Understanding the effect of knowledge management strategies onknowledge management performance: A contingency perspective§

Tae Hun Kim a, Jae-Nam Lee b,*, Jae Uk Chun b, Izak Benbasat c

a Eli Broad College of Business, Michigan State University, East Lansing, MI 48824, USAb Korea University Business School, Korea University, Anam-Dong, Seongbuk-Gu, Seoul 136-701, Republic of Koreac Sauder School of Business, University of British Columbia, Vancouver, BC V6T 1Z2, Canada

1. Introduction

Developing a knowledge management (KM) strategy is impor-tant in effective KM. An appropriate KM strategy enables a firm tocreate, acquire, access, and leverage knowledge in a timelymanner, thereby resulting in better performance [1]. Consideringthis KM strategy impact, the knowledge-based view (KBV) of thefirm has extended the resource-based view (RBV) of the firm,which contends that organizational resources should be valuable,rare, and appropriable to generate a competitive advantage and besustainable over time because of their low substitutability, lowmobility, and low imitability. That is, the KBV contends thatorganizational knowledge is the primary resource for creating andsustaining competitive advantage [2].

AlthoughpriorstudiesonKMhaveimprovedourunderstandingofKM strategy, its roles and impact are fragmented for several reasons.First, studies on KM strategy have primarily adopted the universalis-tic perspective under the assumption that certain KM strategies areconsistently effective regardless of their organizational contexts[3,4].However,overlookingcontextualfactorscreatesavulnerabilityto contingencies under certain conditions because the effects ofdifferent KM strategies on knowledge management performance(KMP) are themselves affected by a firm’s external and internalcontexts. Nonetheless, the alignment of KM strategy with organiza-tionalcontextshasnotbeenfullyaddressed intheKMliterature[5,6].Second, the KBV has devoted substantial attention to KM strategyanalysisbyidentifyingtwomajordimensionsatthefirmlevel:(1)theextent to which knowledge is accumulated by a person or a system(knowledge type) [7,8]; and (2) whether knowledge originates fromwithinoroutsideafirm(knowledgeorigin)[9,10].However,previousstudies examining the effects of KM strategies on KMP have onlyconsidered a single KM dimension—either knowledge type (system/person) or origin (external/internal)—and have neglected thepossible combinations of these two dimensions [e.g., 4,6]. Therefore,these studies do not elucidate the effect of KM strategies because of

Information & Management 51 (2014) 398–416

A R T I C L E I N F O

Article history:Received 25 March 2013Received in revised form 11 January 2014Accepted 7 March 2014Available online 18 March 2014

Keywords:Knowledge managementKnowledge management strategyKnowledge management performanceContingency perspectiveTechnology–organization–environmentframeworkEnvironmental knowledge intensityOrganizational IS maturity

A B S T R A C T

The universalistic perspective research on employing a unidimensional knowledge management (KM)strategy has yielded conflicting findings and recommendations in different contexts. This study proposesa contingency model for investigating the effects of KM strategies on KM performance to resolve thesecontradictions. Drawing on the knowledge-based view (KBV) of the firm, which identifies knowledgetype and origin as two key KM dimensions, this study first defines four KM strategies: externalcodification, internal codification, external personalization, and internal personalization. A multiplecontingency model of KM strategy is then developed based on a technology–organization–environmentframework. This study proposes that the effectiveness of each KM strategy depends on both external andinternal contextual conditions, namely, environmental knowledge intensity and organizationalinformation systems (IS) maturity. To test and validate the contingency model, we analyze data from141 firms to explain the effects of KM strategies on KM performance. Our results reveal three KMstrategies, not including the internal personalization strategy, which have a significant association withKM performance in their hypothesized contexts. This study expands KM strategy research bytheoretically developing an advanced contingency model aligned with external and internal contextsand by providing valuable practical suggestions to managers for selecting a KM strategy based onmultiple contingencies related to the external and internal conditions of a firm.

! 2014 Elsevier B.V. All rights reserved.

§ This work was supported by the National Research Foundation of Korea Grantfunded by the Korean Government (NRF-2013S1A5A2A01014969).

* Corresponding author. Tel.: +82 2 3290 2812; fax: +82 2 922 7220.E-mail addresses: thkim@broad.msu.edu (T.H. Kim), isjnlee@korea.ac.kr

(J.-N. Lee), juchun@korea.ac.kr (J.U. Chun), izak.benbasat@sauder.ubc.ca(I. Benbasat).

Contents lists available at ScienceDirect

Information & Management

jo u rn al h om ep ag e: ww w.els evier .c o m/lo c ate / im

http://dx.doi.org/10.1016/j.im.2014.03.0010378-7206/! 2014 Elsevier B.V. All rights reserved.

their failure to investigate interactions between different strategicdimensions of KM.

To explore these gaps, this study uses the KBV to defineindividual KM strategies that indicate the two dimensions ofstrategic KM approaches: i.e., knowledge type and origin [11].Combining both dimensions, this study proposes four KMstrategies: external codification (external system-oriented), internalcodification (internal system-oriented), external personalization(external person-oriented), and internal personalization (internalperson-oriented). To suggest the optimal choice of KM initiativesgiven multiple contingencies, we propose a contingency modelbased on the technology–organization–environment (TOE) frame-work [12], which identifies environmental knowledge intensity andorganizational information systems (IS) maturity as two keycontextual factors that interact with KM strategy. As proposedby Sambamurthy and Zmud [13], the multiple key contingenciesthat we propose for organizations using the four KM strategiesposit that KMP is determined by a firm’s fit formation of KMstrategies with its external information-processing needs arisingout of its environment and its internal, technology-orientedcapabilities [14]. The major premise underlying this study is thatthe effectiveness of KM strategies depends on their external andinternal contexts [15,16].

Specifically, this study aims to answer the following question:How does the effect of KM strategies on KMP differ depending on afirm’s external and internal contexts, i.e., the degree of environ-mental knowledge intensity and the level of organizational ISmaturity? This study attempts to answer the question using datacollected from 141 Korean firms that have implemented enter-prise-wide KM initiatives. This study expands KM strategyresearch by theoretically developing an advanced contingencymodel aligned with external and internal contexts and byproviding valuable practical suggestions to managers in selectinga KM strategy that will be successful in different external andinternal contexts. We also believe that our two factor-contingen-cies fill another gap in the KM literature, which faces difficulty inintegrating the effect of KM strategy into multiple contingencies,such as business-related environmental and information system(IS)-oriented organizational contexts [14], in an empirical analysis.The gap is primarily due to the tradeoff between the omittedvariable bias among multiple existing contingencies and aparsimonious research design for robust empirical evidence[17]. The findings of this study can be added to existing studiesfrom North America [4,18,19] and Europe [5,6,8] to provide a moreinternational and comprehensive perspective on KM strategies.

2. Knowledge management strategy

Research on the effect of KM strategies on KMP has yieldedconflicting findings in different contexts [e.g., 8,18–20]. Forexample, certain studies [8,19] propose that the internallysystem-oriented KM strategy provides firms with a competitiveadvantage because people can easily access and acquire codifiedknowledge from internal rather than external sources. However,other studies [18,20] indicate that this strategy has the oppositeeffect. The basis for those studies’ conclusions is that overrelianceon codified knowledge-oriented strategy results in internalknowledge losing its integrity and the causal connections betweenorganizational knowledge and firm-specific contexts in whichknowledge is applied because codified knowledge in electronicform primarily contains basic and general information rather thannew insights or creative ideas. Thus, prior studies have not fullyresolved such conflicts, as summarized in Table 1.

The KBV indicates that conceptualizing the type and origin oforganizational knowledge is important to simultaneously explainorganizational learning [19,20]. However, little is known about the

combined functions of knowledge type and origin, despite theirinterrelationship. This insufficient consideration has seriouslylimited the KBV because the functions and effects of differentknowledge aspects have not been understood in an integratedmanner. Although KM researchers have emphasized the need tosimultaneously consider external and internal contexts, prior KBV-based studies have not integrated multiple contexts into a singlestudy [e.g., 5,6]. In prior KM studies, the relationship of a firm’sstrategic effort with its environment was not a key concern inexplaining how the firm improves its KMP [5]. Therefore, the KBVmust be advanced by considering strategic KM alignment in bothexternal and internal contexts.

To fill these research gaps, this study uses the KBV to define KMstrategies based on two major dimensions of KM—knowledge type(person- or system-oriented) and knowledge origin (internal- orexternal-oriented). The KBV suggests that implementing a KMstrategy requires not only firm-specific accumulated knowledgeassets but also knowledge flows within or channeled into a focalfirm, which is assimilated and developed into its accumulatedknowledge [11]. Therefore, we define KM strategy as a logical planwith regard to firms’ decisions about the types and origins ofknowledge to create and sustain a competitive advantage. KMstrategy does not need to be a conscious, unidimensional decision;rather, it may be a manifestation of multiple decisions.

The effects of KM strategies on KMP, the degree to which a firmachieves knowledge-oriented benefits by adopting and implantingKM [21], have been analyzed from various theoretical perspectives,such as the integrative capability view [10], knowledge sourcingtheory [1], organizational learning theory [22], and transactioncost theory [4]. In particular, the KBV posits that organizationalknowledge is the most significant resource that leads to long-term,sustainable, competitive advantage [2]. The main focus of KBV is onvalue creation through the use of knowledge. Thus, its core purposeis to understand how KM should be pursued to improve a firm’scapability and performance.

A critical contribution of the KBV is the recognition of two KMdimensions based on knowledge type: system-oriented (codifica-tion) and person-oriented (personalization) [3,7]. Although severalstudies across disciplines such as economics, psychology, strategicmanagement, and IS have proposed different dimensions ofknowledge, the most enduring distinctions are both explicit andtacit [8]. Codification and personalization provide underlyingmechanisms for creating, accessing, and acquiring both explicitand tacit knowledge. Codification relies on simple and explicitknowledge and attempts to improve firm performance through theuse of KM systems [23]. Personalization deals with complex andtacit knowledge and applies personal contacts and socializationprocesses to increase the effectiveness of KM [3].

Another contribution of the KBV is identifying two distinct KMchoices based on knowledge origin: internal-oriented and external-oriented [9]. Thus, the forces that motivate a firm toward internalknowledge sourcing may not be the same as those motivating itaway from external knowledge sourcing [24]. The internal-orientedapproach attempts to increase firm performance by integratingknowledge within a firm’s boundaries [18]. Knowledge generatedwithin a firm is unique and specific. Thus, competitors may find itdifficult to imitate that knowledge, yielding considerable value forthe firm. By contrast, the external-oriented approach attempts toimport knowledge from outside sources via acquisition orimitation and then transfer that knowledge within the organiza-tion [25]. Thus, firms can obtain fresh ideas to complement theirknowledge bases, thereby leading to higher KMP [4].

Individual KM approaches can improve KMP. However,generating synergies among the four KM dimensions can be morebeneficial to firms. Given the existing synergies among theapproaches based on knowledge type and origin, the KBV suggests

T.H. Kim et al. / Information & Management 51 (2014) 398–416 399

Table 1Review of key extant studies on knowledge management strategy.

Author(s) Study dimensions Theoretical ap-proach

Researchmethodology

Main findings and contributions

Studies focused on knowledge type in organizationsChoi and Lee [3] ! System-oriented

! Human-oriented! RBV! Universalisticview

! 51 firm-leveldata

! This study defines system- and human-oriented along withdynamic (high system- and high human-oriented) and passive(low system- and low human-oriented) KM styles by focusingonly on knowledge type, thus ignoring knowledge origin.! This study’s results indicate that a dynamic KM styleeffectively enhances performance, thus suggesting theimportance of a balance between tacit and explicit knowledge.However, its results are generalized to organizations withdifferent levels of internal capabilities in various industrieswithout consideration of their given contexts.

Gammelgaard andRitter [20]

! System-oriented! Human-oriented

! KBV! Universalisticview

! Conceptualstudy

! This study classifies KM strategy into codification andpersonalization strategies according to knowledge type.! This study suggests that the personalization strategy can becombined with the codification strategy in cases involvinginternationally operating firms. However, this study does notprovide any empirical evidence.

Leiponen [8] ! System-oriented! Human-oriented

! KBV! Universalisticview

! 16 case studies! 167 firm-leveldata

! This study focuses on knowledge types (collectiveness andtacitness) without consideration of contextual factors.! This study’s results indicate that collectiveness isadvantageous to business service improvements/new serviceintroductions, whereas tacitness is disadvantageous toinnovation.

Revilla et al. [6] ! System-oriented! Human-oriented

! KBV! Contingency viewbased on externalfactors

! Survey datafrom 80functional-levelmanagers

! This study considers only knowledge type as a dimension ofKM strategy, implying that the managerial choice of KMstrategy depends on the environment. However, it only focuseson environment as a context.! This study’s results indicate that a more dynamic and morecomplex environment demands both exploitation andpersonalization strategies. However, it sampled only managersof Spanish companies who are engaged in productdevelopment.

Studies focused on knowledge origin in organizationsCassiman and

Veugelers [5]! External sourcing! Internal sourcing

! Complementaryview! Contingency viewbased on multiplecontexts

! 269individual-leveldata

! This study considers only knowledge origin withoutconsideration of knowledge type.! This study’s results indicate that internal R&D and externalknowledge acquisition are complementary to each other inimproving innovation.! This study shows that the mutual relationship (a contextualfactor) between internal and external innovation strategies issensitive to reliance on basic R&D, which reflectsorganizations’ specific contexts.! This study sampled only innovation-active companies in theBelgian manufacturing industry.

De Clercq andDimov [18]

! External sourcing! Internal sourcing

! KBV! Universalisticview

! Longitudinal200 firm-leveldata of venturecapital industry

! This study defines KM strategy as accessing externalknowledge and developing internal knowledge (knowledgeorigin only).! This study’s results indicate that accessing externalknowledge is more beneficial than developing internalknowledge for firms in more knowledge-intensive industries.In addition, if there are gaps between knowledge obtained by afirm and knowledge that the firm needs, interactions betweenaccessing external knowledge and developing internalknowledge are positive.! With respect to empirical evidence, this study focuses only onthe venture capital investment sector in the U.S.

Nevo et al. [4] ! External sourcing! Internal sourcing

! RBV! Universalisticview

! 111 firm-leveldata

! This study focuses exclusively on knowledge origin indefining innovation strategies, explaining the impacts ofexternal and internal IT capabilities on IT productivity withoutany consideration of potential contextual factors.! This study’s results indicate that relying on external ITconsultants provides tangible benefits (e.g., IT productivity),which are moderated by internal IT capabilities.

Zahra and Nielsen [19] ! External sourcing! Internal sourcing

! RBV! Contingencyperspective basedon internalcontexts

! Longitudinal119 firm-leveldata

! This study considers only internal contexts as a moderator(integration mechanism) without consideration ofenvironmental contexts, suggesting that internal knowledgesources are positively associated with successful technologycommercialization.! This study’s results indicate that formal and informalintegration mechanisms moderate the impact of internal andexternal capability sources on technology commercialization.However, it sampled only U.S. manufacturing companies,causing a lack of generalizability.

T.H. Kim et al. / Information & Management 51 (2014) 398–416400

that firms should consider both dimensions in launching KMstrategies for improving KMP, thereby deriving four KM strategies:external codification, internal codification, external personalization,and internal personalization. These strategies have unique features,as summarized in Table 2.

3. The contingency approach to knowledge managementstrategy

A contingency approach shows fit relationships amongmultiple factors and their resulting influence on relevant criterianot only by identifying ideal configurations that generate optimaloutcomes but also by indicating deviations from such fitconfigurations that cause low performance [15]. The contingencyapproach has been taken by prior studies [13,15,16] to reveal suchfit or misfit among multiple factors. However, research onstrategy–structure–performance relationship has severely criti-cized the contingency theory because of its lack of clarity andfailure to specify the form of interaction between factors ofinterest [26]. In addition, the contingency perspective is limitedby its difficulty in exhaustively defining, measuring, and testingfull contingencies without the omitted variable bias amongpotential contingencies [17] because the fit among factorsdepends on the contingencies that are considered. In this sense,our contingency approach addresses such theoretical andempirical issues via the TOE-based identification of multiple

contingencies [13], which are exhaustively (both internally andexternally) relevant to KM in organizations [27].

Based on the contingency approach, this study suggests thatsuccessful KM requires a firm to employ mixed strategies in itsgiven situations [5,6]. A few recent studies have analyzed theeffects of KM strategies within various organizational contexts,but their contingency perspectives have focused only on eitherenvironmental factors [e.g., 6,18] or internal contexts [e.g., 5,19].Those studies’ failure to simultaneously consider external andinternal contingencies has resulted in inconclusive findings andinappropriate practical guidelines that prevent organizationsfrom developing KM strategies suitable for their situations [27].

To overcome that limitation, this study proposes an advancedcontingency model that reflects both the environmental andinternal fits of the best KM strategy, based on the TOE framework[12], which identifies multiple factors related to organizationalinnovations. Following Sambamurthy and Zmud [13], we arespecifically interested in the multiple contingencies of environ-mental knowledge intensity and organizational IS maturity as keyexternal and internal contexts [22]. As shown in Table 3, KBV positsthat environmental knowledge intensity is a key external contextbecause organizational knowledge is the most significant resourcefor strategic advantage [22]. A knowledge-oriented economycharacterizes the business environment as an industrial trendtoward greater dependence on organizational knowledge [28].Knowledge-oriented logic implies that firms are exposed to

Table 1 (Continued )

Author(s) Study dimensions Theoretical ap-proach

Researchmethodology

Main findings and contributions

This study focused on both knowledge type and origin in organizationsThis study ! System-oriented

! Human-oriented! External sourcing! Internal sourcing

! KBV! Contingency viewreflectingenvironmental fitand internal fit

! 141 firm-leveldata frommultiple ratersin organizations

! This study classifies KM strategy, based on both knowledgetype and origin, into external codification, internal codification,external personalization, and internal personalizationstrategies.! A contingency perspective advances our understanding of theimpact of KM strategies on KMP by representing bothenvironmental fit and internal fit.! This study’s results suggest that KM strategies should bechosen and developed according to external and internalcontexts, i.e., environmental knowledge intensity andorganizational IS maturity.

Table 2Two-dimensional classification of knowledge management strategies.

KM Strategy type KM strategic dimensions Motivation Results

Knowledge type Knowledge origin

External codificationstrategy

! Codifying organizationalknowledge through formalinformation systems

! Accessing external knowledgeacross inter-organizations

! To build up external KM systemfor mutual collaborations

! Reduces the time and effort todevelop valuable knowledge! Eliminates redundancy in anorganization’s knowledge

Internal codification strategy ! Codifying organizationalknowledge through formalinformation systems

! Developing internal knowledgewithin an organization

! To elevate user satisfaction byupgrading an organization’s KMsystems

! Standardizes and generalizesexisting knowledge! Elevates users’ convenienceand satisfaction

External personalizationstrategy

! Personalizing knowledgethrough informal humannetworks

! Accessing external knowledgeacross inter-organizations

! To expand an organizationalnetwork to outside knowledgesources

! Raises employees’ level ofknowledge! Gets closer to customers! Wins more trust fromcustomers! Increases customer satisfaction

Internal personalizationstrategy

! Personalizing knowledgethrough informal humannetworks

! Developing internal knowledgewithin an organization

! To establish an organizationalculture for KM

! Forms an open organizationalculture! Speeds up exchanges ofknowledge among internalexperts! Overcomes time–spacelimitations

T.H. Kim et al. / Information & Management 51 (2014) 398–416 401

Table 3Review of key extant studies on external and internal contexts in strategic knowledge management.

1. Environmental knowledge intensity as external context

Author(s) Theory/Method Constructs of interest External context Internal context Findings and implications

Alvesson [36] ! KBV! Theoreticalstudy

! Knowledge-intensivecharacteristics! Personnel loyalty

! Knowledgeintensiveness

! N/A ! This study classifies knowledge-intensiveand non-knowledge-intensive firms accordingto knowledge intensity, thus explainingbehaviors of employees regardingorganizational loyalty and relationships withclients.! This study finds that firms manage andorganize their professional personneldifferently according to knowledge intensity.

Liao et al. [28] ! KBV! Empiricalstudy

! Knowledge sharing! Absorptive capability! Strategic advantage

! Knowledge-intensiveindustries

! Innovationcapability

! This study analyzes the relationship ofknowledge sharing with absorptive capacityand innovation capability.! This study finds that the extent oforganizational knowledge required topreserve a firm’s valuable heritage, learn newtechniques, solve problems, create corecompetences, and initiate new situationsdiffers according to specific businesses andtasks.

Nonaka andTakeuchi [22]

! KBV! Theoreticalstudy

! KM processes! Knowledge creation

! Knowledge density ! N/A ! This study defines knowledge-intensivefirms, which are exposed to an environmentthat is ‘‘a high-density field’’ in whichmembers cooperate with each otherfrequently and intensively in metaphysicaland analogical ways.! The more mature the market, the strongerthe intensity of its firms’ dependency onqualitative types of information/knowledge.

Todtling and Trippl [29]Todtling et al. [30]

! KBV! Empiricalstudy

! Innovation process! Innovation

! Knowledge-intensivesectors

! N/A ! These studies classify knowledge-orientedenvironments into analytic and syntheticknowledge bases by industrial features.! Considering knowledge-orientedenvironments, these studies assume thatinnovation processes, mechanisms ofknowledge exchange, and respective linkagesin analytic base sectors are different fromthose in synthetic knowledge base sectors.

Common interpretations of these studies:! Regarding the external context, this review indicates that prior KM studies have focused on only knowledge-intensive firms and excluded traditional firms that exist inthe current knowledge-oriented era even though the studies have formed an obvious consensus on a significant difference in environmental requirements betweenknowledge-intensive and traditional firms.! The result of the exclusion of traditional firms is that these studies’ implications of the best KM strategy choice are restricted to only knowledge-intensive firms in whichenvironmental knowledge intensity is high.! Although the existing KBV cannot differentiate the external context of firms from the general strategy literature, KM strategy studies should be based on the contingencyperspective, considering the knowledge-oriented external context and thus appropriating the contingency KBV.! For the contingency KBV, we suggest that environmental knowledge intensity is a key alternative that well reflects the knowledge-oriented environmental context.

2. Organizational IS maturity as internal context

Author(s) Theory/Method Constructs of interest External context Internal context Findings and implications

Bhatt and Grover [34] ! RBV! Empirical study

! Organizational learningintensity! Strategic advantage

! N/A ! IT capabilities ! This study is motivated by a controversyabout whether IT can provide differentorganizational capabilities to individual firms.! This study finds that the relationshipbetween IT infrastructure quality andcompetitive advantage is not significant, thusimplying that the quality might not directlyinfluence firm performance.

de Burca et al. [33] ! Contingency! Empirical study

! Best practices! Service/businessperformance

! Servicessector

! IT sophistication ! This study considers IT sophistication (ITmaturity) as a key organizational context anddenies its direct effect on firm performance,thus suggesting instead that it has contingenteffects.! According to this study, a firm’s IT maturitydetermines its capacities to acquire and applyorganizational knowledge.

Drucker [31] ! KBV! Theoretical study

! Information-orientedfirms’ characteristics andrequirements

! Knowledge-intensiveenvironment

! IS in a firm! A firm’s structure

! This study defines ‘‘information-orientedorganizations,’’ which are restructured withhigh IS maturity for KM processes.! In addition to high IS quality, this studyindicates the main characteristics andrequirements of information-oriented firms.

T.H. Kim et al. / Information & Management 51 (2014) 398–416402

analytic or synthetic knowledge bases in their environments[29,30]. By contrast, the RBV emphasizes IS as a set of key internalresources for sustainable competitive advantage [31]. The ISliterature suggests that each firm has different endowments of ISresources at its discretionary disposal to identify, assemble,deploy, and use technologies because IS resource barriers causea firm to achieve a certain level of IS maturity by exploiting itsidiosyncratic set of technologies [32]. Accordingly, a firm estab-lishes its IS resources to a greater (or lesser) degree than itscompetitors because it depends more (or less) on IS resources[33,34]. In this sense, the best KM strategy is contingent on ISmaturity decided as a discretionary matter by a firm’s dependencyon IS assets and resources [35].

This study investigates the contingent effects of different KMstrategies on KMP by simultaneously considering (1) environ-mental knowledge intensity, or the extent to which a firm relies onknowledge inherent in its activities and outputs for a competitivemarket advantage [6,36], and (2) organizational IS maturity, or thedegree of IS effectiveness in supporting organizational decisionsand the contribution of IS to a firm’s desired outcomes [37,38]. TheTOE framework suggests that changes in organizations are causedby a confluence of multiple sources [12]. Thus, this studyintegrates a firm’s strategic choices (i.e., organizational decisionsabout KM strategies), evolutionary forces from outside a firm (i.e.,environmental knowledge intensity), and IS competence-orient-ed forces within a firm (i.e., organizational IS maturity) in a singleresearch model. In the sections that follow, these representationsof knowledge intensity and IS maturity are explained in detail.

3.1. Environmental knowledge intensity

Environmental knowledge intensity reflects knowledge-orient-ed requirements from external circumstances in which a firmmanages and organizes its tasks and businesses differently [36],thereby forming external, context-driven propensities related toKM strategy [6]. Environmental knowledge intensity has beenconsidered as an important key context because the extent oforganizational knowledge required to preserve a firm’s valuableheritage, learn new techniques, solve problems, create corecompetences, and identify new business opportunities differsaccording to industry-specific businesses and tasks [28]. A firm’sinnovation is influenced by its knowledge-oriented environment[39]. Thus, a firm must confront different innovation logics to besuperior to its competitors.

In this sense, a widely accepted assumption is that theinnovation that firms require depends on their environmental

necessities with respect to key knowledge sources, codified andtacit knowledge functions, and types of knowledge links [40],which distinguish between analytic and synthetic knowledgebases for successful businesses [29]. A low knowledge-intensiveenvironment composed of a synthetic knowledge base, which isdominant in some industries in which competition is based on thelogic of incremental innovations (e.g., food products and bev-erages), is principally characterized by the application or novelcombination of existing knowledge, a relatively low R&D level, anda strong orientation toward solving specific problems articulatedby customers [30]. By contrast, firms in a high knowledge-intensiveenvironment compete against each other according to a differentlogic, that is, an analytic knowledge base, which requires thosefirms to realize radical innovations through effective interplaybetween their codified and tacit knowledge in other industries(e.g., pharmaceuticals and electrical machinery and apparatuses)[30]. Thus, this study characterizes the knowledge-orienteddynamic environment as having high knowledge intensity (i.e.,analytic knowledge bases) and the knowledge-oriented stagnantenvironment as having low knowledge intensity (i.e., syntheticknowledge bases).

3.2. Organizational IS maturity

Organizational IS maturity has been addressed in differentterms, including IS success [41], IS maturity [42], IS effectiveness[43], and IS evaluation [44]. This comprehensive concept embracesthe internal development of information resources, the properintegration of computer-based systems, and the ability of users toutilize organizational systems. Based on the RBV, organizationalcapability implies that firms create new knowledge internally andleverage their existing knowledge, thereby achieving favorablepositions in their external contexts [45]. In this sense, IS maturity iscrucial to improving a firm’s ability to use its prior knowledge. ISmaturity also helps firms to recognize the value of newinformation, assimilate it, and apply it to create new knowledge.Therefore, IS maturity affects not only a firm’s ability to designknowledge types and origins that meet its internal needs [37] butalso the quality of key functions that channel and utilizeknowledge both within and outside of the firm [38]. IS maturityis a decisive factor in determining how a firm leverages itsknowledge effectively; thus, a firm’s KM strategy choice shouldadequately align with its IS maturity to form an internal fit forsuccessful KM [41].

However, firms do not necessarily need high IS maturity to besuccessful [7]. On the one hand, high organizational IS maturity

Table 3 (Continued )

2. Organizational IS maturity as internal context

Author(s) Theory/Method Constructs of interest External context Internal context Findings and implications

Raymond [35] ! Contingency! Empirical study

! Organizational contexts! IS success

! N/A ! IS sophistication ! This study suggests that IS maturity(organizational maturity and ISsophistication) is an important internalcontext in firms.! This study finds that IS success is somewhatcorrelated to organizational maturity,whereas the correlation is not significantwhen IS sophistication is controlled.

Common interpretations of these studies:! Prior KBV studies have considered mature organizational IS not as an organizational context but as an organizational capability, which explains KM’s success in firms.! This review shows a controversy related to the strategic role of IS, which enables or inhibits the management of organizational knowledge according to the KM strategicgoal, thus threatening a universalistic expectation that higher organizational IS maturity in firms translates to greater success of KM.! Thus, we suggest a contingency approach in which organizational IS maturity is considered not as a key explainer of KMP, but rather as a key internal context of firms inanalyzing the impact of KM strategy.

T.H. Kim et al. / Information & Management 51 (2014) 398–416 403

provides high-quality, reliable, and fast KM systems by encourag-ing organizational workers to rely on codified knowledge in a‘‘reuse economy,’’ in which investing once in a knowledge assetand recursively using it are required. On the other hand, loworganizational IS maturity generates competitive advantages byreducing unnecessary investments in KM systems and facilitatingconversations and exchanges of tacit knowledge. Regardless of thelevel of organizational IS maturity, individual expertise channelscan provide a firm with creative, analytically rigorous advice aboutand highly customized solutions to its unique tasks, therebyyielding substantial profits with competitive advantages. Ourconsideration of IS maturity as a key internal context reflects thecontingent relationship between a firm’s KM strategy choices andits IS resource endowments. That is, successful KM is only partiallydependent on IS maturity, which indicates IS-oriented resourcesand capabilities [41]. This implies that a firm’s best KM strategy forKMP improvement depends on its IS maturity level.

4. Contingency hypothesis development

Our contingency perspective on the best choice of KM strategytheoretically relies on the TOE framework, which explains thesuccessful adoption of KM systems and practices in organizations[12]. Simultaneously dealing with knowledge intensity from firms’environments and achieving an appropriate level of IS maturity isimportant for firms to be competitive in KM. Both contexts must beconsidered in developing a contingency model to suggest the bestKM strategies across internally and externally different situations.The research model proposed in this study is represented withcontingent expectations for each of the four KM strategies withdifferent levels of environmental knowledge intensity andorganizational IS maturity, as depicted in Fig. 1.

4.1. High knowledge intensity and high IS maturity: Cell 1

External codification strategy refers to a firm’s attempt toaccess specialized and standardized external codified learningsources, such as technical reports, trade journals, patents, andother sources [25]. This strategy provides opportunities for a firmto improve its competitiveness by benchmarking other successfulfirms and conveying stories about best practices to its employees[46]. Such standardized codified knowledge is usually lesssensitive to space than is tacit knowledge that is embedded in aperson. Thus, a firm can easily adopt external codified knowledgeand assimilate it into other activities and processes to create newknowledge.

An organization with high knowledge intensity relies heavily onexternal knowledge sources [30]. Such an organization is more

likely to continuously monitor other firms to identify and imitatesuperior solutions and to combine them with its own knowledge[47]. The strategy also facilitates the creation and sharing ofvaluable organizational knowledge, thereby making the firm moreinnovative. Therefore, accessing external knowledge might haveimportant functions in enhancing knowledge creation and sharingand achieving organizational innovation at a fast pace. In addition,a system-oriented approach focuses on codifying and storingknowledge via IS orientation [48,49]. A firm with high IS maturityaccumulates the specific knowledge of its experts and transferssuch codified knowledge to other firm members through suchtechnologies as video conferencing, groupware, intranet chattingtechnology, virtual reality systems, and online communities forefficient and timely communication [50]. Therefore, firms withhigh IS maturity effectively codify organizational knowledgethrough the use of IS-oriented tools and channels. We suggestthat external codification is the most suitable strategy for firmswith high organizational IS maturity and high environmentalknowledge intensity.

H1. When a firm’s organizational IS maturity and environmentalknowledge intensity are both high, the external codification strat-egy is the most effective way to improve that firm’s KMP.

4.2. Low knowledge intensity and high IS maturity: Cell 2

Internal codification strategy refers to a firm’s effort to manageknowledge residing in its internal IS or documents (e.g., when amanufacturing firm sources knowledge from its enterpriseknowledge portal for the convenience of internal users and sharesvaluable content with internal users). This strategy improves afirm’s ability to access, update, and utilize standardized internalcodified knowledge. Because internal computerized knowledgeprovides ease of access and comprehension [51], this strategyfacilitates knowledge sharing [19] and enables a firm to achieveeconomics of reuse and to decrease search costs by reconfiguringinternal computerized knowledge to fit new situations [52],thereby improving KMP.

An organization with low knowledge intensity is characterizedby synthetic knowledge bases, not analytic ones [29,53]. Environ-mental requirements for synthetic knowledge bases primarilyinclude the application of existing knowledge, the importance ofproblem-related knowledge derived from inductive processes, anda strong orientation toward solving specific but routine problems.Learning to acquire practical skills is also crucial for firms with lowknowledge intensity, resulting in incremental innovations, notradical ones. Given the key features of synthetic knowledge bases,firms should focus on developing their internal knowledge withhigh IS maturity. Therefore, we suggest that internal codification isthe most suitable strategy for improving KMP, in situationsinvolving firms that have high organizational IS maturity and lowenvironmental intensity.

H2. When a firm’s organizational IS maturity is high and itsenvironmental knowledge intensity is low, the internal codifica-tion strategy is the most effective way to improve that firm’s KMP.

4.3. High knowledge intensity and low IS maturity: Cell 3

External personalization strategy refers to a firm’s orientationtoward the use of external personal learning (e.g., a consultingfirm’s request for internal consultants to attend various seminars,conferences, and workshops). For example, users share theirinnovative ideas with outsiders to solve their own and sharedproblems in an open-source software development environment.This situation provides an opportunity for firms to co-createFig. 1. The contingency model of knowledge management strategies.

T.H. Kim et al. / Information & Management 51 (2014) 398–416404

products with the participation of external customers [54].Suppliers can also improve their competitive advantage byexpanding their firms’ knowledge base [25].

External personalization strategy yields higher KMP becausedirect contact with people outside of a firm opens that firm todifferent viewpoints and expands its knowledge base [25]. Asdiscussed above, the externalization of organizational knowledgeis appropriate for a firm with high knowledge intensity. For a firmwith low IS maturity, creating and sharing organizationalknowledge through knowledge personalization is more effectivethan knowledge codification. Human-oriented networks enableorganizational members to identify and share knowledge viainterpersonal interactions with external sources and professionalgroups, along with knowledge transfer via communications withexternal experts [48]. Next, individuals personalize knowledge bysharing it with coworkers through personal contacts. A firm inwhich intellectuals share knowledge through human-orientednetworks can establish knowledge dialogs with external expertsrather than knowledge objects from databases [7]. Personalizedknowledge, which is not yet or cannot be codified, can betransferred through brainstorming sessions and one-on-oneconversations, even in organizations with low IS maturity. Thus,external personalization might be the best strategy for a firm withlow organizational IS maturity and high environmental knowledgeintensity to encourage its members to personalize knowledge viahuman-based external networks.

H3. When a firm’s organizational IS maturity is low and itsenvironmental knowledge intensity is high, the external person-alization strategy is the most effective way to improve thatfirm’s KMP.

4.4. Low knowledge intensity and low IS maturity: Cell 4

Internal personalization strategy refers to a firm’s ability toaccess, acquire, and leverage knowledge from its internalpersonnel. Examples include valuable experience, know-how,and internal expert networks that a firm sources from communi-ties of practice (COPs) in which collective, joint sense-making andcollaborative problem-solving form strong interpersonal tiesamong members and encourage direct reciprocity within acommunity with common interests [55]. Unique knowledge, suchas the skills and experiences of internal personnel, provides a firmwith a competitive advantage [19]. Knowledge resulting from thiscombination cannot be quickly amassed because this knowledgehas firm-specific and tacit characteristics, creating difficulty inimitating and copying this knowledge [22].

Internal personalization strategy helps firms to increase theirperformance by focusing on an internal and person-orientedapproach, in that they not only do they develop their own corecompetencies and appropriate benefits, but also they effectivelyand efficiently control and understand tacit knowledge in the KMprocess [10]. The characteristics of synthetic knowledge basessuggested in H2 indicate that a firm with low knowledge intensityis more likely to cultivate and utilize its internal knowledge toimprove KMP. Moreover, as discussed in H3, a firm with low ISmaturity can be expected to improve its KMP by relying on human-oriented networks. By internally personalizing knowledge, a firmattempts to informally acquire and share its valuable internalknowledge [20] through one-on-one mentoring, face-to-face helpby internal experts, and informal dialogs for knowledge sharingwithin organizational boundaries [49]. Therefore, internal person-alization is proposed as the best strategy for firms with lowenvironmental knowledge intensity and low organizational ISmaturity.

H4. When a firm’s organizational IS maturity and environmentalknowledge intensity are both low, the internal personalizationstrategy is the firm’s most effective way to improve that firm’sKMP.

5. Research methodology

5.1. Sample and procedure

Data for empirical examination were obtained through a surveyconducted in Korea. For the representativeness of the study sampleand the generalizability of the results, we first referred to the ninthedition of the Korean Standard Industrial Classification [56] toselect an initial sample of 154 firms whose industrial compositionwas proportional to that in Korea. In this sampling phase, weensured that the selected firms had implemented KM initiativesbecause it would have been impossible to examine the contextualrelationships between KM strategies and KMP by sampling firmsthat had never attempted KM and thus would have had had norecords of KMP. We communicated with the managers in charge ofKM at the 154 selected firms. We explained the purpose of thisstudy and the contents of the questionnaires, and then asked themanagers to select 5–10 employees with backgrounds that wouldmake them eligible to answer the survey questions. We asked themanagers to choose multiple respondents because KM strategiesand KMP are firm-level phenomena that are better assessed bymultiple raters across different ranks, functions, ages, gender,organizational tenures, and years in the focal industries [57]. Wethen visited the headquarters of the selected firms, personallydistributed 738 survey questionnaires, and collected data on site.To ensure confidentiality and minimize socially desirableresponses, enclosed with each questionnaire was a joint research-er-company cover letter stating that respondents should corre-spond only with researchers when returning the survey, that theemployees’ firms would not have access to individual ratings, andthat aggregated results without firm identification would bereported.

Of the 738 survey questionnaires collected from the 154 firms,we filtered out 78 responses from 13 firms, which either containedunreasonable numbers of missing values for key study variables orfailed to meet the adequate level of index for interrater reliabilitywithin the firm, rwg(j) [58], yielding a total of 660 survey sets from141 firms for hypothesis testing (raters per firm: M = 4.68,SD = 1.09, min = 2, max = 10). KM strategies and KMP are firm-wide decisions, practices, and outcomes. Thus, the unit of analysisfor examining the KM strategy-KMP relationship in specificcontexts was placed at the firm level. In addition, multipleindividual ratings on a study variable in each firm were aggregatedto obtain an average score to represent firm-level properties.Details on the size and age of the 141 firms and demographicinformation for the 660 respondents are presented in Table 4.Analysis of variance (ANOVA) revealed no significant differences instudy variables based on the number of raters in each firm acrossrespondents’ rank, age, gender, organizational tenure, and years inthe industry. All of the final data were pooled for analysis.

5.2. Measurement

An initial version of the survey instrument was vetted through aseries of personal interviews with five academic experts in KM. Asa pilot test, the survey instrument was then administered to 56graduate students in Master of Business Administration programsat two top-tier universities in Korea. The respondents had atleast three years of KM-related work experience. The multi-phase development of the instrument resulted in a number of

T.H. Kim et al. / Information & Management 51 (2014) 398–416 405

refinements, along with restructuring. The process helped toestablish the measures’ initial face and internal validity, whichconfirmed the suitability of the questionnaire for studying real-world phenomena [59]. The Korean version of all of the measuresthat were adapted from prior studies was created by followingBrislin’s [60] translation-back-translation procedure. Unlessotherwise indicated, each measure was assessed on a seven-pointresponse scale ranging from 1 (strongly disagree) to 7 (stronglyagree). The structure of all of the measures that were used in thisstudy is shown in Appendix.

5.2.1. Knowledge management strategiesGiven the absence of an established measure for the four types

of KM strategies, we developed single-item measures for KMstrategies by combining two KM dimensions: knowledge type [7]and knowledge origin [9]. For example, sample items state ‘‘Mycompany relies on a KM strategy that aims at both codifyingorganizational knowledge through systems and accessing externalsources for knowledge’’ for external codification and ‘‘My companyrelies on a KM strategy that aims at both personalizingorganizational knowledge through human networks and develop-ing knowledge from internal sources’’ for internal personalization.

Klein and Rai [61] have identified conditions under whichsingle-item measures are acceptable in the IS research domain: (a)when the concentration of respondents must be maintained byshortening the length of the survey, (b) when additional questionscause unnecessary redundancies, and (c) when constructs areconsidered as unambiguous and narrowly focused. During thesurvey administration phase, respondents in each firm wererequired to attend an explanation and Q&A session beforeanswering the survey. We provided in-depth explanations andexamples of how organizational knowledge is accumulatedthrough systems or personal networks (knowledge type) andhow it is internally or externally regulated (knowledge origin). Thisexplanation was followed by a Q&A session to further ensure thatthe respondents had an accurate understanding of KM strategies.This procedure also helped respondents to focus on the survey byenhancing their precise understanding of its subject and prevent-ing boredom resulting from inter-item redundancy. In addition, in

the survey we presented the four single-item measures in parallelto help respondents differentiate among KM strategies and todetermine the relative dominance of certain KM strategiesover others in a firm [62]. Thus, respondents from each firmindicated the extent to which their firms use each of the four KMstrategies. Using single-item measures for the four types ofKM strategies seemed acceptable when considering the ques-tionnaire’s procedures and design. Moreover, as detailed below,empirical evidence for the reliability of single-item measures wasestimated.

We used various within-firm interrater reliability indices,namely, rwg(j), h2, and intraclass correlation coefficients (ICCs:ICC(1) and ICC(2)), for three reasons. First, because single-itemmeasures were used to assess KM strategies, we could not estimatewithin-rater, internal consistency reliability coefficients, such asCronbach’s alpha. Instead, we used interrater reliability indices forthe single-item measures to determine whether multipleresponses to a single-item measure within a firm were consistentand reliable. Nunnally [59, p. 191] states that measurementreliability represents the extent to which the instrument for aconstruct is intended to be repeatable and stable and to generatethe same results over different conditions (e.g., either acrossdifferent raters or with multiple items or both). Thus, usinginterrater reliability indices for single-item measures is appropri-ate to ensure the reliability of measures in this context. Second, weestimated the within-firm interrater reliability indices to supportthe aggregation of multiple ratings of a KM strategy and itsperformance within a firm for producing their firm-level averagescores and testing their relationships at the firm level. Third, weestimated all interrater reliability indices based on the recom-mendation of Klein and Kozlowski [57] because each index has itsown strengths and weaknesses and there is no single best index.

rwg(j) is an index of within-unit interrater agreement [58]. Thus,the rwg(j) index was calculated separately for each unit. An averagerwg(j) value across units is typically reported [57]. The criterion forrwg(j) for supporting within-unit agreement is above 0.70 [58].Because within-unit agreement does not necessarily representbetween-units variance, other indices comparing within- andbetween-variance, such as h2, ICC(1), and ICC(2), with F-test

Table 4Profile of companies and respondents.

(a) Number of employees (b) Firm age

Range Frequency Percent Year Frequency Percent

Fewer than 50 19 13.5 Less than 10 19 13.551–100 7 5.0 11–20 22 15.6101–500 25 17.7 21–30 20 14.2501–1000 15 10.6 31–40 25 17.71001–5000 48 34.0 41–50 25 17.75001–10,000 10 7.1 51–100 24 17.010,001 and above 17 12.1 101 and above 6 4.3

Total 141 100.0 Total 141 100.0

(c) Demographics of individual respondents (n = 660)

Measure Items Freq. Percent Measure Items Freq. Percent

Gender Male 477 72.3 Gender Female 183 27.7

Position Staff members 270 40.9 Age 21–30 283 42.9Assistant managers 244 37.0 31–40 296 44.8General managers and above 113 17.1 41–50 71 10.8Others: Experts/researchers 33 5.0 51 and above 10 1.5

Years in the industry Fewer than 5 320 48.5 Organizational tenure Fewer than 5 430 65.26–10 175 26.5 6–10 142 21.511–15 92 13.9 11–15 46 7.016–20 52 7.9 16–20 30 4.521 and above 21 3.2 21 and above 12 1.8

T.H. Kim et al. / Information & Management 51 (2014) 398–416406

significance levels derived from ANOVA, were also used supple-mentarily [57].

h2 is an estimation of the relative amount of between-unitversus within-unit variance across an entire sample of units. Thegreater the between-unit variance relative to the within-unitvariance, the higher h2 will be. Bliese [63] notes that h2 isinfluenced by unit size (the number of raters in the unit); thesmaller the unit, the higher the obtainable h2 value. ICC(1) providesan estimate of the proportion of total variance of a measureaccounted for by unit membership. Unlike h2, this index is notinfluenced by unit size [63]. A significant F-test and the suggestedacceptable range of 0.05–0.50 are often used to determine whetherunit-level property exists in individual responses [57]. ICC(2)estimates the reliability of unit mean values and is a function ofICC(1) and unit size [63]. The greater the ICC(1) or unit size, thehigher the obtained ICC(2) value. Like other reliability measures,ICC(2) values are commonly considered to be acceptable foraggregation if they are equal to or exceed 0.70 [57].

Table 5 shows adequate levels of interrater reliability indices forthe single-item measures of KM strategies. These indices supportthe aggregation of individual ratings into firm-level average scores.These estimates of within-firm agreement and between-firmvariance prove the reliability of the single-item measures,indicating that respondents within a firm had similar under-standings of the single-item measures and answered accordingly.Although reliability does not guarantee the validity of thesemeasures, shared understandings among multiple raters and the

ratings within each firm were unlikely to be substantially biasedbecause this study’s sampling and survey administration were notbased on convenience but on a rigorously planned procedure, asdescribed above.

5.2.2. Environmental knowledge intensityTo determine whether a sample firm exists in a knowledge-

intensive external environment, we adapted the OECD’s industrialclassifications [64] and categorized sample firms into environ-ments that were either high or low in knowledge intensity. TheOECD’s classifications [64] are based on R&D expenditures, and itsoutput data differentiates between knowledge-intensive and otherindustries by focusing on ‘‘the leading producers of high-technology goods and on the activities (including services) thatare intensive users of high technology and/or have the relativelyhighly skilled workforces necessary to benefit fully from techno-logical innovations’’ (p. 210). This industrial classification ofknowledge intensity reflects a categorization of industrial sectorsthat distinguishes between analytic and synthetic knowledgebases [29]. The OECD classification also generally accommodatesthe notion that firms in knowledge-intensive industries primarilyrely on intellectual work by well-educated, qualified employees[36].

In the OECD classification, industries are generally divided intotwo categories, as presented in the left-hand side of Table 6: themanufacturing sector and the service sector (divided into marketand non-market services). First, firms in high-technologymanufacturing industries were classified as those with highknowledge intensity. According to the International StandardIndustrial Classification (ISIC), specific sectors corresponding tothese sample firms include pharmaceuticals, medicinal chemicals,and botanical products (2423: the code by ISIC Revision 3), specialpurpose machinery (292), electrical machinery and apparatuses(31), radio, television and communication equipment and appara-tuses (32), and aircraft and spacecraft (353). By contrast, samplefirms in low-technology manufacturing industries were classifiedas those with low knowledge intensity. Specific sectors corre-sponding to these firms are food products and beverages (15),textiles and apparel (17, 18), rubber and plastics products (25),

Table 5Aggregation indices for study variables.

Variables rwg h2 ICC(1) ICC(2) F

External codification 0.812 0.477 0.336 0.703 3.370**

Internal codification 0.818 0.528 0.401 0.758 4.136**

External personalization 0.835 0.442 0.291 0.658 2.922**

Internal personalization 0.844 0.482 0.343 0.710 3.440**

IS maturity 0.987 0.520 0.392 0.751 4.014**

KM performance 0.949 0.442 0.293 0.660 2.939**

** p < 0.01.

Table 6Sector-specific classification of environmental knowledge intensity.

Industries Environmental knowledge intensity

Organizations high in knowledge intensity (N = 62) Organizations low in knowledge intensity (N = 79)

Manufacturing (N = 24) ! Pharmaceuticals, medicinal chemicals and botanical products (N = 1)! Special purpose machinery (N = 2)! Electrical machinery and apparatuses (N = 2)! Radio, television and communication equipment and apparatuses (N = 4)! Aircraft and spacecraft (N = 1)

! Food products and beverages (N = 2)! Textiles (N = 1)! Apparel (N = 1)! Rubber and plastics products (N = 2)! Basic iron and steel (N = 2)! Fabricated metal products, except machineryand equipment (N = 3)! General-purpose machinery (N = 3)

N = 10 N = 14Services (N = 117) ! Telecommunications (N = 3)

! Non-life insurance: reinsurance, fire insurance (N = 2)! Security-dealing activities (N = 2)! Public administration and defense; compulsory social security (N = 13)! Computer and related activities (N = 15)! Research and development (N = 1)! Other business activities

- Legal activities (N = 2)- Accounting (N = 2)- Market research and public opinion polling (N = 2)- Business/management consultancy activities (N = 8)! Education (N = 1)! Hospital activities (N = 1)

! Construction (N = 8)! Wholesale household goods (N = 1)! Non-specialized retail trade in stores (N = 15)! Monetary intermediation (N = 14)! Life insurance (N = 6)! Activities auxiliary to financial intermediation (N = 19)! Other service activities (N = 2)

N = 52 N = 65

T.H. Kim et al. / Information & Management 51 (2014) 398–416 407

basic metals and fabricated metal products (27, 28), andmanufacturers of general-purpose machinery (291). The abovecategorization of sample firms in the manufacturing industry(N = 24) produced 10 companies with high knowledge intensityand 14 firms with low knowledge intensity.

Second, the OECD industrial classification also specifies theservice sectors that involve knowledge-intensive activities. Theseservice sectors include telecommunications (642: ISIC Revision 3code), finance and insurance (65–67), computer-related services(72), R&D (73), and other business services (74), including suchmarket-oriented services as legal services, accounting, marketresearch, and management consultancy, along with non-market-oriented services such as education (80) and health (85). However,the OECD classification does not identify service sectors that do notinvolve knowledge-intensive activities. Furthermore, many KMscholars [e.g., 65,66] posit that a disparity in knowledge-relevantactivities exists among firms in a particular sector of the serviceindustry and that not all service firms depend on know how to offerfairly sophisticated knowledge or knowledge-based services.

To supplement these gaps, we elaborated the OECD classifica-tion of knowledge intensity in the service industry, particularlyfocusing on the finance and insurance sector and other sectorsthat are not specified by the OECD classification but that wereincluded in our sample. Specifically, service firms in business-to-business (B2B) relationships with a few clients, which can create arelatively unpredictable and unstable market environmentshould be considered to exist in a knowledge-intensive businessenvironment [65,66]. By contrast, service firms with business-to-customer (B2C) relationships that offer large-scale services tonumerous customers primarily rely on widely applicable,standardized and routine approaches. Thus, these firms shouldbe classified as being of low knowledge intensity [65,66].Accordingly, institutes in the sector of public services to thecommunity as a whole (752: ISIC Revision 3 code) werecategorized as high knowledge intensity. In the sector of financeand insurance, sample firms in security-dealing activities (6712)and non-life insurance (6603), including reinsurance and fireinsurance, were classified as high knowledge intensity, whereasfirms in the sectors of activities auxiliary to financial intermedia-tion (6719), other monetary intermediation (6519), and life andautomobile insurance (6601) were categorized as low knowledgeintensity. Lastly, firms in the sectors of wholesale householdgoods (513), non-specialized retail sales (521), construction (45),and other service activities (930) were classified as lowknowledge intensity because the primary operation of thosesectors does not depend on intellectual work by highly skilled andwell-educated employees [36,65]. Taken together, this elaboratedclassification of sample firms in the service industry (N = 117)yielded 52 and 65 institutes in high and low knowledge intensityenvironments, respectively. Table 6 presents the results ofknowledge intensity classification and sample size in eachcategory.

5.2.3. Organizational IS maturityWe adapted 10 items from the criteria provided by Kumar [67]

for evaluating organizational IS implementation to tap the degreeof developing information resources, integrating computer-basedsystems, and enabling users to utilize organizational systems, aspresented in Appendix. In the IS literature, organizational ISmaturity has been examined under various terms, such asIS success [41], IS maturity [42], IS effectiveness [43], andIS evaluation [44]. Compared to other criteria for evaluating ISimplementation, Kumar’s [67] instrument is unique because of itspost-implementation evaluation of IS in organizations [44]. Thus,this instrument is consistent with our purpose, which is tomeasure the current state of IS maturity in an organization as the

key internal organizational context, which is neither an anteced-ent nor an outcome of KM strategies and KMP. The results of ourfactor analyses show that the 10 items from Kumar’s [67] criteriaare not cross-loaded on the measures of KMP. A sample item reads,‘‘The IS in my company provides a user-friendly interface.’’Aggregation of multiple ratings on the organizational IS maturitywithin a firm is supported by rwg(j) = 0.99, h2 = 0.52, ICC(1) = 0.39,ICC(2) = 0.75, F(140, 519) = 4.01, p < 0.01, as shown in Table 5. Theinternal consistency reliability estimate of Cronbach’s alpha forthis measure is 0.98.

5.2.4. Knowledge management performanceChong et al. [21] has proposed a comprehensive set of 38 items

that measures KMP and that represent the effectiveness of KM infive areas: systematic knowledge activities, employee develop-ment, customer satisfaction, external relationships, and contribu-tion to organizational success. Of the 38 items, we used five itemsthat represent each of the five dimensions but are distinct fromoverall firm performance. To capture the effectiveness of KM onthese dimensions while ruling out compounding with generalorganizational performance, we specified the referent as ‘‘KM inmy company’’ in the question items, as shown in Appendix. Asample item states, ‘‘KM in my company is effective in enhancingthe value of products and services.’’ Multiple individual ratings onKMP within a firm were aggregated based on rwg(j) = 0.95, h2 = 0.44,ICC(1) = 0.29, ICC(2) = 0.66, F(140, 519) = 4.01, p < 0.01, as pre-sented in Table 5. Cronbach’s alpha for this measure is 0.94.

5.2.5. Control variablesFirm size and age were controlled in the analyses because of

their potential effects on KMP [38]. To measure firm size, we usedthe natural logarithm of the number of employees to correct thediminishing effect of firm size, given the wide variation in thenumber of employees in the study sample [68]. We also controlledthe effect of firm age, which was estimated by the number of yearsthat a firm had existed. Firm age manifests a firm’s externallegitimacy of existence in its relationships with other firms, itsstaying power, and the pervasiveness of internal routines thatinfluence its overall performance [19].

6. Analysis and results

Means, standard deviations, and correlations of variablesincluded in this study are presented in Table 7. A review of thecorrelations shows that KMP is significantly and positively relatedto external codification (r = 0.481, p < 0.01), internal codification(r = 0.561, p < 0.01), external personalization (r = 0.270, p < 0.01),and internal personalization (r = 0.382, p < 0.01) strategies. Weconducted two sets of multiple regression analyses. One set testedthe effects of KM strategies on KMP without considering thecontextual effects of both environmental knowledge intensity andorganizational IS maturity. Another set tested the effects of KMstrategies on KMP contingent on the four contexts.

6.1. Hypothesis test

The first set of regression analyses revealed that all four KMstrategies explained the additional 44.1% of the total variance inKMP as a variance accounted for by firm size and age (DR2 = 0.441,DF (4, 134) = 27.261, p < 0.01). Specifically, except for the internalpersonalization strategy (b = 0.045, ns), the external codification(b = 0.158, p < 0.05), internal codification (b = 0.368, p < 0.01), andexternal personalization (b = 0.186, p < 0.05) strategies weresignificantly related to KMP.

However, these results for the main effects of KM strategies onKMP simply represent the weighted average effects across the four

T.H. Kim et al. / Information & Management 51 (2014) 398–416408

contingency cells. For example, the significant positive relation-ship between external codification and KMP in Cell 1 (high in bothenvironmental knowledge intensity and organizational IS maturi-ty) was most likely strong enough to compensate for theinsignificant effect of external codification on KMP in other cells,resulting in an overall significant positive main effect. Thus,whether a particular KM strategy is effective for a contingency cell,for several cells, or for all cells remains unclear. Accordingly, weconducted additional multiple regressions in which KMP wasregressed simultaneously on all four KM strategies in each cell.These regression analyses allowed us to test whether a focal KMstrategy was significantly related to KMP in its hypothesizedcontext in the presence of other KM strategies.

To test the hypothesized contingency effects, we used amedian-split method to divide sample firms into high and loworganizational IS maturity groups (median = 4.63, SD = 0.81,min = 1.60, max = 6.58). This procedure produced four distinctivecontingency cells: high environmental knowledge intensity andhigh organizational IS maturity (Cell 1: N = 34), low knowledgeintensity and high IS maturity (Cell 2: N = 36), high knowledgeintensity and low IS maturity (Cell 3: N = 28), and low knowledgeintensity and low IS maturity (Cell 4: N = 43). As shown in Table 8,in each of the four cells (except for Cell 4), the hypothesized KMstrategies had significant effects on KMP in the correspondingcontext, whereas other KM strategies had no significant effect.Specifically, the external codification strategy was significantlyrelated to KMP in Cell 1 (H1: b = 0.329, p < 0.05). However, otherKM strategies were not associated with KMP in this context.Likewise, the internal codification strategy was significantlyassociated with KMP in Cell 2 (H2: b = 0.621, p < 0.01). Theexternal personalization strategy had a significant relationshipwith KMP in Cell 3 (H3: b = 0.546, p < 0.01). However, the internalpersonalization strategy was not related to KMP in Cell 4 (H4:b = 0.040, ns). In summary, except for the internal personalizationstrategy in Cell 4, in which both environmental knowledgeintensity and organizational IS maturity were low, all of the otherKM strategies were significantly associated with KMP in theirhypothesized contexts, thus initially supporting H1, H2, and H3.

6.2. Robustness check

The aforementioned tests for contingency effects were based ona relatively small sample size that ranged from 28 to 43 firms ineach context. A recommended statistical remedy for dealing withsmall sample sizes is the bootstrapping method, which does notassume sampling distribution. Although bootstrapping is by nomeans a substitute for inference drawing based on parametric

assumptions, it is an appropriate way to control and check thestability of the results [69]. We conducted the bootstrapping ineach context with the bootstrapping subsample N = 1000 and theestimated bias-corrected and accelerated 95% lower and upperlevels of confidence intervals (CI) for regression coefficients [69].Table 8 shows the results for the 95% lower and upper bounds of CIregarding the regression coefficients in the original multipleregressions. The bootstrapping analyses revealed that the CI ofsignificant regression coefficients in the multiple regressions didnot include any zero points. Specifically, as presented in model 2 ofTable 8, the external codification strategy was significantly relatedto KMP in Cell 1 (b = 0.329, p < 0.05 in the normal theory test; 95%CI = 0.101–0.585 in the bootstrapping test). The internal codifica-tion strategy was significantly associated with KMP in Cell 2(b = 0.621, p < 0.01; 95% CI = 0.305–0.871), as shown in model 4.The external personalization strategy had a significant relationshipwith KMP in Cell 3 (b = 0.546, p < 0.01; 95% CI = 0.110–0.802), asdisplayed in model 6. Finally, the bootstrapping result for therelationship of the internal personalization strategy with KMP(95% CI = "0.334 to 0.608) confirmed the insignificant relationshipbetween the two in Cell 4 (b = 0.040, ns), as shown in model 8.

Additionally, we conducted Cohen’s power analysis [70] todetermine whether the sample size in each context was adequateto obtain a significant effect. Given that the correlations of firm sizeand age with KMP were not significant (r = 0.128 and 0.082,respectively), we looked at the necessary sample size for multipleregressions with four independent variables (i.e., four KMstrategies) aimed at a large effect size (i.e., f2{R2/(1 " R2)} = aboveabove 0.35) at power = 0.80 for alpha = .05. The necessary samplesize to meet these requirements was 38, whereas 28–43 firms ineach context were included in hypothesis tests. However, it shouldbe noted that even in Cell 3 with N = 28 firms, the necessary samplesize for f2{0.485/(1 " 0.485)} = 0.678 at power = 0.80 for al-pha = 0.05 is 23 firms, given the obtained R2 = 0.485. Takentogether, the results of our power analyses and the bootstrappingmethod further support the findings of normal theory testing usingmultiple regressions.

Next, we operationalized organizational IS maturity as adichotomous variable. A two-by-two configuration of high andlow knowledge intensity and IS maturity was created to highlightthe contingency effects concisely and to make them clearlyinterpretable. However, a median-split method used to transforma continuous variable to a dichotomous variable may causeinformation loss [71]. For example, ratings slightly higher andlower than the median value were categorized into two separategroups. Moreover, ratings that were far from and slightly higherthan the median value were classified into another group. Given

Table 7Descriptive statistics and inter-correlations.

Variables 1 2 3 4 5 6 7 8 9

1. Firm size2. Firm age 0.494**

3. External codification 0.250** 0.0904. Internal codification 0.362** 0.024 0.428**

5. External personalization 0.020 0.133 0.391** "0.0916. Internal personalization 0.035 0.006 0.079 0.479** 0.306**

7. IS maturity 0.281** 0.162 0.472** 0.502** 0.048 0.1368. Knowledge intensity "0.255** "0.233** 0.073 0.105 "0.023 0.045 0.0929. KM performance 0.128 0.082 0.481** 0.561** 0.270** 0.382** 0.578** "0.008

Mean 6.787 36.369 4.414 4.846 4.405 4.871 0.500 0.440 4.676SD 2.164 24.167 0.749 0.838 0.664 0.695 0.500 0.498 0.584

Note: IS maturity and knowledge intensity were dummy-coded: i.e., high IS maturity = 1; low IS maturity = 0; high knowledge intensity = 1; low knowledge intensity = 0.N = 141.

** p < 0.01.

T.H. Kim et al. / Information & Management 51 (2014) 398–416 409

Table 8Regression results for knowledge management strategies and their effects in each context.

Variables Cell 1 (N = 34)High in KIa/High in ISMb

Cell 2 (N = 36)Low in KI/High in ISM

Cell 3 (N = 28)High in KI/Low in ISM

Cell 4 (N = 43)Low in KI/Low in ISM

Model 1 Model 2 BCa 95% CIc

at model 2Model 3 Model 4 BCa 95% CI

at model 4Model 5 Model 6 BCa 95% CI

at model 6Model 7 Model 8 BCa 95% CI

at model 8

Lower Upper Lower Upper Lower Upper Lower Upper

Intercepts 5.614**

(0.253)2.555**

(0.874)0.389 4.199 4.811**

(0.397)0.827

(0.845)"0.580 2.592 4.203**

(0.229)2.133**

(0.703)0.520 3.364 4.351**

(0.273)2.084**

(0.667)0.180 3.902

ControlsFirm size "0.090*

(0.037)"0.106*

(0.040)"0.201 "0.019 0.027

(0.058)"0.016(0.045)

"0.107 0.094 0.019(0.043)

"0.015(0.040)

"0.120 0.060 0.021(0.045)

"0.051(0.041)

"0.130 0.022

Firm age "0.003(0.003)

0.001(0.003)

"0.008 0.010 0.002(0.004)

0.001(0.003)

"0.005 0.007 0.003(0.005)

0.002(0.004)

"0.007 0.010 "0.005(0.005)

0.001(0.004)

"0.009 0.010

Main effectsExternal codification 0.329*

(0.146)0.101 0.585 "0.031

(0.133)"0.291 0.223 "0.204

(0.132)"0.480 0.203 0.121

(0.147)"0.153 0.512

Internal codification 0.138(0.116)

"0.196 0.667 0.621**

(0.146)0.305 0.871 0.377

(0.181)"0.145 0.782 0.308

(0.167)"0.069 0.587

External personalization "0.079(0.135)

"0.475 0.425 0.288(0.163)

"0.045 0.618 0.546**

(0.135)0.110 0.802 0.120

(0.136)"0.142 0.366

Internal personalization 0.222(0.115)

"0.051 0.420 "0.012(0.162)

"0.345 0.381 "0.203(0.133)

"0.517 0.276 0.040(0.172)

"0.334 0.608

DR2 0.216 0.288 0.031 0.518 0.031 0.454 0.026 0.378DF 4.274* 3.925* 0.526 8.329** 0.398 4.629** 0.527 5.707**

df 2, 31 4, 27 2, 33 4, 29 2, 25 4, 21 2, 40 4, 36R2 0.216 0.504 0.031 0.549** 0.031 0.485* 0.026 0.404**

Adjusted R2 0.166 0.394 "0.028 0.456 "0.047 0.338 "0.023 0.304Overall F 4.274* 4.579** 0.526 5.883** 0.398 3.296* 0.527 4.063**

df 2, 31 6, 27 2, 33 6, 29 2, 25 6, 21 2, 40 6, 36

Note: Unstandardized regression coefficients are presented with standard errors in parenthesis.a KI, Environmental Knowledge Intensity.b ISM, Organizational IS Maturity.c BCa 95% CI, bias-corrected and accelerated 95% confidence interval from bootstrapping with N = 1000.* p<0.05.** p<0.01.

T.H.

Kim

et a

l. /

Info

rma

tion

& M

an

agem

ent

51

(20

14

) 3

98

–4

16

41

0

these possibilities, we conducted a series of multiple regressionanalyses after dropping firms in the ‘‘top and bottom 10%’’ and‘‘between 10% below and above the median value’’ in frequency.The results are almost identical to the results that include thosefirms, thus indicating that this median-split method did not causeserious information loss and spurious findings. In considering all ofthe results, we find that H1, H2, and H3 are fully supported. Thesefindings show that, except for the internal personalizationstrategy, the effectiveness of a firm’s KM strategies on KMP iscontingent on both environmental knowledge intensity where thefirm is embedded and its organizational IS maturity.

7. Discussion and implications

This study seeks to apply a contingency perspective to theimpacts of KM strategy on KMP. Drawing on the TOE framework,we identify two major contextual factors (i.e., organizational ISmaturity and environmental knowledge intensity) and proposemultiple contingencies with four KM strategies (i.e., externalcodification, internal codification, external personalization, andinternal personalization), based on the KBV, to hypothesize thedistinctive advantage of each strategy in a given situation. Table 9summarizes the results of our hypothesis testing.

This study aims to answer how the effect of KM strategies onKMP differ depending on the degree of environmental knowledgeintensity and the level of organizational IS maturity. On the onehand, knowledge-intensive firms are much more susceptible to thelogic of knowledge economy than are traditional firms. In thisstudy, the KBV is supplemented by external characteristics. On theother hand, organizational IS maturity has been commonlyrecognized as a key enabler. This is because organizationalcapabilities and KM-related infrastructures and processes aredependent on it. Thus, this study theoretically extends the existingKBV by including multiple contextual factors, such as environ-mental knowledge intensity and organizational IS maturity, basedon the TOE framework.

We interpret the results from the contingency perspective. First,as hypothesized, the external codification strategy leads to the bestKMP for firms with high knowledge intensity and IS maturity,whereas the other KM strategies do not have any effect on KMP inthe same context. Our results indicate that the expected benefit ofthe external codification strategy can be realized more effectivelywhen organizations have both high IS maturity and highknowledge intensity. Firms in this situation are able to identifyand absorb external codified knowledge, combine it with internalknowledge, transform tacit knowledge into explicit knowledge,and then transfer the accumulated knowledge to organizationmembers through system-oriented channels [48,50]. Second, ourtest results offer clear evidence regarding internal codification asthe most appropriate strategy for organizations with high ISmaturity but low knowledge intensity. As noted earlier, firms inthis category generally perform routine activities and preferincremental innovations to radical ones [29,53]. Thus, they tend to

develop and accumulate their internal knowledge using highorganizational IS capabilities. Third, the results support ourexpectation that firms with low IS maturity and high knowledgeintensity are more successful when using the external personali-zation strategy. Firms with lower IS capabilities are likely to usehuman-based networks as their primary knowledge sources [24].Moreover, as these firms are working in knowledge-intensiveenvironments, they attempt to import knowledge from outsidesources through face-to-face interactions and transfer thisknowledge across organizations [25]. Finally, the result for theinternal personalization strategy does not support the lasthypothesis from the contingency perspective, as shown inTable 8. One possible explanation for the insignificant effect isthat its role is an underlying strategy in realizing the effects of theother three strategies, regardless of environmental knowledgeintensity and organizational IS maturity [7]. In other words, theinternal personalization strategy is suggested as the first initiativeto effectively implement other KM strategies, rather than as anindependent strategy. This is because emerging firms must focuson the internal personalization strategy as they initiate andfacilitate the creation and sharing of their KM statuses, whichfacilitates the development of internally codified knowledge basesvia technology or the establishment of reliable interfirm alliancesfor external knowledge sources [7]. Another possibility for theunexpected finding lies in understanding the organizationalclimate in Korea, which may be explained by the concept of‘‘stickiness.’’ This concept refers to the difficulty involved intransferring internal knowledge within a firm [72]. The cost ofstickiness is relatively low when organizational members sharecommon values and engage in similar practices, as in the case ofKorean firms that have a high collectivism orientation [73]. In fact,employees in Korean organizations are believed to be even stickierthan those in Japanese firms [74]. Korean firms are skilled atconducting person-based internal sourcing of knowledge withoutspending a significant amount of resources. Thus, the internalpersonalization strategy in the collectivistic context may not beconsidered an independent KM strategy. Instead, this internalpersonalization strategy may be a natural process that provides thefoundation for implementing KM strategies [23].

7.1. Implications for research

Considering the inconsistent findings of prior studies regardingthe effect of KM strategies on KMP, this study emphasizes theimportance of multiple contingencies in developing or selecting aneffective KM strategy. The support that we find for the contingencyperspective is entirely consistent with theories about strategy,human resource management, and interorganizational relation-ships. Given the interdependence between dimensions of knowl-edge type and origin in constituting a KM strategy, the contingencyperspective is more meaningful than the universalistic perspective,which has been adopted by prior studies on KM strategy [3,4]. Ourfindings indicated that the contingency perspective clearly

Table 9Summary of testing results.

No Hypotheses Results

H1 When a firm’s organizational IS maturity and environmental knowledge intensity are both high, the external

codification strategy is the most effective way to improve that firm’s KM performance.

Supported

H2 When a firm’s organizational IS maturity is high and its environmental knowledge intensity is low, the internalcodification strategy is the most effective way to improve that firm’s KM performance.

Supported

H3 When a firm’s organizational IS maturity is low and its environmental knowledge intensity is high, the external

personalization strategy is the most effective way to improve that firm’s KM performance.

Supported

H4 When a firm’s organizational IS maturity and environmental knowledge intensity are both low, the internalpersonalization strategy is the most effective way to improve that firm’s KM performance.

Not supported

T.H. Kim et al. / Information & Management 51 (2014) 398–416 411

explains the KM phenomenon with regard to KM strategies andtheir effect on KMP. Therefore, a contingency explanation isrequired when the effects of strategy dimensions on outcomes aredependent on environmental situations.

Strategic alignment between business environments and IS-oriented capabilities has been considered a critical issue in theliterature on organizational strategy, particularly with respect tostrategic fit with environmental factors in the organizationalresearch [75,76] and strategic fit with IS resources-orientedcapabilities in the IS research [77,78]. In this study, we theoreti-cally reply using the TOE framework to explain the alignment ofKM strategy from the contingency perspective, using key externaland internal contextual factors that bridge the missing link instrategic KM research [79].

We further elaborate the TOE framework through theoreticalsynthesis. This means that the theoretical incorporation of KBVinto the TOE framework in the current study helps to specify theTOE contexts, thus restraining the freedom to differentiate amongorganizational context factors. This integration generates atheoretical advancement that evolves and appropriates the TOEframework in the strategic KM research, while maintaining thesimplicity of the framework’s classification. In addition, this studyfurther extends the TOE framework by including a performance-based outcome from an organizational innovation, that is, KMP. Inpredicting the improvement of KMP, the organizational valuecreated by successfully adopting KM helps to extend theapplication of the TOE framework, instead of merely explainingorganizational KM adoption. We believe that this theoreticalattempt might advance our understanding of a contingencyperspective toward strategic KM alignment. Such an understand-ing can be achieved by predicting KMP that involves somethingmore complex than isolating specific KM strategy factors, such as amore holistic configuration view, to fill the gap caused by themissing link in the strategic KM research [79].

Previous research on organizational effectiveness has focusedon the effects of different organizational contexts, such as firm size,industry type, firm choices, and performance [80]. The legacy ofthis early approach is still apparent in recent strategy research, inwhich such variables are treated as ‘‘controls.’’ Two distinctperspectives later emerged in the intellectual tradition of strategyresearch, namely, the universalistic and contingency perspectives.The former suggests that for all firms under all circumstances,there exists a single best approach to managing the knowledgeadopted by those firms. Accordingly, researchers have attemptedto identify ‘‘best practices’’ or those processes that positively affectfirm performance [81]. The latter suggests that neither structuralfeatures nor firm choices directly affect performance; instead,contextual features moderate the effectiveness of choices or workpractices [82]. Therefore, we account for multiple contingencies byintegrating environmental knowledge intensity and organizationalIS maturity, grounded on the TOE paradigm, to explain thecontingent effects of KM strategies on KMP according to differenthypothesized contexts. Therefore, this study contributes to thecontingency framework in KM literature by considering keyexternal and internal contexts that a firm should consider indeveloping or selecting the best KM strategy.

7.2. Implications for practice

KM strategies are linked to practical implications for managerswho make decisions related to the selection of those strategies.Their KM strategy choices represent alternate methods by whichfirms can increase the value of KM in meeting corporate objectives.This study offers contingent guidance to managers on how to bestuse different KM strategies in different organizational contexts.Our findings indicate that managers must understand the effect of

each KM strategy on KMP from the contingency perspective andcautions firms against focusing on all four KM strategies, whichmight be an exhausting approach to KM. Firms should be moreefficient and realistic in implementing KM and in selecting anappropriate KM strategy, which can be achieved by consideringenvironmental knowledge intensity and organizational IS maturi-ty. In other words, managers must simultaneously considerknowledge type and origin with external and internal contextualsituations so as to reflect the knowledge economy in their businessenvironments under limited resource conditions. The findingsencourage managers to focus on a KM strategy not as a decisionalisland but as a critical link to their organizations’ businessenvironments and resource barriers.

Another implication of this study is that the contingent KMstrategies identified herein provide organizations with a bench-mark against which they can compare their own KM strategies.Unlike prior KM strategy studies focusing on knowledge-intensiveindustries/organizations and excluding traditional industries/organizations, this study relies on an alternative, sector-specificcategorization of firms’ environmental knowledge intensity [28].This approach provides more legitimate contingent guidelines toboth knowledge-intensive and traditional firms for their bestchoice of KM strategy. In investigating the effects of KM strategies,we also recognize organizational IS maturity, not as an explanatoryfactor but as an internal contextual factor related to the contra-dictions of IS effectiveness in managing organizational knowledge.These contradictions could explain why many IS-focused KMefforts face difficulties in building effective KM environments.Therefore, this study offers firms a meaningful lesson, which statesthat the effectiveness of KM strategy in managing organizationalknowledge depends on both environmental knowledge intensityand organizational IS maturity for the external and internal fits[27] of KM success.

The final implication for practitioners is that the internalpersonalization strategy does not have any significant impact onKMP. This finding suggests that the internal personalizationstrategy may not be directly effective as an independent strategy,unlike the other three strategies, and that the former should beconsidered as an underlying requirement to realize the expectedbenefits of other KM strategies [7] across the different contextshypothesized in this study. For example, a firm can generatevaluable knowledge, experience, and networks of internal expertsthrough a KM practice of internal personalization strategy, that is,COPs in which reciprocal members share similar interests relatedto their organizational tasks, mutually interact by meeting face toface, and directly negotiate, communicate, and coordinate withone another [55]. This situation can enable the practical applica-tion of innovative ideas to overcome time–space limitationsthrough the COP’s constant learning activities. Firm members canconveniently discover appropriate knowledge that is customizedfor each worker and task. Eventually, a ‘‘COP culture’’ can become avehicle for organizational innovations. As a result, createdknowledge and workers’ experiences are efficiently codified andstandardized to improve organizations’ KM processes. In this way,the internal personalization strategy might facilitate knowledgeexchange among firms’ internal experts specializing in differentfields, thus serving as a cornerstone for the effectiveness of otherstrategies. In addition, this unexpected finding might compromisethe conflicting findings of prior studies. Some studies havesuggested that firms accumulating and leveraging internal tacitknowledge tend to outperform competitors that pursue other KMstrategies [19,20], whereas other studies argue that externalknowledge sourcing [18] or explicit knowledge creation [4,8] arethe key enablers of a competitive advantage. In this sense, ourfindings suggest that managers who are eager to achieve higherKMP by implementing effective KM strategies must shape their

T.H. Kim et al. / Information & Management 51 (2014) 398–416412

knowledge type and origin to external and internal contextualsituations. Although pursuing only the internal personalizationstrategy could not lead to the expected KMP, a firm could notrealize higher performance through other KM strategies withouteffectively implementing the internal personalization strategy inadvance.

7.3. Limitations and future research directions

We now turn to the limitations of this study, some of whichoffer opportunities for future research. First, the metrics of KMstrategy and KMP clearly require further development. Strongermetrics may evince clearer distinctions across strategies becausethe instruments suggested by earlier studies provide a startingpoint for assessing strategy dimensions and the performance of KMstrategies. For example, expectations regarding the internalpersonalization strategy are not supported by data, thus suggest-ing the need for stronger metrics to assess KMP, an alternatetheory, or the study of our proposed model within an alternateinstitutional environment. As KM becomes more complex,researchers need to develop more sophisticated metrics to assessits success. Nonetheless, the nature of the metrics alone may notcompletely account for the perceived relationships among thedimensions of KM strategy and those between KM strategies andtheir contextual situations. Different KM strategies can eventuallyenhance firm performance in different ways, although appropriatecontextual factors are considered. Longitudinal research on KMstrategies and their effects on KMP can tease apart such cumulativeeffects.

Second, it is possible that the strategy dimensions in this studydrawn from the KBV are inadequate in completely specifying a KMstrategy. Furthermore, in addition to the two contextual factorsidentified in this study, other conditions may have contingenteffects on the relationship between KM strategies and theirperformance. Further theoretical development and organization-al practice can thus stimulate the exploration of such newcontingent variables. A deeper revelation of the interactionbetween KM strategies and their situational factors would also bevaluable in enhancing the understandability of KM practice inorganizations.

Finally, the findings of this study are limited by its cross-sectional design, in which the ratings of both independent anddependent variables were collected from a single source. However,the likelihood of a common source bias might be reduced in thisstudy because testing the hypothesized relationships at theorganizational level by aggregating individual ratings would causemost individual-level random errors and sources of bias to canceleach other out [83]. Nonetheless, replications and extensions ofour findings using longitudinal designs with different ratingsources are needed. These longitudinal studies could also benefitfrom including financial measures of firm performance to examinethe causal relationships among KM strategies, KMP, and firmperformance. This study’s results also contain regional biasesbecause the data were collected only in Korea. Thus, the resultsrequire careful interpretation and replication in other industriesand countries to improve the generalizability of our findings.

8. Conclusions

With increasing attention focused on KM strategy, organiza-tions must effectively recognize the importance of the contingencyperspective to their actions. This study expands KM strategyresearch by theoretically developing an advanced multiplecontingency model that is aligned with both external and internalcontexts. At the same time, our study contributes to this researcharea by providing valuable practical suggestions to managers in

selecting KM strategies that can be successfully applied in differentexternal and internal contexts. The results of this study highlightthe importance of the fact that organizations must considerexternal and internal contextual factors in developing their KMstrategies so as to fully realize the benefits of KM strategies. Theresults highlight distinctive KM strategies that accrue from two KMdimensions (i.e., knowledge type and origin) and comprise twocontingent contextual situations (i.e., environmental knowledgeintensity and organizational IS maturity). KM strategy choicesrepresent alternate ways through which organizations increase thevalue of KM in meeting corporate objectives.

Appendix. Survey questionnaire: procedures, instruments, andstructure

Survey procedures

Before answering the questionnaires, the survey participants ineach firm were required to attend an explanation and Q&A session.The purpose was to clearly convey the meanings of the primaryterminologies used in the survey questionnaires. In addition, asshown below, the front portion of the survey questionnairesincluded detailed instructions and explanations with specificexamples, thus encouraging respondents to pay careful attentionto the questionnaire’s respective requirements. By doing so, wewere able to ensure that the participants were able to sufficientlyunderstand key points when they answered the questionnaires.

Instructions for the respondents

All of this survey’s questions are oriented to company-wideobservations. Thus, in rating each question, you (respondents)should ensure that your answers reflect your company’s overallpractices and specific features. In addition, please answer thequestions only after carefully reading the relevant explanationsand simultaneously considering concrete examples that areprovided in the front portion of the relevant questionnaire.

Explanation of the concept of knowledge management strategies

The purpose of Knowledge Management (KM) strategies is toactivate (1) knowledge accumulation (codified type versus person-alized type) and (2) knowledge regulation (external origin versusinternal origin) in an organization so as to acquire and maintaincompetitive advantages for sustainable organizational growth.

There are four types of KM strategy based on knowledgeaccumulation and regulation. They are separately but simulta-neously implemented to varying extents in your company. Asshown above, an individual KM strategy consists of twocomponents, each based on two different dimensions (Dimension1: knowledge type; Dimension 2: knowledge origin). Each KMstrategy type is determined by the composite components of twodimensions: based on Dimension 1, a component of KM strategycan be a system-oriented (Component 1-1) and person-oriented(Component 1-2) method of accumulating organizational knowl-edge; based on Dimension 2, another component of KM strategycan be one of external sourcing (Component 2-1) and internalsourcing (Component 2-2) to regulate knowledge in organizations.As explained above, the four paired groups of specific examplesinclude the two strategic components of KM strategies. On onehand, the first two paired groups of specific examples explain howa firm accumulates knowledge in a system-oriented manner(Example A, Example B, and Example C) or a person-orientedmanner (Example D, Example E, and Example F). The two ways of

T.H. Kim et al. / Information & Management 51 (2014) 398–416 413

accumulating knowledge in organizations (system-oriented versusperson-oriented) refer to one strategic component of an individualKM strategy that is dominant in firms. On the other hand, the finaltwo paired groups of specific examples explain how a firmregulates knowledge origin through external sourcing (Example G,Example H, and Example I) or internal sourcing (Example J,Example K, and Example L). These two methods of regulatingknowledge origin in organizations (external sourcing versusinternal sourcing) comprise another strategic component of anindividual KM strategy that is dominant in firms. Thus, you shouldbe reminded about the two dimensions of KM strategy by thespecific examples of components when answering the followingquestions about the four different types of KM strategy.

The structure of the survey questionnaire

! Knowledge management (KM) strategy1. My company relies on KM strategy that aims at both codifying

organizational knowledge through systems (Component 1-1)and accessing external sources for knowledge (Component 2-1).

2. My company relies on KM strategy that aims at both codifyingorganizational knowledge through systems (Component 1-1)and developing knowledge from internal sources (Component2-2).

3. My company relies on KM strategy that aims at bothpersonalizing organizational knowledge through human net-works (Component 1-2) and accessing external sources forknowledge (Component 2-1).

4. My company relies on KM strategy that aims at bothpersonalizing organizational knowledge through human net-works (Component 1-2) and developing knowledge frominternal sources (Component 2-2).

! Organizational IS maturity1. The IS in my company improves the timeliness of informa-

tion.2. The IS in my company is considered satisfactory by users.3. The IS in my company is helpful in ensuring project schedule

compliance.4. The information provided by the IS in my company is

adequate.5. The information provided by the IS in my company can be

reused.6. The IS in my company uses high-quality hardware.

7. The IS in my company is fully utilized.8. The IS in my company provides a user-friendly interface.9. The IS in my company uses high-quality programs.

10. The IS in my company fits the organization well.! Knowledge management (KM) performance

1. KM in my company is helpful in solving organizationalproblems.

2. KM in my company is useful in improving communication.3. KM in my company contributes to the development of

employee abilities.4. KM in my company is effective in enhancing the value of

products and services.5. KM in my company is efficient in satisfying customers.

References

[1] P.H. Gray, D.B. Meister, Knowledge sourcing effectiveness, Management Science50, 2004, pp. 821–834.

[2] R.M. Grant, Toward a knowledge-based theory of the firm, Strategic ManagementJournal 17, 1996, pp. 109–122.

[3] B. Choi, H. Lee, An empirical investigation of KM styles and their effect oncorporate performance, Information and Management 40, 2003, pp. 403–417.

[4] S. Nevo, M.R. Wade, W.D. Cook, An examination of the trade-off between internaland external IT capabilities, Journal of Strategic Information Systems 16, 2007, pp.5–23.

[5] B. Cassiman, R. Veugelers, In search of complementarity in innovation strategy:internal R&D and external knowledge acquisition, Management Science 52, 2006,pp. 68–82.

[6] E. Revilla, I.M. Prieto, B.R. Prado, Knowledge strategy: its relationship to environ-mental dynamism and complexity in product development, Knowledge andProcess Management 17, 2010, pp. 36–47.

[7] M.T. Hansen, N. Nohria, T. Tierney, What’s your strategy for knowledge manage-ment? Harvard Business Review 77, 1999, pp. 106–116.

[8] A. Leiponen, Managing knowledge for innovation: the case of business-to-busi-ness services, Journal of Product Innovation Management 23, 2006, pp. 238–258.

[9] T. Menon, J. Pfeffer, Valuing internal vs. external knowledge: explaining thepreference for outsiders, Management Science 49, 2003, pp. 497–513.

[10] V.L. Mitchell, Knowledge integration and information technology project perfor-mance, MIS Quarterly 30, 2006, pp. 919–939.

[11] B. Choi, J.-N. Lee, Complementarities and substitutabilities among knowledgesourcing strategies and their impact on firm performance, Journal of the Associ-ation for Information Systems 13, 2012, pp. 498–545.

[12] L.G. Tornatzky, M. Fleischer, The Processes of Technological Innovation, D.C.Heath & Company, Lexington, MA, 1990.

[13] V. Sambamurthy, R.W. Zmud, Arrangements for information technology gover-nance: a theory of multiple contingencies, MIS Quarterly 23, 1999, pp. 261–290.

[14] O.A. El Sawy, A. Malhotra, Y. Park, P.A. Pavlou, Research commentary: seeking theconfigurations of digital ecodynamics: it takes three to tango, InformationSystems Research 21, 2010, pp. 835–848.

[15] H.P. Andres, R.W. Zmud, A contingency approach to software project coordina-tion, Journal of Management Information Systems 18, 2002, pp. 41–70.

Dimension Component Specific examples

Dimension 1: Knowledge typein organizations

Component 1-1: Codifying organizationalknowledge through systems (system-oriented)

Example A: Accumulating knowledge in systematized materials and documented formsExample B: Accumulating knowledge using organizational computer-aided systemsExample C: Accumulating knowledge through an internal portal and official intranet

Component 1-2: Personalizingorganizational knowledge throughpersonal networks (person-oriented)

Example D: Accumulating knowledge from dialogic communications of interpersonalrelationshipsExample E: Accumulating knowledge through mentoring supervisors who areexperienced in related fieldsExample F: Accumulating knowledge through collective participation in a community ofpractice

Dimension 2: Knowledge originin organizations

Component 2-1: Accessing external sources(inter-firm) for organizational knowledge

Example G: Sourcing knowledge through strategic alliances, benchmarking, outsourcing,consulting services, and external professional conferencesExample H: Sourcing knowledge from esteemed journals and newspapers, externaltechnical reports, and publications of external consultative bodiesExample I: Sourcing knowledge using feedback from suppliers, customers, and competingand cooperating firms

Component 2-2: Developing internalsources (intra-firm) for organizationalknowledge

Example J: Sourcing knowledge from internal experts, predecessors, and pedestal workersExample K: Sourcing knowledge through intra-firm training and development programs,along with internal R&D projectsExample L: Sourcing knowledge from company bulletins, internal conferences andseminars, and internal technical reports

T.H. Kim et al. / Information & Management 51 (2014) 398–416414