Service interaction design: A Hawk-Dove game based approachto managing customer expectations for oligopoly service providers

Yen-Hao Hsieh & Soe-Tsyr Yuan & Hsiao-Chen Liu

Published online: 20 September 2012# The Author(s) 2012. This article is published with open access at Springerlink.com

Abstract In the “experience economy”, effectively deliver-ing memorable and exciting customer experiences has be-come a key issue for service providers. Service experiencedelivery involves service encounters through which interac-tions between service providers and customers can beshaped into interactive artifacts managing customer expect-ations and dynamically delivering suitable services. Serviceinteraction design aims to optimize customer interactionswith services to match customer expectations and yieldsatisfactory service experiences. On the other hand, serviceproviders typically make profits and cost the priority, de-spite knowing that high service quality can maximize satis-faction, particularly in markets served by an oligopoly,resulting in customers only accepting existing limited-value services. Hence, the oligopoly market can be regardedas a value-bounded context. Additionally, understandingcustomer expectations regarding a wide range of interac-tions is crucial to service providers selecting and designingservices that match customer expectations. Therefore, thispaper presents a service interaction design mechanism tohelp oligopoly service providers systematically and effec-tively manage customer expectations in dynamic interac-tions, even in value-bounded contexts. The proposed

mechanism models this service interaction design problemas a series of Hawk-Dove games that approach an evolu-tionary stable state. The evaluation results suggest that oli-gopoly service providers should change their mindsets anddesign service interactions to manage customer expectationsassociated with service delivery, not only to ensure highsatisfaction and profit but also to engage customers in co-creating value.

Keywords Service interaction design . Service experiencedelivery . Hawk-Dove game . Customer expectationmanagement . Oligopoly service provider

1 Introduction

Pine and Gilmore (1999) proposed that the rise of the“experience economy” will become the main determinantof the economic activities of service firms and customerbehaviors in the 21st Century. Figure 1 shows the fourstages in the evolution of economic value, namely commod-ities (agrarian), goods (industrial), services (service), andexperiences (experience). In the agrarian economy, humansuse natural resources (such as water, wheat or vegetables) ascommodities to fulfill their basic needs (thirst or hunger).During this stage, economic activity focuses on satisfyingcustomer physiological requirements. In contrast, the expe-rience economy sees businesses provide customers withparticular ways to experience services or goods. The eco-nomic activities in experience economies not only meetcustomer physiological needs, but also consider their emo-tional and psychological experiences. Accordingly, serviceproviders in the experience economy must understand cus-tomer needs to provide memorable service experiences, andmeeting customer expectations is critical (Meyer andSchwager 2007; Ojasalo 2001; Parasuraman et al. 1991;

Y.-H. Hsieh (*)Department of Information Management, Tamkang University,New Taipei City, Taiwane-mail: yhhsiehs@mail.tku.edu.tw

S.-T. Yuan :H.-C. LiuDepartment of Management Information Systems,National Chengchi University,Taipei, Taiwan

S.-T. Yuane-mail: yuans@mis.nccu.edu.tw

H.-C. Liue-mail: joanneliusc@gmail.com

Inf Syst Front (2014) 16:697–713DOI 10.1007/s10796-012-9386-5

Pine and Gilmore 1999). Managing customer expectationsis thus essential to achieving customer satisfaction. Serviceproviders can define and provide competitive services thatmeet customer expectations to increase market competency(Pitt and Barbara 1994; Schurter and Towers 2006).

Service encounters can be defined as direct interactionsbetween service providers and customers (Lewis andEntwistle 1990). Service experience delivery involves nu-merous service encounters (Holmlid 2007). This studyargues that service experience delivery can be regarded asan interactive artifact that involves process innovation andinteractive technology, and that requires responding to cus-tomer input and possibly changing customer content orbehavior. This study developed an interaction design toinfluence customer behavior by providing customers withtimely and relevant information to enable them to make in-formed decisions, complete their work easily, or simply beentertained (Saffer 2005). This could be further extended toincorporate services (besides information) delivered to cus-tomers. This study claims that interactions between serviceproviders and customers could be shaped into interactiveartifacts to manage customer expectations and dynamicallydeliver suitable services through service encounters. Thisstudy terms this kind of interaction design service interactiondesign. Service interaction design aims to optimize customerinteractions with services to match customer expectations andyield satisfactory service experiences.

On the other hand, an oligopoly is a common marketstructure characterized by imperfect competition. Mergersand alliances resulting from the global division of labor andeconomic sluggishness have seen numerous industries be-come oligopolies. For example, the merger of Google andYouTube is a famous case of an oligopoly involving onlineportal sites. Furthermore, Ingram Micro is the top distributorof information-based products, with many resellers globally(Zhao et al. 2009), and the telecommunications industry canbe considered to be an oligopoly, in which a few informationservice providers (ISPs) dominate the markets for mobile and

network services. Oligopoly markets have fewer sellers thanbuyers. Oligopoly service providers often dominate customerinteractions, and focus on competition rather than customers(Hendricks and McAfee 2000, 2010). Additionally, oligopolyservice providers usually lack a long-term perspective on theco-creation of value with customers, and force customers toaccept existing services (Hendricks and McAfee 2000, 2010).Hence, an oligopoly market can be considered a value-bounded context.

Customer value is highly restricted in an oligopoly market,since oligopoly service providers may only offer customerslimited services and resources. However, oligopoly serviceproviders are often highly competent in delivering positiveservice experiences (Ren 2007). Such service providers con-sider costs and profits over customer satisfaction, despitecustomers expressing numerous needs and making multiplerequests regarding service experiences (for example in re-sponse to the limited range of services available to them).However, oligopoly service providers can increase customersatisfaction and attract more customers to use their servicesthrough effectively managing customer expectations, increas-ing their market share and thus profits. Restated, to maintain aleading position in an oligopoly market, oligopoly serviceproviders can still manage customer expectations to maintaina good image in the minds of their customers, even in a limitedcustomer value-bounded context.

Restated, oligopoly service providers can provide goodexperiences and design effective customer interactions bymanaging customer expectations without significantlyimpacting revenue. This situation differs from the dilemmawhere service providers must choose whether to increaseservices or reduce prices, both of which require consideringcomplex issues like market share and competitors. Thisstudy thus examines the following questions:

& Do service interaction design mechanisms exist thatoligopoly service providers can adopt to manage cus-tomer expectations during a service experience?

& How can oligopoly service providers co-create valuewith customers in a value-bounded context?

This study develops a service interaction design mecha-nism for managing customer expectations by providingoligopoly service providers with advice on how to meetcustomer requirements in service encounters. This workmeets the following two objectives. First, it develops aquantitative mechanism to enable service providers to mon-itor and direct customer interactions during service encoun-ters in a manner that manages customer expectations andensures customer satisfaction. Second, it conducts simula-tion experiments to determine the effectiveness and feasi-bility of applying the proposed mechanism to oligopolyservice providers, and to determine the extent of value co-creation in the value-bounded context.

Differentiated

CompetitivePosition

Undifferentiated

ExtractCommodities

MakeGoods

DeliverServices

StageExperiences

Market Pricing Premium

Fig. 1 The progression of economic value (Pine and Gilmore 1999)

698 Inf Syst Front (2014) 16:697–713

Despite the importance of managing customer expecta-tions, many previous investigations have examined the effectof customer expectations on service satisfaction. These studiesused quantitative methods (such as surveys) to gather datafrom service providers and customers for further analysis. Tothe best of the knowledge of the authors, few studies haveaddressed the issue of customer expectation management inrelation to service experience delivery, especially in oligopolymarkets, by running simulations. However, this study is thefirst to use the Hawk-Dove game to solve the research prob-lems identified above. Researchers have tended not to tacklethe specified problems for three reasons. First, customerexpectations are mental states, and thus clearly defining thelevels of customer expectations and the influences on thoseexpectations is challenging for both researchers and serviceproviders. Second, many factors (service-related operations,employee quality, service encounter design, and so on) influ-ence service provider delivery of service experiences in anoligopolistic market. Researchers and service providers havedifficulty defining research boundaries and simulation param-eters to capture practical service experience delivery in such amarket. Finally, for the above reasons, constructing a serviceinteraction design mechanism to understand customer expect-ations and help deliver dynamically appropriate services inreal-time service contexts is difficult. Therefore, despite manydifficulties in addressing the above research problems andissues, this study proposes both solutions and approaches toovercome these obstacles.

The proposed service interaction design mechanism isbased on Hawk-Dove game theory. In an oligopolistic mar-ket, competition between providers and customers can beregarded as dynamic (Tan 2006). Both customers and pro-viders seek benefits, namely satisfaction with the serviceexperience and maximized revenue, respectively. The situa-tion is one of conflict, and is considered a non- zero-sumgame in which customers and providers can adopt strategiesbased on cooperation or resistance. As indicated above, anoligopolistic market is a value-bounded context in whichboth service providers and customers utilize limited resour-ces to increase values and improve the service experience(Durham, 2004). To optimize the management of customerexpectations in a value-bounded context, the tension be-tween customer satisfaction and cost to service providersshould be coordinated to achieve ecosystem equilibrium(given that both customers and service providers have lim-ited resources to improve the service process and increasecustomer value). This struggle resembles that of animalscompeting for resources. The Hawk-Dove game can thussimulate both customers and service providers to identifythe optimal evolutionarily stable solutions (Berninghaus andEhrhart 2003). Restated, both service providers and custom-ers can effectively use the Hawk-Dove game to plan re-source investments and optimize customer satisfaction.

The remainder of this paper is organized as follows.Section 2 summarizes the literature on customer expectationsand the Hawk-Dove game. Section 3 then presents the logicfor designing a system for managing customer expectationsand the service interaction design mechanism for the Hawk-Dove game. Section 4 performs experiments to validate theproposed mechanism. Section 5 then discusses contributionsto the service interaction design and service experience de-livery. Finally, Section 6 discusses the findings and theirtheoretical and practical implications, and also suggests fu-ture research directions.

2 Background and related work

Customer expectation management is a central concept intheproposed interaction design mechanism. Service providersdesigning interactive encounters develop innovative serv-ices for customers. This study helps oligopoly service pro-viders manage customer expectations using the Hawk-Dovegame during service experience delivery to increase custom-er satisfaction and profits. This section thus discusses andanalyzes the importance of customer expectations and intro-duces the Hawk-Dove game-based approach.

2.1 Customer expectation

A dual-level (desired and adequate levels), dynamic modelof customer expectations provides the optimal method forcharacterizing customer expectations (Parasuraman et al.1991). The desired and adequate expectation levels repre-sent the service levels that the customer hopes for, and thosethat are acceptable, respectively. The desired level combineswhat the customer believes the product “can be” with whatthey believe it “should be”. The adequate level is partiallybased on the customer predicted service level. The tolerancezone varies among customers, and even among situationsfor a single customer. Initial customer expectations are re-lated to expectation disconfirmation, the degree of whichinfluences customer satisfaction. Zeithaml et al. (1993) pre-sented a conceptual framework for specifying serviceexpectations and elucidated customer expectations using11 antecedent factors (transitory service intensifiers, per-ceived service alternatives, customer self-perceived servicerole, situational factors, predicted service, enduring serviceintensifiers, personal needs, transitory service intensifiers,perceived service alternatives, customer self-perceived ser-vice role, situational factors and predicted service), whichdirectly and indirectly affect desired and adequate expect-ations (as listed in Fig. 2) (Zeithaml et al. 1993). For exam-ple, when an automobile provider produces advertisementsor contracts that explicitly promise customers it will provideperfect maintenance service, customer expectations increase

Inf Syst Front (2014) 16:697–713 699

accordingly. Perceived service alternatives are considered,for example: if customers can choose alternative serviceproviders or services, then their levels of adequate servicemay increase highly. Accordingly, this work uses thesedeterminants of expectation to define service componentsand thus manage customer expectations.

Understanding customer expectations requires service pro-viders to successfully reach the customer franchise (Parasuramanet al. 1991; Kurtz and Clow 1993; Rust et al. 1999; Haeckel et al.2003). The conceptualization of multiple expectations influen-ces service provider allocation of resources (Walker and Baker2000). Service providers can fulfill customer expectations andachieve customer satisfaction if they deliver services in amanner that considers the determinants of expectation. Under-standing customer tolerance zone is crucial to ensuring custom-er satisfaction (Johnston 1995). Managing customerexpectations is therefore incorporated into the proposed inter-action design mechanism, affecting service delivery and ensur-ing customer satisfaction.

Competent service providers in an oligopoly market canmanage customer expectations by raising the adequate ex-pectation (from Adequate to Adequate’) and stabilizing thedesired expectation to reduce the tolerance zone (as shownin Fig. 3) and thus increase barriers to competition. Forexample, since Apple uses branding and promotion to raisecustomer expectations of its product, customers expect theproduct to be superior to alternatives. Customer loyaltyproves that Apple has successfully managed customer

expectations by raising them, and has achieved a specialposition in the computer systems market.

2.2 Hawk-Dove game

In nature, individuals must forage and sometimes fight forfood. The Hawk-Dove game represents one application ofgame theory to animal behavior (Kokko et al. 2006). Thegame describes the situation in which contestants can selecteither an aggressive (Hawk) or non-aggressive (Dove) strate-gy to compete for a collective resource. Hawks use everyopportunity to steal or defend food, while Doves never stealor resist food thieves. These two birds thus comprise a poly-morphic population. Mathematically, the Hawk-Dove game isa non-zero summatrix game with the payoff matrix in Table 1,where V denotes the value of the contested resource; C is thecost of an escalated fight (Smith and Parker 1976), and theformer is assumed to be less than the latter (C > V > 0).

The payoff matrix displays four possible fight situations.In the first case, when Hawk encounters Hawk, the winnerobtains a resource of value V while the loser bears the costof the fight, C, or both participants bear the cost of the fightequally, C/2. In the second case, when Hawk meets Dove,the Hawk monopolizes the entire resource and leaves theDove with nothing. In the third case, when Dove meetsHawk, the Dove receives no resource. In the fourth andfinal case, two Doves meet and share the value of theresource with V/2.

ENDURING SERVICE INTENSIFIERS

Derived expectationsPersonal service philosophies

PERSONAL NEEDS

TRANSITORY SERVICE INTENSIFIERS

EmergenciesService problems

PERCEIVED SERVICE ALTERNATIVES

SELF-PERCEIVED SERVICE ROLE

SITUATIONAL FACTORSBad weatherCatastropheRandom over-demand

EXPECTED SERVICE

Adequate Service

Desired Service

Zone of Tolerance

PREDICTED SERVICE

PAST EXPERIENCE

WORD-OF-MOUTHPersonalExpert(Consumer Reports,

publicity, consultants, surrogates)

IMPLICIT SERVICE PROMISESTangiblesPrice

EXPLICIT SERVICE PROMISESAdvertisingPersonal sellingContractsOther communications

Fig. 2 The nature anddeterminants of customerexpectations of service(Zeithaml et al. 1993)

High

Expectation

Low

Expectation

Adequate Adequate’ Desired

Fig. 3 The objective of customer expectation management

Table 1 The basic pay-off matrix of Hawk-Dove game

Hawk Dove

Hawk V�C2 ; V�C

2

� �(V, 0)

Dove (0, V) V2 ;

V2

� �

700 Inf Syst Front (2014) 16:697–713

The relationship between Vand C is associated with differ-ent population equilibria. Equilibrium refers to the situationwhen game participants adopt the optimal combination ofstrategies. By approaching equilibrium, players can optimizeresource value. In the aforementioned classic Hawk-Dovegame, if C > V > 0, then the proportion of individuals in thepopulation that adopts the Hawk strategy tends to equal V/C atequilibrium. (If C < V the environment favors Hawks, but thisfactor is not considered here.) Each individual considers valueand cost in strategy selection. Strategy selection may eventuallyapproach a stable evolutionary state (ESS). The concept of ESSinvolves the characteristics of specific Nash equilibriums,which are immune to the invasion of mutant strategies andwere developed during the 1970s (Smith and Price 1973).ESS equilibrium has been shown strong enough to attract play-ers who could perform better in other equilibriums, such as thedominated equilibrium (Berninghaus and Ehrhart 2003), par-ticularly when players who are not fully informed of the payofffunctions are unaware that they are selecting an ESS.

This study adopts the Hawk-Dove game to simulate thecosts incurred, and the value gained, by oligopoly serviceproviders and customers, to examine strategy selection. Theevolutionary dynamics associated with consecutive gamesare considered to be based on repeated interactions betweenprovider and customers in service delivery, and therefore theESS equilibrium is used to guide strategy selection.

3 Hawk-Dove game based interaction design mechanism

This study attempts to design a service interaction designmechanism that oligopoly service providers can use to man-age customer expectations in service encounters. The inter-action design mechanism is based on the Hawk-Dove gameas an analogy for interactions between providers and cus-tomers (as described in Section 2.2).

The mechanism comprises five modules of the Hawk-Dove game-based interaction design mechanism, namelythe context detection module, the determinant decision mod-ule, the Hawk-Dove game module, the expectation measure-ment module, and the solution selection module (as shownin Fig. 4). In response to customer service requests, theproposed mechanism detects customer behaviors andqueries their preferences. The service provider determinesthe values desired by customers, and considers the circum-stances surrounding the selection of the determinants ofexpectation considered in designing the service interaction.Following determinant selection, the mechanism calculateshow to use the determinants in service delivery and usescustomer expectations to improve service experiences. Oli-gopoly service providers thus can adopt the effective servicecomponents to achieve satisfactory customer service expe-riences. The modules are described below.

This study integrated the Hawk-Dove game into an ex-hibition service system (UESS (Hsieh and Yuan 2009))characterized by assured delivery of a high quality serviceexperience, enabled by different categories of interactiondesign mechanisms when interfacing with exhibition stake-holders (visitors, exhibitors and organizers) in a location-based services-enabled exhibition environment. Each classof stakeholder has various goals and can influence theothers. The Hawk-Dove interaction design mechanism isactivated when the value-bounded context is observed inthe presence of oligopolistic exhibitors.

3.1 Context detection module

The context detection module is designed to identify theenvironment during service delivery. Service operationsand resources must be carefully considered to optimize ser-vice delivery performance and efficacy. The proposed mod-ule must recognize customer feedback and behavior, whichcan be immediately and accurately inputted to the mecha-nism to modify the determinant decision module. The learn-ing approach thus is critical to service delivery in a dynamiccontext.

3.2 Determinant decision module

The determinant decision module uses useful determi-nants to influence customer expectations based on inputsfrom the context detection module, customer preferences(such as age, brand interest, price etc.), and data from theencounter database, including previously collected cus-tomer data, information regarding delivered services,and so on. This module then automatically analyzes andconverts the encounter data, customer preferences and thekey performance indicators of the organizer (KPIs) intocomputable scores to yield the weight of each determi-nant, indicating the strength of its influence on customer

Determinant

Decision

Hawk-Dove Game

Module

Solution Selection

Module

Context detection Module

Expectation Measurement

Module

Encounter Database

Service Component

Database

Journey template

Fig. 4 The Design logic

Inf Syst Front (2014) 16:697–713 701

expectations in the service context. This module selectscandidate determinants of expectation that can be adoptedto alter customer expectations regarding service delivery.Since this work applies the proposed approach to theexhibition service sector, one domain expert and oneresearcher with over a decade of experience of thisservice sector help define a reasonable range for eachweighting and parameter.

3.3 Hawk-Dove game module

The Hawk-Dove game module determines suitable deter-minants of expectation for exploitation in service encoun-ters. After the determinant decision module obtains theweight of each candidate determinant of expectation,these determinants are inputted to the Hawk-Dove gamemodule, which thus calculates the effective interactiondesign for use in service delivery. Figure 5 displays theprocedure followed by the Hawk-Dove game module.Based on the input data from the determinant decisionmodule and information on the ESS goal database andcustomer requirements, the ESS goal identification com-ponent can identify goals for customer expectations. Fi-nally, these determinants are inputted to the solutionselection module, which is outlined below.

The Hawk-Dove game models a competitive situationinvolving a shared resource, in which contestants can adopteither an aggressive (Hawk) or passive (Dove) competitivestrategy. Mathematically, the Hawk-Dove game is a two-player non-zero sum matrix game (Grundman et al. 2009).This study applies the Hawk-Dove game to design interac-tions between an oligopolistic service provider and its cus-tomers. Either the oligopoly service provider or its customersmust attempt to maximize profits while reducing the costs ofservice delivery. Not only the does the oligopoly serviceprovider meet customer needs at relatively high cost, butcustomers must make more effort to obtain excellent service.From this perspective, each interaction can be regarded as anon-zero sum game.

3.3.1 Players

As noted above, experiments involving two main playerswho can adopt either the Hawk or Dove strategy wereperformed. Table 2 clarifies both strategies.

3.3.2 Payoff matrix of Hawk-Dove game

The payoff matrix of this Hawk-Dove game of interactionbetween customers and providers is shown below (as listed inTable 3). When a customer and service provider both selectthe Dove strategy, the payoff is (Coe (p1-c1), V1 + V2). Thevalue V1 + V2 represents the potential benefit to the customerin a service encounter, while the value Coe (p1-c1) representsthe profit to the provider. Table 4 defines the relevantparameters.

As the payoff matrix indicates, a service provider notonly evaluates the profit and cost to all stakeholders in aservice encounter, but also determines the service arrange-ment most suited to a customer. Figure 6 displays the payoffin each encounter when a customer and provider select theHawk and Dove strategies, respectively. The figure alsopresents the overall profits of the customer and provider,which can be calculated by summing all payoffs in theservice encounters.

3.3.3 Environmental parameters

In using the Hawk-Dove game, this work assumes manyenvironmental parameters based on the research of Kokko(Kokko et al. 2006). In each service encounter, customersselect the Hawk or Dove strategy at rate x and the interactionwith degree θ. Simultaneously, the provider adopts a strat-egy at rate y. This interaction gives providers and customersprofits of δ1(P-C) and δ2(V-E) respectively. Consequently,

the function xyθd22 denotes the benefit gained by a customer

in a single service encounter. For service providers, a similar

function xynθd12 specifies the profit from each service en-

counter, given that the encounter involves n customers. Theprovider can design a series of interactions in variousencounters during service experience delivery.

Hawk-Dove Game for

Determinants Selection

ESS Goal

Identification

ESS Goal Database

Solution Selection

Module

Customer Needs

Determinants

Fig. 5 Hawk-Dove game module

Table 2 The Hawk-Dove strategies considered in our interactiondesign

Hawk Providers always propose the serviceinteraction design that cannot accommodatethe needs of customer.

Customers always resist taking the service.

Dove Providers propose the service interaction designthat can accommodate the needs of customer.

Customers accept the given service arrangement.

702 Inf Syst Front (2014) 16:697–713

Both the provider and customer can gain. This study thususes μT to represent the gain to the provider and μF torepresent the gain to the customer. Accordingly, the totalprofit to a provider equals

mT ¼ μT þ xynθd12

ð1Þ

The total gain by a customer is calculated similarly:

mF ¼ μF þ xyθd22

ð2Þ

Table 5 lists the definition of each parameter.

3.3.4 Evolutionary stable state

In the equilibrium state, n customers are assumed to beinvolved in a particular encounter. The oligopoly serviceprovider and customers share limited resources and benefits,as represented by Eq. 3:

mF � nþ mT ¼ k ð3Þwhere k represents the total gain. The service componentsare limited since the oligopoly service provider is willing to

pay a limited cost to supply service components to custom-ers. However, customers can opt out of receiving servicesfrom service providers if the receipt of such servicesrequires excessive effort. Accordingly, the service compo-nents represent a balance that satisfies both customers andservice providers. The limitations on the designed servicecomponents (or on the provision of services) set a bound onthe profits of the customer and oligopoly service provider. Astable evolutionary state suggests a direction for oligopolyservice providers seeking to design interactive serviceencounters. Oligopoly service providers can set differentESSs to manage expectations to induce customers to seekdifferent values. For example, since oligopoly service pro-viders wish to reduce customization costs, customers can beasked to seek related services themselves.

Table 3 The payoff matrix of this Hawk-Dove game

Customer Dove HawkProvider

Dove (Coe(p1-c1), V1 + V2) (Coe (p1-c1), V1)

Hawk (Coe(p2-c2), V1 + V2-E) (Coe(p2-c2), V1-E)

Table 4 The definition of parameters of the payoff matrix

Parameter Definition

V1 The value generated from the service

V2 The benefit generated when a customeradopts the Dove strategy

p1 The profit generated when a provideradopts the Dove strategy

p2 The profit generated when a provideradopts the Hawk strategy

c1 The cost generated when a provideradopts the Dove strategy

c2 The cost generated when a provideradopts the Hawk strategy

E The extra effort that the customerpay for the service excellencewhen the provider play Hawk strategy

Coe The coefficient indicates one over thenumber of service providersin the oligopoly market; This studyassumes one service providerin the context that Coe is to represent 1

Provider

Customer n

Encounter

point

Dove Hawk Dove

Hawk Dove Hawk

V1+V2-E

p2-c2

V1

p1-c1

V1+V2-E

p2-c2

Payoff

Payoff

Fig. 6 The payoff in service encounters

Table 5 The model parameters and notation

Symbol Definition

mT The total profit of the provider

μT The profit of the service tacticsthe provider gained in afore-encounters

x The probability of the customer choosingcertain action strategy

y The probability of the provider choosingcertain action strategy

n The number of customers encountered ina time unit

θ The degree of customer participating incertain service encounter

δ1 The rate of providers making profit incertain service solution (P-C)

mF The total profit of the customer

μF The profit the customer gained inafore-encounters

δ2 The rate of customer gaining profit incertain service solution (V-E)

k The total profit among customersand providers

Inf Syst Front (2014) 16:697–713 703

The three strategies of the Evolutionarily Stable State aredefined by the cost and profit to the provider and customer,as well as by some environmental factors.

(1) The customer always adopts the Dove strategy, “com-pletely respecting” the service agreement and strivingto ensure excellent service.

(2) The provider always adopts the Dove strategy to ac-commodate customer needs and expectations.

(3) In the mixed strategy, the customers “partially respect”the service agreement and sometimes play Hawk toprevent the provider taking their gains, and the provid-er is forced to fulfill certain customer needs andexpectations.

The ESS can supply a provider with direction on thedesign of service interactions with customers. The providerwould set various ESSs as goals in managing the expect-ations of customers with different values. For example,when a customer wants to control service composition, theservice provider may set the goal of the ESS to be the mixedstrategy in which the customer is free to select suitableservices. In contrast, when the provider asks the customersto find information independently, the service providerreduces the cost of customization while accommodatingcustomer desires with some loss of benefit. This studyadopts the mixed strategy (partial respect) for demonstrativepurposes.

Figure 7 illustrates the process of designing a serviceinteraction using the Hawk-Dove game. First, the oligopolyservice provider utilizes the determinant decision module todetermine the combination of determinants of candidateexpectation that are relevant to service delivery. Second,

the oligopoly service provider can calculate the costs andprofits of both sides to determine whether different determi-nants can be set to realize the ESS goals of customers atequilibrium. After identifying the proper determinants, themechanism exploits them in an orderly fashion using thegreedy approach, based on the weights derived from userneeds and preferences as well as the provider KPI. Thegreedy approach involves arranging the determinants se-quentially from highest to lowest weight. When the Hawk-Dove game selects four determinants for use in servicedelivery, the greedy approach arranges them sequentially(as shown in Table 6). Figure 8 lists the results for servicedelivery using the greedy approach. Restated, service pro-viders can deliver required services to customers in eachservice encounter by analyzing the payoffs to both them-selves and customers, and with reference to ESS goals.

3.4 Solution selection module

After the Hawk-Dove game module calculates the determi-nants to be exploited in service delivery, the solution selec-tion module identifies the most suitable service componentsfor managing customer expectations in the value-boundedcontext. The identification is confirmed by predicting theeffects of the chosen determinants using the expectationmeasurement module, which provides both the scores ofthe dual expectation measurements and appropriate expec-tation tactics.

This work assumes that each determinant of expectationis associated with corresponding service tactics (expectationtactics), which are used to select suitable service compo-nents, supporting a flexible method for delivering service

Payoff ( )Tm T 21xyn

Provider

Determinants

(D1,D2,D3,D4)

Encounter point

Equilibrium:

Customer nESS goal:

partial-respect

Payoff ( )

kmnm TF

Fm2

2xy

Dove Hawk Dove

Hawk Dove Hawk

D1, D2 D3 D3, D4

+

+F

+

Fig. 7 Exemplar process of theHawk-Dove game module

704 Inf Syst Front (2014) 16:697–713

components. The recommended product is an example ofservice tactics used to represent the determinant of expecta-tion, Word-of-Mouth.

The effects of the selected determinants on customermental state are evaluated to determine the feasibility ofemploying the expectation measurement module (as de-tailed in section 3.5). The expectation measurement modulecan measure customer expectations in real time and indynamic environments that are affected by the selecteddeterminants.

Based on the chosen expectation tactics, the solution selec-tion module selects service components for implementation.When customers alter their behaviors and service encounterpoints, the proposed mechanism must interact with customersimmediately and dynamically. In this case, the greedy algo-rithm selects the optimal service components for efficientsolution identification. Selecting the optimal combination ofservice components is a knapsack problem in which theoligopoly service provider must select service componentsto maximize benefits given a limited budget. The knapsackproblem is the following maximization problem.

maxPq

j¼1pjyj

s:t:Pq

j¼1cojyj � cot; yj 2 0; 1f g; j ¼ 1 . . . ; q;

ð4Þ

where q denotes the total number of service components, andeach service component j has cost coj and profit pj. Theoligopoly service provider determines the upper bound onthe total cost, cot, which is the sum of the costs of labor,technology and materials. This module sets the values of yjat yj01 if the knapsack contains service component j and atyj00 otherwise. The greedy strategy selects the alternativewith a maximum pj/coj that fits into the knapsack.

The service components identified by the greedy strategymay not be optimal for all implementations, but representthe best combination that the oligopoly service provider canidentify given time constraints. Since all service compo-nents are allocated by customers and service providers, theoligopoly service provider can estimate costs and profits foreach service component during the first stage of servicecomponent design. For example, when the oligopoly serviceprovider adopts advertising, it can evaluate the cost and theeffect of different advertisements (such as a commercialvideo or customer mail-outs). Following service componentselection, this module considers the context factors (such asthe distance between booths or how crowded each booth is,and uses this to arrange the encounters in order) to generatea service delivery template. After the service componentsare exploited in service delivery, the information obtainedby the oligopoly service providers and customers (benefitand cost) is retrieved and fed back to the Hawk-Dove gamemodule.

3.5 Expectation measurement module

The expectation measurement module measures the likelyperformance of the selected determinants identified by theselection solution module by calculating the customer ex-pectation scores. This process helps the solution selectionmodel choose appropriate service components. The expec-tation measurement module is critical to determining cus-tomer mental state (Hsieh and Yuan 2010), and ensuring theintegrity and effectiveness of the Hawk-Dove game-basedinteraction design mechanism. Owing to space limitations,the measurement model is not described in detail here.

The outputs of the expectations measurement model in-clude the adequate expectation score, desired expectationscore and a list of recommended expectation tactics. Oncethe oligopoly service provider determines actual customerexpectations based on these outputs, it can propose suitableservices to help customers achieve their business goals.Furthermore, the list of expectation tactics, derived from areal-time database, provides a reference. This list of appro-priate expectation tactics can be mapped to specific servicecomponents to affect customer expectations via the solutionselection module. After implementing the expectation tac-tics (service components), the solution selection moduleshould store (the values of expectation variations and infor-mation of providers’ capabilities in a real-time database. Theexpectation measurement model can then use feedback con-trol to reflect actual customer expectations.

3.6 Summary

The Hawk-Dove game-based interaction design mechanismis applied to enable oligopoly service providers to deliver

Table 6 Chosen deter-minant and determinantweight

Determinant Determinantweight

Word of mouth 5.2

Explicit service promise 7

Perceived servicealternative

3.8

Situation factor 2

The sequence of service delivery

Word of mouth(5.2)Explicit service

promise (7)Perceived servicealternative (3.8)

Situation factor(2)

Fig. 8 The determinants exercised in accord with the greedy approach

Inf Syst Front (2014) 16:697–713 705

services that ensure customer satisfaction by managingexpectations. Initially, needs and preferences are specifiedto identify customer values, and enable service providers todesign and deliver effective services during service experi-ence delivery. Customer expectations determine customersatisfaction during service delivery in a dynamic context.This study applies an innovative method for enriching cus-tomer experiences that considers customer mental state. Thiswork also utilizes animal behavior to mimic service inter-actions between customers and providers. After selectingthe determinants of expectation, a greedy algorithm is adop-ted to choose appropriate low cost and high efficiencyservice components to enable customers to build a highperformance eco-system. Therefore, the Hawk-Dove game-based interaction design mechanism can be considered asubstantial and theoretical interface design artifact, compris-ing five core modules and using customer expectation man-agement to answer the research questions. The followingsection tests the novel mechanism using experimentalsimulations.

4 Experiments and results

Design assessment is a key step in demonstrating the utility,efficiency, and quality of a designed artifact using welldefined evaluation methods (Hevner et al. 2004). The ex-perimental method is a design evaluation method used toexamine the designed artifact. This work tests the perfor-mance of the Hawk-Dove game-based interaction designmechanism by conducting four sets of experiments (forchecking the convergence of the mechanism, evaluatingperformance in managing customer expectations, measuringcustomer satisfaction, and evaluating mechanism perfor-mance in terms of total payoff).

4.1 Experiments for the mechanism’s convergence check

The first set of experiments studies the number of iterationsin each simulation of the use of determinants of expectation.When the number of iterations exceeds the threshold, themechanism identifies the determinant of expectation as im-practical. This set of experiments determines the optimalnumber of iterations for causing the simulation results toconverge in a particular direction.

The Hawk-Dove game is designed to achieve an evolu-tionary balance between the various ecology approaches. Inthe exhibition scenario, equilibrium is defined as total cus-tomer profit for the scenario involving a stable oligopolyservice provider. Customer reactions to the service contextare important in designing effective customer encounterinteractions. The total payoff for customers and the oligopolyservice provider, and the strategies available for adoption by

customers, thus are considered indicators of game stability.The experiments involve randomly setting game conditions(such as payoff) in various games. This work simulates tengames to examine the number of iterations within which mostgames reach equilibrium. At the beginning of each game,roughly half of all customers use the Dove strategy, whilethe remainder are assigned to the Hawk strategy. This exper-iment thus defines ten games that 1000 customers and oneoligopoly service provider must adopt within 50 iterations.Table 7 indicates the detailed parameter settings.

In Fig. 9, the Y axis represents the percentage of custom-ers adopting the Hawk and Dove strategies and the X axisrepresents the number of iterations. Customer strategies (inten games – including game1, game2 and so on) maychange significantly during the first ten iterations and thengradually evolve toward a single strategy. Consequently, aspecific strategy is best suited to each game, each with itsown conditions. The total payoffs to the customer andoligopoly service provider also change, as shown inFig. 10. However, when the payoff from one strategy clearlyexceeds that from the other, and when the behavior of theoligopoly service provider changes in response to learning,further iterations may be required to achieve ecologicalstability of both customers and oligopoly service provideragents. As shown in Fig. 10, most games cease to evolveafter an average of ten iterations, and all converge within nomore than 40 iterations.

To efficiently simulate the exhibition context and thebehavior of customers and oligopoly service providers, thisexperiment attempts to optimize the number of iterations inthe game simulation to minimize resource wastage in theremaining game runs. From the experimental results, 10 ~40 iterations can be used as the default number of iterationsin subsequent Hawk-Dove game simulation experiments.

4.2 Experiments related to managing customer expectations

The Hawk-Dove game-based interaction is developed tomanage customer expectations with reference to the objec-tives of oligopolistic service providers. Since customerexpectations vary in dynamic environments, customer ster-eotypes and their preferences should be considered. The

Table 7 The parametersettings Item Value

Number of serviceprovider agents

1 (oligopoly)

Number ofcustomer agents

1000

Number ofgames (strategies)

10

Number of simulationiterations

50

706 Inf Syst Front (2014) 16:697–713

second experiment aims to analyze and manage customerexpectations to refine the interaction mechanism by clarify-ing different customer stereotypes and their preferences.Accordingly, this experiment verifies that the Hawk-Dovegame-based interaction design mechanism is an effectivemeans of managing customer expectations and identifyingcustomer stereotypes. Expectation management is designed tohelp oligopoly service providers increase their expectations ofwhat is adequate and stabilize their expectations of what isdesirable. Consequently, the zone of tolerance determined bythe level of desired expectation is narrowed. This experimentexploits this mechanism to examine the effects of differentcustomer stereotypes on customer expectations.

Five factors (including arrival, capability, effort, request,and subjective preference) influence customer behavior (Frei2006). Table 8 lists various customer stereotypes. Stereotype 1involves moderate effort; stereotype 2 makes a serious effortto bargain with the oligopoly service provider, and stereotype3, which is the opposite of stereotype 1, involves little effort.

This experiment not only employs the above three stereo-types to demonstrate how the Hawk-Dove game-based in-teraction design mechanism can be used to managecustomer expectations in service encounters, but also com-pares the measured adequate and desired expectations of

each customer stereotype. For simplicity, this work assumesthat all customers have the same initial scores associatedwith expectation states (where 3.5 means the adequate ex-pectation while 7 indicates the desired expectation). TheHawk-Dove game-based interaction design mechanismorganizes ten service encounters for each customer and usesthe appropriate determinants in each encounter (as listed inTable 9). Each experiment involves three users of eachstereotype to simulate and record the changing expectationsassociated with each encounter in ten runs. The results inFigs. 11–12 are obtained by averaging the variations in userexpectations for a particular stereotype.

Figure 11 demonstrates that the adequate expectation ofstereotype 1 increases with number of encounters. For ster-eotypes 2 and 3, the mechanism slightly increases adequateexpectations. Regarding stabilization of desired expecta-tions, Fig. 12 also shows that customer expectations arestable for each customer stereotype. Furthermore, with re-spect to the goals of increasing the level of adequate expec-tation and stabilizing the desired level, the experimentalresults indicate that the mechanism was successful, particu-larly for stereotype 1.

Stereotype 1 is frequently considered to represent the targetcustomers of the oligopoly market, who are generally able toindependently select the best service type since they haverequisite domain knowledge (Kuusisto 2008; Rowley 1996).Stereotype 1 can also make more effort to become involved inservice experience delivery (Kuusisto 2008; Rowley 1996).When stereotype 1 customers lack service choices, they mayadopt medium or low expectations. The experimental results

100.00%

80.00%

60.00%

40.00%

20.00%

0.00%

20.00%

40.00%

60.00%

80.00%

100.00%

0 10 20 30 40 50 60

per

cen

tag

e o

f ag

ent

ado

pti

ng

po

pu

lar

stra

teg

y

Iteration

game1

game2

game3

game4

game5

game6

game7

game8

game9

game10

Completely Dove Strategy

Completely Hawk Strategy

Fig. 9 The strategy most customers adopt at each iteration

-500000

-300000

-100000

100000

300000

500000

700000

900000

1100000

0 10 20 30 40 50 60tota

l pay

off

of

bo

th s

ides

Iteration

game1

game2

game3

game4

game5

game6

game7

game8

game9

game10

Fig. 10 The total payoff of both sides at each iteration

Table 8 Three customer stereotypes of variability

Customers Stereotype 1 Stereotype 2 Stereotype 3StereotypesIndicators

Arrival often seldom seldom

Capability high medium low

Effort medium much a little

Request growth relation existence

Subjectivepreference

medium, low medium, low high, medium

Table 9 The parameter settings

Item Value

Number of experimented customer stereotypes 3

Number of customers in each stereotype 3

Initial adequate expectation measurement value 3.5

Initial desired expectation measurement value 7

Number of service encounters 10

Number of runs 10

Inf Syst Front (2014) 16:697–713 707

indicate that oligopoly service providers can provide stereo-type 1 customers with the opportunity to reflect on their needsand become involved in service experience delivery usinginteractions designed specifically to manage expectations.The proposed mechanism can modify the determinants whencustomer expectations move in the opposite direction to thatdesired or exceed the expected range. Accordingly, the Hawk-Dove game-based interaction design mechanism not onlyprovides customers with required services, but also reliablymanages their expectations.

4.3 Experiments for customer satisfaction

The third experiment tests whether the Hawk-Dove game-based interaction design mechanism can deliver services thatmeet customer needs. The Hawk-Dove game algorithm drawson a fight between a hawk and a dove and is used to simulateinteractions between oligopoly service providers and custom-ers. This experiment assesses the effectiveness of the Hawk-Dove game-based interaction design mechanism.

Table 10 shows the payoff matrix of the proposed experi-ment based on the concepts listed in Table 3. The recommen-dation determinant denotes the average number of customerswho recommend a particular service component to others. Ahigher recommendation means more customers enjoy andbenefit from the service. Regarding the costs of experiencinga service, customers must spend time and effort to use aservice component. Accordingly, the customer payoff can be

calculated as the determinant profit minus the determinantcost. Furthermore, the payoff of the oligopoly service provideris the benefit of the determinant minus the exercise cost. Afterinteracting with the oligopoly service provider, customeragents can take various steps to increase profits from serviceencounters, as follows.

& Each agent compares its payoff with those of otheragents to confirm the advantage of the selected strategyrelative to those of neighboring agents. If the compari-son reveals that a different strategy yields a larger ben-efit, then the agents learn that strategy and adopt it in thenext iteration.

& Agents react differently in different simulation runs be-cause they know when their payoff is significantly lessthan those of the other agents. For example, after anagent implements a service strategy (such as by follow-ing the Dove strategy) and compares its payoff withthose of other agents, and it sees that it has paid a highercost than other agents that have used the Hawk strategyin the same service encounter, that agent will reject theservice next time around.

& Each agent goes through the processes of interacting andlearning to determine when a strategy should bechanged. These processes enable agents to maximizeprofit. Customers that use the Dove strategy are willingto accept the service. The evolutionarily stable strategybalances the strategies of the population. When mostagents in the population behave identically, the ecologyof the race can be evolutionarily stable, meaning theselected strategy is evolutionarily stable.

& To evaluate the effect of the Hawk-Dove game-basedinteraction design mechanism on customer satisfaction,this experiment selects appropriate service componentsfor three customers of stereotype 1. In simulating thecustomer and provider behavior, the Hawk-Dove inter-action design mechanism dynamically arranges servicecomponents in the visiting journey based on the chosenexpectation determinant. Maintaining generality, theseexperiments used ten encounters in each journey andestablished seven runs of journey arrangements as listedin Table 11.

These experiments collect the customer satisfaction scorefor each service component and average these scores forservice components arranged in a journey to indicate thedegree of customer satisfaction with each journey. Theexperiments use the degrees of satisfaction of the journeyarrangements to evaluate the performance of the serviceinteraction mechanism in examining if the service compo-nents dynamically arranged in service encounters are appro-priate for customers. In contrast with the effect of the Hawk-Dove game interaction design mechanism, the experimentsalso conduct an approach in which the service components

2

3

4

5

6

7

0 2 4 6 8 10 12

Ad

equ

ate

Exp

ecta

tio

n

encounter

Stereotype1

Stereotype2

Stereotype3

Fig. 11 Adequate level of customer expectation

5.5

6

6.5

7

7.5

8

0 5 10

Des

ired

Exp

ecta

tio

n

encounter

Stereotype1

Stereotype2

Stereotype3

Fig. 12 Desired level of customer expectation

708 Inf Syst Front (2014) 16:697–713

are selected randomly during each service encounter of ajourney. In these experiments, the customer reaction asmeasured by satisfaction score is derived based on analysisof the post-report of AutoTronics Taipei 2009 conducted bythe Taiwan External Trade Development Council (TAITRA)which asked visitors to rate their impression for every ser-vice component, with the associated scores becoming thebasis for assessing the effect of service components.

Figure 13 shows the results of customer satisfaction andevery point represents the average degree of satisfaction foreach journey of the three visitors. The curves indicate thatthe degrees of satisfaction for the journeys managed by theHawk-Dove game based interaction mechanism (around80 % of customer satisfaction) are higher than when theservice components are randomly assigned (around 50 % ofcustomer satisfaction). This indicates that the service com-ponents could effectively fulfill customer needs in deliver-ing customer value and services when using the Hawk-Dovegame interaction design mechanism.

4.4 Experiments for high performance of the H&D gamebased interaction design mechanism

This set of experiments is designed to determine how chang-ing visitor expectations increase stakeholder benefits(corresponding to high performance). This set of experi-ments evaluates the overall performance of the Hawk-Dove game-based interaction design mechanism from amacro perspective. A high-performance ecosystem maxi-mizes benefit to all stakeholders (oligopoly service pro-viders and customers). Surplus value theory (Marx 1952;Carlsson and Davidsson 2002) is used to evaluate eco-system performance by measuring stakeholder perceptionsof value. The equation of surplus value is as follows.

S ¼ P � C þ Vð Þ ð5Þwhere S denotes surplus value; P represents total value; C istotal spending on investments and material; V denotesspending on labor, and (C + V) represents total cost. To

model surplus value of an exhibition, the definition ofsurplus value is extended to include customer and oligopolyservice provider value and costs.

This set of experiments uses pareto optimization (Hochmanand Rodgers 1969; Sawaragi et al. 1985; Jin and Sendhoff2008) to evaluate the performance of the Hawk-Dove game-based interaction design mechanism in the target ecosystem.A pareto-optimal outcome is one in which no-one is madebetter off without making someone else worse off. Customerservices are supposed to be pareto-optimal solutions, obtainedvia the Hawk-Dove game-based interaction design mecha-nism, that consider various interaction design factors, includ-ing profit and cost. The measurement model thus measures themaximum surplus value of all stakeholders achieved using theHawk-Dove game-based interaction design mechanism withthe following objective functions.

Maximum : Sp ¼ Pp � C þ Vð ÞpMaximum : Sc ¼ Pc � C þ Vð Þc ð6Þ

Table 12 lists the definitions of the experimental param-eters. Pp represents the ratio of variation in the quantity ofthe zone to the original quantity of the zone, and the totalcost of implementing service components during the jour-ney is (C + V) p. The mechanism represents the customersatisfaction score for all service components experienced inthe delivery of a service as Pc. Finally, the total time spentinteracting with every service component in the serviceencounters is represented by (C + V) c. The Hawk-Dovegame-based interaction design mechanism arranges 25 cus-tomers. The customer satisfaction score for each encounter

Table 11 The parame-ters setting Item Value

Number of customers 3

Number of journeys 7

Number of encounters 10

0102030405060708090

100

0 2 4 6 8

Deg

rees

o f

cu

sto

mer

sa

tisf

acti

on

in s

ervi

ce d

eliv

ery

(%)

Nmubers of journeys

HDGmechanism

randomlyarranged

Fig. 13 Customer satisfaction of each journey

Table 10 The payoffmatrix Customer Dove Hawk

Provider

DoveRecommend�EffortþCost

3 ; Value�EffortþCostost3

� � Recommend�Effort2 ; Value�Cost�Effort

2

� �

HawkRecommend�Effort�Cost

2 ; Value�Cost2

� �(0,0)

Inf Syst Front (2014) 16:697–713 709

minus the effort made by the customer to interact with aservice component is Sc. The variation in the size of the zoneas a proportion of the original quantity of the zone minus thecost of implementing all service components is Sp. Twenty-five customers with randomly assigned Sc and Sp are consid-ered to model all possible decisions made by all other possiblemechanisms. The random rules are taken primarily from do-main experts and practical reports. Comparing these twogroups of customers in terms of overall surplus value clarifiesthe effectiveness of a Hawk-Dove game interaction designmechanism in creating a high-performance eco-system.

Based on the results of the ecosystem evaluation exper-iment (as shown in Fig. 14), each point represents a visitor.(A red square represents a customer served using the pro-posed mechanism and a blue rhombus indicates a customerwith randomly assigned values of Sc and Sp). Except for asingle square point, square points outperform rhombuspoints. If any points in the rhombus group exhibit qualityperformance of Sp, those same points must be associatedwith poor performance of Sc. In conclusion, the simulationresults clearly indicate that the Hawk-Dove game interactiondesign mechanism can not only be used to build a well-performing ecosystem but can also benefit customers andproviders by providing pareto-optimality.

4.5 Discussion

Service interaction design is important in determining theperformance of service experience delivery. This study usesa systematic and quantitative approach, the Hawk and Dovegame, to model service interactions between oligopoly ser-vice providers and customers where customer expectationsare dynamically but systematically managed. Restated, ser-vice providers consider customer expectation managementduring their design of the service interaction aspects ofservice experience delivery.

Traditionally, service providers have used the surveymethod (questionnaires) to determine customer expecta-tions, but only after service completion. Understanding cus-tomer expectations in real time in service contexts isparticularly difficult for service providers who deliver serv-ices in real time. This study presents a service interactionmechanism, which involves customer expectations based onelicited service-related requirements and needs. Service pro-viders can thus provide required services based on thisinformation. Customers can choose to receive availableservices and modify their requirements based on what theyreceive. Service providers thus can dynamically providecustomers with satisfactory services by immediately analyz-ing customer feedback to clearly understand customerexpectations in real-time service contexts. Consequently,both service providers and customers can achieve their goalsby participating in interactive service activities.

As indicated above, customers can be grouped into nu-merous types with different levels of expectation. Serviceproviders can deliver suitable services when they clearlyunderstand customer expectations. For instance, if a serviceprovider knows that a customer has low expectations offrontline services, they can reduce the number of serviceemployees and devote resources to preparing other usefulservices for the customer. More generally, service providerscan dynamically but effectively allocate existing resourcesto provide customers with more suitable services. Serviceproviders do not need to provide customers with all services,necessarily increasing the cost of service provision. Hence,the proposed service interaction mechanism helps serviceproviders who can flexibly modify their service-relatedresources provide services that meet customer expectationsin a cost-effective manner.

One core feature of S-D logic is the co-creation of valueby service providers and customers during service experi-ence delivery. The proposed service interaction mechanismcan be considered a platform on which service providers andcustomers can co-create value. When service providers pro-vide services to customers via the proposed service interac-tion mechanism, customers can participate in the serviceactivities to improve their own service experience. Custom-ers interact with both service providers and other customers

0

500

1000

1500

2000

0 200 400 600 800 1000

SP

SC

not served

served with mechanism

Fig. 14 Performances of the proposed mechanism in surplus value

Table 12 The definitions of parameters in surplus value

Parameter Definition

Sp Surplus value of service providers

Pp The proportion of variation of the zoneof tolerance managed after Hawk-Dovegame based interaction design mechanismto the original zone of tolerance (z-z’)/z

(C + V) p Providers’ participation

Sc Surplus value of customers

Pc Real customers’ responses

(C + V) c Customers’ participation

710 Inf Syst Front (2014) 16:697–713

to share the experiences and knowledge gained during ser-vice experience delivery via the proposed service interactionmechanism. Both service providers and customers createvalue by participating in service activities. Therefore, cus-tomer involvement is essential to successful service experi-ence delivery.

Since the proposed service interaction mechanism isdesigned to be trained and improved by the input of real-world data, the number of training sets and parametersrequires careful consideration. In this work, experts on theexhibition industry helped specify suitable training sets andparameters for the experimental simulations. For example,the numbers of training iterations required to ensure theconvergence of the HDG method and the customer expec-tation measurement model are 40 runs and ten runs, respec-tively. However, since the settings were only impreciselyspecified, to imitate the behavior of real-world exhibitionvisitors, the parameters of the proposed service interactionmechanism must be further tuned by increasing the inputtedvisitor data. Although continuous learning and searching forsuitable parameters in real-time service contexts is necessary,the constraints imposed on service providers by resourcelimitations and costs are the main factors in terminatingtraining iterations.

5 Implications and conclusions

In the experience economy, service providers must designand deliver good customer service based on understandingcustomer needs. Service providers can increase customerloyalty and thus profits by developing their capacity forexperiential design (Pullman and Gross 2004). Interactionsinvolved in service delivery are designed to create function-al, purposeful, compelling, and memorable customer expe-riences (Meyer and Schwager 2007). Additionally, customerinvolvement can be considered important to the co-creationof value with service providers. In an oligopoly market,while service providers dominate, they must consider cus-tomer satisfaction and develop new services. However, busi-nesses face the dilemma of providing additional services tosatisfy customers while maintaining high profits (the pursuitof which can risk customer discontent). This predicamentbecomes serious, particularly in oligopoly markets, whichconstitute a value-bounded context for customers (owing tothe low motivation of oligopoly service providers to investadditional resources in ensuring customer happiness). Toresolve this issue in an oligopoly market, this study presentsthe Hawk-Dove game interaction design mechanism to in-crease value enjoyed by service providers and customers ina value-bounded context using customer expectation man-agement and service interaction design. This study not onlyexamined the factors that influence customer satisfaction

and the quality of the service experience, but also estab-lished a mechanism based on expectation theory and theHawk-Dove game for delivering a high-quality customerexperience.

The Hawk-Dove game-based interaction design mecha-nism not only efficiently manages customer mental states tomaximize customer satisfaction, but also dispatches appro-priate services and provides value to customers in a value-bounded context. Based on analysis of the experimentalresults, the following conclusions can be drawn regardingthe Hawk-Dove game-based interaction design mechanism.

Service experience delivery involves service encountersthrough which interactions between service providers andtheir customers can be shaped into interactive artifacts man-aging customer expectations and delivering suitable services.Service interaction design involves optimizing customer inter-actions with services so theymatch customer expectations andyield satisfactory service experiences.

& The mechanism can help oligopoly service providersdynamically interact with customers to effectively matchcustomer expectations and ensure satisfactory serviceexperiences, achieving provider business goals in avalue-bounded context.

& The mechanism can still deliver satisfactory customerservices tailored to various preferences and needs.

& The mechanism can enable oligopoly service providersand customers to co-create value and design a high-performance ecosystem using a theoretical and system-atic approach.

The Hawk-Dove game-based interaction design mecha-nism can alter the managerial perspective of an oligopolyservice provider to increase the emphasis on customer men-tal states and co-create value with customers, rather thanmerely considering profit in the context of service experi-ence delivery. Oligopoly service providers can consideralternative methods of serving customers based on continu-ous interaction design.

The experimental results also help oligopoly service pro-viders realize that effective service interactive design canincrease both customer satisfaction and profits. Oligopolyservice providers must manage limited resources and under-stand context to deliver appropriate services and ensurecustomer satisfaction - particularly for stereotype 1 custom-ers. Additionally, oligopoly service providers can also ben-efit from increased opportunities to involve customers,enabling them to contribute co-creating value. Oligopolyservice providers can still significantly improve profits with-out changing their business methods and service situations.When the proposed mechanism is used to realize effectiveand interactive service design, oligopoly service providerscan attract more customers, increase their market share andimprove service ecosystem performance.

Inf Syst Front (2014) 16:697–713 711

However, this work still suffers limitations. For instance,to test the feasibility of the proposed mechanism, manyexperimental simulations were conducted, but a real fieldtest still needs to be performed to validate the simulationresults and establish the mechanism as a practical system.Accordingly, environmental variations and the results ofcustomer feedback should be analyzed to iteratively im-prove the proposed approach, thus contributing to designscience. Additionally, numerous directions for possible fu-ture research exist. First, the learning efficiency of theproposed mechanism and ESS recognition mechanism inthe Hawk-Dove game requires further improvement. Thelearning rule could be enhanced by adjusting the mechanismused for neighbor selection and by adding saltant (or mutant)factors to prevent its evolution into an absolute ecologicalstrategy because such an absolute strategy would mean anecological agent was unable to make strategy adjustment.Second, since the Hawk-Dove game-based interaction designmechanism exploits numerous preference indicators involvingthe determinants of expectations, it is necessary to definesuitable rules to enable consideration of more indicators inselecting appropriate determinants. Finally, future investiga-tions should consider system disturbance. The mechanism canbe made to more closely match real-world situations by mod-ifying it to receive immediate user (customer and provider)feedback and reduce information interference

Open Access This article is distributed under the terms of the Crea-tive Commons Attribution License which permits any use, distribution,and reproduction in any medium, provided the original author(s) andthe source are credited.

References

Berninghaus, S. K., & Ehrhart, K. M. (2003). The power of ESS: Anexperimental study. Journal of Evolutionary Economics, 13(2), 161–181.

Carlsson, B., & Davidsson, P. (2002). Surplus values in informationecosystems. International Journal of Information Technology andDecision Making, 1(3), 559–571.

Durham, Y., McCabe, K., Olson, M. A., Rassenti, S., & Smith, V.(2004). Oligopoly competition in fixed cost environments. Inter-national Journal of Industrial Organization, 22, 147–162.

Frei, F. (2006). Breaking the trade-off between efficiency and service.Harvard Business Review, 84, 92–101.

Grundman, S., Komarkova, L., & Rychtar, J. (2009). A Hawk-Dovegame in finite kleptoparasitic populations. Journal of Interdisci-plinary Mathematics, 12, 181–201.

Haeckel, S. H., Carbone, L. P., & Berry, L. L. (2003). How to lead thecustomer experience. Marketing Management, 12(1), 18–24.

Hendricks, K., &McAfee, R.P. (2000). A theory of bilateral oligopoly withapplications to vertical mergers, mimeo, University of Texas, Austin.http://www.mcafee.cc/Papers/Bin/Vertical/Concentration.pdf.Accessed 26 June 2010.

Hendricks, K., & McAfee, R. P. (2010). A theory of bilateral oligopoly.Economic Inquiry, 48(2), 391–414.

Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design sciencein information systems research. MIS Quarterly, 28(1), 75–105.

Hochman, H. M., & Rodgers, J. D. (1969). Pareto Optimal Redistri-bution. The American Economic Review, 59(4), 542–557.

Holmlid, S. (2007). Interaction design and service design: Expanding acomparison of design disciplines. http://www.nordes.org/data/uploads/papers/143.pdf. Accessed 9 July 2010.

Hsieh, Y.H., & Yuan, S.T. (2009). Exquisite technology for serviceinnovation using ubiquitous exhibition service system as exam-ple. 2009 INFORMS International Conference on Service Sci-ence, August 6–8 2009, Hong Kong.

Hsieh, Y. H., & Yuan, S. T. (2010). Design of the customer expectationmeasurement model in dynamic service experience delivery. PacificAsia Journal of the Association for Information Systems, 2(3),1–19.

Jin, Y., & Sendhoff, B. (2008). Pareto-based multi-objective machine learn-ing:An overview and case studies. IEEETransactions on Systems,Man,and Cybernetics, Part C: Applications and Reviews, 38(3), 397–415.

Johnston, R. (1995). The zone of tolerance: Exploring the relationshipbetween service transactions and satisfaction with the overall service.International Journal of Service Industry Management, 6(2), 46–61.

Kokko, H., Lopez-Sepulcre, A., & Morrell, L. J. (2006). From Hawksand Doves to self-consistent games of territorial behavior. TheAmerican Naturalist, 167(6), 901–912.

Kurtz, D. L., & Clow, K. E. (1993). Managing Consumer Expectationsof Services. Journal of Marketing Management, 2(2), 19–25.

Kuusisto, A. (2008). Customer roles in business service production-implications for involving the customer in service innovation, Lap-peenranta University of Technology. http://akseli.tekes.fi/opencms/opencms/OhjelmaPortaali/ohjelmat/Serve/fi/Dokumenttiarkisto/Viestinta_ja_aktivointi/Julkaisut/Kuusisto_Paeaelllysaho_raportti.pdf.Accessed 18 August 2010.

Lewis, B., & Entwistle, T. (1990). Managing the service encounter: Afocus on the employee. International Journal of Service IndustryManagement, 1(3), 41–52.

Marx, K. (1952). Theories of Surplus Value. New York: InternationalPublishers.

Meyer, C., & Schwager, A. (2007). Understanding customer experi-ence. Harvard Business Review, 85(2), 117–126.

Ojasalo, J. (2001). Managing customer expectations in professionalservices. Managing Service Quality, 11(3), 200–212.

Parasuraman, A., Berry, L. L., & Zeithaml, V. A. (1991). Understand-ing customer expectations of service. Sloan Management Review,32(3), 39–48.

Pine, B. J., & Gilmore, J. H. (1999). The Experience Economy. Boston:Harvard Business School Press.

Pitt, L. F., & Barbara, J. (1994). Management of customer expectationsin service firms: A study and a checklist. Service IndustriesJournal, 14(2), 170–189.

Pullman, M., & Gross, M. (2004). Ability to experience design ele-ments to elicit emotions and loyalty behaviors. Decision Sciences,35(3), 551–578.

Ren, F. (2007). Analysis of the double-oligarch’s output-competitiondecisions in relevant region. Asian Social Science, 3(5), 45–49.

Rowley, J. E. (1996). Customer compatibility management: An alterna-tive perspective on student-to-student support in higher education.International Journal of Educational Management, 10(4), 15–20.

Rust, R. T., Inman, J. I., Zahorik, J., & Jia, A. (1999). What you don’tknow about customer-perceived quality: The role of customerexpectation distributions. Marketing Science, 18(1), 77–92.

Saffer, D. (2005). The Role of Metaphor in Interaction Design, Thesis,Carnegie Mellon University, Pittsburgh.

Sawaragi, Y., Nakayama, H., & Tanino, T. (1985). Theory of Multi-objective Optimization. New York: Academic.

712 Inf Syst Front (2014) 16:697–713

Schurter, T., & Towers, S. (2006). Customer Expectation Management:Success without Exception. Florida: Meghan-Kiffer Press.

Smith, J. M., & Parker, G. A. (1976). The logic of asymmetric contests.Animal Behavior, 24(1), 159–175.

Smith, J. M., & Price, G. R. (1973). The logic of animal conflict.Nature, 246, 15–18.

Tan, W. (2006). The effects of taxes and advertising restrictions on themarket structure of the U.S. cigarette market. Review of IndustrialOrganization, 28(3), 231–251.

Walker, J., & Baker, J. (2000). An exploratory study of a multi-expectation framework for services. Journal of Service Marketing,14(5), 411–431.

Zeithaml, V. A., Berry, L. L., & Parasuraman, A. (1993). The natureand determinants of customer expectations of service. Journal ofthe Academy of Marketing Science, 20, 1–12.

Zhao, K., Xia, M., Shaw, M. J., & Subramaniam, C. (2009). Thesustainability of B2B e-marketplaces: Ownership structure, mar-ket competition, and prior buyer-seller connections. DecisionSupport Systems, 47(2), 105–114.

Yen-Hao Hsieh is an Assistant Professor of Department of InformationManagement in College of Business and Management at Tamkang Uni-versity in Taiwan. He received his Ph.D. in Management InformationSystems from National Chengchi University. His research interests in-clude Service Science, Service Experience Design, Customer ExpectationManagement.

Soe-Tsyr Yuan is the Director of Service Science Research Center anda Professor of Department of Management Information Systems inCollege of Commerce at National Chengchi University in Taiwan.She received her Ph.D. in Computer Science from Oregon State Uni-versity in USA. Her research interests include Service Science, ServiceDesign, Intelligent Agents, Data Mining, E/M-Commerce.

Hsiao-Chen Liu received her Master degree from Department ofManagement Information Systems in College of Commerce at NationalChengchi University in Taiwan. She currently works in ChunghwaTelecom Company in Taiwan.

Inf Syst Front (2014) 16:697–713 713