Академический Документы
Профессиональный Документы
Культура Документы
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
M.E. Sharpe, Inc. is collaborating with JSTOR to digitize, preserve and extend access to Journal of
Management Information Systems.
http://www.jstor.org
A KnowledgeManagementSuccess
Model: TheoreticalDevelopmentand
EmpiricalValidation
UDAY R. KULKARNI,SURY RAVINDRAN,AND RONALD FREEZE
KM SuccessModel
We start witha discussion of how priorresearch intheareaofIS canbe built
upontofitthecontextofKM. OurKM successmodelusesideasandconstructs from
fromIS successtoKM success
theIS successmodelsdescribedbelow.Transitioning
requiresthe and
consideration inclusionof organizational
appropriate drawn
factors
fromorganizational
behavior,economics,and otherareasofresearch.
IS SuccessModel
DeLone andMcLean [23] compileda taxonomy of six IS successcategories (Infor-
mationQuality,SystemQuality,IS Use, User Satisfaction, IndividualImpact,and
Organizational Impact)froma comprehensive reviewof different IS successmea-
suresandproposeda modelincluding"temporal and causal" interdependenciesbe-
tweenthesecategories.IS researchers havevalidatedthemeasuresandempirically
testedtheassociations among them [45, 71, 77,79, 80]. D&M [24] maderefinements
totheiroriginalmodelbasedan evaluation oftherichresearchstreamthatemanated
fromtheirinitialmodel.
One significantrefinement ofD&M's modelis presented in Seddon[76]. Seddon
argued that D&M combined process and variance models of IS successdepending
onthreedistinct meanings thatcan be attributedto theIS Use measure. Onemeaning
of IS Use, a variablethatproxiesforthebenefits fromuseygivesriseto a variance
modelthatlinksinformation qualityand systemqualityto IS successdefinedin
termsofbenefits fromIS Use. It is thismeaningthatis mostrelevant toourcontext,
as we elaboratelater.Based on thismeaning, Seddonrespecified D&M's modelus-
ingperceptual measuresof netbenefits of an IS, namely, PerceivedUsefulnessand
UserSatisfaction, as surrogatesforIS success(fora diagrammatic representationof
themodel,see [76]). Rai et al. [71] testedmajoraspectsof Seddon'smodelin the
contextofa semi-volitionaluniversity student IS andfoundevidenceforsupporting
therelationships.
A KNOWLEDGE MANAGEMENT SUCCESS MODEL 3 13
MeasuresofOrganizational
Support
H11 i '
X. / H12 >Th8 /h9
H13 'H10 / Ny /
Knowledge H1 Perceived
ContentQuality ^ 0 p Usefulnessof S
"*
S. KnowledgeSharing 'u6
. .X
My/
KMSystem
Qualjtv
/ H^ User " H7
-
N^
Knowledge
Use
1
Satisfaction I
I I H5 I I I- *|
Measures of General Perceptual
Knowledge Content Measures of Net
and KMS Quality Benefitsof IS Use
1. KM SuccessModel
Figure
Definitions
Table 1. Construct
Construct Definition
Hypothesis ofKnowledgeSharingleads
3: HigherlevelofPerceivedUsefulness
tohigherlevelofUserSatisfaction.
4: HigherlevelofKnowledgeContentQualityleads tohigherlevel
Hypothesis
ofUserSatisfaction.
Hypothesis ofKnowledgeSharingleads
6: HigherlevelofPerceivedUsefulness
tohigherlevelofKnowledgeUse.
Factors
Organizational
Thereis noargument thatIT is an importantenablerofKM efforts.KM systems and
electronicnetworksallowknowledge workerstoshare,store,andretrieve
documents
andotherknowledgeobjectsthatmaybe usedin theirwork.However,KM success
requiresa completesolution;merelyproviding an IT-basedKM systemwithaccess
to knowledgerepositories does notguarantee thatknowledgeworkerswill use the
systemto retrievetheknowledgecontainedtherein or sharetheirknowledgewith
othersbymakingitavailableintherepository. Carefulattentionmustbe paidto the
knowledge attitude
sharing among coworkersandsupervisors,incentives
forcontrib-
utingand usingknowledge, as well as theneed for leadershipand
organizational
3 18 KULKARNI, RAVINDRAN, AND FREEZE
Hypothesis leadstohigherlevelofPerceivedUse-
9: HigherlevelofSupervisor
fulnessofKnowledgeSharing.
IfKM is considered tobe less aboutmanaging knowledge andmoreaboutmanag-
ingknowledge workers whose work depends on what they know andcan learnfrom
others,structuration theory[33] mayprovidean alternative approachaimedatderiv-
ingfurther insights intothisissue.In an organization withmultiple agents(knowl-
edgeworkers) havingmultiple goals(possiblydivergent objectives), is necessary
it to
shareresources(e.g., knowledge)and worktowardsome commongoals; thisre-
quiresinteraction andcoordination amongtheagents.Giddens'sthreedimensions of
structureas a basisofinteraction areSignification,Domination, andLegitimation. Of
these,Domination and Legitimation the
provide insight into how an organization's
leadershipinfluences thequalityof sharedknowledgeand itsreuse.Dominationis
therealizationof whohas theauthority. In theKM context, thepeoplein authority
can influencetheKM-relatedactions(contribution, use, and so on) of individuals
who possess therelevantsharableknowledgeand also of thosewho can possibly
benefitfromreusingavailableknowledge. Domination canmanifest itselfintheform
of strongleadership -
forKM viewingKM as havingstrategic importance, promot-
an
ing organization-wide climate of knowledgesharing, and so on. Legitimation is
knowing what is acceptableand what toexpect. In our setting,organizational leader-
shipsetsthenormsandexpectations withrespecttoknowledge exchangeandreuse.
Legitimation can occurwhenknowledgeworkers receivepositivesignalsaboutthe
and
desirability acceptability ofKM practices and itsbenefits. BecauseKM is a com-
plex issue, it follows thatthe more commitment the senior management showsto
A KNOWLEDGE MANAGEMENT SUCCESS MODEL 32 1
Hypothesis leadstohigherlevelofKnowledgeCon-
12: HigherlevelofIncentive
tentQuality.
leadstohigherlevelofKnowledgeUse.
13: HigherlevelofIncentive
Hypothesis
Finally,forcomparison withtheIS successmodel,it is conceivablethattheKM
successmodelcouldincludelinkageswitha boxlabeled"othermeasuresofnetben-
similarto Seddon[76] and updatedD&M [24] IS success
efitsof KM initiatives,"
models,whichuse an analogoussetofconstructs called"othermeasuresofnetben-
efitsofIS Use."In thelongrun,successful KM initiativeswillresultinbetter
knowl-
edge, KM systems,and of
internalization good knowledgesharing and reusework
practices.Thismaylead to netbenefitsto individualsin theformofmeasurable im-
322 KULKARNI, RAVINDRAN, AND FREEZE
provements in workefficiencies,
productivity,andon-the-job effectiveness, eventu-
allyresulting inhigherprofits.As pointedoutbyearlierresearchers (e.g.,[54]) mea-
suringthatcomponent of performance to KM is nontrivial.
attributable Thereare
somestudiesthatuseperceptual outcomemeasureslikeknowledge satisfaction (e.g.,
[9]), andothersthatuse firmperformance measuressuchas return on assets(ROA)
(e.g., [10]). These latterstudiescan be questionedbecausesometimes theydo not
extract thatpartofROA thatis duetoreasonsotherthanKM.
To theorganization, KM mayresultinnetbenefits thataccrueintheformofintan-
gibleknowledgeassetsthatenhancetheorganization's sustainablecompetitive ad-
vantageand itsvalue.As pointedoutabove,suchintangible andlong-term benefits
cannotbe directly attributedto KM initiativesalone. Controllingforall theother
influences on suchlong-term is a complextaskand is deemedoutsidethe
benefits
scope of thepresentstudy.In fact,theeffectof KM perceptual outcomemeasures
(e.g.,knowledgesharing, knowledgeuse) on firmperformance (e.g.,ROA) has not
beenwell studied[18, 19],in partbecauseitis difficult to empirically establishthe
link; there is an implicitassumption that KM
desirable outcomes lead to desirable
firmperformance outcomes.
ofMeasures
Operationalization
Knowledge is a very broad concept ranging fromtacitto explicit.We believe
thatourmodelis applicableto knowledgeofbothtypes.In thefirstpartofthissec-
typesofknowledge
describethedifferent
tion,we briefly and scopeofourstudy.
the
ourmeasuresanddesigning
We thendescribethebasisforoperationalizing a survey
our
comprising knowledge-related independent -
constructs qualityof knowledge
contentcapturedand retainedwithina firmand qualityof theKM systemsin the
of organizational
We thendescribetheoperationalization
organization. factorsand
thedependent measures.Finally,we reviewthesourcesreferenced forandthepro-
cess ofinstrument
development, andtheexploratoryfactoranalysisresults.
KnowledgeTypes
Therichnessandmultidimensionality ofknowledge has ledresearchers torecognize
thatknowledgeis composedofat leasttwodistinct types, tacitand explicit,andob-
servethateveryorganization maypossess varying levels of capabilityin different
areas.Thedifferenceinemphasisondifferent typesofknowledge couldbe duetothe
industry,thetypeof business(manufacturing, service,and so on), or thebusiness
strategyoftheorganization.
Bothtypesofknowledge cannotbe managedinthesame
manner. Thepersonalizationstrategy[41] reliesextensively on theidentification
of
expertsand theareas of theirexpertise.
This strategy viewsknowledgetransfer as
occurring throughdirectcontact,suchas apprenticeship and mentoring. The most
important are expertknowledge,and theabilityof an organization
ingredients to
facilitate
thecontactand collaborationbetweenthementor and thetrainee.On the
A KNOWLEDGE MANAGEMENT SUCCESS MODEL 323
QualityandKM SystemQuality
KnowledgeContent
The KnowledgeContentQualityconstruct in ourmodelrequiredtherecognition of
thetypeandqualityofknowledge available.The Information Qualitymeasureofthe
IS successmodelfocuseson precisionandrelevanceof information [71]. Whilein-
formation qualityis a multi-attribute
construct, and an important ofresearchin
area
itself(e.g.,[7, 36, 84]),ourfocusis on a morecomprehensive measureofknowledge
quality. Therefore, the of
quality knowledge contentis determined by theabilityto
the via
present knowledge appropriate presentation formats(e.g.,text,graphics,
video),
as wellas theusefulness ofthecontent to theuser.
324 KULKARNI, RAVINDRAN, AND FREEZE
a KM system
In addition, willhavetoindextherepository contentsusinganappro-
schemethatis consistent
priateclassification This is an at-
acrosstheorganization.
tributeof systemqualityand anotherfactorin KM success- initialscale itemsfor
KM SystemQualitywereadaptedfromtheEase of Use construct in priorsurveys
[71] becauseitservedas a surrogateforSystemQuality previousresearchon IS
in
success.Because of ourfocuson explicitknowledge, we use a broadersetof scale
items,includingwhether or nottools and systemswerein place to meetvarying
needs.Oursystemqualityscale itemsaddressedeachofthefollowing: utilizationof
multiple searchcriteria, frommultiplelocations,ability
accessibility to add relevant
documents, andadequacyofdocumentation.
Factors
Organizational
Ourinitialsetof 15 surveyquestionswas designedtomeasuretheunderlying organi-
zationalsupportforKM efforts in termsof thefourcategoriesmentioned earlier:
Supervisor, Coworker,Leadership, and Incentive.The on
questions Supervisor and
Coworker dealtwith( 1) holding
support regularmeetings tosharework-related knowl-
edge, (2) encouragementtoshare effective to
solutions work-relatedissues,and(3) sup-
for
port open communication. The scaleitems onLeadership covered ( 1) understanding
of
aboutKM attoplevels management, (2) seniorlevelparticipation direc-
in setting
tionforKM, (3) seniormanagement's demonstration of commitment to KM, and
(4) periodicreviewofeffectiveness ofKM practices. thequestionson Incen-
Finally,
tiveincludedpromoting knowledge-sharing behaviorby(1) buildingitintoappraisal
systemsand(2) rewarding teamwork.
KM SuccessMeasures
As in previousresearchstudies,PerceivedUsefulnessof KnowledgeSharingcap-
turedtheuser'sperception of theeffectof knowledgesharingon job performance,
and
productivity, job effectiveness,andaskedifknowledge sharing madeiteasierfor
theknowledge worker toaccomplishhisorherjob. We usedfivescale itemstomea-
surePerceivedUsefulness ofKnowledgeSharing.
We usedthreesurvey questionstomeasureUserSatisfaction. Ourscaleitemswere
designed to measure the knowledge workers'beliefson whether (1) theknowledge-
sharingcapabilitieswithin theirbusinessunitmade it easierfor themto obtainthe
neededknowledge, (2) theyweresatisfied withtheknowledge obtained, and(3) the
availableknowledgewas adequatein meeting theirneeds.
To measureKnowledgeUse, we developeda setof questionsto capturerespon-
dents'perceptions of thedegreeto whichknowledge-based decisionmaking(e.g.,
incorporationofdocumented explicitknowledge)andknowledge capture werepreva-
lentintheirwork.Thesequestions alsoincludedthequalityoftheclassificationscheme
forfacilitating
theease ofuse ofknowledge. and
McKinsey Company addressedthis
issuebyencouraging a self-organizing
andevolutionary classificationprocess[57].
A KNOWLEDGE MANAGEMENT SUCCESS MODEL 325
Instrument
Development
Initialscaleitemsforseveralofourconstructs weretakenfrommultiple sources.KM
SystemQualityandKnowledgeContentQualitywereadaptedfroman Ease ofUse
andInformation Qualityconstructs [71].MeasuresforPerceivedUsefulness andUser
Satisfaction, as measures of success, weredrawn from several studies[25, 27, 71].
KnowledgeUse was adaptedfromtheGoodhueand Thompson[35] dependence
measures.Gold et al. [34] measuredan aspectoforganizational culturein theirsur-
vey instrument from whichour structure-related
organizational questionswereadapted.
All itemswereoperationalized as a five-pointLikertscale ranging from1 "strongly
disagree"to5 "strongly agree."
All scaleitemswerediscussedwitha focusgroupofeightexecutive MBA students
withan averageofeightyearsofmiddleto uppermanagerial experience, whowere
also familiar withKM andKM systems. Basedonthefeedbackobtained, someques-
tionswererephrased andsomedropped.Twoconstructs (KnowledgeContentQual-
ityandIncentive) consistedoftwoscaleitemseach- thisis notuncommon, as there
areearlierresearch studies[12,16,37,70, 82] thathave used two scaleconstructs; for
example,extrinsic rewards[ 12] is a two-itemconstruct, as is qualityofinformation
[ 16]. A pilotstudyof65 respondents fromthesamepopulation providedfurther indi-
cationoftheappropriateness ofthequestions.Ourfinalsurveyinstrument is repro-
ducedinTables2, 3, and4, whichshowthescale itemsalongwiththeresultsofthe
exploratory factoranalysisas explainedinthenextsection.
Analysis,andResults
Data Collection,
The survey described in the previous section was administered to a groupof
150 midlevelmanagersenrolledin theexecutiveMBA and part-time professional
MBA programs atone ofthelargesturbanuniversities intheUnitedStates.The par-
ticipantshadan averageofoversixyearsofmanagerial experience distributedacross
variousfunctional areas.Jobpositionsof therespondents includedengineers (e.g.,
software systems, electrical,and project),managers(e.g., project,marketing, pro-
cess, and manufacturing), analysts(market, account,and financial),and directors
(operationsandsoftware development). Scrutiny oftheir showed
job responsibilities
thattheywouldbe routinely involvedwithknowledge work.Therewas also substan-
tial cross-industry representation by way of firms,includingCharles Schwab,
Honeywell, Intel,Motorola,PinnacleWest,andTheVanguard Group.After eliminat-
ingincomplete surveys,there were 111 usableresponses. From ourscrutiny ofthe39
unusableresponses,we notethat(1) 22 of therespondents had no KM program in
theirfunction 1
and(2) theother 7 had omitted tofill
out one or anothersection ofthe
survey - forexample,theresponseson incentives weremissingor thesectionon
leadershipwas blank. Ratherthanimputing responsestothesesurveys, we decidedto
drop themfrom the sample.
326 KULKARNI, RAVINDRAN, AND FREEZE
- FactorLoadings
Table2. ExplicitKnowledge
Content System
Surveyquestion quality quality Use
ResultsofFactorAnalysis
We conducted a preliminaryexploratoryfactoranalyseson thefirstsetof70 usable
responses to testthevalidityofthe constructs
in ourtheoreticalmodel.Thisresulted
inidentification ofseveraldistinct
factors tothedependent
relating andindependent
variableslaid outin ourmodel(see Tables2, 3, and4). In general,all itemsloaded
abovea valueof0.7 on thepredicted factorandbelowa valueof0.35 on anyother
dimension(withtheexceptionof KnowledgeContentQuality,as notedin Table2).
Also,inthecase ofKnowledgeUse,therewas oneitemthatloadedon morethanone
- thiswasdroppedfromsubsequent
factor analysis.Wenotethatoverall,Cronbach's
alpha values were exceedingthecutofflevelof 0.7 recommended
satisfactory, by
Nunnally[65],withtheexception notedinTable2. Finally,inall cases,item-to-total
correlationswereabove0.6.
The knowledge-related surveyitemsgroupedintothreedimensions representing
Knowledge Content Quality,KM SystemQuality,andKnowledgeUse as expected.
Table2 presents theresultsofthefactoranalysis.The nextpartofthesurvey, which
327
> T- Tf O
w N CVI <0
C 00 1^. 00 CD
8 cid on
.&
- LO CO CO CO ^-
2 LO 1^. ^ CO O
D CO N 00 00 O) Tf
C
^
8 CVI (D O) i-
'uJ LO "i- t- h-
O CO 00 CO CO O)
^ odo co
u
lo oo co lo
&2
; r^ N; co co -^ -d
odd 3
| dg
i J
I f bt I si i |1 I
! ! Hi H | It
i J. i
C/3
Iti lllil.lil
Hi ilililU i!l *
I
a
CL.
l lllfflt S
13
.9
a*
sllliuilfsillli !
N
'Sa
s llills,s,"lsiilis,l fi i
?
O
CO
* II lllillillllilll
X) C "LUOLU-LULULUgl- h- CO CO cD^^o"^
OOS 2 X < CL ^
328 KULKARNI, RAVINDRAN, AND FREEZE
Perceived User
Surveyquestion usefulness satisfaction
fable5. FactorAnalysis
Variance
explained Standard
Construct (percent) Mean deviation Minimum Maximum
Perceived
usefulness 82.6 3.96 0.94 1 5
User
satisfaction 75.4 3.23 0.92 1 5
Supervisor 76.3 3.51 1.02 1 5
Coworker 79.5 3.66 0.90 1 5
Leadership 77.8 2.85 1.03 1 5
Incentive 79.4 2.71 1.08 1 5
Knowledge
contentquality 70.3 3.41 0.88 1 5
Knowledge
management
systemquality 70.4 3.02 1.07 1 5
Knowledgeuse 8O5 a36 1L01 1 5
0.18*** !psl |^
, , 0.91*** ^
-
0.22*** H ps2 L ' y '
0.89*** X Perceived X
- ^'^
0.16*** !ps3 U- o.92***^^( Usefulness 1
I 1 0.86*** s' '^ ^sS '
0.26*** ]ps4Y s' '
i x^s 0.85*** 0.57***I
0.27*** [ps5p /
^*~ ^^. /
0.45*** [usi |^^o.74*** User >v /
f
, , m^^l Satisfaction V
0.25*** [us2 'k-0.87***^^ V /
1 ^0.79*** ^
I
0.38*** [us3Y^
0.33*** Icml
I L^
1 1/ ' * I
0.82***
i-cm2i
-^ / /
0.40*** - ^^ ] Z^^n/ -52***
^^7 Leadership /
, /
[ I+0.78*** ^^C jp
. o'90***^x^^ ^^**-^ - *^**^ ' //
i
0.19*** ) cm3 W^
1 ' ' ^^ '
0.60*** '
/
1/
i 1^/0.86*** 1/
0.26*** |cm4Y 'ff
-
0.50*** - Incentives j/
[di [+0.71*** --^(^
i 1 0.83***
0.31*** ci2 ^
0.65*** !
I kbl L 0.76*** I'
l^v. 0.76*** '
0.56*** !kb2L ^S. ^ ^^ / '*
I ' 0.83***'. y^ ^' /
I 1 ^"^^ / '/ 0.73***
0.63*** W kb3 M-0.78***- Z^'
KMS Quality
Jr
.
0.89***s'
0 47*** J, kb4,W S' '^_ ^^' ' /I
1 s' 0.63*** /
0.70*** d~fcb5~Va72*** W
- Know|edge Use
0.73*** [cm2 4r0.69*** ~^' J
I 1 0.83***/^ ^^_ ^^
0.56*** [cm3 Y^ y^
Table 6. ReliabilityMeasures
Constructname Reliability
ResultsofModelEstimation
Usingtheseemingly unrelated (SUR) simultaneous
regression equationestimation
procedure,we testedthevalidityoftheKM successmodel.SUR allows forthepossi-
of
bility correlatederrortermsacrossregression We
equations. also rantheestima-
tionprocedure usingLISREL 8.54 to verifytherobustnessoftheresults,whichare
discussedinthefollowing Themodelappearstobe robustbecausethese
paragraphs.3
resultsareverysimilar.
332
!s 5
o
II 25
cd
t/)
> g <O 00 r-
8i ?p
1g sN;sq q ?co
o o c>
c/3 00 CO CM G) 00 d
I^ fS 8 SI S -S
d d d d d -g
J o
O
s g m m co co in ^
S g dddddd
u
t !
in
8 s 1
|fe 88S5?8|
I Oh
3
Ncorw^qw
OOOOOOO^
t
T3
I
CO >
u i^^3 NO5t-CDCX)C'|00CO M
teed (OOr-cOCMOCVI^; e
^gn OOS
ed
>
a
0)
CO ^v
o
^
J
lOCOCNJT-COCNiqCNj^f
ddddddddd g
| 1e
1
a 8 1
t1
>, S
i= -e c -S
J ^ CO= g
S S E > g | S S S
e2
A KNOWLEDGE MANAGEMENT SUCCESS MODEL 333
Measures ofOrganizationalSupport
Measures of GeneralPerceptual
KnowledgeContent Measures ofNet
and KMS Quality Benefitsof IS Use
5. KMSuccessModelSUREstimation
Figure Results
**and***indicate
significance and1 percent
atthe5 percent levels,
respectively.
S
+
"^ 45 ai
< | (3
I * ? i
1 I s "-
liti
I ^ I +"
^ ^a
Q)
8 I
"I
S
1
I* &
"co
l I f f
I
i^
i
*
fi
i *
*S 8 I
^w g g
i2 11 3 ^
(S i i.
III
.^ p
8? 5
! -
u |
f I |
1
'S
i-
c
o
co
O"
0
cvj
c
o
co
O"
0
-g
i-
w
co
c
o
co
O"
0
^r
c
o
co
O"
0
6 o o o o
o </></)(/)(/)
I O)
0
C
O)
0
CC
O)
0
CC
O)
0
DC
335
l ss
0 "i S S q
M M CO O CM O <D
e IS
.2 S
S 0 W CO 'S
J g c m co o
8 o o 'S
I
g S
rt *c3
8
NOOOOOONCNO
g S3 8 |
"? , oo,d*-'o,c'o, 2
1
ro ?a)
c o
2 ^
* 3
2- es (MCOOtO
O^Ct lOCDOCO >
^ O O *" CO '**
'3
1
S dodo
I
S S 2
<u o
| 5B I
t> cviqcoqcop o .g
?? t CD^CO go
1
3 u * 11
s
S-"S"cd cm to in ^c
P 'J3 ^J" T- CO J3 D
O.
g
I ^*
I
^
o - ^ _ _
^ I | SJI 8 1 SSS8
-s s S
3 ? IB
I SE
i 8 8.
i2
1^ajos^^ S i coevi iwo
*g
pjccoco oco *
g " c'j p cvjcM y
<=>? do (o 2 S
Si!
Il -35 s
a o g
t S fi
I li s *
2 c >, ^ a
ri-
-o
- -."^
o 3OJ 3
.- <D -n
^ SS e
C
o 3
^ m
c to
W O" O) > t
r Supervisor j Coworker
j
'o.l7* /
V / 0.37***
^TT^^ZTN 052***^^KnowiedgT^'
(Leaoersnipi > tentent Quality^"" - " .^^Srceived^N
-^^Usefulnessv* ^
/ 0.22*V ' I Ns
/ '^ 0.52*** v
0.47***X A^ s
X 3" UsT^N
^hcentiveTi ^M Systei^y ^T^owled^N
^incentives' ^^Satisfection^y' 0 17*
^^^ Jj^^^^y
I 0.45*** ^/^ /
0.28*** /
Conclusion
We developed and tested a KM success model derivedfromtheIS success
modelofDeLone andMcLean [23,24] andSeddon[76].Ourmodelwas enriched by
researchin thearea of KM byAlavi and Leidner[3], Davenportand Prusak[19],
Davenport et al. [20], andothers.Thusfar,theemphasisin KM-related IS research
hasbeenonimproving KM applicationsandsystems andtheirimplementation across
corporateintranets - thatis,the focushas been on technology.Although ampleanec-
dotalevidenceexists, KM researchhas paid limitedattentionto creatinga formal
empiricalmodelwithorganizational factorsthatcan complement thetechnology.
One objectiveof thisstudyis to incorporate bothknowledgecontributions (in the
formofsharedknowledge and
quality) knowledge use as outcomesof KM initiatives
undertaken bya firmina modelinvolving theirantecedents- comprising bothtech-
nologyandorganizational factors.
Contributions
An important elementofourstudyis theidentification oftheorganizationaldimen-
sionandmeasuresthatenableknowledgesharingandreuse,a stepbeyondthecul-
factorsof earlierresearch(e.g., [34]). Anothercontribution
turalinfrastructure of
thisstudyis the of
integration approaches from social,organizational,andeconomic
toshowthattheyconverge
theories toprovideconsistent directionsforKM research.
The development of ourmodel'sconstructs is based on theoriesdrawnfromthese
diversedisciplines;we showhow insightsfromsocial exchangeand structuration
theoriescan be reconciledwiththosefrommicroeconomic (agency)theoryin the
contextofKM.
Mostpriorresearchin KM focuseson the"supply"sideofknowledge, whichin-
volvesresources andeffortsneededto stimulate knowledge creation
and storage;for
340 KULKARNI, RAVINDRAN, AND FREEZE
Implications
Ourfindings tofurther
contribute theunderstanding ofthewayin whichKM efforts
shouldbe implemented in organizations. The statisticalresultsconfirm anecdotal
evidencethatorganizational factorsinvolving people(namely,leadershipcommit-
mentandsupervisor andcoworker support forreinforcing KM initiatives) areas im-
portantas the that
technology supports these KM Without
initiatives. these "people"
factors,whatmayhappenis thateventhemostenthusiastic knowledge worker may
eventually dismissthepotential benefits of KM ifhe or shedoes notsee otherswith
thesamelevelofenthusiasm.
Ourresultsclearlyindicatethatthecommitment exhibited bytheseniorleadership
affectsqualityofsharedknowledge as wellas theextentofknowledge use. Someof
theKM practitioner comments we encountered were"thereasonsomegroupsinour
unitaremoresuccessful thanothersinknowledge sharingandreuseis theirfocuson
customerratherthanproduct, people viewed as assets not costs,and emphasison
opennessnotsecrecy." We interpret theseremarks tomeanthatthepracticeofdefin-
ing desirable behaviorand enticing staffinto exhibiting thatbehaviormaylead to
conformance butnotto commitment. Therefore, seniormanagement shouldtakeon
theroleofexemplarandnotthatofa merecoach.In fact,ourresultscloselymirror
theseviewsandsuggestconcrete stepsfirms cantakeinthisregard. Someofthesteps
as expressedby thesesame seniormanagerswere(1) havesenior-level KM advo-
cates;(2) KM
associate with unit,group, and individual goals and objectives; (3) cul-
tivatecommunities ofpracticeandinterest; and(4) use feedbackto improveKM.
One waybywhichtheorganizational leadership can demonstrate commitment to
KM is byhavingtopmanagement assumethevisibleroleofknowledge champions.
The knowledgechampionsshouldspearheadthetasksofcrafting a KM strategy for
thefirm, goals,andemphasizing
setting thepotential benefits of KM. Otherimpor-
tantactionsincludeinstituting policies and procedures forrewards, recognition, and
incentives, andpromoting internalization ofknowledgesharingandreusepractices.
In firmswhereKM responsibility is decentralized and distributed amongbusiness
units,thereshouldbe consistency intheactionsofmultiple champions. Thechampi-
ons mustenlistparticipation of supervisors in theinitiatives in orderto shapeem-
A KNOWLEDGE MANAGEMENT SUCCESS MODEL 34 1
Limitations
andFutureResearch
One limitation of thisstudyis thatit consideredonlyexplicitknowledge. To study
differences acrossindustries or businesstypes,it maybe necessaryto distinguish
betweenexplicitandtacitknowledge andmeasureKnowledgeContent Quality,KM
and
SystemQuality, Knowledge Use levels in thesedifferent of
types knowledge. For
example, one can argue that
success in thehigh-tech industry may be more dependent
on thequalityof its technicalexpertise(a formof tacitknowledge),whereasthe
transportation industry mayrelymoreon operational knowledgeavailablethrough
lessonslearned(a formofexplicitknowledge). Similarly, maybe observed
variations
in businessesof different -
types manufacturing versus In
service. a semiconductor
manufacturing plantthatuses highlysophisticated machinery, we maysee thatles-
sonslearnedarecapturedanddocumented bywayof "bestknownmethods," which
becomecrucialinreducing downtime inordertoattainthetarget yields.On theother
hand,service-oriented businesses,suchas resortsandcasinosthatdeal witha large
number ofcustomers, needtocreateandreusemodelsofcustomer profilestoexcelin
theirbusiness.Harrah'sCasinos,a servicebusinessin thegamingindustry, forex-
ample, builds data-mining models and tests theireffectiveness field
through experi-
ments;thebestmodelsarestoredandreused[55].Thisis anexampleoftheir emphasis
on effective use ofexplicitknowledge. Suchextensions mayallowformorespecific
recommendations.
Ourmodelstudiesknowledge sharing andusefroma knowledge worker's perspec-
tiveas an indication of successof a KM initiative. In thisview,theknowledge pro-
cesses aretreated at a highlevelofabstraction. A moredetailedapproachis totreat
knowledge processesata muchmoregranular levelas someoftheotherresearchers
(e.g.,[34]) havedone,forexample,bytreating thenatureofidentification andvetting
processes, andbyanalyzing workflowstepsthatfacilitate captureofidentifiedknowl-
as
edge separate constructs.Future research can include these variables
to understand
theantecedents ofKM success.
Thereis nodoubtthatobtaining objectivemeasuresofactualperformance improve-
mentsdirectly attributable to KM initiatives wouldhavestrengthened thestudy.In
theabsenceofsuchmeasures, itis perhapsbetter togatherusers'perceptions thatact
as proxiesforperformance. Moreover, in a cross-sectional studysuch as ours,when
one is studying thegeneralized effect ofmultipleKM initiatives acrossa number of
A KNOWLEDGE MANAGEMENT SUCCESS MODEL 343
theunitsofmeasureforbusinessbenefits
organizations, can be problematic. Differ-
entbusinesseswouldbe invariably usingtheirownmetrics forbusinessperformance
(e.g.,cycletime,backorders, bidswon,customer satisfaction).
Aggregation ofsuch
is if not
diversemeasures difficult, impossible, in a cross-sectionalstudy.Future re-
searchcan be aimedat an in-depth studyofone KM initiative in a particularsetting,
whichcouldbe further enhancedbylongitudinally measuringthe business benefits.
The natureofourempirical modelallowsmultiple avenuesforfurther researchas
outlinedhere.Another
partially issue worthy of empiricalexamination and one we
arecurrentlyexamining is thatofcomplementarities betweenKM factors andorgani-
zationalfactors.Intuitively, it is feasiblethata higherlevelof knowledgequality
combinedwitha higherlevel of KM SystemQualityor Incentivemay lead to
supermodular benefits [62]. Similarly, complementarities can also existbetweenca-
in
pabilities different knowledge areas.Verifying complementarity through super-
modularbenefits is nontrivial andrequiresrigorousstatistical methods[6]. Another
worthwhile direction is toexaminetheaforementioned beliefsaboutdifferences gov-
ernedbythecharacteristics oftheparticular industry orbusiness.A largernumber of
from
respondents multiple industrieswill be needed forsuchanalysis.Resulting find-
ingscan makemorespecificrecommendations allowingbusinessesto investmore
prudently inresourceswhileplanningKM initiatives.
Accordingto someKM practitioners, forKM to be effective, one mustnotbegin
andendwithimproving howwellworkgetsdone.It shouldalso improvewhatgets
done.Further, an organization shouldreexamine theprocessesfordiscovering and
creatingnew knowledgeas well as refining existingknowledge.In otherwords,
businessprocessesprovidethecriticalconnecting factorsthatbridgeKM andbusi-
nessresultsorperformance. Thus, KM effortsmust includeidentificationofknowl-
edge-intensive workprocessesand workflowsthatare deemedimportant forthe
typeof business,and theIT and systemssupportneededto facilitateknowledge
sharing.Such infrastructural changescan eventuallyaid in transforming knowl-
edge-intensive business processes. These qualitativeinsightsprovide rich avenues
forfuture research.
Notes
is seenbysomeas conforming
1. AnIT artifact tothe"tool"viewoftechnology described
by Kling:kkA computing resource[that]is bestconceptualized pieceofequip-
as a particular
ment, ortechnique
application whichprovidesspecifiableinformationprocessingcapabilities"
[50,p. 308].
2. Wereiteratethatthereis a processcomponent toKM, representingtheprocessesembed-
ded in an organizationforcapture,sharing, and retrieval
of knowledge, and thisis partially
captured in ourKnowledge Use construct.
344 KULKARNI, RAVINDRAN, AND FREEZE
3. The valueofcoefficients
fromtheLISREL procedure maydiffer fromtheSUR
slightly
estimationcoefficients also,thelevelsofsignificance
thereis no difference;
but,qualitatively,
arevirtuallythesame.
References
1. Ackerman, M.S.,andHalverson, C. Considering anorganization's memory. InS. Poltrock
andJ.Grudin(eds.),Proceedings ofthe1998ACM Conference on Computer Supported Coop-
erativeWork. NewYork:ACM Press,1998,pp. 39-48.
2. Adams,G.L., and Lamont,B.T. Knowledgemanagement systemsanddeveloping sus-
tainablecompetitive advantage. JournalofKnowledge Management, 7, 2 (2003), 142-154.
3. Alavi,M., and Leidner,D. Knowledgemanagement and knowledgemanagement sys-
tems:Conceptual foundations andresearch issues.MIS Quarterly,25, 1(March2001), 117-136.
4. Anderson, J.C.,andGerbing, D.W. Structural equationmodelingin practice:A review
andrecommended two-step approach.Psychology Bulletin,103,3 (1988),41 1-423.
5. Argote,L.; McEvily,B.; and Reagans,R. Managingknowledgein organizations: An
integrativeframework andreviewofemerging themes. Management Science,49,4 (April2003),
571-582.
6. Athey,S., andStern,S. Anempirical framework fortestingtheoriesaboutcomplementarity
in organizational design.Working Paper,Department of Economics,Massachusetts Institute
ofTechnology, Cambridge,1998.
7. Barua,A.; Kriebel,C.H.; andMukhopadhyay, T. MIS andinformation economics:Aug-
menting richdescriptions withanalytical rigorininformation systems design.In J.I.DeGross,
J.Henderson, andB. Konsynski (eds.),Proceedings oftheTenth International Conference on
Information Systems. NewYork:ACM Press,1989,pp. 327-339.
8. Bates,K.A.; Amondson, S.D.; Schroeder, R.G.; andMorris,W.T.The crucialinterrela-
tionship betweenmanufacturing strategy andorganizational culture.Management Science,41,
10(1995), 1565-1580.
9. Becerra-Fernandez, I., and Sabherwal,R. Organizational knowledgemanagement: A
contingency perspective.JournalofManagement Information Systems, 18, 1 (Summer2001),
23-55.
10.Bierly,P.,andChakrabarti, A. Genericknowledge management strategies inthepharma-
ceuticalindustry. Management
Strategic Journal,17, 10 (Winter1996) 123-135.
11. Blau,P.M.Exchangeand Powerin Social Life.NewYork:JohnWiley,1964.
12. Bock,G.W.; Zmud,R.W.; Kim,Y.G.; and Lee, J.N.Behavioralintention formation in
knowledgesharing:Examiningtheroleof extrinsic motivators, social-psychological forces,
andorganizational climate.MIS Quarterly, 29, 1 (March2005), 87-1 11.
13. Brajnik,G.; Mizzaro,S.; Tasso,C; andVenuti,F. Strategic helpin userinterfaces for
knowledge retrieval.JournaloftheAmericanSociety forKnowledge Scienceand Technology,
53, 5 (2002), 343-358.
14.Brown,J.S.,andDuguid,P. Organizing knowledge. California Management Review, 40,
3(1998), 90-111.
15. Carlsson,S.A.; El Sawy,O.A.; Driksson,I.; andRaven,A. Gainingcompetitive advan-
tagethrough sharedknowledge creation:Searchofa newdesigntheory forstrategic informa-
tionsystems.In J.Dias Coelho,W. Konig,H. Krcmar,R. O'Callaghan,and M. Saaksjarvi
(eds.),Proceedings oftheFourthEuropeanConference on Information Systems. Lisbon:USA
Publishing, 1996,pp. 1067-1076.
16. Chang,J.C.J.,and King,W.R. Measuringtheperformance of information systems: A
functional scorecard.Journalof ManagementInformation Systems, 22, 1 (Summer2005)
85-115.
17. Cole, R.E. Specialissueon knowledge andthefirm - Introduction. California Manage-
mentReview,40, 3 (Spring1998), 15-21.
18. Davenport, T.H.,andBeck,J.TheAttention Economy:Understanding theNewCurrency
ofBusiness.Boston:HarvardBusinessSchoolPress,2001.
19.Davenport, T.H.,andPrusak,L. Working Knowledge. Boston:HarvardBusinessSchool
Press,1998.
A KNOWLEDGE MANAGEMENT SUCCESS MODEL 345
The followingStatement of Ownership, Management, and Circulationis provided in accordance withthe requirements
contained in 39 USC 3685. Journalof Management informationSystems (Publication no.0742-1222) appears quarterly
and the annual subscription price is $869.00. It is owned, managed, and published by M.E. Sharpe, Inc., which is
located at 80 Business Park Drive,Armonk,Westchester County,NY 10504-1 71 5. The Publisher is M.E. Sharpe, at the
same address. The Editoris VladimirZwass, c/o M.E. Sharpe, Inc. 80 Business Park Drive, Armonk,NY, 10504. During
the preceding twelve months the average number of copies printedforeach issue was 1 ,125; the average paid circula-
tion (by mail subscription) was 805; the average free distribution,61 ; the average number of copies distributed,866.
Corresponding figuresforthe issue published nearest to the filingdate: total number of copies printed,1 ,100; total paid
circulation(by mail subscription), 719; totalfree distribution,60; total distribution,839. Filed by Vincent Fuentes, Senior
Vice President.