor Service OperationsGABRIEL BITRAN, Sloan School of Management, MITLUIS PEDROSA, Sloan School of Management, MIT

In this paper we review the literature on productdevelopment from a services perspective. We ident-ify similarities in the creation and evolution of pro-ducts and services, and discuss three types ofknowledge that are commonly required in a devel-opment process: the sequence of steps or proceduralplan that must be followed; the understanding ofwhat components integrate the design and howthey interact (architectural knowledge); and theprinciples and models that describe physical orhuman behavior in the system that is beingdesigned. For each step of a generic developmentprocess we review the methods and tools that are

European Management Journal Vol 16 No 2 April 1998 169

widely used in product development and may besuccessfully applied to service development. Toillustrate the notion of architectural knowledge inthe service context, we introduce an example of aservice operation structure and discuss importantaspects of its components. Finally we explain therole of models in the development of products andservices and argue how they can help design intan-gible elements. We conclude the paper by ident-ifying gaps in the literature and suggesting direc-tions for future research. 1998 Elsevier ScienceLtd. All rights reserved

A STRUCTURED PRODUCT DEVELOPMENT PERSPECTIVE FOR SERVICE OPERATIONS

Introduction

In this paper we review the literature on productdevelopment from a services perspective. For the ser-vices community we review methods and tools thatare widely used for product development and maybe successfully applied to service development. Forthe manufacturing audience we explore some aspectsthat might be considered in the development of ser-vices as a competitive edge for products.

Although products and services have tangible andintangible elements, most of the product develop-ment literature deals mainly with tangible character-istics of both. In a world of increasing competition,however, manufacturing companies are beingrequired to not only design better products but alsodesign the appropriate supporting services. Oneexample of this new trend is the recent introductionof a line of cars with a differentiated network ofdealerships and supporting services. On the servicesside, companies are being required to look for waysto make their operations more reliable, consistent,and replicable at the global level. This is more likelyto be accomplished if a systematic, product-likeapproach for service development is employed.

The rest of the paper is organized as follows. In thesecond section we discuss the creation and evolutionof products and services to point out similaritiesbetween the two contexts. In the third section wereview development methodologies and three typesof knowledge that are commonly required in thedevelopment of a product or service. The first isabout the sequence of steps or procedural plan thatmust be followed in the development process. Thesecond is the understanding of the components thatintegrate the design and how they interact(architectural knowledge). Finally, the third type ofknowledge concerns the principles and models thatdescribe physical or human behavior in the systemthat is being developed. Such principles are the foun-dation over which designers build models to assesstradeoffs and check the feasibility of design solutions.

Within this framework in the fourth section wereview each step of a generic development processtogether with the tools and methods that can beemployed at that step. In the fifth section we illus-trate the notion of architectural knowledge in thecontext of services. We do so by introducing anexample of a service operation structure with ninecomponents and discussing important aspects ofeach. In the sixth section we discuss the role of mod-els in a development process and argue how they canhelp design intangible elements. We conclude thepaper by identifying gaps in the literature and sug-gesting directions for future research.

European Management Journal Vol 16 No 2 April 1998170

Product and Service Innovation

The greater the intangible component of a product orservice, the more difficult it is to understand whatthe customers want, why they want it, and how todeliver it. The evaluation criteria of intangibleelements are in general subjective, multi-dimensional(e.g., pleasure, courtesy, convenience, hope) and notalways clearly defined. High intangibility is thereforeat the root of most difficulties in dealing with inno-vation. It makes the precise definition of a conceptand its subsequent design difficult at best.

The classical model for product evolution suggestedby Abernathy and Utterback (1978) argues that whena major product innovation is first introduced to themarket, its production processes and equipment arefor general purpose, its materials are widely avail-able, and ‘performance criteria are vague and littleunderstood.’ As the product adoption grows, compe-tition stimulates experimentation in both productand process, and a variety of designs compete forcustomer acceptance. Successive innovations aretested in the product until a dominant designemerges, influenced by a clearer definition of per-formance. Thereafter innovation is incremental forboth product and process. Production process andequipment become dedicated to reap economies ofscale, and special materials or technologies are intro-duced to improve the performance of the product.

Although the distinction between product and pro-cess is not always clear in services, the Abernathy–Utterback model provides insight for services as well.Sasser et al. (1978) discuss the evolution of a serviceoperation in terms of cost, range of services, and geo-graphical reach as it goes through stages of evolutionsimilar to those described by the Abernathy–Utterback model: entrepreneurial, multi-site rational-ization, growth, maturity, and decline/regeneration.A dominant design in services is the set of tangibleand intangible elements implemented by functionsthat are essential for that service (core functions) andfunctions that support the service experience(peripheral functions). Once a firm offers a new fea-ture that enhances the service experience, be it tan-gible (e.g., in-flight meal) or intangible (e.g., con-venience, speed), competition soon follows and thenew feature becomes part of the industry standard.Customers incorporate the new feature into theirexpectations and it becomes part of the dominantdesign. The existence of dominant designs in servicescan be observed by the similarity in operations andformats of both regulated and competitive servicebusinesses, across a wide range of industries such asfast-foods, hotels, airlines, consulting, and so forth.As a dominant design appears and the industrymatures, companies tend to use specialized tech-nology or equipment whenever possible to improveperformance of their service operation — fast-foodrestaurants, airlines, and banks are typical examplesof such phenomenon.

Architectural variations

often lead to new concepts


Clark (1985) further elaborates on the emergence ofa dominant design, arguing that the design processleads to the identification and analysis of the compo-nents of a product, and experimentation with theirinterrelations. Since some components are more rel-evant to the product concept, their design choicesimpose constraints on the design of other compo-nents, creating a hierarchy of design. The physicalstructure of a product and the interrelations amongits parts are therefore the result of an interplayamong customer requirements, technical constraints,and technical options during the design process.There seems to be a similar phenomenon in services,in which some parts of the operation are more rel-evant to the concept than others — a supermarket,for example, may be less dependent on its phone sys-tem than a pizza delivery service. Also, the outcomeof the service design activity is constrained bychoices made along the design process, depending onhow early a decision is made concerning parts of theservice operation, and how relevant those parts areto the service concept.

Anderson and Tushman (1990)suggest that the core tech-nology of a product or process‘evolves through long periodsof incremental change punctuated by technologicaldiscontinuities.’ Technological discontinuity is aninnovation that either advances the price–perform-ance frontier of the product by an order of magni-tude, or changes the traditional process of makingthat product in such a way that it improves its cost orquality by an order of magnitude. They characterizediscontinuities as either competence enhancing orcompetence destroying. While competence-enhanc-ing discontinuities promote small improvements inperformance, competence-destroying discontinuitiesallow order-of-magnitude improvements in perform-ance and make obsolete the older process or tech-nology.

Tracing a parallel to the service industry, the creationof discount brokerage of securities can be viewed asa technological discontinuity since it changed the tra-ditional process of brokering securities, whileadvancing the price–performance of that service.Interestingly, this innovation was achieved by offer-ing a narrower mix of services (i.e., no financialadvice or research), and often a wider mix of pro-ducts (i.e., broader choice of investment vehicles).The creation of category killers and wholesaler clubsin the retail industry are other similar examples.

Henderson and Clark (1990) provide a link betweenClark’s hierarchy of design and Anderson and Tush-man’s technological discontinuities. They argue thatproduct design requires two types of technicalknowledge: component knowledge (knowledgeabout the concept and functions of each component)and architectural knowledge (knowledge about howto integrate the components into a consistent whole).


The authors propose the term ‘architectural inno-vation’ to refer to those ‘innovations that change theway in which components of a product are linkedtogether, while leaving the core design concepts (andthus basic knowledge underlying the components)untouched.’ When a major new feature is added tothe dominant design of a product, both componentand architectural knowledge about this new featureare integrated into the underlying knowledge of thedominant design of the product. Take, for example,the introduction of airbags as standard equipment inmost passenger cars. It requires knowledge not onlyabout the equipment itself, but also about how itaffects other elements: steering wheel, dash board,fixtures for collision sensors, etc.

Services also provide good examples of architecturalinnovation. Consider, for example, a gym thatdecides to add a major piece of equipment to itsfacility (e.g., sauna, or swimming pool). By doing so,the gym has slightly changed its concept because it

now offers the capabilities thatthe new equipment provide.Moreover, the new equipmentbrings into the gym newknowledge about its operation(how it works), and how it

interacts with the rest of the gym through, forexample, vibration, heat exchange, space and powerrequirements. Another interesting example is pro-vided by stores that begin to use an Internet link toprovide information and order entry interfaces. Thisinnovation is mostly on the interface between twocomponents of the system: the store and the cus-tomer. Notice that the store has neither changed itsappearance or inventory nor has the customer beenrequired to understand a whole new concept. All thatchanged was the link between the two. Exploring thisexample further, a completely virtual store does notneed a physical facility, nor inventory on hand sinceit can ship orders directly from the vendors. This newconcept eliminates the front-office and also changesthe way customers interact with the ‘store.’ It is bothan architectural innovation and a component inno-vation.

The example above suggests architectural variationsoften lead to new concepts, that is, identifyingcomponents and analyzing links between them canalso be a source of innovation.

Development Process

The development of a new product or service gener-ally starts with a coarse, information-poor format(e.g., an idea) and gradually evolves to a detailed,information-rich format (e.g., charts, blueprints). Ateach stage of the process, knowledge is added to thedesign with the help of tools, methods, models, andarchitectural knowledge. Starting from the identifi-


cation of a need or opportunity, designers seek tounderstand desirable attributes of the prospectiveproduct and generate an initial concept. Those attri-butes are translated into component attributes, sothat components can be separately designed and laterintegrated. Such translation, however, requires thatcomponents be known and their interaction be takeninto account. When a product or service is the firstof its kind, no previous architectural knowledgeexists to help its design. Thus, an initial architectureis proposed and evolves over time until a dominantdesign emerges. At this point, the architecturalknowledge becomes part of the methodology todevelop other similar products or services.

Due to physical and technological limitations, not allsets of attributes are feasible, so that tradeoffs ariseamong the attributes. Designers need therefore toassess such tradeoffs to ensure the feasibility of adesign solution and to make design refinements.They do it by using analytical or physical models thatroughly replicate relevant behaviors of the systemthat is being designed. Models therefore have a piv-otal role in the design quality: while complexbehaviors in general require complex models fortheir reproduction, too complex a model may becumbersome to use, and too simple a model maymiss important behaviors. Designers must find a bal-ance between model complexity and accuracy thatfits their needs.

To illustrate the modeling process in the context ofservices, consider a simplified task of designing thefront office of a bank. In deciding how much spaceto allocate to the waiting line for tellers, a tradeoffbetween two design variables arises (line space vsnumber of tellers). Fewer tellers imply less laborcosts, but will generate longer waiting lines thatwould require more costly floor space. On the otherhand, allocating a small area to the waiting line usesless floor space, but will require that more tellers beavailable, implying higher labor costs. To assess sucha tradeoff we may use queueing theory to link thenumber of tellers with the number of customers inthe waiting line, and then use estimates of labor costsand floor space to make the decision. Note that thescope of this abstraction need not be limited to theperspective provided by queueing theory. Allocatinga small area to the waiting line assumes that waitingtimes will be shorter, which is related to customersatisfaction. Thus, we could for example, link waitingtimes and customer satisfaction through anothermodel and assess the tradeoff between operationalcosts and customer satisfaction.

One of the difficulties in building models that sup-port product or service design is that humanbehavior is difficult to predict or to model. Neverthe-less the authors believe that human participantsshare, at least partially, common behaviors that couldbe described by general principles and appropriatemodels, which we introduce and discuss later in this


paper. In the next sections we present three differenttypes of knowledge used in the development process(procedural plan and methods; architectural knowl-edge; and principles and models).

Procedural Plan and Methods

Most product and service development method-ologies proposed in the literature are presented in theform of a sequence of steps similar to those proposedby Urban and Hauser (1980) and the consulting com-pany Booz-Allen and Hamilton (1982). There seemsto be a consensus around the general sequence ofstages that development projects evolve through (seethe six stages shown in Figure 1).

Each stage represents a group of activities with acommon objective as indicated by the label of thestage. Since the output of a stage is also an input forthe next one, there is a degree of precedence betweenstages that imposes a time sequence. The processhowever moves forward and backward as solutionsare proposed, tested and refined. We also indicate inFigure 1 the steps that define the architecture, andthe central role of methods and models in the devel-opment process. Next we describe in more detaileach stage and the methods available to performthem.

Strategic Assessment

The development of new products or services is stra-tegic because of its lasting impact on the firm’sprofitability and growth. Innovative companies use itas a strategic vehicle to expand into new geographicregions, market segments, or to develop new capa-bilities. It is natural therefore for the development ofa new product or service to start with a strategicassessment. Hax and Majluf (1991) propose a stra-tegic planning framework with four phases that wediscuss next: definition of mission; external analysis;internal analysis; and strategic analysis. Figure 2shows an overview of the process.

The mission of a development project is generallydetermined by the firm’s long-term strategy. It can beanalyzed across two dimensions: degree of conceptinnovation, and degree of market innovation. Itmeans that firms may develop new products and ser-vices for reasons ranging from a cosmetic change toserve a known market to a complete new concept toserve a new market. Each type of project has differentuncertainties and impose different demands on acompany’s capabilities, logistical system, organiza-tional structure, marketing, and so on, that must betaken into account in the development project.

The external analysis identifies trends, threats andopportunities in the industry by gathering data about


Figure 1 General Sequence of Stages of Development Projects.

Figure 2 Activities of Strategic Assessment.

suppliers, customers, competitors, substitutes, regu-lations and other relevant aspects. In the case of amajor innovation, an exploratory market researchshould help the company appraise the potentialdemand for the new product or service (see Kinnearand Taylor, 1991 for a thorough exposition on mar-ket research).

The internal analysis determines how the new pro-duct or service fits in the firm’s current offering andhow it will impact the firm’s operations. It must alsodetermine what skills, resources, and capabilities areneeded in the development effort.

Finally, the strategic analysis weights the informationgenerated by the internal and external analysis togenerate marketing requirements (e.g., businessopportunities, market segment and positioning); stra-tegic requirements (e.g., use of a certain technology,expansion into new markets); and regulatory require-ments (e.g., health, safety requirements) for the newproduct or service.

Concept Development

Once the company determines the marketing require-ments (e.g., geographic expansion, target segments)for the new product or service, they must collect,through appropriate methods and tools, the voice ofthe customer. Thus, concept development starts withthe identification of customer requirements (whatcustomers need and expect to find in the product).They are then combined with other requirements


identified in the strategic assessment stage (e.g.,safety, regulation) and translated into attributes forthe new product or service. Burchill and Shen (1992)propose a methodology they call ‘Concept Engineer-ing’ to translate the voice of the customer into newconcepts. We present a similar sequence of steps inFigure 3, calling attention also to other requirementsthat must be considered to generate a concept.

In the case of services, concept development requiresspecial attention to intangibles such as ambience orsocial needs, that may be part of the service conceptand therefore essential to shape the service experi-ence. In the design of a hair salon, for example, acustomer opinion such as ‘I like to go to a salonwhere I can talk to other people,’ may be translatedinto a requirement for the salon, that later becomesan attribute of its concept. In this example, an attri-bute for the salon could be that ‘it must allow easysocial interaction.’ In the system design stage, thatsalon attribute may be translated into attributes forparticipants of the service delivery such as employeesand the physical layout of the store. Possible trans-lations might be: ‘Employees must be friendly andreceptive to conversation’; and ‘the store layout mustencourage interactions among customers.’ Needlessto say, one customer requirement may unfold intoseveral attributes for a number of participants of theservice system.

As we can see in Figure 3, the process starts with thecollection of the voice of the customer. Griffin andHauser (1993), thoroughly review methods availablefor that purpose. They compare techniques to ident-


Figure 3 Activities of Concept Development.

ify customer needs (e.g., individual interviews, focusgroups); techniques to structure or cluster customerneeds; and techniques to estimate or measure needimportance. They also suggest that the discussionwith customers should identify basic needs (whatcustomers expect from the product), articulatedneeds (what the customers would like to see in theproduct), and exciting needs (needs that if fulfilledwill delight the customer). Burchill and Shen (1992)identify customer needs by clustering similar labelscontaining customers’ verbatim opinions. They usequestionnaires developed by Kano (1982) to classifyneeds in five categories: must-be (minimumrequirement); attractive; one-dimensional (the moreof the attribute the better); indifferent; and reverse(the more of the attribute the worse).

Andersen (1983) argues that a description of needsshould distinguish between rational and emotionalneeds. It should also include a time dimension, thatis, a description of needs in terms of existing andfuture (or latent) needs. Future or latent needs, how-ever, are not easy to identify because customers aregenerally not aware of them yet. One possible wayto identify them would be, for example, to monitorenvironmental trends (e.g., raising energy prices).von Hippel (1986) proposes that those needs areexperienced by what he calls ‘lead users.’ The termwas introduced to refer to ‘users whose currentstrong needs will become general in a marketplacemonths or years in the future.’ Since lead usersexperience needs ahead of the general market, theycan also serve as a ‘need-forecasting laboratory.’ Healso discusses how to identify lead users and toincorporate their input into marketing researchanalyses.

Note that design requirements may have differentnatures, such as physical (e.g., colorful, portable),functional (e.g., must provide food, laundrycapability), intangible (e.g., degree of silence), oper-ational (e.g., 24-hour operation, parking facility) andso on. Since some requirements are more critical thanothers, it is necessary to rank them and focus the con-cept generation on those that are most important.Conjoint analysis and utility functions are commontools employed for this purpose (refer to Shocker andSrinivasan, 1979 or Hauser and Urban, 1979 for goodreviews). Wind et al. (1989), for example, describe indetail an application of conjoint analysis togetherwith statistical and other marketing tools to thedesign of a new hotel concept.


Once design requirements are identified and ranked,a rough sketch of the attributes, functions, productsand services that may meet them is generated. Thisis the concept generation step, when designers startto consider what attributes and functions of thephysical system are likely to fulfil the design require-ments. Suppose, for example, that in the hair saloncase the following requirements are identified: ‘itmust allow easy social interaction;’ ‘it must provideinnovative hair styles:’ and ‘it must convey an imageof a futuristic place.’ Those requirements may betranslated into attributes for the whole system orparts of it, thereby beginning to shape the architec-ture of the service system. For example, the require-ments might be translated into the following attri-butes: layout resembling a living-room; brightlycolored; background music; neon decorations; state-of-the-art seats; bar; video-games; CAD system forhair design; manicure; make-up; etc. Depending onthe emphasis put on each of these attributes, andtheir ranking, different concepts may be generated.Some screening procedure must therefore beemployed to separate the most promising conceptsfor further development.

Ulrich and Eppinger (1995) defines concept selectionas ‘the process of evaluating concepts with respectto customer needs and other criteria, comparing therelative strengths and weaknesses of the concepts,and selecting one or more concepts for further inves-tigation or development.’ They list several methodsfor concept selection, and also propose a two-stagemethodology for it. Their methodology starts withconcept screening, using a method developed byPugh (1990) that enables combination of concepts toimprove the solution. After concept screening, thebest concepts are given scores based on the assess-ment of how well they meet their design attributes,weighted by the relative importance of the attributes.The one or two concepts with the best scores proceedto the next stage.

System Design

During the concept development stage, the designrequirements identified in the strategic assessmentare translated into a concept described by attributesand functions that are likely to fulfil the require-ments. In this stage, the concept is translated intoattributes and functions for each component of thesystem, so that they can be designed separately for


later integration. Thus, the purpose of system designis basically to find possible ways (or processes) toimplement the concept with the components determ-ined from architectural knowledge. It comprises notonly the determination of what components areneeded but also the design of the processes throughwhich the components will interact.

All attributes and functions are analyzed individu-ally to determine how they can be implemented bycomponents of the system. The House of Quality(Hauser and Clausing, 1988) or Quality FunctionDeployment (QFD) helps this process by mappingthe relationships between design attributes andcomponent attributes in a ‘relationship matrix’ (seethe step ‘Parts Deployment’ in Hauser and Clausing,1988). Behara and Chase (1981) proposes the use ofthe House of Quality for service design, or ‘ServiceQuality Deployment.’ They call it ‘House of Service,’and refer to service components as ‘the service com-pany facets,’ which are divided into three major cat-egories: Planning, Procedures, and Personnel. Noticethat architectural knowledge is crucial in the appli-cation of this tool, for the design components mustbe known in advance to map product attributes intocomponent attributes.

Pimmler and Eppinger (1994) suggest a methodologyto determine the architecture of a system by studyingexpected interactions among its parts. In their words:‘Selecting the proper architecture of the product is anextremely influential decision which must be madeduring the concept development and system-leveldesign phases of the project; the architecture definesthe sub-systems upon which the team will work forthe bulk of the development effort.’ They argue thatthe choice of product architecture not only affectsproduct performance but also the firm’s developmentcapability and manufacturing expertise. Rechtin(1991) explains that the architecture of a system isdetermined through two opposite processes: par-titioning and aggregating. While partitioningimposes structure to the design problem by breakingit down into smaller subproblems, too many par-titions make the design intractable. On the otherhand, while aggregating similar functions and fea-tures helps simplify the design, too much aggre-gation may cause important details to be left out ofthe design. The author suggests that aggregating andpartitioning functions often suggests a structure orarchitecture of the system itself. Thus, another wayto decompose the design problem is through its‘function structure,’ that is, to determine what means(or sub-functions) are necessary to implement a goalor function (see also Pahl and Beitz, 1984).

During the concept generation stage, attributes andspecifications can be treated in a fairly loose waybecause the whole concept is roughly defined. How-ever, as the design proceeds, they are successivelyrefined to a level considered proper for implemen-tation. If the specifications are too strict, the cost of


the product will most likely increase unnecessarily.On the other hand, if the specifications are too loose,the integration of components will likely presentproblems which may lead to redesigns and delays.That is because every physical system has limitations,and the design implementation can rarely meet allspecifications without tradeoffs. Ulrich and Eppinger(1995) suggest the use of competitive benchmarkingwhenever possible to determine target specificationsfor a product. This is generally the case of services,where competition through imitation is fairly com-mon.

Specifications are essential to system design becausethey allow components to be designed separately forlater integration into a consistent whole. Integratingcomponents, however, requires special attention toeffects resulting from expected or incidental interac-tions. Pahl and Beitz (1984) argue that interactionsamong the components of a physical system happenthrough flows of energy (e.g., mechanical, thermal,electrical), materials (e.g., gas, liquid, dust), andinformation signals (e.g., data, control pulses). Inter-estingly, in service systems information may beexchanged through a multitude of ways: ambience,emotion, noise and smells are just a few examples.To further illustrate, consider a hotel in which a con-ference room is placed close to the kitchen to enablethe hotel to cater for executive dinners and recep-tions. In this case, the interaction between the kitchenand the conference room is expected and designed toallow, for example, easy flow of waiters and food. If,however, this interaction is not well understood, itmay yield a poor design due to incidental, unantici-pated interactions — noise from the kitchen may dis-turb people at the conference room.

The example above shows that although the level ofdetail at this stage is still coarse, designers need tostart analyzing interactions among components tominimize integration problems later in the process.The analysis is supported by simple models that areemployed to roughly check the feasibility of thedesign. Figure 4 shows a schematic of system design.

Component Design

Similarly to the previous step, where system attri-butes and specifications are mapped into componentattributes and specifications, component design isalso an iterative process of refining attributes andspecifications. The design of each component pro-

Figure 4 Activities of System Design.


ceeds by successively decomposing it into sub-sys-tems (as long as it is necessary), generating specifi-cations for the sub-system, checking their feasibility,and then designing each sub-system in the same way.Figure 5 shows an overview of the process of compo-nent design.

Three difficulties are commonly faced by designersin this step. The first is to manage the sequence ofdesign decisions taking into account interdepen-dencies among different components. The second isto translate non-numerical specifications into thephysical design of each component. Finally, the thirddifficulty is to assess tradeoffs among non-numeri-cal specifications.

The first difficulty arises because the design of acomponent generally constrains the design of others,imposing a precedence relationship among compo-nents. Since not all components can be designed con-currently or sequentially, designers need to deter-mine, up to a certain level, dependencies amongthem and integrate their design accordingly. Sincesome design decisions are more restrictive thanothers, and may make unfeasible possible solutionsto the problem, they should be taken as late in theprocess as possible to keep the design flexible tochange. The notion of architectural knowledge dis-cussed in the fifth section may help the assessmentof component interdependencies by providing aframework to investigate the impact of the design ofone component on the others.

The second difficulty arises from the existence ofnon-numerical attributes. In this case designers needto attach a metric to each attribute so that they canbe measured and transformed into a specification.Ideally such metric should define the attribute asclearly and objectively as possible. This however isnot always possible. For example, if ‘having gooddiction’ is an attribute for a phone operator, a metricthat qualifies diction into degrees such as ‘poor,’‘fair,’ or ‘good’ transforms ‘good diction’ into a speci-fication for the operator. The assessment howevermay be subjective and dependent on the person wholistens to the operator.

Finally, non-numerical specifications make theassessment of tradeoffs between design variablesmore difficult. Although tradeoffs among measurablevariables are traditionally assessed with the help ofanalytical models that establish a relationship

Figure 5 Activities of Component Design.


between them (refer to our discussion about the useof principles and models in the sixth section of thispaper), those relationships can also be described bycausal-loop diagrams (see, for example, Senge, 1990).The difficulty involved in this assessment lies inunderstanding the nature of each non-numericalattribute and how they impact each other.

Concept Testing and Implementation

The output of the previous stage is a set of docu-ments and blueprints that describe in detail everyelement of the product or service so that a productionline or service system can be set up and begin oper-ation. In both cases, however, this description maybe inaccurate due to the impossibility of models toreproduce reality. In service systems, for example,the description of human elements requires not onlya description of their physical appearance (uniforms,name tags, etc.), but also a description of expectedbehaviors and skills. Since human behavior is notfully understood, actual behavior in the system maybe quite different from what is expected.

Due to the possibility that the product or service doesnot meet its design requirements, before a companycommits resources for the production ramp-up orservice implementation, the design is tested to makesure that the outcome of the development effortyields a product or service with the attributes andfunctions as originally planned. This test involvesbuilding prototypes (for tangible elements) or pilotoperations (for processes); measuring their perform-ance; comparing with expected performance; andrefining the design so that the product or serviceexhibits the expected behavior.

There are established methodologies available forprototyping and market testing. Ulrich and Eppinger(1995), for example, discuss types and objectives ofprototypes, and propose a plan for prototyping.Wheelwright and Clark (1992) advocate that to beeffective, prototyping and market testing must bedone under market conditions as close as possible tothe ones that the new product or service will faceafter its launching. This seems to be speciallyimportant when the main benefit of the new conceptis intangible. In this case customers must be able toexperience the product or service to understand theconcept. That is the reason why some service compa-nies (e.g., hotel chains) implement a pilot operationbefore full-scale launching. Also, some companiesallow customers to experience their products for acertain period without obligation to buy. Under-standing that Murphy and Robison (1981) definedmarket testing for a new service as a technique‘designed to evaluate whether a prospective user (1)understands the idea of the proposed service; (2)reacts favorably to it, and (3) feels it offers benefitsthat answer unmet needs.’


Once the concept is tested in the market and itsdesign is refined, the company must start planning alaunching strategy for the new product or service. Atthis stage equipment is purchased and installed,physical facilities are built, decorated, furnished, andemployees are recruited and trained. Recruiting andtraining must provide employees with the necessaryskills to perform their functions satisfactorily. In ser-vices that require the participation of customers,employees must be trained not only to perform theirfunctions, but also to induce customers to behave indesired manners. Note that human behavior is gener-ally a consequence of not only training, but also thephysical setting, reward systems and operating pro-cedures. Although reward criteria and operating pro-cedures can be taught through training and docu-mentation, employee behavior is shaped over time bythe organization’s culture and internal marketing.

Figure 6 shows concept testing and implementation,with the activities of the former arranged in a cycleto indicate that they must be done interactively.

Feedback and Learning

The adoption of a development methodology facili-tates learning because it enables the systematizationof knowledge: modifications in the procedures andmethods used during the project may be documentedand incorporated into the methodology. The samehappens with new methods and models that mustalso be formally included in the methodology forfuture use.

Although learning at the individual level starts thefirst day of the development project and continuesthroughout the product or service life cycle, it is com-mon to dedicate a separate stage for the purpose ofsystematic learning. A company that does not use theknowledge generated in the development of a newproduct forfeits most benefits that a methodologybrings. Wheelwright and Clark (1992) argue thatlearning for product development goes beyondchanging procedures, methods, or models, andrequires careful systematic effort. Individuals thatparticipate in a development project naturallyacquire new knowledge, but learning must also beextended to the organization. When a complex sys-tem is being designed, a number of individuals with

Figure 6 Activities of Service Implementation


diverse expertise interact to generate the design.Learning in this environment therefore requires acoordinated effort because no single individual hasthe knowledge needed to solve design problemsalone. The authors list crucial items for a successfulsystematic learning that are similar to the three typesof knowledge discussed in this paper: team processes(sharing mental models), a model of the developmentprocess (or a procedural plan), data analyses, pat-terns, and root causes (architectural knowledge, prin-ciples and models). They report that changing thedevelopment procedure, and adapting tools andmethods is a possible way to capture insights andlearning generated by the project. Also, using a ‘pro-ject audit — a formal review conducted by a cross-functional team’ appears to be of great help toorganize the learning process.

The daily operation of the production or service sys-tem also generates learning that must be used inimprovements and to develop future line extensionsor similar services. Once the new product or serviceis introduced, customer suggestions and complaintsbecome a source of valuable information that mustbe used as feedback or generate new insights. Whilemost products carry manufacturing warranties thatare generally claimed if the product is defective, ser-vice companies have a hard time to collect feedbackfrom customers because only a few (the most dissatis-fied or satisfied) ever return a feedback form (seeHart, 1988). The author advocates that serviceguarantees lead to a better collection of data on ser-vice failures, resulting in more opportunities forimprovement. The service guarantee, however, mustbe designed so that it is easy to claim and promotesimprovement in the system.

Architectural Knowledge

The notion of architecture is based on the assumptionthat a system can be divided into subsystems orcomponents which interact to achieve a certain pur-pose. From the perspective of product or servicedevelopment, establishing an architecture improvesthe development process in at least three ways. First,it forces a systematic search of alternatives in the sys-tem-design stage of the development effort. Second,it facilitates the design task by partitioning it into


smaller subproblems (the design of each componentseparately) for later integration. Finally, it requiresthat interactions among components be identifiedand analyzed, which by itself may lead to a deeperanalysis and better results.

Given a concept, there may be several different archi-tectures that map design attributes into functions andattributes of its components. This mapping is done atthe system design stage with the help of architecturalknowledge, that is, knowledge about what compo-nents integrate the system, and how they interlink.Some architectures however will likely emphasizesome attributes in detriment to others. The attributeranking done in the concept development stage mayhelp determine which architecture best fits the con-cept. Similarly, there may be several differentimplementations of the same architecture. Prototyp-ing and market testing support the search for the bestimplementation, given the input of customers.

To illustrate, let’s consider a simplified case of a res-taurant concept defined by the following attributes:exotic ambience; entertaining experience; traditionalsteak-house menu. There are several different archi-tectures that map those attributes into a restaurant.We roughly describe three of them here. One wouldbe a restaurant decorated in Middle Eastern style andbelly dancers. Another would be a hibashi-type res-taurant in which the food is prepared in front of thepatrons with a hot plate in each table. A third poss-ible architecture could be a similar restaurant butwith a single central hot plate surrounded by a barcounter where patrons would sit. Note that the tra-ditional steak-house menu somehow influences thearchitecture in the sense that it makes possible thesecond and third architectures — if the menu wasmainly composed of baked food, the second andthird architectures would be less attractive in thesense that it is not entertaining to watch food bakinginside an oven. There are endless possibilities for theimplementation of any of those architectures. Someimplementations, however, fit better the concept thanothers. For example, hibashi-type restaurants are tra-ditionally decorated with wood shaped with orientalmotives — any implementation too different fromthat might not be approved by prospective cus-tomers.

In order to discuss architectural knowledge in thecontext of services, we use an example of a serviceoperation structure with nine high-level componentsthat are pervasive to most service operations and thatmay be designed and implemented in a new service.They are: internal organization; external organiza-tion; technology (soft or hard); customers; front-lineemployees; support employees; product mix; servicemix; and customer interface, as shown in Figure 7.We use circles to depict each component, and inter-sect circles to illustrate cases of participants withcharacteristics of more than one component. Forexample, we define support employees as those who


Figure 7 Components of a Service System Structure.

mainly do not have any kind of interaction with cus-tomers, and front-line employees as those who do.Still, support employees may eventually have contactwith customers in some service situations, and front-line employees may eventually have back-office func-tions.

The physical structure of service organizations iscomposed of the internal organization, the externalorganization, and technology. The internal organiza-tion is the part of a service operation that is out ofsight or reach of the customer. Although similar tothe notion of ‘Back Office,’ the term ‘internal organi-zation’ seems more appropriate in the sense that itmay refer to several locations, physical facilities, andequipment utilized for service support and pro-duction (out of the sight of customers). Similarly, theexternal organization is the part of a service oper-ation that can be reached or seen by customers. Itincludes facilities, vehicles, uniforms, and externalsigns, among other elements. Common examples arefront-offices of banks or retail stores, parking lots,cafeterias, elevators, rooms or public areas of hotels,resorts, and hospitals. Technology participates in thephysical structure as equipment or ‘hard technology,’but it can also be viewed in a broader context asoperating procedures, flows of activities, or infor-mation systems, often referred to as ‘soft technology.’Thus, we define it as a distinct component that inter-sects both the internal and external organizations.

People are represented in a service system by the cus-tomers, front-line employees, and supportemployees. This two-type classification of employeesdoes not consider function or hierarchical level, onlythe degree of contact with customers: ‘Front-lineemployees’ are those who interact directly with cus-tomers, while ‘support employees’ are those whocontribute to the production of service without hav-ing direct contact with customers. Under this classi-fication, managers who do not interact with cus-tomers are also viewed as supporting employees. It

Although it is impossible

to control individual actions,

people can be induced to

behave in desirable ways by

artifices


emphasizes the importance of having the right set ofskills at all positions, specially those that interactdirectly with customers.

From a design perspective, we consider people ascontrollable (to some extent) participants in the sys-tem. In fact, training and selection allow service com-panies to shape the skills and personal characteristicsof employees, and market positioning allows compa-nies to target segments in which customers exhibitdesired behaviors. As we will see, although it isimpossible to control individual actions, people canbe induced to behave in desirable ways by artificeslike, for example, environmental cues and tech-nology.

In the product and service offering we have a mix oftangibles and intangibles that are ‘sold’ to customers.A hotel, for example, may have the products and ser-vices traditionally offered by most hotels (room ser-vice, laundry, restaurant, etc.), but can also offer tem-porary offices, conference rooms, health clubs, and soon. Observe that the mix of tangibles and intangiblesintersect to indicate those cases in which the distinc-tion between them is blurry.For example, in the case ofcable TV, while the cable signalis intangible and the decoderused for reception is tangible,one has no use without theother.

Finally, the interface is the setof all possible interactions (anycontact through any medium)between customers and the ser-vice organization. It is perhaps the most complexcomponent to design because successful service inter-actions are often the result of an interplay amongvarious components. That is the reason why wedepict it in the center of Figure 7, surrounded by allthe other components.

It is important to note that just like the case of pro-ducts, service quality is only achieved when allcomponents work well and are properly coordinated.For example, no matter how politely an employeeapologizes for service failures at the back office, theservice experience may still be not as satisfactory asit could be. Although there are tactics for servicerecovery (see Hart et al., 1990), it should be clear thatthe role of service design is to carefully considerevery possible detail in the new service so that fail-ures are not caused endogenously.

In the sequel we further describe the components,clarifying their role in a service system and how theyinteract with each other. We point out issues thatmight be considered in their design as well as theirfunctions and attributes.


Internal Organization

The internal organization is the part of a service sys-tem that is out of sight of customers and competitors.It may be distributed in several locations (back-offices) or may have dedicated facilities for R & D,data processing, production, and other functions. Themain consideration in its design is to provide servicequality with as much efficiency at cost as low as poss-ible.

Thompson (1967) proposes that organizations shouldbe able to operate more efficiently by isolating theirtechnical core from external disturbances. He refersto minimizing customer contact as ‘decoupling.’Chase and Tansik (1980) elaborate on this notion,arguing that since the customers are a source ofuncertainty, the less contact customers have with theservice system, the more possible it should be tooperate efficiently. The fundamental assumptionadvocated by the Contact Model is that customersare a source of uncertainty that should preferably bedecoupled from the technical core for the sake ofefficiency. Decoupling in this context, however,means reducing the physical presence of customersin the internal organization without necessarily limit-

ing the exchange of informationbetween customers and the ser-vice organization. Therefore,we define the internal organiza-tion as being decoupled fromthe external environment byvisual contact only, and inter-acting with customers throughany media other than personalcontact.

Although economies of scale are commonly regardedas not very significant in services, the design of theinternal organization must consider opportunities forsuch. Several firms have managed to achieve econ-omies of scale through the use of technology; by cen-tralizing back offices and decentralizing front offices;or by centralizing activities that require expensivepersonnel or equipment and decentralizing othersthat do not (refer to Quinn and Gagnon, 1986; Upah,1980; and Porter, 1980, respectively).

Like any other service component, the internalorganization must be designed based on its desiredattributes and functions. The voice of the customerin this case is mainly given by the customers of theinternal organization, that is, operations personnel(for facilities and equipment) and supportemployees, that may help determine the attributes ofthe internal organization. Those may include accessi-bility, illumination, ventilation, location of restroom,laundry facility for employees, directional or infor-mational signs (reminding employees of messages),among others.


External Organization

The external organization is the part of the structurethat is visible to customers and competitors. It is simi-lar to the ‘front-office’ found in the literature, butwith a broader definition. The external organizationis composed by facilities and physical evidence usedfor service delivery. It may have several front-offices,or use intermediaries to perform some function,depending on the nature of the service business andits distribution channels. We discuss three aspects ofthe external organization: exterior design, interiordesign, and design of physical evidence.

Ward et al. (1992) suggest that customers start to formexpectations about the service even before enteringthe facilities. They argue that customers tend to com-pare the external appearance of a service businesswith their mental models or ‘prototypes’ of what theyexpect that service facility to be. Upon entering theplace, customers compare the interior appearancewith their prototype and decide whether to stay ornot. This mental process of categorizing the servicesetting prompts expectations and inferences aboutthe type of service, quality, price, and behaviors. Theauthors conclude that more ‘typical’ outlets seem tobe preferred by customers, and that the importanceof exterior typicality or interior typicality was likelyto depend on the service category — they are per-ceived by customers to be part of the dominantdesign of that category of services.

Upon entering a service facility for the first time, cus-tomers try to match that service business to a knowncategory, and adopt certain behaviors which arebased on how they perceive their roles in the servicesystem (Solomon et al., 1985). When customers do notknow their role in the system, they rely on other cus-tomers or on environmental cues to understand whatbehaviors are proper to that situation. In this case,the service setting and environmental cues can havegreat impact on the ability of customers to achievetheir objectives. The floorplan, layout of equipment,and signage can help or hinder employees and cus-tomers from performing their tasks (Kelly et al., 1990).

The service setting and environmental cues can alsoinfluence behavior in subtle ways. For example, thefurnishings of some fast food restaurants are pur-posefully a little uncomfortable to discourage cus-tomers from lingering. On the other hand, certaincafes try to provide an environment as comfortableas possible to their customers, encouraging them tostay as long as they please — it is just part of theirservice concept. Milliman (1982) reports that thetempo of background music in supermarkets and res-taurants can affect traffic flows and revenues. Bitner(1992) argues that the service setting may encourageor block certain forms of social interactions amongcustomers or between customers and employees. Itmay even encourage more consumption of the firm’sservice offering. An interesting example are res-


taurants that perform theatrical cooking in front ofpatrons. Since unknown customers usually sittogether at the same table, the restaurant keeps allpatrons of a table occupied all the time so that theydo not feel embarrassed to share the table with some-one they don’t know. Also, a bar in the receptionallows patrons to consume drinks before beingdirected to their dinner table.

When service companies try to communicate properbehaviors to customers, they do it through tech-niques known in the literature as ‘organizational soci-alization’ (Kelly et al., 1990). Organizational socializ-ation is achieved through a number of means such asinformation signs, recorded messages, environmentalcues, literature, differentiated fees, or explicitly train-ing customers and employees. It is used to shape cus-tomers’ expectations, communicate organizationalvalues, teach how to best use facilities, guide queue-ing behavior, transmit reminders, control demand,etc. — the external organization may implement anyof those functions. It has such a strong influence oncustomers’ behavior and perception of the servicequality that special attention must be given to itsdesign. Recognizing its importance, Shostack (1985)states that the service setting should be deliberatelydesigned to shape the ‘service reality’ rather than‘leave it to default.’

The unconscious characterization of service settingsby customers reinforces the need of using their inputto identify attributes of the external organization.Using the appropriate techniques (as discussed inConcept Development), service designers can extractimages and feelings of customers, which are veryimportant in the determination of the attributes of theexternal organization. Moreover, customers may alsoprovide information about operational aspects of theexternal organization such as accessibility, store-hours, layout, etc. From that point onward, special-ists must translate that information into design speci-fications.

Technology (Soft and Hard)

Quinn et al. (1990) define technology as ‘the system-atic application of knowledge — particularly aboutphysical, chemical and systems phenomena — to use-ful purposes.’ Levitt (1972) distinguishes between‘hard’ technology (e.g., hardware, equipment), and‘soft’ technology (e.g., software, operatingprocedures).

Technology may be used to enhance the capabilitiesof both employees and customers. With the help ofwell designed, easy-to-use systems, relatively inex-perienced people can perform sophisticated tasksquickly, and without much training. Employees canhave their scope of action enhanced in capability andflexibility to provide better service or more effectiveselling. It can also induce customer participation in


the service process by enhancing the capability ofcustomers to perform some activities. For example,service concepts that were unfeasible previously (e.g.,home banking) are now a reality thanks to the use ofdata communication technology to link customersand the service organization. On the other hand,technology may also restrain operating proceduresfor the sake of uniformity and standardization, ordue to regulations (e.g., airplane maintenance, bankaccounting). By standardizing equipment andoperating procedures, service organizations havebeen able to replicate the service concept and expandgeographically through franchising, for example.

Chase and Stewart (1994) propose the use of both softand hard technology to prevent service failures fromhappening. This application of technology is knownas poka-yoke and has been widely used in manufac-turing systems. They show practical poka-yokes inservice systems and discuss similarities and differ-ences of their manufacturing counterparts. Theauthors argue that service poka-yokes need often toinduce customers to act in desirable ways, in whichcase they contribute to the process of ‘organizationalsocialization’ discussed in the previous component(External Organization).

Establishing an organizational climate for servicequality may be regarded as soft technology in theform of organizational culture or compensation sys-tem that rewards customer-driven behavior by ser-vice employees. Such behavior is often perceived bythe customers as service quality. It is important tounderstand, however, that when customers oremployees are required to operate a piece of equip-ment or abide by procedural changes, the implemen-tation of such innovation is dependent on its accept-ance by them. Simple and repetitive tasks can beshifted to machines, resulting in cost savings, pro-ductivity and wage gains, as well as improved qual-ity, and more fulfilling jobs.

Quinn and Paquette (1990) argue that in the sameway that technology allows standardization, it can beused as entry barrier for competitive imitation. Thisis achieved by having customers using technologyprovided by the service organization, which mayincrease switching costs and lock customers into arelationship with the company (a common exampleis that of customer companies that link their infor-mation systems with the service provider’s systemfor order entry, inventory status, etc.).

Bharadwaj et al. (1993) suggest that technology maycreate opportunities to exploit demand synergy bycross-selling services and cost synergy by centraliz-ing administrative activities. A good example is thefinancial services industry, where companies useinformation technologies to distribute a variety ofservices to a diffused customer base.

As we can see, technology is used throughout a ser-


vice organization, serving a number of functions andpurposes. When designing this component therefore,it is necessary to determine what functions it willimplement and what should be the correspondingattributes. In the design of ATMs, for example, it isnecessary not only to determine the breadth anddepth of functions they will allow customers toexecute, but also attributes such as the layout, easi-ness of operation, clarity of instructions, and so forth.In the case of soft technology such as a process oroperating procedure, many activities are performedsimultaneously by employees, customers, and equip-ment. It is important therefore to determine what willbe performed, where, how, and by whom, and howinteractions among components affect the design ofother components, so that the process can beexecuted smoothly.

Support Employees

Support employees are those who do not interactwith customers but are necessary to implement coreor peripheral functions. Although their impact onservice quality is less straightforward to perceive,they have a fundamental role in the system. That isbecause customers usually go through routine ser-vice encounters without much awareness of the pro-cess, a state of mind referred to as ‘mindlessness’(Solomon, 1985). Only when something deviatesfrom the script, either positively or negatively, iswhen customers become aware of it. Airlines providea good example: if everything goes perfectly in allsteps of the process — reservations, check-in, inboardservice — but one suitcase strays, that may be thestrongest impression a customer will keep from theservice. If the airline is not able to recover and deliverthe suitcase, the effort of hundreds of otheremployees to assure service quality and customer sat-isfaction will most probably be wasted.

As in any other component, the attributes and func-tions of all support employees are determined at thesystem design stage. Then, the set of tools and skillsneeded to perform those functions is identified. Inthis process, operations personnel and employees canbe of much help to simplify procedures and estimateservice capacities needed to perform well the func-tions allocated to the position. Support employeescan also be helpful in the design of technology, elicit-ing the attributes of tools or equipment that theywill use.

Front-Line Employees

Front-line employees are those who interact directlywith customers, either personally or through anycommunication media. Bowen and Schneider (1985)refer to front-line employees as ‘Boundary-Spanning-Role’ employees because they act as interfacebetween the service organization and its customers.


They not only work with customers to deliver or cre-ate the service, but are also a major factor in shapingthe customer’s perception of service quality — anyslip in the manners of a front-line employee can spoilthe customer’s overall impression about the serviceexperience, even if everything else is perfect. Interest-ingly, the contact with customers is perceived byemployees as enriching their jobs, which leads togreater job satisfaction.

Front-line employees receive immediate feedbackabout the service they provide through a variety ofmeans. Thus, they perceive service quality andorganizational effectiveness similarly to their cus-tomers. When management orientation or a badlydesigned system hinders their ability to perform theirjobs well, they experience stress that generally leadsto frustration, dissatisfaction, and ultimately indiffer-ence or willingness to quit.

As we discuss in the design of the customer interface,each interaction between customers and the serviceorganization has different characteristics and attri-butes. Thus, depending on the interactions that front-line employees participate, they are required to per-form different functions and to have different sets ofskills. If front-line employees interact with customersover the phone, for example, their physical appear-ance is of less importance than good verbal and list-ening skills. On the other hand, when there is per-sonal contact with customers, front-line employeesmust not only have good interpersonal skills but mayalso be ‘packaged’ to transmit or reinforce the com-pany’s image. Blackman (1985) discusses how thepackaging of front-line employees goes beyond thesimple visual consistency. Putting employees in uni-forms helps convey an image that service will be thesimilar every time the customer experiences it. More-over, uniforms is a constant reminder of the com-pany’s expectations about the employee.

The design of a front-line position therefore involvesthe determination of the right set of attributes andskills that the job demands, that is, the appropriate‘packaging,’ and range of functions that theemployee should have to provide good service.

Customers

In many service businesses the customer not onlyparticipates in the service production, but is alsophysically present in the facilities. In this situation,the behavior of a customer can affect the serviceexperience of others. That is the reason why it is soimportant to monitor and manage customer behaviorin the service setting. Noisy guests near your roomin a hotel, or a queue delay in a check-out counterof a supermarket are two common examples thataffect our everyday lives.

The service organization benefits from customer par-


ticipation in a number of ways. It not only improvesthe efficiency of a service operation but may also beperceived by customers as better quality. By shiftingsome of the tasks required in the service productionto customers, service firms reduce the labor required,and consequently reduce their costs. Since customersprovide help just when it is required, the capacity toserve varies with demand, reducing the need for idlecapacity. An interesting example is provided by thefast food technology. Through appropriate signs andthe observation of other people, customers under-stand that they are supposed to dispose of their garb-age and return the tray after the meal.

Involving customers in the service production alsoleads to their acceptance of some responsibility forthe overall service quality. Companies, however,must determine whether customers have the appro-priate knowledge to do what is expected from them,and whether customers are able to perform a task.Companies must also provide customers with toolsand skills necessary to execute the role assigned tothem. It involves training customers to use new tools,and educating them to reduce resistance to change.Mills and Morris (1983) suggest that in services thatrequire relatively complex skills of the customers(e.g., consulting, tax accounting), companies shouldconsider the possibility of educating or training cus-tomers in a pre-service encounter stage. In fact, thisis commonly employed by hospitals and doctors toparents-to-be before the delivery of babies.

Considering customers in the design of service sys-tems is therefore a process of defining the set of func-tions and behaviors expected from them. Naturally,their input is tantamount to define those functionsand behaviors. The set of functions is then used todefine the skills and tools required for successful per-formance. Moreover, the set of expected customerbehaviors can provide information for the design ofenvironmental cues, signs, technology, and objectsthat can be used to induce them.

Customer Interface

Bitran and Lojo (1993) suggest that the customerinterface may be systematically analyzed andimproved. They argue that service encounters hap-pen through one or more phases dedicated to: access,check-in, diagnosis, service delivery, disengagement,and follow-up. Shostack (1985) observes that not onlyservice encounters are manageable and controllable,but also that the best practices place a lot of emphasison controllable details. In other words, outstandingservice firms design quality into their delivery sys-tem and interface. To do so, it is necessary to deter-mine not only the interactions that are necessary forthe operation of the service, but also to anticipateinteractions that may happen in an uncontrolled way.

Chase (1981), for example, analyses interactions in


which the customer is in direct contact with the ser-vice facility. He suggests that direct contact with thecustomer may be reduced or enhanced with advan-tages for both the company and the customer. Heproposes strategies for contact reduction (e.g., byautomating the interface) and for contact improve-ment (e.g., use of signs to indicate queueing patternsand take-a-number systems).

Designing the customer interface therefore consistsin systematically designing all interactions betweencustomers and the service system consistently withthe service concept. Each interaction has attributesdetermined by five dimensions, which we discussnext: purpose; duration and time delay; breadth anddepth of options; nature of contact; and mediaemployed.

Since service interactions are in general ‘purposiveand task oriented’ (Solomon, 1985), there are notmany different objectives or purposes for customersto interact with service organizations. That simplifiesthe design of the interface because there is only a lim-ited set of functions that it may accomplish. Interac-tions with different purposes, however, may havedifferent attributes. Problem-solving interactions, forexample, may have their attributes related to assert-iveness and responsiveness whereas the socializationinteractions may not. We list some possible interac-tion purposes in Figure 8.

The duration and time delays of a service interactionare also important dimensions in determining itsattributes. For example, Lovelock (1984) notes thatthe longer it takes for the service delivery to be com-pleted, the more likely it will be for customers torequire information about the work in progress, suchas completion date or time, costs, etc. Maister (1985)observes that the passage of time in service interac-tions is felt in different ways, depending on whethercustomers are kept entertained, are treated fairly, arekept informed, etc. That is the reason why servicecompanies try to entertain or keep customers busy

Figure 8 Service Interaction Purposes.


while they are waiting. The rule of thumb is to keepcustomers busy (or entertained) while they wait andkeep customers informed when there are delays. It isimportant to notice however that a longer durationof customer involvement is not necessarily bad whenthe customer is properly treated — it may expose thecustomer to more of the company’s service offering,and motivate more sales.

During service interactions menus of options areoften made available explicitly or implicitly to cus-tomers, so that they can choose the course of actionto take. Those options however have an enormousimpact on the interaction. Each option that is madeavailable requires that the service organization havethe adequate capacity to respond properly to thatdemand. By carefully designing the breadth anddepth of options in the menu made available to cus-tomers, the service company is able to control serviceinteractions so that only certain types of demands areimposed to appropriate parts of the organization. Inthis way the service system has to cope with lessuncertainty, and customers also get a better servicebecause each interaction can be more focused to thetype of demand it is designed for. Focusing (orlimiting) the menu of options on service interactionsis widely employed by airlines, supermarkets, hotels,banks, and other service businesses, in the form ofcounters assigned to a narrow set of activities, suchas ticketing, problem solving, express check-outcounters, and so forth.

We must also consider other types of contact beyondperson-to-person interactions: person-to-machine,person-to-object, and person-to-environment. Serviceinteractions occur between customers and machines(e.g., automatic answering systems), between cus-tomers and objects or any physical evidence used bythe service organization (e.g., stickers, catalogs,windows), and between customers and the environ-ment through background music, smells, etc.Although service encounters may take a number ofdifferent forms, they are always experienced through


one or more of the five senses that also shape theattributes of each interaction. To illustrate, considertwo situations of a person-to-machine interactionwith the same purpose of exchanging information(account balances, for example). The customer maycall a number and navigate through an automaticresponse system, or operate an automatic tellermachine. In each case different menus of choices aremade available and different senses are used — eachinteraction has a distinct set of attributes but thesame purpose.

The interface may also have long-term, strategicimportance to the business. It may make switchingcosts high enough to keep customers hooked up tothe service provider’s information systems, and itmay also be used to keep customers aware of a ser-vice even when they do not have to interact with theservice organization for a long period of time (e.g.,major surgery, central heating repair, etc.). Blackman(1985) proposes the use of a series of short interac-tions to shape expectations, educate potential users,or remind customers of previous service experiencesto avoid a situation in which ‘the forgetting curvedooms a business to deal primarily with new cus-tomers.’

Service Offering

Service organizations frequently offer a menu of ser-vices and products, generally called ‘service offering.’Examples are abundant: restaurants offer a set of dis-hes in their menus, hospitals specialized in certainhealth problems, and insurance companies offer poli-cies with different coverage.

Gronroos (1987) argues that the ‘basic service pack-age’ is a combination of core service, facilitating ser-vices, and supporting services. Core service is the pri-mary purpose of the service company, whereassupporting and facilitating services complement thecore service (e.g., reservation services in airlines) ormake it more attractive (e.g., in-flight movies). Simi-larly, a service organization also has core productsas well as supporting and facilitating products. Coreproducts are those related with the primary com-petence of the company whereas supporting and faci-litating products are tangibles that either help theconsumption of the core products and services ormake them more attractive.

When the service offering encompasses a portfolio ofcore products and core services, they may share sup-porting and facilitating products and services to achi-eve economies of scope. For example, a school mayoffer executive education programs together withundergraduate and graduate programs. Each pro-gram may have dedicated supporting and facilitatingservices as well as share them with the other pro-grams.


The objective of explicitly considering the serviceoffering as a component of the service architectureis to identify the core products and services of thecompany as well as the supporting and facilitatingproducts and services that should be offeredtogether. The determination of the offering generatesspecifications for other parts of the organizationbecause each product and service requires that theorganization have adequate capacity to respondproperly to the corresponding demand — when theimpact of adding new products or services to anexisting service offering is overlooked, parts of theorganization may become congested decreasing theoverall service quality.

Since the service offering is crucial to position thecompany vis-a-vis its competition, its depth and bre-adth are typically decided based on marketing andstrategic considerations. For example, companiesmay be able to reap economies of scale or scope byfocusing their product and service offering. In theretail industry category killers offer a wide choice ofproducts within a narrow market segment (e.g., officesupplies, do-it-yourself, etc.). On the other side of thespectrum there are convenience stores that offer alarge number of different products ranging from foodto office supplies, but limited brand choice for eachproduct.

The service offering is an important facet of the ser-vice concept, and as such must have its elementsaligned with the service concept (it would not makemuch sense for example, to design a Chinese res-taurant with a menu of French specialties). Inno-vations in the offering, however, have little patentprotection. In spite of that, service firms can stillimpose barriers to imitation by using resources thatcannot be traded either because their value is not welldefined or because they are endogenous, specific tothe firm (Dierickx and Cool, 1989).

Models and Principles

Modeling is a fundamental part of the designactivity. In manufacturing, engineering models havetraditionally been used to design new products. Sincemodels are abstract representations of the systemthey seek to reproduce, they can be used for experi-mental investigation instead of the real system, atlower cost and less time. By constructing models andanalyzing their behavior, designers can develop abetter understanding of the real system and improvethe design. Models can also help designers to assesstradeoffs among attributes of the system, predictfuture behavior, or understand unexpected interac-tions among components of the system.

In sciences such as Physics and Chemistry, there arefundamental principles that describe phenomena inthe context of those sciences by means of equations.

The forgetting curve

dooms a business to deal

primarily with new

customers


In other sciences like Psychology, principles notstated through mathematical equations can neverthe-less be used as building blocks for research in thefield. No matter the format through which principlesare expressed, they are the theoretical foundationover which models are built. The types of behaviorsthat can be represented by a model, therefore, arelimited by the model’s theoretical perspective. More-over, the quality of the final design depends not onlyon the ability of designers to apply models to evalu-ate design decisions and carry out design refine-ments, but also on the quality of the available models(Hoover et al., 1991).

When developing new models, designers use vari-ables to represent cause and effect factors of real-world phenomena. Since the relationship amongthose variables are in general non-linear and com-plex, designers have to restrict the scope of theirmodels so that they are complex enough to capturethe desired behavior, but also simple enough thatthey do not become too cumbersome. The appropri-ate level of aggregation of a model must also be care-fully considered. For example, it is certainly inappro-priate to consider theindividual behavior of mol-ecules to model the (macro)behavior of a gas in a con-tainer — such a model wouldbe too complex to be of anypractical use. Still, the behaviorof a gas is very predictable, andcan be described by thermo-dynamic models. In the sameline of reasoning, modeling human behavior at theindividual level is sometimes inappropriate fordesign decisions.

Observe that the design activity requires models thatare able to reproduce as well as explain behavior inthe system. The difficulty of building such modelsfor service design is that human behavior is not aspredictable as the behavior of physical objects undernatural laws. Thus, few models and principles havebeen proposed that help assess tradeoffs for servicedesign. Therefore, identifying what principles andmodels are appropriate for service design is animportant first step to make possible their appli-cation. We try to do it in the sequel when we discusssome principles and types of models available in theliterature, that may help service design.

Maister (1985) proposes a set of principles thatdescribe the psychology of waiting. One of the prin-ciples, for example, states that unoccupied waitingfeels longer than otherwise. The passage of timetherefore has a psychological factor attached to it thatmust be considered in the design of services. Larson(1987) relates the psychology of waiting to customersatisfaction and queueing phenomena. He suggestsdesirable attributes of a service system, given thatcustomers will have to wait for service. Taylor (1994)


discusses the impact of delays on the customer’sevaluation of the overall service for airlines. Shereports that ‘delays trigger a negative mood’ in cus-tomers resulting in a ‘negative-mood bias’ in theirevaluation of service. She finds consistent differencesin the evaluation of service attributes (other thanpunctuality) between delayed and non-delayed pass-engers. See Figure 9 for a schematic of Taylor’s modelof the wait experience.

The psychology of waiting and its relationship withcustomer satisfaction provide a model to understandthe behavior of customers subject to delays. Servicedesigners should therefore use such model to designthe service system accordingly. That is, given a theor-etical foundation and corresponding models, design-ers can use the models to achieve better designs. Inthe case of possible delays, a service designer mayeither try to minimize the burden of waiting by enter-taining, enlightening, and informing, or add capacityto the service facility. The difference between the twocourses of action is in the models that must be usedto support each. In the first case, the model is anattempt to represent the psychological mechanisms

that operate in customers’minds while they wait. In thesecond case, the model mustconsider the physical structureof the service system, customerarrival and service rates, to pro-vide an estimate of the appro-priate capacity necessary tomatch demand and supply. Ineither case the analysis helps a

better understanding of the set of attributes for theservice interaction and the place where it occurs.

Mehrabian and Russell (1974) provide a set of prin-ciples that help understand human behavioralresponses to stimuli from the physical environment.Supported by empirical findings, they argue thathuman response to the physical environment can beof either approach or avoidance, moderated by threeemotional dimensions: pleasure–displeasure, arou-sal–non-arousal, and dominant–submissive. Batesonand Hui (1987) revise the Mehrabian–Russell modelto consider dominance as antecedent of pleasure andarousal. They conclude that dominance and per-ceived control are ‘closely associated theoretical con-structs,’ and directly correlate with pleasure. Crowd-ing, for example, inversely correlates with pleasureand perceived control, and directly correlates witharousal. They conclude that negative responsescaused by loss of control due to crowding can beminimized by returning some control to the cus-tomer. A greater degree of choice (for example,choice about where to wait for a long job) can lowerthe customer’s ‘perceived crowding’ and conse-quently improve customer satisfaction.

Donovan and Rossiter (1994) use the Mehrabian–Rus-sell model to understand the effects of store environ-


Figure 9 Taylor’s Model of the Wait Experience.

ment on shopping behavior and patronage. Theyargue that pleasure and arousal interact with eachother in such a way that arousal intensifies pleasureor displeasure reactions. Moreover, although cogni-tive factors (e.g., location, price — quality of mer-chandise, variety) largely account for store selectionand planned purchases, emotional affect and store-induced pleasure are important determinants ofbehavior within the store (e.g., time, affiliation, andspending behavior). They advocate the use of exper-imentation to determine which specific types of in-store stimuli (e.g., color arrangement, layouts, noiselevels, lighting) evoke desired emotional responses.That is the reason why prototyping and testing mayalso be so important in services.

Wener (1985) discusses the effect of disorientation oncustomers and employees. He argues that disorien-tation may be the result of lack of previous experi-ence, confusing designs, and lack or misplacement oforientation aids. Lack of previous experience or fam-iliarity with the service setting is also related to lackof control, which also leads to increased stress, per-ceived crowding, or other negative responses. That isone of the reasons why fast food restaurants andother service firms look so similar from the outside.Customers start to buy a service even before enteringthe facility, and feel more comfortable in a familiarenvironment. This model is certainly useful in thedesign of the external organization, as discussed inthe architecture of services.

It is now easy to understand the mechanisms behindthe rule of thumb of entertaining and informing thecustomer. By doing so, a service company is generat-ing pleasant stimuli or providing more control to cus-tomers, which in turn translates into positive serviceexperiences, and by association into a positive viewof the firm.

Somewhat related to perceived control is the issue ofrisk in the consumption of services. Guseman (1981)argues that services are generally perceived to beriskier because of their inherent intangibility, hetero-geneity in quality, and fuzzy quality standards.While a faulty product can be promptly returned,services often cannot, particularly those with highdegree of intangibility (e.g., art performances, pro-fessional services). The risks associated with the pur-chase of a product are discussed by Roselius (1971)


in terms of four types of losses: time, hazard, ego,and money losses. In the purchase of a product orservice customers try to employ (consciously or not)risk reduction methods that are different whether it isa product or service. Guseman (1981) concludes thatcustomers are more prone to rely on word-of-mouth,and on tangible factors such as appearance andlocation of buildings, internal design of facilities, andappearance and demeanor of employees as a surro-gate of service quality (or risk reduction method). Hesuggests that service guarantees can be offered toshift some control to customers and reduce the per-ceived risk associated with purchase of services.

The principles and models discussed so far representpossible psychological mechanisms that explainparticular individual behaviors in service settings.They establish simple direct relationships that can beobserved in real settings. There are models howeverthat seek to reproduce (and explain) more complexbehaviors with time delays and feedback loops. Theyare represented by causal-loop diagrams and systemdynamics models that map relationships of cause andeffect in a service system (refer to Forrester (1961) fora good exposition on System Dynamics).

System Dynamics models are useful in the design ofservice systems because they allow a better under-standing of how intangible variables (e.g., frustration,fatigue, excitement) affect behavior of human parti-cipants in the service setting, and ultimately its qual-ity. By employing them, service designers can buildinto the service system mechanisms that minimizeundesirable effects, or maximize desirable effects. Forexample, once companies understood that entertain-ing customers makes them more comfortable aboutwaiting, several firms installed televisions in theirwaiting rooms.

Schlesinger and Heskett’s Cycle of Failures Model isa causal-loop diagram which shows how problemstraditionally regarded as exogenous to the servicesystem are in fact structurally embedded in its design(Schlesinger and Heskett, 1991). Pue (1996) uses asystem dynamics model to simulate the effects ofmanagement policies in a service setting. He investi-gates (among other relationships) the effects of lackof service capacity on work pressure, employee turn-over and customer satisfaction to propose a dynamicmodel of service quality. Finally, Senge and Sterman


(1992) report a case study of applying systemdynamic models for improving quality and cost per-formance in the insurance industry.

Another class of models that may be used in servicedesign is operations research models. Since a largenumber of references is available in the literature, wecite just a few for the sake of brevity. Generally, theiruse in the context of service design address two basictypes of problems: performance evaluation and opti-mization. Performance evaluation problems eitherestimate the capacity necessary to achieve a targetperformance level, or estimate performance given atarget capacity level. Ittig (1994), for example, investi-gates the impact of having demand as a function ofaverage waiting times in designing the number ofcheck-out counters of a supermarket; Sze (1984)describes a telephone operator system through aqueueing model to predict average delays; and Albin(1990) uses a queueing network model to analyze thecauses of delays in a health center appointment clinic.Optimization problems try to maximize some per-formance measure at the appropriate planning level(strategic, tactical or operational). Bitran andMondschein (1995), for example, propose a model fordynamically determining optimal allocation of hotelrooms, subject to the current occupancy, reservations,and expected demand.

The research cited above shows that with the appro-priate theoretical base and modeling techniques it ispossible to approach the design of services from amodeling perspective. In fact, much of the knowl-edge incorporated in the models discussed in thissection has already been used in practice by leadingservice organizations. Models allow a systematiz-ation of this knowledge by expressing it in a formatthat is simpler to learn and to apply.

Conclusion and Future Research

The literature on product development encompassesseveral frameworks that deal with different aspectsof the problem. We attempted to integrate the vari-ous contributions in the course of this paper and dis-cussed them from a service perspective.

Similarly to the concept of design for manufactur-ability (see Whitney, 1988), in services there is theneed of designing for reproducibility. When a servicecompany plans to expand domestically or inter-nationally, reproducibility of the service conceptbecomes a fundamental attribute in creating a com-pany identity, reducing costs, and ensuring quality.The challenge is to design a service system for repro-ducibility but with flexibility to recognize and adaptto differences in culture, legal systems, regulations,and local infrastructure, among other factors. Howthe development effort should accommodate thisrequirement is an important topic for future research.


The discussion of methods and models available forproduct development indicates the inability of somemodels to support certain aspects of service develop-ment. For example, it would be desirable to developmodels that address in greater depth the connectionbetween tangible design variables (capacity, location,etc.) and intangible ones such as customer satisfac-tion and customer retention, among other variables.It would also be desirable to develop appropriatemetrics that enable the translation of intangible attri-butes into specifications that can be used for ser-vice design.

In order to illustrate the notion of architecture in ser-vices, we introduced in Figure 7 a service operationstructure with components that we believe are per-vasive to most service companies. It is sometimespossible, however, to describe the service system indifferent ways (e.g., through functions and processes)that may provide different insights and enrich thedevelopment effort. A study of alternative represen-tations of service systems is a topic that requiresfurther research. It may help designers understand,among other issues, how design decisions freezeparts of the final design and drive the search for sol-utions in the development effort.

Finally, this paper has attempted to show thatmethods available for product development can bevery useful in service design when complemented toconsider specific service characteristics. However, asin the case of any methodology, it must be adaptedto the particular needs of each company. Not all stepsand methods should be applied in the developmentof every service. Trying to do so can be cumbersomeand may slow down the development process.

Acknowledgements

This work was partially supported by the Brazilian Depart-ment of Education — CAPES Foundation Grant #125-92/2.The authors would like to thank Julio Gomes, Prof. RogelioOliva, Prof. Steve Eppinger, and Prof. Richard Chase for theirvaluable contributions and references.

References

Abernathy, W. and Utterback, J. M. (1978) Patterns of indus-trial innovation. Technology Review 80, 41–47.

Albin, S. L. (1990) A queueing network analysis of a healthcenter. Queueing Systems 7, 51–61.

Andersen, S. (1983) Need assessment: a way of improving thevalue of new products. Design Studies 4, 183–187.

Anderson, P. and Tushman, M. L. (1990) Technological dis-continuities and dominant designs: a cyclical model oftechnological change. Administrative Science Quarterly 35,604–633.

Bateson, J.E. and Hui, M.K. (1987) A model for crowding inthe service experience: empirical findings. In The ServicesChallenge: Integrating for Competitive Advantage, eds Czep-iel, J., Congram, C., and Shanahan, J., pp. 85–89. Proceed-ings Series of the American Marketing Association.

Behara, R. and Chase, R.B. (1981) Service quality deployment:


quality service by design. In Perspectives in OperationsManagement: Essays in Honor of Elwood S. Buffa, ed. Sarin,R. Kluwer Academic Publishers, Norwell, MA.

Bharadwaj, S. G., Varadarajan, P. R. and Fahy, J. (1993) Sus-tainable competitive advantage in service industries: aconceptual model and research propositions. Journal ofMarketing 57, 83–99.

Bitner, M. J. (1992) Servicescapes: the impact of physical sur-roundings on customers and employees. Journal of Mar-keting 56, 57–71.

Bitran, G. R. and Lojo, M. (1993) A framework for analyzingthe quality of the customer interface. European Manage-ment Journal 11, 385–396.

Bitran, G. R. and Mondschein, S. (1995) An application ofyield management to the hotel industry considering mul-tiple day stays. Operations Research 43, 0–00.

Blackman, B.A. (1985) Making a service more tangible canmake it more manageable. In The Service Encounter, edsCzepiel, J., Soloman, M., and Suprenant, C. LexingtonBooks, Lexington.

Booz-Allen and Hamilton (1982) New Product Management forthe 1980s. Booz-Allen and Hamilton, New York.

Bowen, D.E. and Schneider, B. (1985) Boundary-spanning roleof employees and the service encounter. In The ServiceEncounter, eds Czepiel, J., Soloman, M., and Suprenant,C., pp. 124–147. Lexington Books, Lexington.

Burchill, G.W. and Shen, D. (1992) Concept Engineering: TheKey to Operationally Defining your Customer’s Requirements.Center for Quality Management, Cambridge, MA.

Chase, R. B. (1981) The customer contact approach to services:theoretical bases and practical extensions. OperationsResearch 30, 698–706.

Chase, R. B. and Stewart, D. M. (1994) Make your service fail-safe. Sloan Management Review 36, 35–44.

Chase, R. B. and Tansik, D. A. (1980) The customer contactmodel for organizational design. Management Science 30,1037–1050.

Clark, K. B. (1985) The interaction of design hierarchies andmarket concepts in technological evolution. Research Pol-icy 14, 235–251.

Dierickx, I. and Cool, K. (1989) Asset stock accumulation andsustainability of competitive advantage. ManagementScience 35, 1504–1511.

Donovan, R. J. and Rossiter, J. R. (1994) Store atmosphere andpurchasing behavior. Journal of Retailing 70, 283–294.

Forrester, J. (1961) Industrial Dynamics. MIT Press, Cam-bridge, MA.

Griffin, A. and Hauser, J. R. (1993) The voice of the customer.Management Science 12, 1–27.

Gronroos, C. (1987) Developing the service offering — asource of competitive advantage. In Add value to yourservice. 6th Annual Services Marketing Conference Proceed-ings of the American Marketing Association, ed. Suprenant,C., pp. 81–85.

Guseman, D.S. (1981) Risk perception and risk reduction inconsumer services. In Marketing of services. ProceedingsSeries of the American Marketing Association, eds Donnelly,J., and George, W.R., pp. 200–204.

Hart, C. W. (1988) The power of unconditional guarantees.Harvard Business Review 66, 54–62.

Hart, C. W., Heskett, J. L. and Sasser, W. (1990) The profitableart of service recovery. Harvard Business Review 68,148–156.

Hauser, J. R. and Clausing, D. P. (1988) The house of quality.Harvard Business Review 66, 63–73.

Hauser, J. R. and Urban, G. L. (1979) Assessment of attributeimportance and consumer utility functions. Journal ofConsumer Research 5, 251–262.

Hax, A. and Majluf, N. (1991) The Strategy Concept and Process:A Pragmatic Approach. Prentice Hall, Englewood Cliffs,NJ.

Henderson, R. M. and Clark, K. B. (1990) Architectural inno-vation: the reconfiguration of existing product techno-logies and the failure of established firms. AdministrativeScience Quarterly 35, 9–30.


Hoover, S. P., Hinderle, J. R. and Finger, S. (1991) Models andabstractions in design. Design Studies 12, 237–245.

Ittig, P. T. (1994) Planning service capacity when demand issensitive to delay. Decision Sciences 25, 541–557.

Kano, S. (1982) Miryokuteki hinshitsu to atarimae hinshitsu(attractive quality and must-be quality). Journal of theJapanese Society of Quality Control 14, 0–00.

Kelly, J. P., Donnelly, J. H. and Skinner, S. J. (1990) Customerparticipation in service production and delivery. Journalof Retailing 66, 315–335.

Kinnear, T.C. and Taylor, J.R. (1991) Marketing Research: AnApplied Approach. McGraw Hill Co., New York, NY.

Larson, R. C. (1987) Perspectives on queues: Social justice andpsychology of queueing. Operations Research 35, 895–905.

Levitt, T. (1972) Production line approach to services. HarvardBusiness Review 51, 42–52.

Lovelock, C.H. (1984) Developing and implementing new ser-vices. In Developing new services. Proceedings Series ofthe American Marketing Association, eds George, W. andMarshall, C., pp. 44–64.

Maister, D.H. (1985) The psychology of waiting lines. In TheService Encounter, eds Czepiel, J., Soloman, M., and Supr-enant, C. Lexington Books, Lexington.

Mehrabian, A. and Russell, J.A. (1974) An Approach to Environ-mental Psychology. MIT Press, Cambridge, MA.

Milliman, R. (1982) Using background music to affectbehavior of supermarket shoppers. Journal of Marketing46, 86–91.

Mills, P. K. and Morris, J. H. (1983) Clients as ‘partial’employees of service organizations: role development inclient participation. Academy of Management Review 11,726–735.

Murphy, P.E. and Robison, R.K. (1981) Concept testing ser-vices. In Marketing of services. Proceedings Series of theAmerican Marketing Association, eds Donnelly, J., andGeorge, W.R., pp. 217–220.

Pahl, G. and Beitz, W. (1984) Engineering Design. The DesignCouncil, London.

Pimmler, T. and Eppinger, S.D. (1994) Integration analysis ofproduct decomposition. ASME Design Theory and Method-ology Conference, Minneapolis, MI.

Porter, M.E. (1980) Competitive Advantage. The Free Press,Macmillan Inc.

Pue, R.O. (1996) An examination of the profit generatingpotential of customer service: a modeling approach. PhDthesis, Massachusetts Institute of Technology.

Pugh, S. (1990) Total Design. Addison-Wesley, Reading, MA.Quinn, J. B. and Gagnon, C. E. (1986) Will services follow

manufacturing into decline?. Harvard Business Review 64,95–103.

Quinn, J. B. and Paquette, P. C. (1990) Technology in services:creating organizational revolutions. Sloan ManagementReview 32, 67–87.

Quinn, J. B., Doorley, T. L. and Paquette, P. C. (1990) Tech-nology in services: rethinking strategic focus. Sloan Man-agement Review 32, 79–87.

Rechtin, E. (1991) Systems Architecting: Creating and BuildingComplex Systems. Prentice Hall, Englewood Cliffs, NJ.

Roselius, T. (1971) Consumer rankings of risk reductionmethods. Journal of Marketing 35, 56–61.

Sasser, W.E., Olsen, R.P. and Wyckoff, D.D. (1978) Manage-ment of Service Operations. Allyn and Bacon.

Schlesinger, L. A. and Heskett, J. L. (1991) Breaking the cycleof failure in services. Sloan Management Review 33, 17–28.

Senge, P. (1990) The Fifth Discipline. Doubleday/Currency,New York.

Senge, P. and Sterman, J. (1992) Systems thinking and organi-zational learning: acting locally and thinking globally inthe organization of the future. European Journal of Oper-ational Research 59, 137–150.

Shocker, A. D. and Srinivasan, V. (1979) Multiattributeapproaches for product concept evaluation and gener-ation: a critical review. Journal of Marketing Research 16,159–180.

Shostack, G.L. (1985) Planning the service encounter. In The


Service Encounter, eds Czepiel, J., Soloman, M., and Supr-enant, C. Lexington Books, Lexington.

Solomon, R. (1985) A role theory perspective on dyadic inter-actions: the service encounter. Journal of Marketing 49,99–111.

Sze, D. (1984) A queueing model for telephone operatorstaffing. Operations Research 32, 229–249.

Taylor, S. (1994) Waiting for service: the relationship betweendelays and evaluation of service. Journal of Marketing 58,56–69.

Thompson, J.D. (1967) Organizations in Action. Mc Graw HillCo., New York, NY.

Ulrich, K.T. and Eppinger, S.D. (1995) Product Design andDevelopment. McGraw Hill Co., New York, NY.

Upah, G. D. (1980) Mass marketing in service retailing: areview and synthesis of major methods. Journal ofRetailing 56, 59–76.

Urban, G.L. and Hauser J.R. (1980) Design and Marketing ofNew Products. Prentice Hall, Englewood Cliffs, NJ.

GABRIEL BITRAN, LUIS PEDROSA, SloanSloan School of Manage- School of Management,ment, Massachusetts Massachusetts Institute ofInstitute of Technology, Technology, Cambridge,Cambridge, MA 02139, MA 02139, USA.USA.

Luis Pedrosa is a doctoralGabriel Bitran is Professor candidate in Operationsof Management Science at Management at MIT,MIT Sloan School of Man- primarily interested in theagement. He is also Edi- design of service oper-tor-in-Chief of Manage- ations. He has researched

ment Science, and a member of the editorial boards product development techniques in the serviceof several management journals including the Eur- context, and is now investigating the applicationopean Management Journal. His research interests lie of stochastic models to service design. Formerly ain the field of operations management in manufac- consultant in information technologies, he is cur-turing and the service industry, and he has pub- rently Director for Marketing Development of thelished and consulted widely in operations man- retail bank division of UNIBANCO S.A., Saoagement. Paulo. Luis expects to complete his doctorate in

1998.


Von Hippel, E. (1986) Lead users: a source of novel productconcepts. Management Science 32, 791–805.

Ward, J. C., Bitner, M. J. and Barnes, J. (1992) Measuring theprototypicality and meaning of retail environments. Jour-nal of Retailing 68, 194–220.

Wener, R.E. (1985)The environmental psychology of serviceencounters. In The Service Encounter, eds Czepiel, J., Solo-man, M., and Suprenant, C., pp. 101–112. LexingtonBooks, Lexington.

Wheelwright, S. and Clark, K.B. (1992) Revolutionizing ProductDevelopment: Quantum Leaps in Speed, Efficiency, and Qual-ity. The Free Press, New York.

Whitney, D. E. (1988) Manufacturing by design. Harvard Busi-ness Review 0, 83–91.

Wind, Y.J., et al. (1989) Courtyard by Marriott: Designing a HotelFacility with Consumer-based Marketing Models. Addison-Wesley, Reading, MA.

Documents

European Management Journal Vol. 16, No. 2, pp. 169–189 ...€¦ · help design intangible elements. We conclude the paper by identifying gaps in the literature and sug-gesting