G. Technological evolution and radical innovation

152/JournalofMarketing,July2005JournalofMarketingVol.69(July2005),152–168AshishSood&GerardJ.TellisTechnologicalEvolutionandRadicalInnovationTechnologicalchangeisperhapsthemostpowerfulengineofgrowthinmarketstoday.Toharnessthissourceofgrowth,firmsneedanswerstokeyquestionsaboutthedynamicsoftechnologicalchange:(1)Howdonewtech-nologiesevolve?(2)Howdorivaltechnologiescompete?and(3)Howdofirmsdealwithtechnologicalevolution?Currently,theliteraturesuggeststhatanewtechnologyseemstoevolvealonganS-shapedpath,whichstartsbelowthatofanoldtechnology,intersectsitonce,andendsabovetheoldtechnology.Thisbeliefisbasedonscat-teredempiricalevidenceandsomecirculardefinitions.Usingnewdefinitionsanddataon14technologiesfromfourmarkets,theauthorsexaminetheshapeandcompetitivedynamicsoftechnologicalevolution.Theresultscon-tradictthepredictionofasingleS-curve.Instead,technologicalevolutionseemstofollowastepfunction,withsharpimprovementsinperformancefollowinglongperiodsofnoimprovement.Moreover,pathsofrivaltechnolo-giesmaycrossmorethanonceornotatall.AshishSoodisanassistantprofessor,GoizuetaSchoolofBusiness,EmoryUniversity(e-mail:[e-mail]).GerardJ.TellisisNeelyChairofAmericanEnterprise,DirectoroftheCenterforGlobalInnovation,andProfessorofMarketing,MarshallSchoolofBusiness,Uni-versityofSouthernCalifornia(e-mail:[e-mail]).Thestudybene-fitedfromcommentsofparticipantsatthe2003AmericanMarketingAssociationWinterEducators’Conference,the2004MarketingScienceConference,the2004HotThoughtsonInnovationConference,the2004UtahConferenceonProductProcessInnovation,the2004WorkshoponMarketingStrategyatFreeUniversityAmsterdam,the2002MSITrusteesMeeting,andseminarsatCambridgeUniversity’sJudgeSchoolofMan-agementandtheWhartonSchoolofBusiness.ThestudywassupportedbyagrantfromtheMarketingScienceInstitute.Understandingtechnologicalinnovationisvitalformarketersforseveralreasons.Technologicalchangeisperhapsthemostpowerfulengineofgrowth.Itfuelsthegrowthofnewbrands(e.g.,Gillette’sMach3),createsnewgrowthmarkets(e.g.,digitalvideorecorders),andtransformssmalloutsiders(e.g.,Intel)intomarketlead-ers(ChandyandTellis1998;Christensen1997;Foster1986).Todate,thetopicoftechnologicalevolutionhasbeenstudiedprimarilyinthetechnologymanagementliter-ature.Acentralpremiseisthatperformanceofanewtech-nologystartsbelowthatofanexistingtechnology,crossestheperformanceoftheoldertechnologyonce,andendsatahigherplateau,thustracingasingleS-shapedcurve(seeFigure1).ThereisscatteredempiricalsupportforthepremiseandlimitedtheoreticalsupportforvariousaspectsoftheS-shapecurve(e.g.,Foster1986;Utterback1994a).Beliefinthispremiseissostrongthatithasbecomealmostalawinthestrategyliterature,fromwhichauthorshavederivedstrongmanagerialimplications.Forexample,theyhavewarnedthateventhoughmanagersmightbeabletosqueezeoutimprovementinperformancefromamaturetechnologyatthetopofitsScurve,improvementistypi-callycostly,shortlived,andsmall.Thus,aprimaryrecom-mendationinthestrategyliteratureandthetradepressisthatmanagersshouldabandonamaturingtechnologyandembraceanewonetostaycompetitive(e.g.,Christensen1997;Foster1986).Acentral,practicalproblemthatman-agersfaceiswhentoshiftinvestmentsfromthecurrenttothefuturetechnology.IftheScurveisindeedvalid,theappropriatetimewouldbetheinflectionpointoftheScurve.Afterthispoint,performanceimprovesatadecreas-ingrateuntilmaturity.NewproductdevelopmentandmajorinvestmentsinresearchdependonacorrectunderstandingoftechnologicalevolutioningeneralandoftheS-shapedcurveinparticular.Tofosterthisunderstanding,thisstudyaddressesthefol-lowingquestions:•Howdonewtechnologiesevolve?DotheyfollowtheS-shapedcurveorsomeotherpattern?Aretechnologicalchangespredictable?Istherateoftechnologicalchangeincreasing?•Howdorivaltechnologiescompete?Whataretheperfor-mancedimensionsofcompetition?Whatarethetransitionsbetweentechnologicalchanges?•Whichfirmscarryoutandsurvivetechnologicalevolution?Whointroducesradicalinnovations?Doincumbentssurvivethechange?Theprimaryfocusofthecurrentstudyisempirical.Wetesthypothesesderivedfromprevailingliteratureandexam-inetheevolutionof14technologiesinfourmarketsorindustries.Inthenextthreesections,wepresentthehypotheses,method,andresults.Inthefinalsection,wedis-cussthefindings,limitations,andimplicationsoftheresearch.HypothesesDevelopmentThefielddoesnotenjoyasingle,strong,andunifiedtheoryoftechnologicalevolution.Toguideourempiricalwork,wereviewedavailabletheoryfromtheliteratureandderivedtestablehypothesesaboutthepath,shape,source,andspeedoftechnologicalevolutionandthecompetitionamongrivaltechnologies.Findingsinthisareahavebeenpartlycon-foundedbytheuseofcirculardefinitions.Thus,webeginTechnologicalEvolutionandRadicalInnovation/153TimePerformanceInflection12TimeOldNewSinglecrossingPerformanceFIGURE1TechnologicalEvolutionA:TechnologicalSCurveB:MultipleSCurvesbydefiningtypesoftechnologicalinnovationsindependentoftheireffects.DefinitionsBeginningwithSchumpeter’s(1939)earlystudy,researchershaveusedawidevarietyoftermstodescribeinnovations.Manyterms,suchas“revolutionary,”“disrup-tive,