company innovation system: a conceptualization
TRANSCRIPT
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page1of19
COMPANYINNOVATIONSYSTEM:ACONCEPTUALIZATION
ERIKDENHARTIGH
ÖzyeğinUniversity,FacultyofBusiness,Istanbul,[email protected]
ABSTRACT
Weconceptualizethecompanyasaninnovationsystemthatconsistsofcomponents,relationshipsand attributes,with the purpose to produce innovation. The systems approach to innovation hasreceived limited attention at the company level.While it is widely accepted for nations, sectors,regionsandtechnologies,andwhilesomecompany-levelfoundationsandbuildingblockshavebeenproposed,thedominantapproachatcompanylevelistoregardinnovationasaprocess.
Components of a company innovation system are actors or resources. Relationships refer to theconfiguration of these components: an innovation process now becomes one of the possibleconfigurations of components in a system. Attributes of a company innovation system arecapabilitiesandothersystemproperties,suchasinnovativecultureorinfrastructure.
Weexploretheconceptofcompanyinnovationsystembyanalyzingandcomparingcaseexamplesof ABB Group, Adobe Systems, Amazon.com, eBay, Hitachi, HTC, Lockheed Martin, Philips,Qualcomm,Salesforce.comandSouthwestAirlines.
Wefindthatusingthecompanyinnovationsystemapproach,wecanmapinnovationsystemsatthecompany level. We can identify the components, such as R&D departments, labs, ventureorganizations,teams,employees,C-levelofficesandfacilitatingtools.Wecanidentifyrelationshipssuch as single or multiple configurations, simple or complex configurations, technology-driven,market-driven or interactive configurations, and open or closed configurations. We can identifyattributes such as creativity versus efficiency emphasis, systematic versus non-systematicapproaches,adaptivenessofthesystem,andlargeproject-focusversusexperimentationfocus.
The findings indicate that companiesdesign, configureandcoordinate their innovation systems indifferent ways. Our current findings are tentative and preliminary and only provide descriptiveinsights of the case examples. A well-conceptualized and validated company innovation systemapproachmaygivemanagersrelevantinsightstoaddresstheproblemsofdesigning,configuringandcoordinating their company innovation systems. Academically, the company innovation systemapproachprovidescomplementaryinsightstotheexistingcompany-levelinnovationapproaches.
Key words: Company Innovation System; Corporate Innovation System; Innovation; Systemsapproach
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page2of19
INTRODUCTION
Weconceptualizethecompanyasaninnovationsystemthatconsistsofcomponents,relationshipsand attributes, with the purpose to produce innovation. The systems approach to innovationmanagementhas received limitedattentionat the company level. Innovation systemsapproacheshave been successfully applied to countries as National Innovation Systems (Lundvall, 1992), toregions as Regional Innovation Systems (Cooke et al., 1997), to sectors as Sectoral InnovationSystems (Malerba, 2002), and to technologies as Technological Innovation Systems (Carlssson &Stankiewicz,1991).Ineachofthosefields,theinnovationsystemsapproachiswidelyaccepted.
We propose the company innovation system approach as a relevant addition to the existinginnovation system approaches and to the innovation process approach. We base the companyinnovation system approach on foundations provided by Van de Ven (1986), who addressed thequestion of part-whole relationships in innovation management, by Granstrand (2000), who firstcoinedthetermcorporateinnovationsystem,byCarlssonetal.(2002),wholaidoutamethodologytoconceptualize innovationsystems,andbyO’Connor(2008),whousedthesystemsapproachformakingpropositionsaboutmajorinnovationinfirms.Weusemanyofthebuildingblocksthathavebeen proposed in the literature (Teece, 1996; Granstrand, 1998; 2000; Sigurdson& Cheng, 2001;Coriat&Weinstein,2002;O’Connor,2008;Steiber&Alänge,2013;Taylor&Wagner,2014;Chenetal.,2015).
At the company level, thedominant textbook approach is to regard innovation as a process (see,e.g.,Trott,2011;Tidd&Bessant,2013).TheintroductionoftheStage-GatesystembyCooper(1985)greatly contributed to the acceptance of the process view. Initially, many companies regardedinnovation processes as linear, with sequential steps, but soon they recognized that using cross-functionalmechanismsandparallelandnon-linearprocessesenhancedeffectivess(see,e.g.,Cooper1990). Cooper (2008) himself debunkedmanyof themyths that the Stage-Gate systemwould berigidandsequential,whilestillrecognizingthatmanycompaniesimplementitassuch.
With the advent of open innovation, companies and researchers have increasingly adopted aninnovationnetworkapproachinadditiontotheprocessapproach(Chesbrough,2003).Researcherswhouseanetworkapproachemphasizethestructureofasystem,countingthenumbersofnodesand linksandtheconnectednessbetweenthenodes.Researcherswhodevelopedtheconceptsofbusinessecosystemsandplatforms(e.g.,Moore,1993; Iansiti&Levien,2004;Gawer&Cusumano2014)haveextendedthisnetworkapproachintoasystemsapproach,inwhichthecompanyisoneof the actors within a technological (eco)system or platform. Researchers who use a systemsapproach look beyond systems structure, emphasizing the exchange relationships between theactors and the emergent properties at the system level. A systems approach to innovationemphasizes interaction, learning, and knowledge creation (Edquist, 1997). It allows inclusion of awide array of institutional attributes that may be important in explaining innovation, such asinnovativeculture,top-downversusself-organizedcoordination,oranopenversusclosedmindset.
The scope of this paper is limited: we only look inside the company, we use strictly a systemsapproach andwe only aim to describe. First, we only look at the company level, andwe do notventure into the links that clearly exist between the company and wider innovation systems or
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page3of19
networks (cf., Sigurdson & Cheng, 2001). Second, we do not make international comparisons ofcompany innovation systems in different national innovation or policy contexts (cf. Granstrand,2000; Sigurdson&Cheng, 2001). Third,wedonot aim to explainwhy a company is successful ininnovationorwhy somecompaniesaremore successful thanothers (cf., Steiber&Alänge,2013).Webelievethatthecomponents,relationshipsandattributesofthecompanyinnovationsystemareamong the factors that influence innovation success, but any attempt to include the companyinnovationsystemconcepttoexplaininnovationsuccesswillhavetowaitforthedevelopmentofabodyorempiricalresearch.
Empirically, too, our scope is limited to descriptive case analyses and a cross-case analysis of thecompany innovation systems of ABB Group, Adobe Systems, Amazon.com, eBay, Hitachi, HTC,LockheedMartin,Philips,Qualcomm,Salesforce.comandSouthwestAirlines.Theonlyaimoftheseanalysesistoshowthattheconceptcanbeusedtodescribecompanyinnovationsystemsandthattherearedifferencesbetweencompanies.Werecognizethatfurtherconceptualwork isnecessaryto fleshout theconceptand its sub-concepts, toenable testingand falsifying theconcept,and toclarify the connections with related concepts. Further empirical work is necessary to test theusefulnessoftheapproachforanalyzinginnovationincompanies.
COMPANYINNOVATIONSYSTEM
Foundationsandmethodologicalaspects
Granstrand(2000)coinedtheconceptofcorporateinnovationsystemanddefineditas“…thesetofactors,activities,resourcesandinstitutionsandthecausalinterrelationsthatareinsomesenseimportantfortheinnovativeperformanceofacorporation.”(p.14),adefinitionthatisinlinewiththeconceptsofnational,regional,sectoralandtechnologicalinnovationsystems.Hestudiedsuchsystemsindifferentcountycontexts,onanaggregatecountrylevel,identifyinganumberofimportantcharacteristicsanddevelopments,suchasthegrowingimportanceofexternaltechnologyacquisitionandtheincreasingdiversificationofcompanies’technologybase.Healsoinvestigatedimplicationsforgrowthandperformance.Grandstrand’s(2000)isthemostcomprehensivestudyonthistopictodate,butresultsarepresentedatanaggregatelevel,andcouldbemoreinformativefordevelopingtheconceptatthecompanylevel.
VandeVen(1986)providedafoundationforcompanyinnovationsystemsinhisdiscussionon‘problemsinthemanagementofinnovation’.Oneofthemainproblemsininnovation,heargues,isthemanagementofpart-wholerelationships.Atemptingandmuch-usedapproachforachievingmaximumproductivityistosegmentinnovationintoasequenceofstagesandtodividethelaboramongspecialistdepartments,likeR&D,productionormarketing.Suchapproacheshaveturnedouttobeinadequateforcomplex,interdependentactivitieslikeinnovationbecausetheefficiencyofthemicro-structurestoooftenleadstomacrononsense(VandeVen,1986).Analternative,heproposes,wouldbetousesimultaneouscouplingofbusinessfunctions,basedonthehologram/brainsmetaphorofMorgan(1986).Thisrequiresradicallydifferentdesignprinciplesfortheorganizationofinnovation,specifically:
i. allowingthecollectionofactorsresponsibleforinnovationtoself-organizeii. creatingredundantfunctions,ratherthannarrowspecialisms
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page4of19
iii. assuringrequisitevariety(Ashby,1962),meaningthatthecomplexityoftheinternalsystemshouldbelargeenoughtodealwiththecomplexityoftheenvironment
iv. usingtemporallinkage,meaningthatactorscanconfigureintogroups,changeconfigurations,eliminateconfigurationsandreconfigureintodifferentgroups,basedonthedemandsoftheirinnovationtask.
Fortheseprinciplestowork,VandeVen(1986)continues,thesystemneedsthegovernance,institutionalcharacteristicsandinfrastructurethatenableittolearn.Thisrequiresnetwork-buildinginsideandoutsidetheorganization.
O’Connor(2008)usedasimilarapproachtomakepropositionsabouthowfirmscouldorganizeandbuildcapabilitiesforgeneratingmajorinnovations.Sheproposesfourrequirementsforthevariousaspectsofacompanyinnovationsystemtobedefinedasa‘system’:
i. thesystemshouldbeidentifiableanditselementsshouldbeinterdependentii. thewholeshouldbegreaterthanthesumofthepartsiii. thereshouldbeinternalandexternalrelationshipsforthesystemtobeindynamicbalanceiv. itshouldhaveaclearpurposeinthelargersysteminwhichthesystemisembedded
Carlssonetal(2002),intheirpaperonanalyticalandmethodologicalissuesforinnovationsystems,indicatethatsystemsconsistofcomponents,therelationshipsamongthem,andtheircharacteristicsorattributes.Componentsareactors,artifacts(cf.,theresourcesmentionedbyGranstrand,1998),andinstitutions(suchaslaws,traditionsandnorms).Therelationshipsbetweenthecomponentsareessentialfortheformationofasystem:thepartsinfluenceeachother;thepartsinfluencethewhole;andthewholeinfluencestheparts.Suchrelationships,Carlssonetal.(2002)argue,canbemarket-basedonnon-marketbased.Thefeedbackloopsintherelationshipsprovidethedynamicsofthesystem.Attributesarepropertiesofcomponentsandrelationships,suchascapabilitiesforselectingmarkets,technologiesandorganizationmodes,organizationalcapabilitiesforcoordinatingandintegratingactivities,functionalcapabilitiesforexecutingtasksefficiently,andadaptivecapabilitiesthatallowthesystemtolearnfromsuccessandfailure.Nexttothis,thesystemhasdynamicproperties,suchasrobustness,flexibility,theabilitytogeneratechangeandtheabilitytorespondtochanges.Suchchanges,theyarguecanbeendogenouslyorexogenouslyinduced.Carlssonetal.(2002)definethreemajormethodologicalissuestoresolveforconceptualizingasystem:
i. whatisthelevelofanalysis?ii. whatisthedefinitionofthesystemboundary?iii. whatconstitutessystemperformance?
First,thelevelofanalysisforacompanyinnovationsystemisfairlyclear:itwillbeeitherthecorporateorthebusinessunitlevel.Weneedtobecarefulnottomixthelevels,althoughtheycanbeanalyzedtogetheraslongasitisclearonwhichlevelweare.Second,wedefinethesystemboundaryforthemomentverystraightforwardasthelegalboundaryofthecompany.Apossiblealternativewouldbetouseastakeholder-baseddefinitionthatwouldinclude,forexample,subcontractorsworkingwithinthecompanyortemporarylaborers.Furthermore,asarguedbyCoriat&Weinstein(2002)andasshownintheconceptsofSigurdson&Cheng(2001)andChenetal.(2015)thecompanyinnovationsystemmaybestronglyintertwinedwithwiderecosystemsor
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page5of19
regional,sectoralornationalsystemsofinnovation,anditmayinpracticenotbesoeasytodefinewhatis‘inside’andwhatis‘outside’.Thedynamicsofthesystemmeanthatthecompanyboundarymayevolveovertime.Nexttoorganicevolutionofthecompany,thismaybetheresultofmergersoracquisitions,orofallkindsofspin-outandsource-inmodesthatresultfromopeninnovation.Third,wedefinesystemperformanceorpurposeofacompanyinnovationsysteminabroadsense:itsoutputsincludedifferenttypesofinnovation(product,technological,businessmodel,organizational)withdifferentratesofinnovativeness(e.g.,incremental,radical),differentratesofsuccessorfailureanddifferentratesofimpactonthecompany,themarketortheworld.
Components
Followingthe‘template’forinnovationsystemsasputforwardbyCarlssonetal.(2002),wedefineacompanyinnovationsystemas‘thesystemofcomponents,relationshipsandattributeswithintheboundaryofacompany,thathasthepurposeofproducinginnovation.’
Startingwiththecomponents:thesecanbeactors,suchasindividuals,teams,departmentsorbusinessunits,businessorcorporatedecisionmakers(executivessuchasChiefTechnologyOfficersorChiefInnovationOfficers)incubators,centralordecentralR&Ddepartments,laboratoriesorfacilitatinggroups.Granstrand(1998,p.475),inhisconceptualizationofthetechnology-basedfirm,viewsafirmas“…alegallydefined,dynamichumansystem,consistingofasetofheterogeneousresourcesinaninstitutionalsetting…”.Heidentifiesresourcesasthemostimportantcomponentsofthesystemandheprovidesadetaileddiscussionoftheseresources,namelyphysicalcapital,financialcapital,intellectualcapital,relationalcapitalandhumanembodiedcapital.Thelatter,is,ofcourse,embeddedintheactorsandtheirrelationships.
Relationships
Thecomponentsoftheinnovationsystemcanbeconfigured,bytop-downcoordinationorbyself-organization,intoconfigurationsthataddressspecifictasks.Forexample,oneconfigurationfocusesoncomingupwithnewideas,anotherconfigurationfocusesondevelopingandlaunchingnewproducts(seefigure1).Wearguethattheseconfigurationswillnormallymatchthecomplexityofthecompany’senvironment(‘requisitevariety’).Wefurtherarguethattheseconfigurationscanrangefromtemporarytopermanent(‘temporallinkage’).Successfulconfigurationsaroundrecurringtasksareexpectedtobemorepermanent,reflectingthecompany’sexploitationofexistingresources,activitiesandcapabilities.Configurationsmeanttodiscovernewcombinationsorunsuccessfulconfigurationswillbemoretemporaryandreflectacompany’sexplorationactivitiesandthebuildingofnewresourcesandcapabilities(seealsoO’Connor,2008).Atthesystemlevel,wewillthereforeseeidentifiablestructuresandinterfaceswiththemainorganizationandtheenvironment(O’Connor,2008).
Notallthesystem’sactorsandresourcesneedtobeinvolvedineveryconfiguration.Indeed,differentconfigurationscanexistsimultaneously,partlyoverlappingandusingthesameactorsandresources,whileperhapsleavingotheractorsandresourcesunused(‘reducndancy’).Theuseofactorsandresourcesfordifferenttasks,orthe(temporal)lackofuseofsomeactorsandresourcesalignwithVandeVen’s(1986)designprincipleof‘redundancyoffunctions’:thecapacityofthesystemislargerthanwhatitactuallyneedsforanyspecificconfiguration,butpreparesitforwiderneeds.
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page6of19
Figure1:CompanyInnovationSystem
Weknowfrominnovationprocessliteraturethatprocessstructurescanrangefromlinearandsimple(inalimitedinterpretationofStage-Gate),tononlinearandinvolvingcomplexcoordinationmechanisms.Lookedatinthisway,aninnovationprocessbecomesaspecificconfigurationmodeofcomponentsoftheinnovationsystem.Thecompanyinnovationsystemapproachdoesthereforenotreplaceorcompetewiththeprocessapproach,butcomplementsandgeneralizesit.AccordingtoOrtt&VanderDuin(2008),acompanymaychooseadifferentconfigurationofitsinnovationprocess,contingentuponthetypeofinnovation,thetypeofbusiness,theresourcesavailable,ortheexternalenvironment.Differentconfigurationsmayexistsidebyside.Someconfigurationsaresuccessfulandwillbemorepermanentlylinked,meaningthatthecompanywillusethemoverandover.Otherconfigurationsmaybemoretemporarylinked.Thisalsopointstothecoordinationfunctioninacompanyinnovationsystem.Amorepermanentconfiguration,onceinplace,mayrequirerelativelylesscentralcoordination.Temporalconfigurationsandreconfigurationswillrequirecontinuouscoordination,eitherself-organizedandcentrallyfacilitatedorcentrallymanagedandcontrolled.
Whatdosuchconfigurationslooklike?Differentconceptshavebeenproposed.Teece(1996)identifieddifferentarchetypesofsystemrelationshipgovernancebasedontheinstitutionalcharacteristicsofexternallinkages,hierarchicaldecisionmaking,changeculture(whichheequateswithinformalstructure),scope(multi-orsingleproduct),andverticalintegration.AsarchetypesTeece(1996)identifiedthemultiproductintegratedhierarchy,thehighflexSiliconValleytype,thevirtualcorporationandtheconglomerate.Eacharchetypefacilitatesspecifictypesofinnovationandthecreationoforaccesstospecifictypesofcapabilities.
Chenetal.(2015)useRothwell’s(1994)fiveinnovationgenerationsasastartingpointtoidentifyanumberofarchetypesofinnovationsystems.TheuseofRothwell’s(1994)generationsalsoimpliesa
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page7of19
genericevolutioninhowcompaniesstructureandleadtheirinnovationsystemsovertime.Specifically,Chenetal.(2015)identify:
i. theinternalR&D-orientedinnovationsystem,withadominanttechnology-pushroleofinternalR&D
ii. theinternalandexternalcollaborativeinnovationsystem,withinterconnectedR&D,marketingandmanufacturingfunctions
iii. thehighlystrategy-orientedinnovationsystem,ledfromthebusinessstrategybytheCEOorChiefInnovationOfficer
iv. theecologicalinnovationsystem,whichdepartsfromthecompanylevelandseesthecompanyasanactorinabusinessecosystem
Coriat&Weinstein(2002)identifytwomainquestionsregardingsuchconfigurations.Thefirstis“Howcanoneunderstandboththediversityoforganizationalpatternsandtheexistenceofdominantmodesoforganization?”(p.276).ThisquestionisrelatedtoVandeVen’s(1986)designprincipleofrequisitevariety-differentenvironmentsrequiredifferentlevelsofsystemcomplexity-andtotheprincipleoftemporallinkage-someorganizationmodesaresuccessfulacrossenvironmentsandovertime,andwilltendtomorepermanentlinking,exploitingexistingresourcesandcapabilities.Coriat&Weinstein’s(2002)secondquestionis“Howcanorganizationalpatternsevolvetogivebirthtonewprinciplesandorganizationalsystems?”(p.277).ThisquestionisrelatedtoVandeVen’s(1986)designprincipleofself-organization(givenanewandunknowntask,thesystemwillreconfiguretotryandsolvethetask)andalsototheprincipleoftemporallinkage(unsuccessfulmodesormodesthatarenotcontinuallyrequiredwilldissolveandbereplacedbyothermodes,exploringnewroutinesandbuildingcapabilities).Coriat&Weinstein(2002)warnagainsttreatingthecompanyasaclosedsystem,explainingthatthecompany-and,consequently,itsdynamics-isapartofthewiderinstitutionalenvironment.Theyinterpretthisinstitutionalenvironmentmainlyasthenationalorsectoralsystemsofinnovationthatthecompanybelongsto.
Likeanysystem,acompanyinnovationsystemhasanenvironment,i.e.,thatwhichisoutsidethecompany,anditinteractswiththisenvironment,exchanginginputsandoutputswithit.Likethesystem’sownresources,suchinputsandoutputscanbephysical,financial,intellectual,relationalandhuman.Externalinnovationsystemssuchasplatforms,businessecosystems,technologysystems,regional,sectoralandnationalinnovationsystemsandthepublicsectorareimportantpartsofthatenvironment(cf.,Sigurdson&Cheng,2001).Companyinnovationsystemscanbemoreorlessopentothatenvironment(Chesbrough,2003).
Attributes
Finally,wecanidentifysystemattributes.Thelistofpossibleattributesislongandinvolveseverycharacteristicthatpotentiallyinfluencessystembehaviorandperformance.Sigurdson&Cheng(2001)providetenelementsofcorporateinnovationsystems,ofwhichsevencanbecharacterizedasattributes:organizationalabilityandstrategy,arrangementsforadvancedlearning,humanresourcemanagement,competitivestrategy,managementofintellectualpropertyrights,networkingabilityandstrategy,andfinancingstrategy.O’Connor(2008)proposesrequisiteskillsandtalentdevelopment,governanceanddecision-makingmechanisms,performancemetrics,andcultureandleadershipcontext.Steiber&Alänge(2013),inanin-depthcaseanalysisofGoogle’s
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page8of19
innovationsystem,identifysevensystemcharacteristicsthatarecoretocontinuousinnovation:innovation-orientedculture,selectionofindividuals,leadersandfacilitators,organizationalinfrastructure,performanceandincentives,organizationallearning,andexternalinteraction.Apartfromtheexternalinteraction,whichisanaspectofsystemrelationships,allaresystemattributes.
Manyoftheseattributeshaveincommonthattheyaredifficulttomeasure,somethingthatSteiber&Alänge(2013)forexampledobyinterviewswithcompanyemployees.Basedontheattributesproposedandtakingmeasurabilityintoaccount,weidentifyfiveattributesofcompanyinnovationsystems:
i. innovationasanexceptionversusinnovationasday-to-daybusiness;thisisrelatedtoinnovationculture
ii. non-systematicversussystematicinnovation;thisisrelatedtoorganizationalinfrastructureforinnovation
iii. dramaticturnaroundsversusadaptableorganization;thisisrelatedtoinnovationcultureiv. largebreakthroughprojectsversus(massive)experimentation;thisisrelatedtomechanisms
oflearning(see,e.g.,Thomke,2001;Taylor&Wagner,2014)v. theabilitytocombinecreativityandefficiency;thisisrelatedexploration-exploitation,to
mechanismsoflearningandtoorganizationalinfrastructure(see,e.g.,Reevesetal.,2013)
METHOD
Weexploretheconceptofcompany innovationsystembyanalyzingcaseexamplesofABBGroup,Adobe Systems, Amazon.com, eBay, Hitachi, HTC, Lockheed Martin, Philips, Qualcomm,Salesforce.com and Southwest Airlines. To construct the case examples,we used a case protocolthatcontainsrelevantaspectsofthecompanyinnovationsystem(seetable1).Fortheanalyses,onlypubliclyavailabledatawasused,suchasthecompanywebsiteanditsannualreports,pressreleases,mediacoverage,academicarticlesandotheravailablecasestudies.Thisdataputslimitsonwhatwecanmeasure;itisnearimpossible,forexample,tomeasureinnovativeculture,leadershiporhumanresourcemanagementpractices.
Table1:Caseprotocol
Question AspectofCIS
Analyzethecompanyaccordingtothecharacteristicsof“moderninnovationmanagement”(useasemanticdifferentialscale):
i. InnovationasexceptionversusInnovationasday-to-daybusiness Attributes
ii. NonsystematicversusSystematic Attributes
iii. DramaticturnaroundsversusAdaptableorganization Attributes
iv. LargebreakthroughprojectsversusMassiveexperimentation Attributes
v. Closed,withincompanyversusopen,innetworks Relationships
vi. InnovationasthebusinessofR&DversusInnovationaseverybody’sbusiness
Components
Didthisinnovationappearfromatechnology-pushmodel,fromamarketpullmodel,fromaninteractivemodel,orfromopeninnovation?
Relationships
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page9of19
Describetheinternalinnovationsystemofthecompany:
i. Isinnovationrepresentedattheexecutivelevel?Doesthecompanyhavea‘chiefinnovationofficer’or‘chieftechnologyofficer’?
Components
ii. Whatarethemaincomponents(actors,departments,units,incubators,centralorde-centralR&Ddepartmentsorlaboratories,etc.)involvedininnovation?
Components
iii. Howdothesecomponentsworktogethertocreateinnovation? Relationships
iv. Drawapictureofthecomponentsoftheinnovationsystemandhowtheyworktogether
Components
Relationships
Analyzewhichmechanismsthecompanyusestocombinecreativitywithefficiency
Attributes
Based on this case protocol junior researchers, i.e., senior undergraduate students, analyzed 152casesof largecompanies.Wecheckedtheseinitialcasesforqualityandcompleteness,forinternalconsistencyandformutualconsistency.Weselectedthe11bestandmostcompletecasesanddidacross-caseanalysisofthose,makingadditionalinterpretationswherenecessary.
RESULTS
Wepresenttheresultsofthecaseanalysesinthetables2,3,and4below.
Table2:Cross-caseanalysisofcoordinationandsystemcomponents
Company Overallcharacter
Principleofcoordination
Components:actors
Components:C-level
Components:R&Dvseverybody
ABBGroup Complex, linear, technology-driven
Key individuals connecting R&D with business; collaborations facilitated through software systems
R&D global and division level, ventures department
Chief Technology Officer overseeing all aspects
R&D is core
AdobeSystems
Individual-based, decentralized, rule-based
Informal, complex, highly de-centralized, but strictly rule-based
All employees (individual)
Chief Technology Officer and a Chief innovation and creativity officer
Everybody's business
Amazon.com Many small teams, informal, complex, decentralized
Informal, complex, highly de-centralized
Many small, independent teams; Lab 126
Chief Technology Officer overseeing key projects
Everybody's business
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page10of19
Company Overallcharacter
Principleofcoordination
Components:actors
Components:C-level
Components:R&Dvseverybody
eBay Multi-incubation, decentralized, technology driven
Self-organizing under strong C-level leadership
Incubation programs worldwide; hackatons; Innovation demo expo; R&D department
Chief Technology Officer leading technology vision and strategy
Everybody’s business, but technology driven
Hitachi Complex and holistic; project-based
Project-based and rule-based
Sister companies; de-central business development; technology strategy office; centers for social and for technological innovation; center for exploratory research
Chief Technology Officer and many decentral business development executives
Moderate, leaning toward everybody’s business
HTC Technology-driven and open
Self-organizing under strong C-level leadership
Developers (internal and external); HTC Design studio; Incubators; Business partners
No, but innovation is integral part of C-level responsibilities
Internal R&D and external partners
LockheedMartin
Multi-modal Depends on the type of innovation
Specialized R&D departments; Skunk works; de-centralized business; ‘lighthouse’ for open innovation
No; C-level officers of the business are responsible
Specific departments for specific breakthroughs combined with everybody’s business for improvements
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page11of19
Company Overallcharacter
Principleofcoordination
Components:actors
Components:C-level
Components:R&Dvseverybody
Philips Interactive, systematic, complex
Collaboration between various departments
Decentral country-based and business-based; central group innovation department; Incubators; Design center
Chief Technology Officer
Decentral: not everybody, but also not a single department: multiple departments are leading
Qualcomm Linear, technology push
R&D driven, linear process
R&D facilities around the world: labs and incubators; decentralized units, each with own unique knowledge
Chief Innovation Officer
Mainly R&D
Salesforce.com Simple, customer data-driven
Initiation de-central; implementation centralized; more impactful innovation coordinated by executive level
Individual employees; tools and labs
No, but innovation is integral part of C-level responsibilities
Mostly business development, driven by the business
SouthwestAirlines
Lean and flexible; decentralized yet efficiently coordinated
Strategy and innovation department initiates and coordinates
Multiple functional business departments
Chief Technology Officer and Chief Strategy & Innovation Officer
No R&D department: multiple departments involved
We find that using a company innovation systemapproachwe canmap themain components ofcompanies’ innovation systems, such as executive level representation, central innovationdepartments, de-central departments in regions or attached to the business units and businessteams. Companies differ in their centralization of innovation efforts, in the emphasis they put oninnovationbydepartments (mostlytheolder, industrialones)versus innovationby individualsandsmall teams (mostly theyounger, IT-basedones),and inseeing innovationas thebusinessofR&Dversusthebusinessofeverybodyinthecompany.Executiverepresentation isremarkablyuniform,witheitherspecificC-levelinnovationortechnologyofficersorclearinnovationresponsibilitieswithgeneralC-levelofficers.
Large differences can be seen in coordination mode: from top-down via R&D driven to self-organizing. We added an overall characterization of the company innovation system, which also
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page12of19
showsconsiderabledifferencesandwhichdoesnotimmediatelyshowanyarchetypesasproposedbyTeece(1996)orChenetal.(2015).
Table3:Cross-caseanalysisofsystemrelationships
Company Relations:innovationgeneration
Relations:closedversusopen
Relations:singleormultipleconfigurations
Relations:simple,linearversuscomplex
Relations:fixedversusreconfigurable
ABBGroup Interactive model, leaning toward technology-driven
Mostly closed; working with universities on R&D
Single: combining technology with customer input
Complex with interdisciplinary cooperation
Reconfigurable in idea and testing stage; fixed in development stage
AdobeSystems
Market pull Open or closed when and where necessary
Many parallel processes
Moderate complexity (small teams)
Extremely configurable
Amazon.com Interactive Mostly closed
Many teams ('pizza teams') working parallel on different projects
Moderate complexity (small teams)
Extremely configurable
eBay Open innovation, but technology is core driver
Leaning toward open
Multiple processes (internal and external paths)
Simple process (technology-driven), but with complex coordination
Reconfigurable within limits
Hitachi Between interactive and technology push
Moderate: not open but working with many alliances and partners
Multiple processes (project-based)
Complex, continuous collaboration for holistic solutions (project-based)
Project-based reconfigurable
HTC Open innovation, but technology is core driver
Open, with partners and external developers (open source)
Single Linear process, but complex coordination
Depending on the project different actors involved
LockheedMartin
Contextual: technology push, interactive and semi-open all exist
Multiple: closed and semi-open
Multiple processes: technology-driven and combining technology with customer input
Linear for breakthrough projects; complex for other projects
N/A
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page13of19
Company Relations:innovationgeneration
Relations:closedversusopen
Relations:singleormultipleconfigurations
Relations:simple,linearversuscomplex
Relations:fixedversusreconfigurable
Philips Interactive Moderate not open but working with many alliances and partners
Single, combining technology with customer input
Very complex, many departments and units involved
Reconfigurable within limits: depending on the project different actors involved
Qualcomm Technology-push
Mostly closed, but cooperation with universities in early stages
Single Simple linear Fixed
Salesforce.com Between demand-pull and interactive
N/A Multiple: data analytics-driven and customer insight-driven
Cross-functional teams
Fixed
SouthwestAirlines
Demand-pull
Moderate: open to input from networks, but innovation internal
Dual: initiation and implementation
Dual: initiation is more complex, implementation is more linear
Reconfigurable for implementation (‘relaxed structure’)
Wefindthatwecanindicatetherelationshipsbetweenthecomponentsofthecompanyinnovationsystems. This is no surprisewith the innovationgenerations concept (Rothwell, 1994;Chenet. al,2015), althoughherewe see companies that are inbetweengenerations and companies thatusemultiplegenerationsinparallel.
Internal and external innovation systems are intertwined (Sigurdson & Cheng, 2001; Coriat &Weinstein,2002;Chenetal.,2015),henceweshouldbecareful to include theeffectsofexternalinnovationsystemsonthecompanyinnovationsystem.Theopenversusclosedaspectcapturespartof these relationships, andhere, too,we seedifferences that transcenda simpleone-dimensionalscale:somecompaniesusemultiplemodes,beingopenforsomeinnovationproblemswhileclosedforothers.
We find that many companies use multiple configurations, either multiple processes or multiplenetworksorsubsystems,withintheircompanyinnovationsystems.Thisshowsthevalueofcompanyinnovationsystemconceptversustheinnovationprocessapproach.Processesexist,buttheyareaspecific typeofconfigurationof thesystem,even if somecompanies limit their systemtoasingleconfiguration.
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page14of19
Singleormultipleconfigurationsdoesnotequalsimple,linearversuscomplexconfigurations.Thereare companies with single, yet complex configurations and companies with multiple yet simpleconfigurations.
Anotherinsightfromthecaseexamplesisemphasisonreconfigurationsand‘newcombinations’(cf.,Edquist, 1997).We can detectwhere new configurations come from, e.g., from individuals, fromteams,fromcentraloffices,orfromprojects.Intheinnovationsystemsapproach,themechanismofhownewcombinationshappenisbuiltintotheanalysisratherthanexternallyassumed.Admittedly,alotofconceptualandempiricalworkneedstobedonetoclearlydemonstratethisprinciple.
Table4:Cross-caseanalysisofsystemattributes
Company Attributes:combiningcreativityandefficiency
Attributes:exceptionversusday-to-day
Attributes:systematicversusnon-systematic
Attributes:turnaroundsversusadaptable
Attributes:breakthroughsversusexperimentation
ABBGroup Multipleapproaches:separation;scientistinitiativewithbusinessselectioncriteria;localbusinesstouseglobalresources
Day-to-day Systematic N/A N/A
AdobeSystems
Self-organizationandself-selection
Day-to-day Moderatelysystematic
Adaptable Massiveexperimentation
Amazon.com Multipleapproaches:massiveexperimentationcombinedwithquicklearningfrommarketfeedback;separation(Lab126)
Day-to-day Moderatelysystematic
Adaptable Massiveexperimentation
eBay Internalandexternalecosystem(eBayinnovationdemoexpo;incubatorsandhackatons)
Day-to-day Systematic Adaptable Regularmajoradaptations(e.g.,Paypal,Skypeintegration)
Hitachi Self-organizationdeliversideasandexperiments;centralmanagementsetsboundaries
Day-to-day Systematic N/A Experimentation
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page15of19
Company Attributes:combiningcreativityandefficiency
Attributes:exceptionversusday-to-day
Attributes:systematicversusnon-systematic
Attributes:turnaroundsversusadaptable
Attributes:breakthroughsversusexperimentation
HTC Externalecosystem Day-to-day Systematic Adaptable Experimentation
LockheedMartin
Multipleapproaches:separation(Skunkworks);withinbusinessself-organizationandexternalecosystem
Day-to-day Systematic Bothadaptableandturnarounds
Breakthroughs
Philips Multipleapproaches:separation(PhilipsResearch);self-organizationthroughcomplexcooperation;ecosystemwithpartners
Day-to-day Systematic Adaptable N/A
Qualcomm R&Dgeneratesideaincooperationwithuniversities;thenlinearprocesstoselectandimplement
Day-to-day Systematic Continuousadaptation
N/A
Salesforce.com Culturesupportsideageneration;veryselectiveinimplementation
Day-to-day Systematic Moderate,leaningtowardadaptable
Regular well-functioningchanges; noradicalchanges
SouthwestAirlines
Separationandself-organization
Moderate,notday-to-day,butalsonotexceptional
Systematic Moderate:standardizedprocessesnotadaptable,butregularupdates
Towardexperimentationand fastlearning; noradicalchanges
Wefindthatwecan identifycompany innovationsystemattributes.Thecombinationofcreativityand efficiency (exploration-exploitation) shows different approaches taken by companies: fromhavingseparateunitsforcreativity/ideagenerationandefficiency/implementationmodes;viaself-organization,wheretheindividuals,teamsordepartmentsthemselvesdecidewhichmodetofocusonand/orwhentomaketheswitchbetweenthemodes;totheuseofexternalecosystems,whereexternal networks are configured for generating ideas or for efficient collaboration (Reeves et al,2013). The attributes of day-to-day innovation versus innovation as exception, systematic versusnonsystematicinnovation,andadaptableorganizationversuslargeturnaroundsdonotshowalotof
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page16of19
variance.Thismaybebecauseourselectionisbiasedtoinnovativeandlargercompanies,orbecausetheseattributesarenot sufficientlydifferentiating.Furtherconceptualdevelopmentandempiricaltestingseemsnecessaryhere.Thelastattribute,breakthroughsversusexperimentation–alearningmechanism(e.g.,Thomke,2001)–againshowsdifferencesfromoneendofthescaletotheother.
CONCLUSION
Objectiveandfindings
The objective of this paper was to conceptualize the company as an innovation system and toexplore this concept using descriptive case examples. The company innovation system approachbuildsonexistingtheoreticalfoundationsandsomerelevantbuildingblockshavebeenproposedintheliterature.Itprovidesadditionalandcomplementaryinsightstoexistingapproaches,specificallytotheinnovationprocessapproach.
Wefindthatusingthecompanyinnovationsystemapproach,wecanmapinnovationsystemsatthecompany levelandshowthedifferencesbetweenthem.Wecan identify thecomponents,suchasR&D departments, labs, venture organizations, teams, employees, C-level offices and facilitatingtools. We can identify relationships such as single or multiple configurations, simple or complexconfigurations, technology-driven,market-drivenor interactive configurations, andopenor closedconfigurations.Wecanidentifysystemcharacteristicssuchascreativityversusefficiencyemphasis,systematicversusnon-systematicapproaches,adaptivenessofthesystem,andlargeprojectversusexperimentationfocus.
Academicandpracticalimplications
This research has academic and practical implications. Academically, we propose that innovationmanagementcanbeanalyzedusingthecompanyinnovationsystemapproach.Wealsoproposethatcertainproblemsandcharacteristics,suchascross-functionalcooperation,learningandknowledge,theemergenceof newcombinations, and coordinationof the innovation functions, canbebetteranalyzedanddeeperunderstoodbyusingacompanyinnovationsystemapproachthanbyusinganinnovationprocessapproach.
Practically,companiesneedtoaddressproblemsofinnovationsystemdesign(whoorwhichpartofthecompany innovationsystem is responsible forwhat), innovationsystemstructure (howdothedifferent parts of the company innovation system work together) and innovation systemcoordination(howtoensurethatthecompanyinnovationsystemisproductive,fulfillsitsobjectives,and is stable). A well-conceptualized and validated company innovation system approach mayprovidemanagerswithinsightstoaddressthoseproblems.Whichbringsustofutureresearch.
Scopeforfutureresearch
The implications stated above are, at thismoment, tentative and preliminary. The current paperprovides merely an initial conceptualization and descriptive exploration of the company as aninnovationsystem.Furtherconceptualworkisneededtofleshouttheconceptanditssub-concepts,to integraterelevantadditionalconcepts(e.g.,variousattributes), toensuretheabilitytotestandfalsifytheseconcepts,andtoclarifytheconnectionswithanddistinctionfromrelatedconcepts.All
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page17of19
the analytical and methodological aspects of innovation systems as identified by Carlsson et al.(2002)shouldbefurtheraddressedandclarified.Specificissuesthatcometomindare:
i. the system definition/boundary, e.g., a strictly legal definition of the company versus astakeholder involvement definition, embedding in environment and larger systems (e.g.,Sigurdson&Cheng,2001;O’Connor,2008)
ii. the roleof the systemdesignprinciples (e.g.,VandeVen, 1986;Morgan1986;O’Connor,2008)
iii. the possible configurations of the system, archetypical, permanent, or temporary (e.g.,Teece,1996;Chenetal.,2015)
iv. the governance and institutional characteristics of the system (e.g., Van de Ven, 1986;Teece,1996;Sigurdson&Cheng,2001;Steiber&Alänge,2013)
v. adefinitionoftheelementaryunitsofthesystem(e.g.,Granstrand,1998)
vi. theroleofresources(Granstrand,1998)andcapabilities(Coriat&Weinstein,2002)
vii. thedynamicandevolutionaryaspectsoftheconcept(Carlssonetal.,2002)
Furtherempiricalworkisneededtodotheactualtestingandtodemonstratetheusefulnessoftheapproachforanalyzinginnovationincompanies.Suchempiricalworkcouldstartwithmappingtheinnovation systems of companies using the case method, making cross-sectional comparisonsbetweencompanies,or followingtheevolutionofspecificcompany innovationsystemsovertime.Specifically,duetothecontinuousfeedbackinthesystemproducedbyitsinteractions,weshouldbecarefulwith‘snapshots’(Carlssonetal.,2002).Longitudinalresearchisthereforestronglypreferred.
Upon availability of a sufficient basis of empirical observations, further questions could beempirically tackled, such as the contingency between system structure, governance and theenvironment, or the relationship between system structure, governance and innovativeperformance.
ACKNOWLEDGEMENTS
We thank our students Aarush Jaggi, Buğra Bilgili, Deniz Tunalı, Ecenaz Başaran, Emre Özdemir,GökçeTurna,KerimAlikemal,TahirGündüz,TracyMolho,UktuKuranandYuşaAtışfortheireffortsindoingtheinitialcaseanalysesandwethankDr.MarcZegveld(IBM)forhisinitialideasandworkontheconceptofcompanyinnovationsystemattheDelftUniversityofTechnology.
REFERENCES
Alikemal,K.(2017),QualcommWiPower.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Ashby,W.R.(1962),Principlesoftheself-organizingsystem.InPrinciplesofSelf-Organization:TransactionsoftheUniversityofIllinoisSymposium,H.VonFoerster&G.W.Zopf,Jr.(eds.),pp.255–278.London:PergamonPress.
Atış,Y.(2017),LockheedMartinF35LightningIIJetFighter.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page18of19
Başaran,E.(2017),HitachiSmartCityAnalysis.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Bilgili,B.(2017),HTCViveVR.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Carlsson,B.,Jacobsson,S.,Holmén,M.,&Rickne,A.(2002),InnovationSystems:AnalyticalandMethodologicalIssues.ResearchPolicy,31(2),233–245.
Carlsson,B.&Stankiewicz,R.(1991),Onthenature,functionandcompositionoftechnologicalsystems.JournalofEvolutionaryEconomics,1(2),93–118.
Chen,J.,Huang,S-f.,&Xu,Q-r.(2015),FirmInnovationSystems:PerspectivesofResearchesonState-ownedKeyEnterprises.FrontiersofEngineeringManagement,2(1),64–70.
Chesbrough,H.(2003),Thelogicofopeninnovation:managingintellectualproperty.CaliforniaManagementReview,45(3),33–58.
Cooke,Ph.,Uranga,M.G.,&Etxebarria,G.(1997),Regionalinnovationsystems:Institutionalandorganisationaldimensions.ResearchPolicy,26(4–5),475–491.
Cooper,R.G.(1985),Selectingwinningnewproductprojects:UsingtheNewProdsystem.JournalofProductInnovationManagement,2(1),34–44.
Cooper,R.G.(1990),Stage-gatesystems:anewtoolformanagingnewproducts.BusinessHorizons,33(3),44–54.
Cooper,R.G.(2008),Perspective:TheStage-Gate®idea-to-launchprocess—Update,what’snew,andNexGensystems.JournalofProductInnovationManagement,25(3),213–232.
Coriat,B.&Weinstein,O.(2002),Organizations,firmsandinstitutionsinthegenerationofinnovation.ResearchPolicy,31(2),273–290.
Edquist,Ch.(1997),Systemsofinnovation:technologies,institutions,andorganizations.London:Routledge.
Gawer,A.&Cusumano,M.A.(2014),Industryplatformsandecosysteminnovation.JournalofProductInnovationManagement,31(3),417–433.
Granstrand,O.(1998),Towardsatheoryofthetechnology-basedfirm.Researchpolicy,27(5),465–489.
Granstrand,O.(2000),CorporateInnovationSystems:AComparativeStudyofMulti-TechnologyCorporationsinJapan,SwedenandtheUSA.PapersubmittedtotheDynacomproject.
Gündüz,T.(2017),AdobeSystems–AdobeCreativeCloud.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Iansiti,M.&Levien,R.(2004),Strategyasecology.HarvardBusinessReview,82(3),68–81.
Jaggi,A.(2017),AmazonEcho.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Kuran,U.(2017),ABBKNXI-BUS.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Lundvall,B-A.(1992),Nationalsystemsofinnovation:Ananalyticalframework.London:Pinter.
Malerba,F.(2002),Sectoralsystemsofinnovationandproduction.ResearchPolicy31(2),247–264.
InternationalAssociationforManagementofTechnologyIAMOT2018ConferenceProceedings
Page19of19
Moore,J.F.(1993),Predatorsandprey:anewecologyofcompetition.HarvardBusinessReview,71(3),75–83.
Molho,T.(2017),eBay.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Morgan,G.(1986),Imagesoforganization.BeverlyHills:SagePublications.
O’Connor,G.C.(2008),Majorinnovationasadynamiccapability:Asystemsapproach.JournalofProductInnovationManagement,25(4),313–330.
Ortt,J.R.&VanderDuin,P.A.(2008),Theevolutionofinnovationmanagementtowardscontextualinnovation.EuropeanJournalofInnovationManagement,11(4),522–538.
Özdemir,E.(2017),Salesfoce.comandCloudComputing.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Reeves,M.,Haanaes,K.,Hollingsworth,J.&ScognamiglioPasini,F.L.(2013),Ambidexterity:theartofthrivingincomplexenvironments.BostonConsultingGroupPerspectives.
Rothwell,R.(1994),Towardsthefifth-generationinnovationprocess.InternationalMarketingReview,11(1),7–31.
Sigurdson,J.&Cheng,A.L-P.(2001),Newtechnologicallinksbetweennationalinnovationsystemsandcorporations.InternationalJournalofTechnologyManagement,22(5/6),417-434.
Steiber,A.&Alänge,S.(2013),Acorporatesystemforcontinuousinnovation:thecaseofGoogleInc.EuropeanJournalofInnovationManagement,16(2),243-264.
Taylor,A.&K.Wagner(2014),Rethinkingyourinnovationsystem.BostonConsultingGroupPerspectives.
Teece,D.J.(1996),Firmorganization,industrialstructure,andtechnologicalinnovation.JournalofEconomicBehavior&Organization,31(2),193–224.
Thomke,S.(2001),Enlightenedexperimentation:thenewimperativeforinnovation.HarvardBusinessReview,79(2),67-75.
Tidd,J.&Bessant,J.(2013),ManagingInnovation:IntegratingTechnological,MarketandOrganizationalChange-5thEdition.Hoboken:Wiley.
Trott,P.(2011),Innovationmanagementandnewproductdevelopment-5thEdition.Harlow:PearsonEducation.
Tunalı,D.(2017),Southwest.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
Turna,G.(2017),PhilipsLumea.AssignmentforENTR401/MGMT401TechnologyandInnovationManagement,ÖzyeğinUniversityBusinessSchool.
VandeVen,A.H.(1986),Centralproblemsinthemanagementofinnovation.ManagementScience,32(5),590–607.