information systems and technology sourcing strategies of e-retailers for value chain enablement

KCCeIsM

arrieeniRiffgi

ARTICLE IN PRESSG ModelPEMAN-828; No. of Pages 18

Journal of Operations Management xxx (2013) xxx–xxx

Contents lists available at ScienceDirect

Journal of Operations Management

j o ur na l ho mepage: www.elsev ier .com/ locate / jom

nformation systems and technology sourcing strategies of e-Retailersor value chain enablement

uliana Y. Tsai, T.S. Raghu ∗, Benjamin B.M. Shaoepartment of Information Systems, W.P. Carey School of Business, Arizona State University, Tempe, AZ 85287, United States

r t i c l e i n f o

rticle history:vailable online xxx

eywords:ontingency theoryomplementary sourcing-Retail value chainnformation systems and technologyourcing strategyake versus buy

a b s t r a c t

In the e-Retail industry, a well-designed IT infrastructure is essential in creating a tightly integratedvalue chain and delivering high quality service. With intense competition for market share and profits,information systems and technology (IST) sourcing decisions are becoming increasingly important to e-Retail firms to support continued growth and market responsiveness. Drawing on the contingency theory,we examine organizational and environmental factors that influence an e-Retailer’s IST sourcing strategyof make versus buy in enabling its value chain activities, and we also look at firm-level performance impactsof IST sourcing decisions that involve bundling across value chain activities. We test the proposed modeland hypotheses using a panel data set of 307 firms over the period of 2006–2010. The results showthat firms that make transformative IT investments tend to source a smaller portion of IST for their e-Retail value chain activities than firms that pursue automate or informate as their strategic role for ITinvestment. Capabilities are positively associated with IST sourcing. Firms experienced in e-Retail aremore likely to build rather than buy their IST. In addition, we find mimicking behavior for IST sourcing

. Introduction

The e-Retail industry has grown rapidly over the last five yearsnd is projected to continue its upward trend. In March 2010, For-ester Research forecasted a 10% yearly growth rate for onlineetail sales over the next 5 years with e-Retail sales amount-ng to $249 billion by 2014. Technology is an inseparable part of-Retailers’ value chain activities. E-Retailers constantly seek tonhance consumer experience by updating their virtual stores withew features and capabilities, such as mobile commerce, dynamic

maging, social networking, site personalization, and videocasts.ecent industry surveys show 52.4% of respondents making more

nvestments in their e-Commerce platforms and the look andeel of their websites, indicating that the focus on building new

Please cite this article in press as: Tsai, J.Y., et al., Information systemsenablement. J. Operations Manage. (2013), http://dx.doi.org/10.1016/

eatures and capabilities and the resulting demand for technolo-ies have created an “arms race” in the e-Retail industry. Fornstance, the 2011 Internet Retailer Conference and Exhibition, the

∗ Corresponding author. Tel.: +1 480 965 8977; fax: +1 480 727 0881.E-mail addresses: juliana.tsai@asu.edu (J.Y. Tsai),

aghu.Santanam@asu.edu (T.S. Raghu), ben.shao@asu.edu (B.B.M. Shao).

largest annual e-Commerce event, chose the theme, “E-CommerceShifts into Overdrive, the Race is On” and focused on e-Retailers’demand for the latest “off-the-shelf” technology solutions frome-Commerce vendors (Love, 2011). Although buying off-the-shelfsystems speeds up acquisition of new features, some e-Retailershave expressed concerns about the increased commoditization andreduced differentiation of store-front features.

For some e-Retailers (e.g., Amazon, Netflix, etc.), technologysupports core competency. For others, technology and related infra-structure remain peripheral to the core business of merchandizingand selling. As a result, e-Retailers face the classic decision prob-lem of whether to make or buy technology solutions. Technologysourcing as a research topic has received very little attention inInformation Systems (IS) and Operations Management (OM) lit-erature. Although extant OM literature has extensively studiedsourcing strategies and decisions in manufacturing, IS literaturehas mainly focused on IT services and infrastructure and soft-ware development contexts to examine sourcing strategies and

and technology sourcing strategies of e-Retailers for value chainj.jom.2013.07.009

decisions (Handfield et al., 1999; Leiblein et al., 2002; Wadeand Hulland, 2004). Both literature streams draw from similartheoretical frameworks and enrich our understanding of makeversus buy decisions by separately exploring distinct contingencies.

ING ModelO

2 ions M

OrbIdBes“tcao“csec

eeelbbtCit(oerttsc

i(fo1faTpcqs

stopmetf

ARTICLEPEMAN-828; No. of Pages 18

J.Y. Tsai et al. / Journal of Operat

utsourcing, a term most commonly used by practitioners and ISesearchers, represents a special case of sourcing in which make oruy decisions pertain to the procurement of business process andT infrastructure services as well as to the acquisition of softwareevelopment expertise from vendors (Quinn and Hilmer, 1994;ardhan et al., 2006, 2007; Loh and Venkatraman, 1992a; Whitakert al., 2010). However, software product sourcing differs from out-ourcing in that a firm seeks to build its capabilities by usingoff-the-shelf” technology solutions from external vendors ratherhan to turn over its IST functions to another firm. Thus, one canonsider technology sourcing as similar to the sourcing of productsnd components in manufacturing. In comparison to the sourcingf products and components in manufacturing contexts, sourcingoff-the-shelf” technologies often poses distinct challenges due toonstantly evolving compatibility, integration, and interoperabilitypecifications. As a result, make or buy decisions can either decel-rate or accelerate e-Retailers’ ability to acquire new features andapabilities.

Drawing on the contingency theory, we propose a model toxplore the organizational and environmental factors that influ-nce e-Retailers’ IST sourcing strategy of make versus buy innabling their value chain activities and to examine the firm-evel performance impacts of IST sourcing decisions that involveundling across value chain activities. Our model opens up the blackox of internal firm operations by introducing a granular view ofhe IST sourcing decisions hitherto unexplored in the literature.ontingency theory states that a firm’s choices are dependent upon

ts internal and external environments and stresses the impor-ance of the alignment between organizational setting and strategyBurns and Stalker, 1961; Lawrence and Lorsch, 1986). In devel-ping our contingency framework as intended contribution to thextant literature, we follow the typical approach of contingencyesearch outlined in Sousa and Voss (2008) by identifying impor-ant contingent variables for different contexts and by examininghe alignment of internal decisions and external context. We con-ider multiple dimensions of context and different scenarios ofontingency in constructing our research model.

Our primary objective is to enhance knowledge of IST sourc-ng among e-Retailers, a topic which has received limited attentionKishore et al., 2004). Extant research on sourcing has mainlyocused on antecedents of the sourcing decisions but not so muchn their performance effects (Kauffman and Tsai, 2009; Smith et al.,998). In this regard, our study develops new knowledge on the per-ormance impacts of IST sourcing decisions by explicitly identifyingnd aggregating sourcing decisions across value chain activities.his enables us to examine complementarities between differentarts of the value chain from a sourcing perspective in the e-Retailontext. This research specifically addresses the following researchuestions to better understand the emerging issues of IST sourcingtrategies among e-Retailers:

How do organizational and environmental characteristics affectmake versus buy IST sourcing decisions for particular activities inthe e-Retail value chain?How do complementarities in IST sourcing choices impact firmperformance?

Our findings, based on a panel analysis of e-Retail firms’ ISTourcing decisions and performance, reveal that firms that makeransformative IT investments tend to source a smaller portionf IST for their e-Retail value chain activities than do firms thatursue automate or informate as the strategic role for IT invest-

ent. Capabilities are positively associated with IST sourcing. Firmsxperienced in e-Retail activities are more likely to build ratherhan buy their IST, and evidence exists of mimicking behavioror IST sourcing among firms in the same merchandizer category.

PRESSanagement xxx (2013) xxx–xxx

Further, our findings reveal partial evidence of performance effectswhen alignment occurs between IT strategic role and IST sourcingdecisions. We find that IT strategic role is strongly associated withgrowth metric, whereas sourcing decisions predominantly impactoperational performance measures. E-Retail firms face trade-offsin adopting an overarching buy approach across the value chain;that is, any benefits of synergies between marketing and sales haveto be weighed against the negative effects of sourcing logistic andoperations technologies.

2. Theoretical model and research hypotheses

2.1. E-Retail value chain and IT infrastructure

To set the context and scope for this research, we begin witha typological overview of e-Business, e-Commerce and e-Retail,and discuss the role of technology in e-Retailers’ value chain.E-Business is the coalescence of the Internet and supply chain inte-gration and captures all processes involving customers, employees,vendors, and business partners (Johnson and Whang, 2002). E-Commerce, on the other hand, is a subcategory of e-Business andrefers to the purchasing, selling, and exchanging of goods andservices over the Internet. It includes business-to-business (B2B),business-to-consumer (B2C), consumer-to-business (C2B), andconsumer-to-consumer (C2C) transactions. E-Retail, also known aseTail, focuses on the selling of retail goods and services on the Inter-net to consumers and refers solely to business-to-consumer (B2C)transactions of e-Commerce.

The many types of e-Retail firms range from Web only e-Retailers to traditional “brick-and-mortar” retailers that offeronline storefronts (i.e., “click-and-mortar”). By transitioning to aclick-and-mortar business approach and creating stronger cooper-ation across channels, retail chains, catalog/call centers, and brandmanufacturers are able to achieve benefits including cost savings,improved differentiation, enhanced trust, and market extensions(Steinfield et al., 2002). This study focuses on how e-Retailers’ ISTsourcing strategies enable the primary value chain activities ofinput logistics, operations, output logistics, marketing, and sales.

Several systems and technologies are required to enable ane-Retailer’s value chain activities. Examples include customerrelationship management, business intelligence, supply chain man-agement, content management, e-Commerce platform, and Webanalytics, among others. One can capture the complex enterprise ofe-Commerce using a three-level hierarchical framework that placesinfrastructure at the lowest level, followed by services at the mid-level, and products and structures at the top (Zwass, 1996). Theinfrastructure consists of all the hardware, software, databases, andtelecommunications required to establish the technological infra-structure for e-Commerce. The services level, which provides thebusiness with the infrastructure for e-Commerce, includes securemessaging and service enablement. The products and structureslevel of e-Commerce is focused on enabling the value chain activi-ties of the e-Retailer (Gunasekaran et al., 2002).

We utilize Porter’s (1985) value chain framework to developa conceptual understanding of interrelationships in the e-RetailerIST infrastructure. Fig. 1 illustrates the value chain model proposedby Porter (1985). Highlighted in gray are the value chain activitiescovered in our study. In essence, the value chain is “a model thatdescribes a series of value-adding activities connecting a company’ssupply side (raw materials, inbound logistics, and production pro-cesses) with its demand side (outbound logistics, marketing, and

sales)” (Rayport and Sviokla, 1996). The five primary activities ofthe value chain include inbound logistics, operations, outboundlogistics, marketing and sales, and service. Inbound logistics referto activities associated with receiving, storing, warehousing, and

ARTICLE IN PRESSG ModelOPEMAN-828; No. of Pages 18

J.Y. Tsai et al. / Journal of Operations Management xxx (2013) xxx–xxx 3

logisti

icatoatafspr

tpoeaepftvti

clmFStitbeeisv(tc

Fig. 1. Technology development for

nventory control of input materials. Operations include such value-reating activities as manufacturing, production, packaging, andssembly, which transform inputs into the final product. Activi-ies that focus on distributing the finished products to buyers, likerder fulfillment, make up outbound logistics. Marketing and salesctivities help buyers to purchase the product and include adver-ising, promotion, channel relations, and pricing. Finally, servicectivities—e.g., customer support, installation, and repair—are per-ormed to maintain and enhance the value of the product after theales. These primary value chain activities are facilitated by the sup-ort activities of procurement, technology development, humanesource management, and firm infrastructure.1

Porter (1985) introduces the value chain as a model to identifyhe sources of competitive advantage that enable a firm to out-erform its competitors using, for instance, technology to carryut primary and support activities better, faster, and cheaper. For-Retailers, the possible use of technology to enable competitivedvantage could mean attracting and retaining customers, low-ring coordination cost with producers, or reducing the cost ofhysical distribution to buyers. Technology serves as the backboneor value chain success because it is the touch point for all activi-ies. Not only does the back-end IT infrastructure support the entirealue chain by coordinating all activities, but all participants inhe value chain, ranging from suppliers to retailers to customers,nteract with one another through these technologies.

The substantial impacts of IT architecture and decisions on valuehain activities in e-Commerce, especially from supply chain andogistics perspectives, have been duly noted in the literature in

any different e-Commerce contexts (Clark and Hammond, 1997;rohlich and Westbrook, 2001, 2002; Evans and Wurster, 1999;ubramani, 2004). For example, Keen and Williams (2013) highlighthe importance of delivering a robust value architecture focused onnnovation and sustained growth for digital businesses. In additiono the impacts on supply chain activities, activities related to theusiness-to-consumer (B2C) portion are equally important (Smitht al., 2000; Gunasekaran and Ngai, 2005). Utilizing technology tongage consumers in online transactions can bring about variousntangible benefits. For example, creating an e-Commerce site thatupports customization, contact interactivity, community, and con-enience can foster e-loyalty of consumers in the B2C marketplace

Srinivasan et al., 2002). Given the extensive and complex role ofechnology, many firms will struggle to develop core competen-ies in technology enablement of value chain activities. As a result,

1 Procurement refers to the function of acquiring raw materials and other inputssed in the firm’s value chain. Technology development captures process automa-ion and other technology development used to support the value chain activities.ctivities of recruiting, hiring, and training of employees make up human resourceanagement. The firm infrastructure consists of activities such as finance, legal,

ccounting, and quality management (Porter, 1985).

cs, operations, marketing, and sales.

off-the-shelf e-Commerce technology solutions have emerged inrecent years to support value chain activities. Thus, firms in the e-Retail industry face strategic sourcing choices from a technologyperspective.

2.2. Make versus buy for IST

The decision to make or buy materials, assets, or solutions is aclassic acquisition problem, which has been explored in multiplecontexts. The fundamental theoretical argument guiding a firm’smake versus buy decision arises from the transaction and coor-dination costs associated with that decision. For example, buyingenables a firm to transfer such risks as integration and project fail-ure costs to large enterprise software vendors that can more easilyabsorb them (Kauffman and Tsai, 2009). Extant literature recog-nizes that transaction and coordination costs manifest themselvesfrom the contingencies of organizational and environmental con-texts. Contingency analysis suggests that the costs of managing afirm are subject to the nature of the environment in which firmscontinuously make tradeoffs between the benefits associated witha corporate strategy and the bureaucratic costs tied to its imple-mentation (Jones and Hill, 1988). As a result, organizational andenvironmental contingencies tend to act as moderators of transac-tion and coordination costs. For instance, Kogut and Zander (1992)propose that a firm should evaluate three elements in its decision-making process for make versus buy: (1) the firm’s present abilityto perform the task, (2) the learning curve involved in develop-ing specific capabilities, and (3) the value of these capabilities tocreate new markets for the firm. Therefore, the experience of afirm and its need to acquire features and capabilities should be fac-tored into its evaluation process for make versus buy. The need toreduce time to market also motivates firms to buy rather than make(Handfield et al., 1999). Buying systems solutions confers such ben-efits as the ability to leverage a vendor’s expertise, greater flexibilityin acquiring new technologies and systems, avoidance of coordina-tion inefficiencies, compression of product development lifecycletime, and the sharing of risks related to technology developmentsamong a firm’s suppliers (Leiblein et al., 2002; Quinn and Hilmer,1994). However, the decision to source from an outside vendorcan introduce its own unique risks, including the loss of criticalskills, the development of wrong skills, a decline in cross-functionalcapabilities, and the handover of control to the vendor.

Risks associated with transaction and coordination costs fac-tor into competency related arguments as well. Quinn and Hilmer(1994) recommend that firms invest their resources in their corecompetencies and outsource non-core activities for which theyhave neither a critical need nor special capabilities. Firms vary in

terms of their competency focus and hence differ in their choicesof which business solutions to outsource and which to insource.Quinn and Hilmer (1994) further note that most companies targetthe two to three value chain activities they deem most critical to

ING ModelO

4 ions M

fcphtacrobti

ddFsKtrbe(aa

opzatr2JficrO2

do(utaStcm(t

uture success. Teece (1986) presents multiple factors for a firm toonsider when deciding if it should integrate or contract for com-lementary assets related to technological innovation. The factorse recommends include the strength of the appropriability regime,he necessity of specialized assets for profitable commercialization,nd the positioning of other relevant players like imitators andompetitors.2 There is evidence that firms may elect to concur-ently source, i.e., to simultaneously make and buy similar goodsr services (Parmigiani, 2007). This suggests that the make versusuy decision is not a simple dichotomous choice but lies on a con-inuum, especially when all value chain activities are weighed formplementation.

Outsourcing literature specifically studies make versus buyecisions from the perspectives of services, processes and softwareevelopment, and highlights the continuum approach to sourcing.or example, Lacity and Willcocks (1998) conduct in-depth casetudies on the IT sourcing of firms in the United States and Unitedingdom and propose a focus on selective outsourcing rather than

otal outsourcing or total insourcing. Selective sourcing, however,equires a coordinated approach to leverage complementaritiesetween activities (Nolan and McFarlan, 2005). Strategic IT lead-rship of a firm could be a key factor in its IST sourcing decisionsFeeny and Willcocks, 1998). In essence, selective sourcing calls for

closer examination of the contingencies associated with the firmnd its technology contexts.

.3. Contingency factors and the IST sourcing model

The contingency theory argues that a fit between strategy andrganizational settings is important and that firms achieve bettererformance when their strategies are aligned with their organi-ational structures and environmental conditions (Venkatramannd Prescott, 1990). Contingency frameworks in sourcing litera-ure have explored the links between internal integration, supplierelationships, internal structures, and practices (Sousa and Voss,008; Flynn et al., 2010; Guide et al., 2003; Germain et al., 2008;ohnson and Whang, 2002). In the IS literature, factors includingrm size, structure, maturity, resources, knowledge, IS sophisti-ation, technology, risk, and environment have all been studied inelation to the contingency perspective (Raymond, 1990; Weill andlson, 1989; Osei-Bryson and Ngwenyama, 2006; Whitaker et al.,010).

Contingency theory literature also recognizes the multi-imensional nature of contingencies and the associated diversityf impacts on organizational configurations and decision-makingGresov, 1989; Sambamurthy and Zmud, 1999). Looking at taskncertainty and horizontal dependence, Gresov (1989) identifieshree possible scenarios for contingencies: reinforcing, conflicting,nd dominating.3 Building upon the study of Gresov (1989),ambamurthy and Zmud (1999) examine contingency factors andheir impacts on IT governance. They find that for a given set ofontingency factors, the impact of individual contingency forces

ay be amplified, dampened, or overridden. Andres and Zmud2001–2002) explore how the alignment and fit of conflicting con-ingencies impact software design and coding outcomes.

2 Appropriability regime is defined as “environmental factors, excluding firm andarket structure, that govern an innovator’s ability to capture the profits generated

y an innovation” (Teece, 1986, p. 287).3 Reinforcing contingencies refer to the situation where all the salient contingen-

ies result in similar influences in IT decision-making. In instances where salientontingencies create contradictory dispositions toward IT decision-making respon-ibilities, conflicting contingencies are observed. Dominating contingencies occurhen a single contingency appears to override other contingencies so only a single

actor appears to be at work.

We utilize the basic ideas of conflicting and reinforcing con-tingencies in considering the impacts of IST sourcing decisions onperformance. According to Gresov (1989), conflicting contingenciesarise in two situational types: Type 2 describes a situation in whichunits experience low task uncertainty and high horizontal depend-ence, and Type 3 refers to units that encounter high task uncertaintyand low horizontal dependence. In e-Retail value chains, individualvalue chain activities may be associated with different types of con-tingencies. For example, a Type 2 conflicting contingency is morelikely for logistics and operations in e-Retail firms due to low taskuncertainty but a high level of horizontal dependence. In essence,many of the tasks in logistics and operations (identified in Table 2)are well defined but depend on each other greatly for informationand hence require significant coordination. In contrast, marketingand sales activities may exhibit reinforcing contingency due to lowtask uncertainty and low horizontal dependence since the tasksperformed are both well-defined and mainly customer-facing. It isalso possible that horizontal dependence assumes greater signifi-cance in e-Retail (especially across value chain activities) becausethe technologies required to support the value chain activities arerapidly evolving, and the alignment between value chain activitiesis critical to overall success. As such, marketing and sales activitiescan benefit from similar sourcing approaches.

The e-Retail industry also presents an interesting context inwhich the strategic role of IT contributes to reinforcing contingen-cies. This is especially true for e-Retail firms that aim to transformtheir industry and value chain activities in innovative ways and, asa result, face high task uncertainty (Schein, 1992). In effect, thesetransformative e-Retail firms may face reinforcing contingenciesthat lead to clearer choices for decision-making on IST sourcingstrategy (Gresov, 1989).

Contingency theory frameworks explicate the relationshipbetween contextual factors, internal practices, and performance(Sousa and Voss, 2008). In our framework, we limit internal prac-tices to the sourcing decisions made in value chain activitiesand choose contextual factors that show variance within the e-Retail industry context. Thus, our approach is consistent with therecommendation of Sousa and Voss (2008) for developing industry-specific objective measures in a contingency framework. To providea comprehensive view of the contingency theory, we also considerimitative behavior (Sousa and Voss, 2008), which is an environmen-tal factor commonly observed in technology practices ranging fromIT outsourcing activities (Loh and Venkatraman, 1992b) to the dif-fusion of manufacturing practices (Ketokivi and Schroeder, 2004)to US hospitals’ adoption of electronic medical records (Angst et al.,2010). We propose a two-stage IST sourcing model for the e-Retailvalue chain as shown in Fig. 2. We first look at five organizationaland environmental factors: experience, capabilities, IT strategic role,strategic IS/IT leadership, and sourcing propensity, all of which areexpected to influence e-Retailers’ make versus buy IST sourcingstrategies for value chain activities. Then, we examine the perfor-mance impacts of make versus buy IST sourcing decisions and studythe impacts of bundling across value chain activities by evaluatingthe effects of complementary IST sourcing. In the following subsec-tion, we explicitly review the literature pertaining to each constructin Fig. 2 and elucidate its hypothesized relationship with sourcingdecisions and performance.

2.3.1. Organizational contextual factorsOrganizational contextual factors have played a strong role in

contingency frameworks from both practice adoption and per-formance perspectives, and yet organizational factors have been

shown to have differing influences based on the practices and con-texts in question (Shah and Ward, 2003; McKone et al., 1999).Existing technology infrastructure is known to create inertia inchanging prevailing practices or adopting new technologies. Early

us buy

ausattraicolbset

ebaGfidffolmcacbOgFt

Fig. 2. Research model of make vers

dopters of technologies often create custom solutions due to thenavailability of off-the-shelf applications. However, by doing so,uch firms find it increasingly difficult for shift to off-the-shelfpplications when adding new features and capabilities. Moreover,hey tend to develop distinct technology capabilities early on thathey may be reluctant to give up later. As a result, highly expe-ienced IT firms are likely to become entrenched and to exhibitversion to sourcing technologies from outside vendors due tonertia and compatibility concerns as well as the specificity ofustom-built technologies made in-house. Alternatively, becausef the rapidly changing pace of the e-Retail industry, firms thatack the experience may not have the time to play catch up byuilding customized solutions; thus, buying becomes the logical, orometimes the only choice. In essence, we argue that experienced-Retail firms are likely to build technology-based competence ando strive to retain that competence.

Contingency models in manufacturing settings have consid-red plant characteristics, variability, and equipment types toe important organizational factors that drive practice adoptionnd performance (McKone et al., 1999; Shah and Ward, 2003;onzalez-Benito, 2007). In e-Retail, the technical capability of therm constitutes an analogous organizational context factor. Therive to add new technical capabilities to storefronts quickly mayorce e-Retail firms to buy off-the-shelf technologies. E-Retail firmsace unique challenges related to differences in customer types,perations of order fulfillment, service quality expectations, andogistical requirements (Johnson and Whang, 2002). E-Retailers

ust be able to adapt swiftly on the technology front becausehanges in consumer demands and preferences spur the rapidnd constant evolution of the industry. Integration of e-Commerceapabilities and functionalities with a firm’s IT infrastructure haseen shown to impact firm performance positively (Zhu, 2004).

utsourcing is one fast way for firms to acquire flexibility andain IT functions and resources (Agarwal and Sambamurthy, 2002).irms often leverage outsourcing to fill gaps in IT capabilities, andhose that pursue aggressive strategies in fulfilling resource gaps

IST sourcing strategy of e-Retailers.

will outsource more (Cheon et al., 1995). For instance, McLellanet al.’s (1995) case study of the banking industry reveals that,through outsourcing vendors, banks were able to acquire newtechnologies and the associated capabilities at a faster pace. Wesummarize our hypotheses related to the organizational contextfactors of experience and IT capabilities as follows:

H1. Experience Favors Make Strategy. An e-Retailer’s e-Commerce experience has a negative association with the degree ofIST sourcing for its e-Retail value chain activities.

H2. E-Commerce Capabilities Favor Buy Strategy. An e-Retailer’se-Commerce capabilities have a positive association with the degreeof IST sourcing for its e-Retail value chain activities.

2.3.2. IT strategic roleStrategic role of IT has been identified as a major contin-

gency factor in Information Systems literature (e.g., Armstrong andSambamurthy, 1999; Dehning et al., 2003). Strategic IT role canmoderate transaction and coordination costs arising from associ-ated technology risks (Sousa and Voss, 2008). IT strategic role isdefined as the shared, aspired state of the role that IT is expectedto play in the firm, and it includes three categories: automate,informate, and transform (Schein, 1992). For automate, the roleof IT is to replace inefficient human labor with information tech-nology. For informate, IT is used to provide information to higherand lower levels of the organization to aid decision-making andempower employees with relevant information and knowledge.Finally, for transform, IT is used to alter the structure and com-petitive forces of the industry or market segment in which thefirm operates or competes. Firms with the transform vision forIT have the strongest relationship between knowledge and sys-tems of knowing (Armstrong and Sambamurthy, 1999; Ke and Wei,

2008). Additionally, firms that use IT in a transform strategic roleare prone to introducing radical business models to gain compet-itive advantages (Chatterjee et al., 2001; Dehning et al., 2003).This suggests that they are willing to take on higher risks with IT

ING ModelO

6 ions M

icfitd

(ahcrorbsvWptfi

tiItescalata(sa(IaI

ifiH(utwiitbttlc

nnovations, such as making their own solutions rather than pur-hasing commoditized off-the-shelf products. On the other hand,rms pursuing informate and automate strategies may prefer toransfer technology-associated risks to their vendors through buyecisions on IST sourcing.

Multi-dimensional contingency theory as proposed by Gresov1989) further supports our argument that transform firms choose

make strategy. The transformative nature of business createsigh task uncertainties in firms’ IST-related functionalities andore value chain activities. As a result, transform firms face theeinforcing contingencies of high task uncertainty and a high levelf horizontal dependence between value chain activities. Theseeinforcing contingencies lead to the need for organic and flexi-le organizational configurations (Gresov, 1989). We argue that ISTourcing strategies that lean toward creating technological inno-ations in-house is the dominant choice for firms in such contexts.hen an organization develops the capabilities for technology and

rocesses in-house, it can better align value chain activities withhose technologies. Therefore, we hypothesize the following forrms with a transform vision:

3. Transform Firms Elect Make Strategy. E-Retailers with theransform IT strategic role have a lower degree of IST sourcing thanutomate or informate firms.

Since e-Commerce firms encounter unique challenges relatedo IT architecture and capabilities, strategic IS/IT leadership is crit-cal to the success of e-Commerce firms. In this context, the Chiefnformation Officer (CIO) can play an important role in addressinghe issues related to technology strategy. Strategic IS/IT lead-rs such as the CIO provide technical insight and expertise tohape an organization’s e-strategy. For example, a firm with ahief e-Commerce officer (CeCO), who leads e-business initiativesnd oversees all aspects of the e-business value chain, is moreikely to establish an organizational structure for a valiant virtualpproach (Pinker et al., 2002). Senior IS managers are also knowno possess the “empire builder” syndrome, in which desire for powernd more resources drives them to build large IT organizationsGurbaxani and Whang, 1991). Political skills, which involve self-erving behaviors to enhance one’s position and build a power base,lso have a strong impact on managerial effectiveness and successPavett and Lau, 1983). The main way for a CIO to create a largeT organization is to produce in-house. Therefore, the presence orbsence of a strategic IS/IT leader like a CIO can influence a firm’sST sourcing decisions. Thus, we propose the following hypothesis:

4. Strategic IS/IT Leader Chooses Make Strategy. An e-Retailerhat has a strategic IS/IT leader has a lower degree of IST sourcing forts e-Retail value chain activities.

.3.3. Mimicking behaviorMimetic isomorphism, the achievement of conformity through

mitation, can explain why firms evolve to be more like otherrms in their environment over time (DiMaggio and Powell, 1983).aunschild and Miner (1997) identify three modes of imitation:

1) frequency-based imitation refers to the adoption of practicessed by a large number of firms (e.g., banks’ use of ATMs); (2)rait-based imitation occurs when a firm adopts the actions of firmsith certain characteristics (e.g., firms adopting the practices of

ndustry-leading companies like Apple); and (3) outcome-basedmitation involves choosing practices that have resulted in posi-ive outcomes for other firms (e.g., a firm chooses IT outsourcingecause other firms have been able to significantly reduce costs by

hat means). Mimetic pressures have been observed to influencehe adoption of financial electronic data interchange, particu-arly in cases in which the innovation is perceived to be highlyomplex (Teo et al., 2003). In the e-Retailing context, firms face

similar pressure when considering IST sourcing for their IT infra-structure. Possible explanations for the influence of peers on afirm’s IS outsourcing decisions relate to institutional forces or socialpressures toward conformity (Ang and Cummings, 1997). As aresult, mimicking behavior among competing firms is common fortechnology practices (Loh and Venkatraman, 1992b; Ketokivi andSchroeder, 2004). Therefore, we argue that e-Retailers are likely toconsider the IST sourcing strategies of their peers when formulatingtheir own IST sourcing strategy. Hence, we propose the followinghypothesis:

H5. Mimicking Behavior Influences Sourcing Strategy. Theindustry sourcing propensity based on a firm’s merchandizer categoryis positively associated with the degree of IST sourcing for its e-Retailvalue chain activities.

2.3.4. Alignment and performanceMultiple studies on IT projects have confirmed the importance

of strategic fit in achieving high firm performance (Nidumolu, 1996;Barki et al., 2001). A study on the relationship of IT strategic roleand firm value reveals that IT investment types provide differentimplications for firm performance (Dehning et al., 2003). Tanriverdiand Ruefli (2004) concur that the performance effects of comple-mentarities can be better understood if one distinguishes differenttypes of complementarities and the roles of IT in realizing each type.Anderson et al. (2006) explore the interaction of industry medianY2K spending with the strategic role of IT and find strong positivevalue implications of Y2K spending in industries where IT plays atransforming role. Other studies show that firms with the trans-form IT strategic role are more able to achieve positive changes inmarket value (Chatterjee et al., 2001; Dehning et al., 2003). Align-ment between IT strategic role and technology investment choicesleads to better firm performance. As explained in earlier sections,we do not expect to see conflicting contingencies for transformfirms—unlike for automate and informate firms—since transformfirms face high levels of task uncertainty and horizontal depend-ence (Gresov, 1989), which suggests their ability to achieve betterperformance through organic design and through avoidance ofstrategies that lead to standardization via commoditized IST pur-chases. Since firms with a transform vision are expected to make (ashypothesized in H3), we postulate that transform firms that chooseto buy have a misfit and will hence show poorer performance:

H6a. Buy Strategy for Transform Firms Results in Poorer Per-formance. A less positive association exists between performance andthe degree of IST sourcing for e-Retailers with a transform IT strategicrole than for automate or informate firms.

Teece (1986) stresses the importance of acquiring complemen-tary assets and argues that incumbents’ possession of such assetscan discourage new entrants from competing. Tripsas (1997) findsthat the commercial performances of incumbents and new entrantsare influenced by the balance and integration of three factors:investment, technical capabilities, and specialized complementaryassets. He also finds that incumbents with access to complementaryassets are able to sustain a high level of commercial performance.However, not all studies on complementary assets show positivefirm performance. Swink and Nair (2007) find mixed results formanufacturing performance in their exploration of complemen-tary effects of processes and technologies, design-manufacturingintegration, and advanced manufacturing technologies. Harrisonet al. (2001) argue that, in order to be successful when integratingcomplementary resources, firms must seek potential synergy and

understand what actions are necessary to achieve it. Although spe-cific empirical evidence for sourcing complementarities betweenvalue chain activities is scant, certain primary activities in e-Retailvalue chain have strong alignment possibilities. For example, Porter

IN PRESSG ModelO

ions Management xxx (2013) xxx–xxx 7

wcemaaiomfmdibapcfitsmcatcecat(f

TmscnGUfiecppf(Fkd2

Table 1E-Retailers by merchant type and merchandizer category.

Merchant type Number of firms Percentage

Catalog/call center 56 18.24%Brand manufacturer 31 10.10%Retail chain 109 35.50%Web only 111 36.16%

Total 307 100%

Merchandiser category Number of firms Percentage

Apparel/accessories 77 25.08%Automotive parts/accessories 2 0.65%Books/music/video 18 5.86%Computers/electronics 36 11.73%Flowers/gifts 7 2.28%Food/drug 12 3.91%Hardware/home improvement 11 3.58%Housewares/home furnishings 47 15.31%Jewelry 8 2.61%Mass merchant 21 6.84%Office supplies 8 2.61%Specialty/non-apparel 29 9.45%Sporting goods 18 5.86%

1985) groups marketing and sales into a single activity to recognizehe potential benefits of coordinating the decisions in the two valuehain activities. Manufacturing studies also stress the tight interre-ationship between marketing and sales (Hayes and Wheelwright,984).

We conjecture that multi-dimensional contingency will be atork, leading to complex inter-relationships between the value

hain activities. While there is considerable heterogeneity among-Retailers, some generic observations can be made regarding theulti-dimensional contingencies at play in each of the value chain

ctivities. For instance, logistics and operations activities in e-Retailre likely to be associated with low task uncertainty and high hor-zontal dependence because their tasks are well defined and relyn information from other related tasks. Therefore, an inherentisfit in strategic choice exists and may contribute to poor per-

ormance. Marketing in the e-Retail setting, on the other hand, isore likely to encounter low task uncertainty and low horizontal

ependence since tasks are not only well defined but also primarilynvolve interaction with customers; thus, higher performance cane achieved through sourcing. As such, buy appears to be a betterpproach for marketing value chain activity, while make may bereferred for logistics and operations. Sales technologies—whichonsist of the e-Commerce platform, order management, and ful-llment, among others—are critical to e-Retailers’ success. Theseechnologies often align better with marketing technologies, andimilar sourcing strategies across the two value chain activitiesay be beneficial (O’Leary-Kelly and Flores, 2002). Thus, theoreti-

ally there are differing contingencies at play across the value chainctivities that would make it difficult to predict how complemen-arities between value chain activities may manifest. We assert thatoordinated decisions across different value chain activities cannable an e-Retail firm to exploit internal capabilities more effi-iently. When sourcing decisions are consistent across value chainctivities, they enable the scaling of operations from both perspec-ives of internal development (i.e., make) and external sourcingi.e., buy). Recognizing the potential of synergy, we propose theollowing hypothesis:

6b. IST Sourcing Complementarities Contribute to Better Per-ormance. IST sourcing complementarities in value chain activitiesositively impact e-Retailer performance.

. Data and methodology

.1. Data collection

For this study, we collected data from Internet Retailer’sop500Guide.com. Internet Retailer is a monthly national businessagazine first launched in March 1999. It has more than 43,000

ubscribers consisting chiefly of the senior executives of retailhains, independent stores, catalogs, virtual merchants, brand-ame manufacturers, and wholesalers/distributors. The Top 500uide provides an annual ranking of the largest e-Retailers in thenited States and Canada based on annual online sales. The top 500rms account for a sizable portion of the e-Retail market share. Forxample, the listed firms for 2007 represent approximately 61 per-ent. We used the ranking lists from 2007 to 2011 to construct aanel data set of 307 firms for the period of 2006–2010. To com-are the sales volume of these 307 firms with the total sales volumeor the US e-Retail market, we have provided Forrester Research’swww.forrester.com) US Online Retail Forecast for 2009–2014 in

ig. 3. It shows the forecasted sales volume for the US e-Retail mar-et at $155.2 billion for 2009 and $172.9 billion for 2010. In ourata set, the sales volume of the 307 firms totaled $101.5 billion for009, $111.3 billion for 2010, and $131.2 billion for 2011. Therefore,

Toys/hobbies 13 4.23%

Total 307 100%

these firms comprise the lion’s share of the e-Retail market, and ourfindings should apply to the majority, if not all, of the market.

Each firm in our sample falls into one of four mer-chant types: catalog/call center, brand manufacturer, retailchain, and Web only. Each firm also belongs to one of thefollowing merchandizer categories: apparel/accessories, automo-tive parts/accessories, books/music/video, computers/electronics,flowers/gifts, food/drug, hardware/home improvement, house-wares/home furnishings, jewelry, mass merchant, office supplies,specialty/non-apparel, sporting goods, and toys/hobbies. Table 1displays the breakdown of the firms by merchant type and mer-chandizer category.

For each firm, Internet Retailer supplies data for financial, oper-ations, customer satisfaction, marketing, and firm performance. Italso provides data on the vendors used, shopper profile, websitefeatures and functions, payment systems, social networks used, sitesearch capabilities, shopping engines and marketplaces used, andcustomer service features offered by the firms. Internet Retailercompiles data about retailers’ Web traffic from comSource Inc. andNielson Online, and Web sales data from each company. In casesin which data were not available for Web sales, Internet Retailerestimated the values based on traffic and the assumed conversionrate for that retailer’s category as well as on analyst interviews.Other related data are estimated using comScore, Nielsen Online,or Internet Retailer sources. For other figures, like the conversionrate and average ticket, Internet Retailer researchers used categorydata and analyst interviews to formulate estimates. The retailershave opportunities to review and respond to these estimates. Todetermine if a firm has a CIO, we searched the Jigsaw databaseand cross-checked the information using the corporate websites,Internet search, and LinkedIn.

3.2. Variable definitions

The variables are grouped into four categories: organizationalcharacteristics, environmental factors, make versus buy strategyfor the value chain, and firm performance. Table 2 lists the vari-

ables and their descriptions. For organizational characteristics, thevariables are SKU, monthly visits, IT strategic role, experience, capa-bility index, and CIO. SKU refers to the total stock-keeping units ofthe firm for the year. We took the natural logarithm of this number

8 J.Y. Tsai et al. / Journal of Operations Management xxx (2013) xxx–xxx

a(vaIvabIu

Fig. 3. Forrester research US o

nd used it as a control variable for the complexity of product mixBendoly et al., 2007). Monthly visits refers to the average monthlyisitors for the year. We took the natural logarithm of this numbernd used it as a control variable. IT strategic role classifies the role ofT investments for each e-Retailer based on its merchant type. Theariable shows a value of 1 for automate, 2 for informate up/down,

nd 3 for transform (in actual estimation, a dummy variable woulde created and used for each corresponding type). To determine theT strategic role for each merchant type, we applied the methodsed by Chatterjee et al. (2001) and Dehning et al. (2003). We

able 2ariables.

Variable Operationalization

Organizational characteristicsSKU (natural logarithm)

Monthly visits (natural logarithm)

IT strategic role Role of IT investment

Experience e-Retail proficiencyCapability index Capabilities

CIO Strategic IS/IT leader

Environmental factorsYear

Merchandizer category

Sourcing propensity

Make versus buy strategy for the value chainContent delivery Logistics and operationsContent management

Site design

Web analytics

Web hosting

Web performance monitoring

Degree of Sourcing Logistics and Operations

Affiliate marketing MarketingEmail marketing

Search engine marketing

Degree of Sourcing Marketing

Rich media SalesSite search

E-Commerce platform

Order management

Fulfillment

Degree of Sourcing Sales

Degree of Sourcing All All

Firm performanceWeb sales (natural logarithm)Conversion rate

Growth rate

retail forecast for 2009–2014.

designed and sent the instrument shown in Appendix A to a panelof 3 judges, composed of IS scholars. Each judge was requestedto code each merchant type as automate, informate up/down, ortransform. All of them coded catalog/call center as automate, brandmanufacturer and retail chain as informate up/down, and Web onlyas transform. We used this categorization to assign the IT strategic

role for each e-Retailer. Experience, which indicates the number ofyears since the e-Retailer launched its website and established itsonline store, measures the e-Retailing proficiency of each firm. Tocreate a value for capability index, which reflects the intensity of

Description

Natural logarithm of the total number of stock-keeping units (SKU)Natural logarithm of monthly average visitors for the year1 = automate, 2 = informate up/down, and 3 = transformNumber of years since e-Retailer launched its website and online storeIntensity of the capabilities of the firm relative to other firms1 if firm has a Chief Information Officer

1 = 2006, 2 = 2007, 3 = 2008, 4 = 2009, and 5 = 20101 = 448, 2 = 453, 3 = 451, 4 = 454, 5 = 443, and 6 = othersAverage degree of sourcing for the firm’s merchandizer category

1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwiseRatio of sourced to total IST for logistics and operations1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwiseRatio of sourced to total IST for marketing1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwise1 if technology is sourced, 0 otherwiseRatio of sourced to total IST for salesRatio of sourced to total IST for all

Natural logarithm of total Web sales for the yearPercentage visitors who take desired actionPercentage change in growth of Web sales from the previous year

ING ModelO

ions M

ttorltniohcgfipco

fiydyoIc4(5psTt

eeotteTaotgmperacrPmtdsdmu

ctvdaty

accounts for complementary IST sourcing, is as follows:

he capabilities of the firm relative to other firms, we first tookhe ratio of 1 (if the firm has the feature) over the total numberf firms that have the same feature, and then summed up suchatios for 27 features (see Appendix B, which shows the completeist of e-Retailer features and functions). This ratio sum number ishen normalized to a value between 0 and 1. To avoid the weak-ess of using a general index, we leveraged a weighted technology

ndex, which quantifies the rarity versus commonality of a technol-gy (Spetz and Baker, 1999; Blank and Van Hulst, 2009). Firms thatave rarer technologies to differentiate them should have a higherapability index than other firms that only have common technolo-ies. To indicate the presence of a strategic IS/IT leader within therm, we used the variable CIO, which received a value of 1 if theosition exists and 0 otherwise. IT strategic role, experience, andapability index also serve as control variables in the second stagef our hypothesis testing on performance impacts.

Under environmental factors, we have the variables year andrm category. Year, which serves as a control variable, shows theear value of 2006, 2007, 2008, 2009, or 2010 (in actual estimation, aummy variable would be created and used for each correspondingear). Merchandizer category refers to the industry of the firm basedn the types of products it sells; this also acts as a control variable.t has a value of 1 for e-Retailers with a NAICS (www.naics.com)ode starting with 448 (clothing, shoe, and jewelry stores), 2 for53 (florists, office supplies, specialty, and gift stores), 3 for 451sporting goods, hobby, toys, and books), 4 for 454 (mass merchant),

for 443 (computer and electronics), and 6 for others. The sourcingropensity for each industry is determined based on the averageourcing propensity of firms in the same merchandizer category.he average value, which ranges from 0 to 1, is determined usinghe degree of sourcing variable, which is defined next.

In terms of make versus buy strategy for the value chain, westablished a degree of sourcing variable for each of the three differ-nt primary activities of the value chain being studied: logistics andperations, marketing, and sales. For each firm, we first counted theotal number of sourced vendor technologies for each activity andhen divided the number by the total number of technologies forach firm in order to obtain the proportion of sourced technologies.his ratio helps normalize the degree of sourcing in each value chainctivity to be in the range [0,1]. We also computed an overall degreef sourcing variable that is a sum of the degree of sourcing over thehree activities. The highest number of sourced vendor technolo-ies is 6 for logistics and operations (i.e., content delivery, contentanagement, site design, Web analytics, Web hosting, and Web

erformance monitoring), 3 for marketing (i.e., affiliate marketing,mail marketing, and search engine marketing), and 5 for sales (i.e.,ich media, site search, e-Commerce platform, order management,nd fulfillment). To classify each of the 14 technologies to the valuehain activity of logistics and operations, marketing, or sales, weeferred to the Process Classification Framework of the Americanroductivity and Quality Center (APQC, 2012), which enabled us toap each technology to a process activity (see Appendix C). The

echnologies that fall under marketing are intended to support theevelopment and management of channels as well as marketingtrategy and plans. The sales technologies are meant to support theevelopment of sales strategy and the development and manage-ent of sales plans. The logistics and operations technologies are

sed to deliver products and services.To assess firm performance, three variables are used: Web sales,

onversion rate, and growth rate. The variable Web sales representshe natural logarithm of a firm’s total Web sales for the year. Con-ersion rate captures the percentage of visitors who perform theesired action, whether the action is buying a product, filling out

form, or some other goal on the Web page. Growth rate shows

he percentage change in growth of Web sales from the previousear.

PRESSanagement xxx (2013) xxx–xxx 9

3.3. Model specification

In this section, we introduce the first-stage and second-stageestimation models. We first describe the basic model structure andpresent model specific tests in the Results section. To study thefactors that influence make versus buy decisions, the first-stagemodel treats degree of sourcing as the dependent variable for each ofthree value chain activities: logistics and operations (DS LogisOpsit),marketing (DS Mktgit), and sales (DS Salesit). Additionally, we alsoregress the aggregate degree of sourcing over the three activities(DS All). This gives us four models for degree of sourcing:

DegreeSourcingit = ˇ0 + ˇ1 ln SKUit + ˇ2 ln MonthlyVisitsit

+ ˇ3Experienceit + ˇ4CapbilityIndexit + ˇ5CIOit

+ ˇ6SourcingPropensityit + ˇ7Year 2006it + ˇ8Year 2007it

+ ˇ9Year 2008it + ˇ10Year 2009it

+ �1ITStrageticRole Automatei + �2ITStrageticRole Informatei

+ �3MerchadiserCategory 448i + �4MerchadiserCategory 453i

+ �7MerchadiserCategory 443i + ˛i + εit

The parameters ˇ0 to ˇ10 and �1 to �7 are to be estimated. Thesubscripts i and t index the firm and the year, respectively. The twoerror terms are ˛i, which is a time-invariant firm i random effect,and εit, which is different for each firm at each point in time. Sincewe drew our sample of firms from a larger population, a randomeffects model was considered more appropriate (Greene, 2008).We further discuss the empirical appropriateness of random effectsspecification in the Results section.

In the second stage, we first evaluated the effects of IT strate-gic role and total degree of sourcing on firm performance. Thenwe looked at the impacts of make versus buy decisions and com-plementary IST sourcing on firm performance. We have threemodels based on the different performance measures for the threedependent variables: Web sales (lnWebSalesit), conversion rate(ConversionRateit), and growth rate (GrowthRateit). The first esti-mation model for firm performance based on IT strategic role is asfollows:

PerformanceMetricit = ˇ0 + ˇ1 ln SKUit + ˇ2 ln MonthlyVisitsit

+ ˇ3Experienceit + ˇ4CapabilityIndexit + ˇ5DS Allit

+ ˇ6ITStrageticRole Automate ∗ DS Allit

× ˇ7ITStrageticRole Informate ∗ DS Allit + ˇ8Year 2007it

+ ˇ9Year 2008it + ˇ10Year 2009it + �1ITStrageticRole Automatei

+ �2ITStrageticRole Informatei + �3MerchadiserCategory 448i

+ ˛t + εit

The parameters ˇ0 to ˇ10 and �1 to �7 are to be estimated. Thesubscripts i and t index the firm and the year, respectively. The twoerror terms are ˛i, which is a time-invariant firm i random effect,and εit, which is different for each firm at each point in time.

The second estimation model for firm performance, which

PerformanceMetricit = ˇ0 + ˇ1 ln SKUit + ˇ2 ln MonthlyVisitsit

+ ˇ3Experienceit + ˇ4CapabilityIndexit + ˇ5DS Mktgit

IN PRESSG ModelO

1 ions Management xxx (2013) xxx–xxx

mstcib

vacaOiottalh

amsweadt

fiitfct

−0.0

−0.1

−0.0

−0.1

−0.0

−0.1

09−0

−0.0

−0.6

−0.1

−0.0

−0.1

−0.0

4 −0

−0.1

8 −0

−0.0

−0.1

−0.2

−0.1

−0.0

−0.3

−0.0

−0.2

−0.0

0 J.Y. Tsai et al. / Journal of Operat

+ ˇ6DS Salesit + ˇ7DS LogisOpsit + ˇ8DS Mktg ∗ DS Salesit

+ ˇ9DS Mktg ∗ DS LogidOpsit + ˇ10DS Sales ∗ DS LogisOpsit

+ ˇ11Year 2007it + ˇ12Year 2008it + ˇ13Year 2009it

+ �1ITStrageticRole Automatei + �2ITStrageticRole Informatei

+ �7MerchadiserCategory 443i + ˛i + εit

The parameters ˇ0 to ˇ13 and �1 to �7 are to be estimated. Theubscripts i and t index the firm and the year, respectively. The tworror terms are ˛i, which is a time-invariant firm i random effect,nd εit, which is different for each firm at each point in time. In bothodels, degree of sourcing is considered an endogenous variable

nd instrumented.

. Results

Table 3 provides the descriptive statistics and the correlationatrix. As would be expected, DS All is highly correlated with other

ourcing variables. However, DS All is not used in models wherehe other sourcing variables are included. In testing for multi-ollinearity, we checked the variance inflation factor (VIF) for allndependent variables and confirmed that all of the values areelow 10 (Greene, 2008).

In the first stage, with degree of sourcing as the dependentariable, we chose the random effects model because it enablesdjustment of the standard error estimates for the within-firmorrelation in the repeated measurements of the dependent vari-ble. An alternative specification, given our sample, is a pooledLS model where firm categories can be used to control for time

nvariant fixed effects. Treating degree of sourcing as a continu-us variable, we used the Breusch and Pagan Lagrange Multiplierest for existence of random effects (i.e., test the null hypothesishat variance ˛i = 0). For all models, we were able to support thessumption that variance ˛i > 0 (at p < 0.0001). We used individualevel robust standard errors that are consistent in the presence ofeteroskedasticity and autocorrelation (Wooldridge, 2003).

Since degree of sourcing can be considered a count variable,n alternative specification would be to use a Poisson regressionodel. Using the same random effects specification and robust

tandard errors, we also ran Poisson regression models.4 The resultsere consistent with that of continuous regression models. How-

ver, since we want to aggregate normalized degree of sourcingcross the value chain activities and our second-stage model treatsegree of sourcing as a continuous variable, we retained the con-inuous regression model for presentation.

In the second-stage model, the dependent variables are per-ormance variables, and the degree of sourcing is treated as anndependent variable. However, there is the potential of endogene-ty between firm performance and IST sourcing decision. We firstest for endogeneity using a cross-lagged model specification to testor Granger Causality. Further, we instrument degree of sourcing toontrol for any potential endogeneity. Details of the specificationests are given in Section 4.2.

.1. Make versus buy

Please cite this article in press as: Tsai, J.Y., et al., Information systems and technology sourcing strategies of e-Retailers for value chainenablement. J. Operations Manage. (2013), http://dx.doi.org/10.1016/j.jom.2013.07.009

Table 4 reports the first-stage analysis results for factorshat influence make versus buy decisions. Model 1 shows the

4 We used Xtreg and Xtpoisson in Stata V. 12 for the analyses. Tab

Table 4Factors on degree of sourcing for activities of the e-Retail value chain.

Variable Model 1 (logistics and operations) Model 2 (marketing) Model 3 (sales) Model 4 (All)

Intercept 0.7856*** (0.1319) −0.1637 (0.2856) 0.0528 (0.2395) 0.3576* (0.1394)lnSKU −0.0003 (0.0031) −0.0038 (0.0070) −0.0121* (0.0059) −0.0035 (0.0035)lnMonthlyVisits −0.0093† (0.0052) 0.0145 (0.0115) 0.0113 (0.0097) 0.0027 (0.0057)Experience −0.0155*** (0.0037) −0.0143† (0.0086) −0.0167* (0.0074) −0.0157*** (0.0046)CapabilityIndex 0.0352 (0.0293) 0.1859** (0.0633) 0.1449** (0.0529) 0.0918** (0.0305)CIO −0.0110 (0.0145) −0.0328 (0.0313) 0.0057 (0.0261) −0.0153 (0.0149)SourcingPropensity 0.3063† (0.1709) 0.9818** (0.3633) 0.8371** (0.3013) 0.5807*** (0.1708)

Base year: 2010Year 2006 −0.0586*** (0.0169) −0.0955* (0.0382) −0.1013** (0.0327) −0.0724*** (0.0201)Year 2007 −0.0527*** (0.0140) −0.0727* (0.0313) −0.0671* (0.0266) −0.0569*** (0.0162)Year 2008 −0.0319** (0.0120) −0.0024 (0.0263) −0.0537* (0.0222) −0.0336* (0.0132)Year 2009 −0.0164† (0.0087) −0.0067 (0.0186) −0.0336* (0.0156) −0.0202* (0.0091)

Base type: transformITStrategicRole Automate 0.0441† (0.0242) 0.2718*** (0.0557) 0.1465** (0.0481) 0.1211*** (0.0301)ITStrategicRole Informate 0.0574** (0.0208) 0.2104*** (0.0478) 0.1389*** (0.0412) 0.1181*** (0.0257)

Base category: othersMerchandiserCategory 448 −0.0137 (0.0300) −0.0519 (0.0677) 0.0326 (0.0580) −0.0030 (0.0356)MerchandiserCategory 453 0.0079 (0.0284) −0.0521 (0.0653) −0.0004 (0.0563) −0.0078 (0.0353)MerchandiserCategory 451 −0.0206 (0.0286) −0.0424 (0.0656) 0.0474 (0.0566) −0.0052 (0.0354)MerchandiserCategory 454 −0.0639† (0.0387) −0.0326 (0.0887) −0.0307 (0.0765) −0.0495 (0.0477)MerchandiserCategory 443 −0.0297 (0.0300) −0.0165 (0.0690) 0.0103 (0.0595) −0.0199 (0.0372)Likelihood ratio �2 756.33 842.79 908.09 1031.1p-value <0.0001 <0.0001 <0.0001 <0.0001

Note: Standard errors in parentheses† Significant at p < 0.1.

EaaafiseMkamd

CoMeM(sem

mTMmpaca

* p < 0.05.** p < 0.01.

*** p < 0.001.

egree of sourcing results for the logistics and operations activ-ty (DS LogisOpsit), Model 2 for the marketing activity (DS Mktgit),

odel 3 for the sales activity (DS Salesit), and Model 4 for all activ-ties (DS Allit).

To test Hypothesis H1, we refer to the coefficient estimate forxperience in each of the models. Hypothesis H1 is supported, as

negative and significant relationship exists between experiencend the degree of IST sourcing for logistics and operations, sales,nd all activities. For Model 1 (logistics and operations), the coef-cient estimate (ˇ3 = −0.0155, p-value = <0.0001) is negative andignificant. Similarly, we see a negative and significant coefficientstimate for Model 3 (sales) (ˇ3 = −0.0167, p-value = 0.0239) andodel 4 (all) (ˇ3 = −0.0157, p-value = 0.001). For Model 2 (mar-

eting), the coefficient estimate (ˇ3 = −0.0143, p-value = 0.0947) islso negative but marginally significant. These findings suggest thatore experienced e-Retailers are generally associated with a lower

egree of IST sourcing.To test Hypothesis H2, we refer to the coefficient estimate for

apabilityIndex in each of the models. For Model 1 (logistics andperations), the coefficient estimate is positive but not significant.odel 2 (marketing) shows a positive and significant coefficient

stimate (ˇ4 = 0.1859, p-value = 0.0034), and the same is found forodel 3 (sales) (ˇ4 = 0.1449, p-value = 0.0063) and Model 4 (all)

ˇ4 = 0.0918, p-value = 0.0027). These results reveal that Hypothe-is H2 is supported and that a positive relationship exists between-Commerce capabilities and the degree of IST sourcing for thearketing and sales value chain activities of e-Retail firms.To test Hypothesis H3, we refer to the coefficient esti-

ates for ITStrategicRole Automate and ITStrategicRole Informate.he reference point in all models is ITStrategicRole Transform. Inodel 1 (logistics and operations), coefficient estimates for Auto-ate (�1 = 0.0441, p-value = 0.0690) and Informate (�2 = 0.0574,

-value = 0.0060) are both positive and marginally significant forutomate and significant for informate. For Model 2 (marketing),oefficient estimates for Automate (�1 = 0.2718, p-value = <0.0001)nd Informate (�2 = 0.2104, p-value = <0.0001) are both positive

and significant. Model 3 (sales) also shows significant and posi-tive associations of Automate (�1 = 0.1465, p-value = 0.0024) andInformate (�2 = 0.1389, p-value = 0.0008) with the degree of ISTsourcing. We observe similar results for Automate (�2 = 0.1211, p-value = <0.0001) and Informate (�2 = 0.1181, p-value = <0.0001) inModel 4 (all). Hypothesis H3 is thus supported by all the models.This suggests that for all three activities of the e-Retail value chain,an e-Retailer with the transform IT strategic role is associated witha lower degree of IST sourcing.

To test Hypothesis H4, we refer to the coefficient estimate forCIO in each of the four models. Consistently, we find no evidence ofCIO effect on the degree of IST sourcing. Thus, Hypothesis H4 is notsupported by our data.

To test Hypothesis H5, we refer to the coefficient esti-mates for Sourcing Prospensity in each of the models. ForModel 1 (logistics and operations), the coefficient estimate(ˇ6 = 0.3063, p-value = 0.0735) is positive and marginally signif-icant. Model 2 (marketing) shows a positive and significantcoefficient estimate (ˇ6 = 0.9818, p-value = 0.0070), and the sameis found for Model 3 (sales) (ˇ6 = 0.8371, p-value = 0.0056) andModel 4 (all) (ˇ6 = 0.5807, p-value = 0.0007). These results revealthat for the marketing and sales activity, Hypothesis H5 issupported.

4.2. Firm performance

Tables 5–8 report the second-stage estimation results of theeffects of alignment between IT strategic role and make versus buy,plus those of complementary IST sourcing decisions, on firm perfor-mance. The performance metrics used are Web sales (WebSales) forModel 1, conversion rate (ConversionRate) for Model 2, and growthrate (GrowthRate) for Model 3.

To begin with, we checked for endogeneity in sourcing decisions(i.e., whether performance in the previous period influenced degreeof sourcing in this period and whether degree of sourcing in the pre-vious period influenced performance in this period). The preferred

Table 5Effects of aligning IT strategic role and IST sourcing on performance.

Variable Model 1 (web sales) Model 2 (conversion rate) Model 3 (growth rate)

Intercept 6.5660*** (0.403) 0.0178 (0.0123) 0.2490* (0.0985)lnSKU 0.0166 (0.0143) −0.0005 (0.0005) −0.0066† (0.0038)lnMonthlyVisits 0.7210*** (0.0270) 0.0001 (0.0.0008) 0.0220** (0.0067)Experience 0.0832*** (0.0114) 0.0026*** (0.0006) −0.0145*** (0.0039)CapabilityIndex 0.399* (0.158) −0.0270*** (0.006) −0.0534 (0.0391)DS All 0.228 (0.232) 0.0196* (0.0081) −0.151** (0.0537)

Base type: transformITStrategicRole Automate −0.0706 (0.298) 0.0284* (0.0126) −0.117* (0.0469)ITStrategicRole Informate −0.191 (0.236) −0.0018 (0.0067) −0.142** (0.0445)DS All*ITStrategicRole Automate 0.398 (0.440) −0.0360† (0.0184) 0.120 (0.0742)DS All*ITStrategicRole Informate 0.490 (0.373) −0.0095 (0.0106) 0.159* (0.0710)

Base year: 2007Year 2008 −0.0962 (0.0846) −0.0077* (0.0031) −0.0832*** (0.0191)Year 2009 −0.183* (0.0824) −0.0096** (0.0031) −0.131*** (0.0178)Year 2010 −0.125 (0.0869) −0.0119*** (0.0033) −0.0424* (0.0169)

Base category: NAICS 448MerchandiserCategory 453 0.0215 (0.104) 0.0136*** (0.0038) −0.0522*** (0.0152)MerchandiserCategory 451 −0.503*** (0.0843) −0.00460* (0.0022) −0.0150 (0.0175)MerchandiserCategory 454 0.217* (0.106) 0.00329 (0.0030) −0.0193 (0.0217)MerchandiserCategory 443 0.359** (0.110) −0.0101*** (0.0022) −0.0817*** (0.0169)MerchandiserCategory Others 0.139† (0.0757) 0.0105*** (0.0032) −0.0114 (0.0194)R-squared 64.80% 18.20% 17.40%

Note: Standard errors in parentheses.†

p < 0.1.* p < 0.05.

** p < 0.01.*** p < 0.001.

pproach to examining this relationship is the Granger causality

est. In conducting the Granger causality test, we follow the rec-mmendations of Hood et al. (2008, p. 331). However, because ofhe large number of cross sections, we do not examine Granger

able 6ffects of make versus buy and complementary sourcing on web sales.

Variable Model 1 Model

Main effects Interac

Intercept 6.2960*** (0.3525) 6.2830lnSKU 0.0112 (0.0132) 0.0110lnMonthlyVisits 0.7280*** (0.0244) 0.7290Experience 0.0842*** (0.0123) 0.0843CapabilityIndex 0.3150† (0.1864) 0.3160

Make versus buy and complementary IST sourcingDS Mktg 0.4870** (0.1673) 0.4860DS Sales −0.1630 (0.1429) −0.1700DS LogisOps 0.4260* (0.1821) 0.4290DS Mktg*DS Sales −0.1170DS Mktg*DS LogisOps

DS Sales*DS LogisOps

Base year: 2007Year 2008 −0.2360* (0.0951) −0.2330Year 2009 −0.3090*** (0.0933) −0.3060Year 2010 −0.2670** (0.0985) −0.2650

Base type: transformITStrategicRole Automate 0.1510† (0.0812) 0.1530ITStrategicRole Informate 0.1060 (0.0721) 0.1080

Base category: NAICS 448MerchandiserCategory 453 0.0707 (0.0955) 0.0690MerchandiserCategory 451 −0.4140*** (0.0954) −0.4140MerchandiserCategory 454 0.3000* (0.1357) 0.2960MerchandiserCategory 443 0.4240*** (0.0988) 0.4220MerchandiserCategory Others 0.2010* (0.0859) 0.2010R-squared 65.00% 65.00%

ote: Standard errors in parentheses† p < 0.1.* p < 0.05.

** p < 0.01.*** p < 0.001.

causality for each cross section. The Granger causality test involves

an F-test that compares the estimations of the unrestricted model(with both performance and sourcing lag) against those of therestricted model (with the variable’s own lag).

1.1 Model 1.2 Model 1.3

tion terms

*** (0.3852) 6.1890*** (0.3823) 6.2310*** (0.3767) (0.0143) 0.0111 (0.0140) 0.0050 (0.0140)*** (0.0265) 0.7360*** (0.0264) 0.7410*** (0.0264)*** (0.0114) 0.0840*** (0.0113) 0.0820*** (0.0115)* (0.1580) 0.2360 (0.1605) 0.2810† (0.1597)

** (0.1693) 0.4610** (0.1707) 0.3630* (0.1704) (0.1504) −0.1650 (0.1513) −0.1580 (0.1477)* (0.1915) 0.4560* (0.1900) 0.5580** (0.1944) (0.2786) 0.4490 (0.3277) 0.8860** (0.3306)

−1.4300*** (0.4039) −0.7860† (0.4030)−2.1720*** (0.0414)

* (0.0928) −0.2260* (0.0922) −0.2190* (0.0901)*** (0.0930) −0.2850** (0.0925) −0.2960** (0.0905)** (0.0956) −0.2360* (0.0963) −0.2490** (0.0946)

* (0.0769) 0.1360† (0.0768) 0.1280† (0.0762) (0.0729) 0.0976 (0.0723) 0.0831 (0.0716)

(0.1078) 0.0733 (0.1062) 0.0595 (0.1044)*** (0.0864) −0.4100*** (0.0856) −0.4320*** (0.0859)** (0.1125) 0.3130** (0.1130) 0.2550* (0.1094)*** (0.1137) 0.4400*** (0.1122) 0.4000*** (0.1139)* (0.0800) 0.1950* (0.0789) 0.1290† (0.0786)

65.50% 66.60%

Please cite this article in press as: Tsai, J.Y., et al., Information systems and technology sourcing strategies of e-Retailers for value chainenablement. J. Operations Manage. (2013), http://dx.doi.org/10.1016/j.jom.2013.07.009

Table 7Effects of make versus buy and complementary sourcing on conversion rate.

Variable Model 2 Model 2.1 Model 2.2 Model 2.3

Main effects Interaction terms

Intercept 0.0222* (0.0112) 0.0232* (0.0113) 0.0212† (0.0112) 0.0221* (0.0112)lnSKU −0.0004 (0.0005) −0.0004 (0.0005) −0.0004 (0.0005) −0.0006 (0.0005)lnMonthlyVisits −0.0000 (0.0008) −0.0000 (0.0007) 0.0001 (0.0008) 0.0003 (0.0008)Experience 0.0026*** (0.0006) 0.0026*** (0.0006) 0.0026*** (0.0004) 0.0025*** (0.0006)CapabilityIndex −0.0302*** (0.0061) −0.0303*** (0.0062) −0.0319*** (0.0062) −0.0307*** (0.0063)

Make versus buy and complementary IST sourcingDS Mktg 0.0222*** (0.0059) 0.0222*** (0.0059) 0.0217*** (0.0060) 0.0191** (0.0060)DS Sales −0.0022 (0.0050) −0.0018 (0.0051) −0.0017 (0.0053) −0.0014 (0.0051)DS LogisOps −0.0023 (0.0061) −0.0024 (0.0063) −0.0019 (0.0061) 0.0006 (0.0062)DS Mktg*DS Sales 0.0080 (0.0099) 0.0198 (0.0127) 0.0312* (0.0124)DS Mktg*DS LogisOps −0.0299* (0.0140) −0.0125 (0.0145)DS Sales*DS LogisOps −0.0577*** (0.0157)

Base year: 2007Year 2008 −0.0143*** (0.0033) −0.0146*** (0.0033) −0.0144*** (0.0033) 0.0143*** (0.0032)Year 2009 −0.0157*** (0.0033) −0.0159*** (0.0033) −0.0154*** (0.0033) −0.0158*** (0.0033)Year 2010 −0.0186*** (0.0035) −0.0187*** (0.0035) −0.0181*** (0.0035) −0.0185*** (0.0035)

Base type: transformITStrategicRole Automate 0.0059† (0.0032) 0.0058† (0.0032) 0.0054† (0.0031) 0.0052† (0.0032)ITStrategicRole Informate −0.0070** (0.0023) −0.0071** (0.0023) −0.0074** (0.0023) −0.0078*** (0.0023)

Base category: NAICS 448MerchandiserCategory 453 0.0141*** (0.0039) 0.0142*** (0.0039) 0.0143*** (0.0039) 0.0139*** (0.0039)MerchandiserCategory 451 −0.0027 (0.0022) −0.0027 (0.0022) −0.0026 (0.0022) −0.0031 (0.0023)MerchandiserCategory 454 0.0030 (0.0031) 0.0033 (0.0031) 0.0036 (0.0032) 0.0020 (0.0031)MerchandiserCategory 443 −0.0082*** (0.0021) −0.0081* (0.0021) −0.0077*** (0.0021) −0.0088*** (0.0022)MerchandiserCategory Others 0.0136*** (0.0033) 0.0136*** (0.0033) 0.0135*** (0.0033) 0.0117*** (0.0032)R-Squared 20.50% 20.50% 20.90% 22.20%

Note: Standard errors in parentheses† p < 0.1.* p < 0.05.

** p < 0.01.*** p < 0.001.

Table 8Effects of make versus buy and complementary sourcing on growth rate.

Variable Model 3 Model 3.1 Model 3.2 Model 3.3

Main effects Interaction terms

Intercept 0.1500 (0.1006) 0.1580 (0.1019) 0.1630 (0.1006) 0.1670† (0.1012)lnSKU −0.0069 (0.0042) −0.0067 (0.0042) −0.0067 (0.0042) −0.0072† (0.0043)lnMonthlyVisits 0.0266** (0.0084) 0.0260** (0.0084) 0.0256** (0.0084) 0.0261** (0.0084)Experience −0.0152*** (0.0039) −0.0153*** (0.0039) −0.0152*** (0.0039) −0.0154*** (0.0039)CapabilityIndex −0.0646 (0.0452) −0.0652 (0.0453) −0.0610 (0.0452) −0.0571 (0.0449)

Make versus buy and complementary IST sourcingDS Mktg −0.0370 (0.0381) −0.0365 (0.0384) −0.0352 (0.0378) −0.0437 (0.0376)DS Sales 0.0037 (0.0463) 0.0083 (0.0461) 0.0080 (0.0444) 0.0086 (0.0453)DS LogisOps −0.0549 (0.0439) −0.0569 (0.0438) −0.0582 (0.0438) −0.0494 (0.0433)DS Mktg*DS Sales 0.0750 (0.0635) 0.0456 (0.0724) 0.0834 (0.0758)DS Mktg*DS LogisOps 0.0744 (0.0845) 0.1300 (0.0818)DS Sales*DS LogisOps −0.1880† (0.1087)

Base year: 2007Year 2008 −0.0741** (0.0232) −0.0764** (0.0236) −0.0767** (0.0235) −0.0761** (0.0234)Year 2009 −0.1210*** (0.0219) −0.1230*** (0.0222) −0.1240*** (0.0220) −0.1250*** (0.0221)Year 2010 −0.0311 (0.0213) −0.0325 (0.0214) −0.0340 (0.0213) −0.0352 (0.0215)

Base type: transformITStrategicRole Automate −0.0429* (0.0169) −0.0442** (0.0169) −0.0434** (0.0168) −0.0441** (0.0167)ITStrategicRole Informate −0.0464*** (0.0130) −0.0481*** (0.0130) −0.0475*** (0.0129) −0.0488*** (0.0130)

Base category: NAICS 448MerchandiserCategory 453 −0.0464** (0.0150) −0.0453** (0.0150) −0.0455** (0.0150) −0.0467** (0.0152)MerchandiserCategory 451 −0.0173 (0.0186) −0.0175 (0.0186) −0.0177 (0.0184) −0.0196 (0.0185)MerchandiserCategory 454 −0.0232 (0.0232) −0.0210 (0.0233) −0.0218 (0.0232) −0.0268 (0.0237)MerchandiserCategory 443 −0.0811*** (0.0172) −0.0800*** (0.0173) −0.0809*** (0.0175) −0.0844*** (0.0177)MerchandiserCategory Others −0.0056 (0.0224) −0.0054 (0.0225) −0.0051 (0.0232) −0.0108 (0.0216)R-squared 16.30% 16.40% 16.40% 16.80%

Note: Standard errors in parentheses.† p < 0.1.* p < 0.05.

** p < 0.01.*** p < 0.001.

Table 9Results of hypotheses.

Degree of sourcing

Hypothesis Logistics and operations Marketing Sales All

H1. Experience favors make strategy Supported Partially supported Supported SupportedH2. E-Commerce capabilities favor buy strategy Not supported Supported Supported SupportedH3. Transform firms elect make strategy Partially supported Supported Supported SupportedH4. Strategic IS/IT leader chooses make strategy Not supported Not supported Not supported Not supportedH5. Mimicking behavior influences sourcing strategy Partially supported Supported Supported Supported

Performance

s(cnfadicvTtvrTatovwAtmac

noIHpreiIdaeLtncpam

Hypothesis

H6a. Buy strategy for transform firms results in poorer performance

H6b. IST Sourcing complementarities contribute to better performance

For each of the models, the results showed that the F-test wasignificant for reverse causality in the case of conversion rate alonep-value = 0.02). In other words, lagged degree of sourcing was asso-iated with higher conversion rate. Reverse Granger causality wasot an issue in all the other cases (for Web sales, p-value = 0.34 and

or growth rate, p-value = 0.8). Thus, endogeneity appeared to ben issue in only one case. The Granger causality model, however,oes not allow for the testing of contemporaneous effects of sourc-

ng on performance. Since e-Retailer sourcing decisions can haveontemporaneous effects on performance, we used an instrumentariable approach to control for endogeneity of sourcing decisions.herefore, we employ the instrument variable (IV) method usinghe degree of sourcing from the previous period as the instrumentariables to address the potential bias arising from endogeneity. Weeport the endogeneity-corrected estimation results in Tables 5–8.he 2SLS approach was used for all models and heteroskedasticity-djusted standard errors are reported.5 Post estimation, we usedhe Durbin–Wu–Hausman test for endogeneity. In the two casesf Web sales and growth rate, we rejected the hypothesis that theariables are exogenous (at p-value < 0.10). In all the other cases,e could not reject the null hypothesis that sourcing is exogenous.s a result, instrument correction may be inefficient in some of

he regressions. Therefore, we reran the analysis without instru-ent correction and verified that the results from the two models

re consistent across the board. We have retained the instrument-orrected regression results for reporting purposes.

In Table 5, since we test the interaction effects of an endoge-ous variable with an exogenous variable, the second-stage effectf IT strategic role may be confounded with the first-stage effect ofT strategic role. To alleviate this confounding issue, we used theeckman and Vytlacil procedure suggested by Wooldridge (2003,. 144) for endogenous variables with interactions as follows. Weegressed the endogenous variable (DS All) on all instrument andxogenous variables and obtained predicted values. We creatednteraction variables using predicted DS All with Automate andnformate. The instrument variable regression is done with the pre-icted value, and the interaction term with the predicted values additional instruments. Further, in testing for complementaryffects, we employed the same method used by Tiwana (2008) andance (1988), which uses a residual centering procedure to correcthe problem of partial coefficient distortion faced in the simulta-eous analysis of main effects and interaction terms due to theirorrelation. This involved a two-step procedure: (1) regress eachroduct term (e.g., DS Mktg ∗ DS Sales) on its components and (2)

pply the resulting residual instead of the interaction term in theodel. Further, we introduced the interaction terms one by one

5 Stata’s ivregress method was used for the estimation.

eb sales Conversion rate Growth rate

ot supported Partially supported Partially supportedpported Supported Not supported

sequentially to show more precisely how each interaction mani-fests in the full model.

To test Hypothesis H6a, we refer to the coefficient esti-mates in Table 5 of interaction terms of DS Mktg ∗ DS Sales andDS Mktg ∗ DS LogisOps. Overall, we find partial evidence thatalignment between IT strategic role and IST sourcing decisionsresults in better performance effects. For conversion rate, we findpartial evidence for automate and no evidence for informate. Forgrowth rate, we find significant evidence for informate but noevidence for automate. The only case in which we do not find anysupport is for Web sales.

To test Hypothesis H6b, which explores the effects of com-plementary IST sourcing of e-Retail value chain activities on firmperformance, we refer to the coefficient estimates of DS Mktg ∗DS Sales, DS Mktg ∗ DS LogisOps, and DS Sales ∗ DS LogisOps inTables 6–8. For Model 1.3 (Web sales) in Table 6, the coefficientestimate for DS Mktg ∗ DS Sales (ˇ8 = 0.8860, p-value = 0.0075) ispositive and significant, and similar results apply for Model 2.3(conversion rate) in Table 7 for DS Mktg ∗ DS Sales (ˇ8 = 0.0312,p-value = 0.0123). These results support H6b, which states thatcomplementary IST sourcing of synergistic value chain activitieslike marketing and sales positively impacts a firm’s performancebecause the two functions are closely linked and are typically per-formed together. Interestingly, we see opposite results for Model1.3 (Web sales) in Table 6 when a similar sourcing approachis used for DS Sales ∗ DS LogisOps with a negative and signifi-cant coefficient estimate (ˇ8 = −2.1720, p-value = 0.0001) and forDS Mktg ∗ DS LogisOps with a negative and marginally signifi-cant coefficient estimate (ˇ8 = −0.7860, p-value = 0.0515). Similarresults are observed for Model 2.3 (conversion rate) in Table 7with a negative and significant coefficient estimate (ˇ8 = −0.0577,p-value = 0.0003) for DS Sales ∗ DS LogisOps. Partially significantopposite results are also found for Model 3.3 (growth rate) inTable 8 with a negative and marginally significant coefficient esti-mate (ˇ8 = −0.1880, p-value = 0.0840) for DS Sales ∗ DS LogisOps.We summarize the results of all hypotheses in Table 9 and discussthe managerial and research implications in the next section.

5. Discussion

In this study, we explored IST sourcing of e-Retailers for threeprimary activities: logistics and operations, marketing, and sales.Our evaluation of complementary IST sourcing reveals that dif-fering performance impacts depend on the combination of valuechain activities chosen for sourcing. We considered combinationsof complementary IST sourcing across the activities of logistics andoperations, marketing, and sales. The study findings demonstrate

the influence of organizational and environmental characteristicson e-Retailers’ IST sourcing decisions for value chain enablementand the resulting impacts on firm performance. We discuss theimplications and limitations of our findings below.

ING ModelO

ions M

Fpocbiaiarcmc

iAacsicciabtea

tmATanswubdcsielssn

gfccsidt

.1. Implications

This study offers two key implications for research and practice.irst, the complexity of the value chain requires consideration of theossibility that optimal IST sourcing strategies could differ for vari-us activities of the value chain. The empirical findings in this studyaution that a one-size-fits-all strategy for make versus buy may note appropriate for this e-Retailing context. Hence, future IST sourc-

ng research should focus on the activity level of a value chain. Inddition to the value chain activities examined here, future sourc-ng models should incorporate additional service-focused activitiesnd observe their relationship to firm performance. The findingselated to the complementarities in IST sourcing between valuehain activities are especially helpful for managers and shouldotivate them to rethink the role of IST sourcing strategies in value

hain enablement.The empirical findings provide strong arguments against adopt-

ng an overarching buy approach across the entire value chain.lthough we find strong complementarities between marketingnd sales IST sourcing decisions, we consistently see negativeomplementarities of logistics and operations with marketing orales. One possible reason for this finding is that synergistic activ-ties (such as marketing and sales) are in essence more tightlyonnected; hence, such synergistic activities should be betteroordinated by simultaneous sourcing whereas coordinated sourc-ng may unnecessarily hamper less synergistic activities. Logisticsnd operations activities represent core functions within theirusinesses and define customer experience in an e-Retailer’s por-al. Thus, by acquiring logistics and operations IST assets fromxternal vendors, e-Retailers may lose their core competitivedvantage.

A key theoretical contribution of the study is its demonstra-ion of how organizational and environmental factors influence

ake versus buy IST sourcing decisions for the e-Retail value chain.pparently, there is a technology arms race in the e-Retail industry.o remain competitive in this game, less experienced e-Retailersre relying on vendors to launch and maintain their online busi-esses. On the other hand, more experienced e-Retailers eitherhow a preference for insourcing or, more likely, may be persistingith technologies they have developed earlier. The e-Retailers thatse IT simply to automate or informate also show a preference foruying over making, while the ones that use IT to transform tend toevelop their solutions in-house. For e-Retailers in the same mer-handizer category, there is evidence of imitative behavior for ISTourcing decisions. Firms employing a higher degree of IST sourc-ng also have greater reliance on their vendors for enabling their-Retail value chain. However, it is not clear if this serves the firms’ong-term best interests. There may be potential downsides of highwitching costs and substantial lock-in in the future. Thus, to under-tand the sourcing implications more clearly, future studies willeed to focus on long-term effects.

Additionally, this study shows that multi-dimensional contin-encies influence IST sourcing antecedents. Our results provideurther empirical evidence to support the notion that reinforcingontingencies are associated with better performance whileonflicting contingencies can create challenges for firms in theearch for more suitable organizational configurations. Our find-ngs point to the need for further research in understanding theominating factors associated with the conflicting contingencieshat contribute to a firm’s make versus buy decisions.

.2. Limitations and future research

The sample of firms is one limitation of this study. Resultsrawn from data about the top 500 e-Retailers may not be rep-esentative of medium and poorer performers. Additionally, our

PRESSanagement xxx (2013) xxx–xxx 15

results on e-Retail industry firms may not be generalizable to otherindustries. The data also prevented us from being able to dividethe technologies for logistics and operations into input logistics,operations, and output logistics, which would have given us amore granular list of value chain activities to study. These arethe challenges one typically encounters when studying complexsystems like e-Retailer value chains, which span across multiplefunctions.

Furthermore, we found limited evidence that alignmentbetween IT strategic role and IST sourcing decisions results in bet-ter performance effects. Studies that have explored the firm-levelperformance impacts of outsourcing reveal the difficulty in find-ing direct significant impacts of outsourcing on firm performance.Gilley and Rasheed (2000) explain that this could be because thereal effects of outsourcing occur at the functional level, so it maybe better to study the direct impact of outsourcing on firm per-formance at a more granular level. Strategic fit, however, usuallyoccurs at a higher organizational level. This may explain why, hav-ing conducted our analysis at the value chain activity level, wefound only partial evidence to support our hypothesis on the effectof alignment on performance.

Another limitation is that, in our effort to maximize the numberof firms studied, we restricted ourselves to using the organizationalcharacteristics provided by Internet Retailers. It would be insightfulto perform a follow-up study on publicly traded firms to exam-ine whether similar results can be obtained with a broader set offirm specific variables. Another interesting perspective for futureresearch would be studying how a particular technology’s impor-tance weighs into a firm’s overall performance. Our degree of ISTsourcing is derived based on equal weights for the various technolo-gies; however, different technologies can add unique values to thee-Retail value chain and hence carry different weights. Therefore,it may be more critical to acquire certain IST assets than others.In this instance, different objective or subjective weights shouldbe assigned to various technologies to determine the value of thesourced assets for different e-Retailers.

6. Conclusion

This study develops new knowledge on the performanceimpacts of IST sourcing decisions by explicitly identifying andaggregating sourcing decisions across value chain activities. Basedon the contingency theory, the research model identified thekey organizational and environmental factors that influence e-Retailers’ IST sourcing strategy and enabled us to examine thefirm-level performance impacts of IST sourcing decisions thatinvolve bundling across value chain activities. Our model opens upthe black box of internal firm operations by introducing a granularview of IST sourcing decisions, which had been hitherto unexploredin the literature. The focus on value chain activities enables usto examine the complementarities between different parts of thevalue chain from a sourcing perspective in the e-Retail context.The panel analysis of IST sourcing decisions and performance ine-Retail industry reveals several key and theoretically importantinsights. We find that strategic role of IT within the industry is akey determining factor in sourcing decisions. Our value chain levelanalysis reveals the pitfalls of following an overarching make orbuy approach at the organizational level. It is imperative for orga-nizations to focus on the synergies between value chain activities inmaking IST sourcing decisions. The fast-paced change in the tech-nology and service offerings of e-Retail industry requires further

understanding of the challenges inherent to synergistic IST sourcingdecisions. Future research will need to continue to focus on func-tional and value chain contingencies to develop further insightsinto the strategic role of IST sourcing in the e-Retail industry.

ING ModelO

1 ions M

shTMiITa

have identified, to the extent possible, a specific process asso-ciated with the activities. Where this is not feasible, we haveused a higher level in the classification hierarchy for mappingpurposes.

cknowledgments

The authors are grateful to the anonymous referees for their con-tructive comments on the earlier version of the manuscript, whichelped to improve the presentation of the paper considerably.hey also would like to acknowledge the insightful suggestions ofichael Goul, Uday Kulkarni, Robert St. Louis, and Trent Spauld-

ng. Juliana Tsai thanks the W.P. Carey School of Business and thenformation Systems Department for doctoral research funding.he authors are entirely responsible for the contents of this paper,nd the usual disclaimer applies.

ppendix A. IT strategic role of merchant type

The below instrument was sent to three IS scholars.

Automate: companies that use IT to automate human labor gen-erally invest in IT in order to improve the efficiency of existingbusiness processes.Informate-up and informate-down: companies that involve theuse of IT to induce decision-making and decision-taking at,respectively, higher and lower organizational levels and providedata/information to empower management and employees.Transform: companies that use IT to introduce radical businessmodels that disrupt industry practices (e.g., bypassing selectvalue chain participants) and market structures (e.g., creationof new market spaces) as a means to position themselves morefavorably within an industry. They alter traditional ways of doingbusiness by redefining business processes and relationships.

For each e-Retailer type below, classify as automate, informater transform based the definitions above:

Type Definition Examples StrategicIT role

Catalog/call center Historically, goodsare sold primarilyby phone or viamail-order catalog

• Crutchfield Corp.• American Girl LLC• L. L. Bean Inc.• HSN Inc.• ShopNBC.com

Brand manufacturer Markets a good orfamily of goodsunder its ownbrand name andsells products toconsumers througha direct channel

• Adidas Inc.• HP Home & Office• Callaway Golf• Coach Inc.

Retail Sells goods toconsumers throughboth online andphysical store

• Staples Inc.• Office Depot Inc.• Walmart.com• OfficeMax Inc.• Sears Holding Corp.

Web only Pure online • Amazon.com Inc.

merchant and onlysells goods andservices over theInternet

• Newegg Inc.• Netflix Inc.• eBags.com• Overstock.com Inc.

Appendix B. E-Retailer features and functions

The following list of e-Retailer features and functions is used toderive the capability index variable.

• Affiliate Program• Auction• Catalog Quick Order• Coupons/Rebates• Customer Reviews• Daily/Seasonal Specials• E-mail a Friend• Enlarged Product View• Frequent Buyer Program• Mapping• Mobile Commerce• Online Circular• Online Gift Certificates• Outlet Center• Pre-Orders• Product Comparisons• Product Customizations• Registry• Site Personalization• Social Networking• Store Locator• Syndicated Content• Top Sellers• Videocasts• Wish List• Advanced Search• What’s New

Appendix C. Mapping of IST to value chain activity

The Process Classification Framework of the American Produc-tivity and Quality Center (APQC, 2012) is used to map informationtechnologies to each of the three value chain activities. Notethat APQC framework is a generic framework intended to applyto all industries. Therefore, industry specific interpretation ofthe processes needs to be made. While APQC has published afew industry specific process classification frameworks, no suchframework is available for e-Retailer firms (or Retail firms). We

IN PRESSG ModelO

ions Management xxx (2013) xxx–xxx 17

Process classification framework

ebsite to dispersedo customers

4.0 Deliver Products and Services (10005)4.3 Deliver Service to Customer (10218)

ing and editing of

l to deliver products

ake products

ation of Website to 4.0 Deliver Products and Services (10005)Manage 4.1 Demand for Products andServices (10222) e-Retailers to offer

s to identify 3.0 Market and Sell Products and Services(10004)3.2.3 Define and manage channel strategy(10122)

uire new customers

ertising on different

eo and graphics 3.0 Market and Sell Products and Services(10004)3.5.1 Generate leads (10182)

d contents on an

at allows e-Retailers 3.5.3 Manage customer sales (10184)

rs to process orders 3.5.4 Manage sales orders (10185)ucts such as routing 3.3.2 Manage sales partner/alliance (10212)

Value chain activity Technology enablement

Logistics andoperations—activitiesthat manage inputmaterials and finishedproducts plusvalue-creatingactivities thattransform the inputsinto the final products

Content delivery—copies pages of a Wservers thus enabling faster delivery tContent management—allows publishproduct contentSite design—provides the look and feeand services to customersWeb hosting—enables e-Retailers to mavailable via the WebWeb analytics—manages visitors navigunderstand trafficWeb performance monitoring—allowsproducts via the Web

Marketing—activitiesthat identifies,anticipates, andsatisfies customerrequirementsprofitability

Affiliate marketing—leverages affiliatecustomer leadsEmail marketing—utilizes email to acqand prospectsSearch engine marketing—enables advsearch engines

Sales—activities thatperform the exchangeprocess of goods andservices to customersin return of money

Rich media—publishes product via vidSite search—lets users find products ane-Retailer’s siteE-Commerce platform—application thto provide e-Commerce to customersOrder management—enables e-RetaileFulfillment—supports shipping of prodof orders to suppliers

eferences

garwal, R., Sambamurthy, V., 2002. Principles and models for organizing the ITfunction. MIS Quarterly Executive 1 (1), 1–16.

nderson, M.C., Banker, R.D., Ravindran, S., 2006. Value implications of investmentsin information technology. Management Science 52 (9), 1359–1376.

ndres, H.P., Zmud, R.W., 2001–2002. A contingency approach to software projectcoordination. Journal of Management Information Systems 18 (3), 41–70.

ng, S., Cummings, L.L., 1997. Strategic response to institutional influences on infor-mation systems outsourcing. Organization Science 8 (3), 235–256.

ngst, C.M., Agarwal, R., Sambamurthy, V., Kelly, K., 2010. Social contagion and infor-mation technology diffusion: the adoption of electronic medical records in U.S.hospitals. Management Science 56 (8), 1219–1241.

PQC, 2012. Process Classification Framework. APQC, Houston, TX.rmstrong, C.P., Sambamurthy, V., 1999. Information technology assimilation in

firms: the influence of senior leadership and IT infrastructures. InformationSystems Research 10 (4), 304–327.

ardhan, I., Mithas, S., Lin, S., 2007. Performance impacts of strategy, informationtechnology applications, and business process outsourcing in U.S. manufactur-ing plants. Production and Operations Management 16 (6), 747–762.

ardhan, I., Whitaker, J., Mithas, S., 2006. Information technology, production pro-cess outsourcing, and manufacturing plant performance. Journal of ManagementInformation Systems 23 (2), 13–40.

arki, H., Rivard, S., Talbot, J., 2001. An integrative contingency model of softwareproject risk management. Journal of Management Information Systems 17 (4),37–69.

endoly, E., Citurs, A., Konsynski, B., 2007. Internal infrastructure impacts onRFID perceptions and commitment: knowledge, operational procedures, andinformation-processing standards. Decision Sciences 38 (3), 423–449.

lank, J.L.T., Van Hulst, B.L., 2009. Productive innovations in hospitals: an empiri-cal research on the relation between technology and productivity in the Dutchhospital industry. Health Economics 18, 665–679.

urns, T., Stalker, G.M., 1961. The Management of Innovation. Oxford UniversityPress, London, Tavistock.

hatterjee, D., Richardson, V.J., Zmud, R.W., 2001. Examining the shareholder wealtheffects of announcements of newly created CIO positions. MIS Quarterly 25 (1),43–70.

heon, M.J., Grover, V., Teng, J.T.C., 1995. Theoretical perspectives on the outsourcingof information systems. Journal of Information Technology 10 (4), 209–219.

lark, T.H., Hammond, J.H., 1997. Reengineering channel reordering processes toimprove total supply-chain performance. Production and Operations Manage-ment 6 (3), 248–265.

iMaggio, P.J., Powell, W.W., 1983. The iron cage revisited: institutional isomor-phism and collective rationality in organizational fields. American SociologicalReview 48 (2), 147–160.

ehning, B., Richardson, V.J., Zmud, R.W., 2003. The value relevance of announce-ments of transformational information technology investments. MIS Quarterly

27 (4), 637–656.vans, P., Wurster, T.S., 1999. Blown to Bits: How the New Economics of Information

Transforms Strategy. Harvard Business School Press, Boston, MA.eeny, D.F., Willcocks, L.P., 1998. Core IS capabilities for exploiting information tech-

nology. Sloan Management Review 39 (3), 9–21.

Flynn, B.B., Huo, B., Zhao, X., 2010. The impact of supply chain integration on per-formance: a contingency and configuration approach. Journal of OperationsManagement 28 (1), 58–71.

Frohlich, M.T., Westbrook, R., 2001. Arcs of integration: an international studyof supply chain strategies. Journal of Operations Management 19 (2),185–200.

Frohlich, M.T., Westbrook, R., 2002. Demand chain management in manufactur-ing and services: web-based integration, drivers and performance. Journal ofOperations Management 20 (6), 729–745.

Germain, R., Claycomb, C., Droge, C., 2008. Supply chain variability, organizationalstructure, and performance: the moderating effect of demand unpredictability.Journal of Operations Management 26 (5), 557–570.

Gilley, K.M., Rasheed, A., 2000. Making more by doing less: an analysis of out-sourcing and its effects on firm performance. Journal of Management 26 (4),763–790.

Gonzalez-Benito, J., 2007. A theory of purchasing’s contribution to business perfor-mance. Journal of Operations Management 25 (4), 901–917.

Greene, W.H., 2008. Econometric Analysis. Prentice Hall, Upper Saddle River,NJ.

Gresov, C., 1989. Exploring fit and misfit with multiple contingencies. AdministrativeSciences Quarterly 34 (3), 431–453.

Guide Jr., V.D.R., Jayaraman, V., Linton, J.D., 2003. Building contingency planning forclosed-loop supply chains with product recovery. Journal of Operations Man-agement 21 (3), 259–279.

Gunasekaran, A., Marri, H.B., McGaughey, R.E., Nebhwani, M.D., 2002. E-commerceand its impact of operations management. International Journal Economics 75(1–2), 185–197.

Gunasekaran, A., Ngai, E.W.T., 2005. Build-to-order supply chain management: aliterature review and framework for development. Journal of Operations Man-agement 23 (5), 423–451.

Gurbaxani, V., Whang, S., 1991. The impact of information systems on organizationsand markets. Communications of the ACM 34 (1), 59–73.

Handfield, R.B., Ragatz, G.L., Petersen, K.J., Monczka, R.M., 1999. Involving sup-pliers in new product development. California Management Review 42 (1),59–82.

Harrison, J.S., Hitt, M.A., Hoskisson, R.E., Ireland, R.D., 2001. Resource complemen-tarity in business combinations: extending the logic of organizational alliances.Journal of Management 27 (6), 679–690.

Haunschild, P.R., Miner, A.S., 1997. Modes of interorganizational imitation: theeffects of outcome salience and uncertainty. Administrative Science Quarterly42 (3), 472–500.

Hayes, R.H., Wheelwright, S.C., 1984. Restoring our Competitive Edge: CompetingThrough Manufacturing. Wiley, New York, NY.

Hood III, M.V., Kidd, Q., Morris, I.L., 2008. Two sides of the same coin: employ-ing Granger causality tests in a time series cross-section framework. PoliticalAnalysis 16, 324–344.

Johnson, M.E., Whang, S., 2002. E-business and supply chain management: anoverview and framework. Production and Operations Management 11 (4),

413–423.Jones, R.J., Hill, W.L., 1988. Transaction cost analysis of strategy-structure choice.

Strategic Management Journal 9 (2), 159–172.

ING ModelO

1 ions M

auffman, R.J., Tsai, J.Y., 2009. The unified procurement strategy for enterprise soft-ware: a test of the move to the middle hypothesis. Journal of ManagementInformation Systems 26 (2), 177–204.

e, W., Wei, K.K., 2008. Organizational culture and leadership in ERP implementa-tion. Decision Support Systems 45 (2), 208–218.

een, P., Williams, R., 2013. Value architectures for digital businesses: beyond thebusiness model. MIS Quarterly 37 (2), 643–648.

etokivi, M.A., Schroeder, R.G., 2004. Strategic, structural contingency and insti-tutional explanations in the adoption of innovative manufacturing practices.Journal of Operations Management 22 (1), 63–89.

ishore, R., Agrawal, M., Rao, H.R., 2004. Determinants of sourcing during technologygrowth and maturity: an empirical study of e-Commerce sourcing. Journal ofManagement Information Systems 21 (3), 47–82.

ogut, B., Zander, U., 1992. Knowledge of the firm, combinative capabilities, and thereplication of technology. Organization Science 3 (3), 383–397.

acity, M.C., Willcocks, L.P., 1998. An empirical investigation of information technol-ogy sourcing practices: lessons from experience. MIS Quarterly 22 (3), 363–408.

ance, C., 1988. Residual centering, exploratory and confirmatory moderator analy-sis, and decomposition of effects in path models containing interactions. AppliedPsychological Measurement 12 (2), 163–175.

awrence, P.R., Lorsch, J.W., 1986. Organization and Environment: Managing Differ-entiation and Integration. Harvard Business School Press, Boston, MA.

eiblein, M.J., Reuer, J.J., Dalsace, F., 2002. Do make or buy decisions matter? Theinfluence of organizational governance on technological performance. StrategicManagement Journal 23 (9), 817–833.

oh, L., Venkatraman, N., 1992a. Determinants of information technology outsourc-ing: a cross-sectional analysis. Journal of Management Information Systems 9(1), 7–24.

oh, L., Venkatraman, N., 1992b. Diffusion of information technology outsourcing:influence sources and the Kodak effect. Information Systems Research 3 (4),334–358.

ove, J., 2011. IRCE 2011: The E-Commerce race is on, Internetretailer.com,available at http://www.internetretailer.com/commentary/2011/01/10/irce-2011-e-commerce-race

cKone, K.E., Schroeder, R.G., Cua, K.O., 1999. Total productive maintenance: a con-textual view. Journal of Operations Management 17 (2), 123–144.

cLellan, K., Marcolin, B.L., Beamish, P.W., 1995. Financial and strategic motivationsbehind IS outsourcing. Journal of Information Technology 10 (4), 299–321.

idumolu, S.R., 1996. A comparison of the structural contingency and risk-basedperspective on coordination in software-development projects. Journal of Man-agement Information Systems 13 (2), 77–113.

olan, R., McFarlan, F.W., 2005. Information technology and the board of directors.Harvard Business Review 83 (10), 96–106.

’Leary-Kelly, S.W., Flores, B.E., 2002. The integration of manufacturing andmarketing/sales decisions: impact on organizational performance. Journal ofOperations Management 20 (3), 22–240.

sei-Bryson, K., Ngwenyama, O.K., 2006. Managing risks in information systems out-sourcing: an approach to analyzing outsourcing risks and structuring incentivecontracts. European Journal of Operational Research 174 (1), 245–264.

armigiani, A., 2007. Why do firms both make and buy? An investigation of concur-rent sourcing. Strategic Management Journal 28 (3), 285–311.

avett, C.M., Lau, A.W., 1983. Managerial work: the influence of hierarchical leveland functional specialty. Academy of Management Journal 26 (1), 170–177.

inker, E.J., Seidmann, A., Foster, R.C., 2002. Strategies for transitioning ‘old economy’firms to e-business. Communications of the ACM 45 (5), 77–83.

orter, M.E., 1985. Competitive Advantage: Creating and Sustaining Superior Per-

formance. The Free Press, New York, NY.uinn, J.B., Hilmer, F.G., 1994. Strategic outsourcing. Sloan Management Review 35

(4), 43–55.aymond, L., 1990. Organizational context and information systems success: a con-

tingency approach. Journal of Management Information Systems 6 (4), 5–20.

Rayport, J.F., Sviokla, J.J., 1996. Exploiting the virtual value chain. The McKinseyQuarterly 1, 21–36.

Sambamurthy, V., Zmud, R.W., 1999. Arrangements for information technology gov-ernance: a theory of multiple contingencies. MIS Quarterly 23 (2), 261–290.

Schein, E.H., 1992. The Role of the CEO in the Management of Change: The Case ofInformation Technology. Oxford University Press, Oxford.

Shah, R., Ward, P.T., 2003. Lean manufacturing: context, practice bundles, and per-formance. Journal of Operations Management 21 (2), 129–149.

Smith, M.A., Mitra, S., Narashimhan, S., 1998. Information systems outsourcing:a study of pre-event firm characteristics. Journal of Management InformationSystems 15 (2), 61–93.

Smith, M.D., Bailey, J., Brynjolfsson, E., 2000. Understanding Digital Markets: Reviewand Assessment. Book chapter in Understanding the Digital Economy: Data,Tools, and Research. MIT Press, Cambridge, MA.

Sousa, R., Voss, C.A., 2008. Contingency research in operations management prac-tices. Journal of Operations Management 26 (6), 697–713.

Spetz, J., Baker, L.C., 1999. Managed Care and Technology Growth. NBER workingpapers (6894), NBER, Cambridge, MA.

Srinivasan, S.S., Anderson, R., Ponnavolu, K., 2002. Customer loyalty in e-commerce:an exploration of its antecedents and consequences. Journal of Retailing 78 (1),41–50.

Steinfield, C., Bouwman, H., Adelaar, T., 2002. The dynamics of click-and-mortarelectronic commerce: opportunities and management strategies. InternationalJournal of Electronic Commerce 7 (1), 93–119.

Subramani, M., 2004. How do suppliers benefit from information technology use insupply chain relationships? MIS Quarterly 28 (1), 45–73.

Swink, M., Nair, A., 2007. Capturing the competitive advantage of AMT: design-manufacturing integration as a complementary asset. Journal of OperationsManagement 25 (3), 736–754.

Tanriverdi, H., Ruefli, T.W., 2004. The role of information technology in risk/returnrelations of firms. Journal of the Association for Information Systems 5 (11–12),421–447.

Teece, D.J., 1986. Profiting from technological innovation: implications for integra-tion, collaboration, licensing and public policy. Research Policy 15 (6), 285–305.

Teo, H.H., Wei, K.K., Benbasat, I., 2003. Predicting intention to adopt interorganiza-tional linkages: an institutional perspective. MIS Quarterly 27 (1), 19–49.

Tiwana, A., 2008. Does technological modularity substitute for control? A study ofalliance performance in software outsourcing. Strategic Management Journal 29(7), 769–780.

Tripsas, M., 1997. Unraveling the process of creative destruction: complementaryassets and incumbent survival in the typesetter industry. Strategic ManagementJournal 18 (SI), 119–142.

Venkatraman, V., Prescott, J.E., 1990. Environment-strategy coalignment: an empir-ical test of its performance implications. Strategic Management Journal 11 (1),1–23.

Wade, M., Hulland, J., 2004. Review: the resource-based view and informationsystems research: review, extension, and suggestions for future research. MISQuarterly 28 (1), 107–142.

Weill, P., Olson, M.H., 1989. An assessment of the contingency theory of manage-ment information systems. Journal of Management Information Systems 6 (1),59–85.

Whitaker, J., Mithas, S., Krishnan, M.S., 2010. Organizational learning and capabilitiesfor onshore and offshore BPO. Journal of Management Information Systems 27(3), 11–42.

Wooldridge, J., 2003. Econometric Analysis of Cross Section and Panel Data, 2nd ed.MIT Press, Cambridge, MA.

Zhu, K., 2004. The complementarity of information technology infrastructure ande-commerce capability: a resource-based assessment of their business value.Journal of Management Information Systems 21 (1), 167–202.

Zwass, V., 1996. Electronic commerce: structures and issues. International Journalof Electronic Commerce 1 (1), 3–23.

information systems and technology sourcing strategies of e-retailers for value chain enablement

Documents

a marketer’s guide to sales enablement making the most out...

best practices for managing ethical sourcing & vendor...

sourcing enablement

cloud enablement

microcontroller software and services enablement ·...

leveraging social media to optimize sourcing-to-shelf...

enablement technologies

strategic sourcing without supplier enablement[1].ppt ·...

retailer enablement

quality enablement - agile practices with quality...

patient identification, enablement and centered...

strategic sourcing without supplier enablement[1].ppt ·...

best-cost country sourcing a next-generation approach · pdf...

responsible sourcing of colored gemstones · jewelry...

solutions for retailers - mastercard | a world beyond cash |...

enablement main

change enablement

what happens when retailers get political? ·...

what we do/media/corporate/pdf-general/pa... · 2016. 10....

organic vegetable storage in wales – opportunities and...