首页 Big Data-The Management Revolution 2012 October…

Big Data-The Management Revolution 2012 October.pdf

Big Data-The Management Revolut…

上传者: 用户2275419063 2013-08-22 评分5 评论0 下载9 收藏0 阅读量11 暂无简介 简介 举报

简介:本文档为《Big Data-The Management Revolution 2012 Octoberpdf》,可适用于战略管理领域,主题内容包含HBRORGOctOBeRreprinTrCSpotlightonBigDataBigData:TheManagementRevolutionExp符等。

HBRORGOctOBeRreprinTrCSpotlightonBigDataBigData:TheManagementRevolutionExploitingvastnewflowsofinformationcanradicallyimproveyourcompany’sperformanceButfirstyou’llhavetochangeyourdecisionmakingculturebyAndrewMcAfeeandErikBrynjolfssonartworktamarCohenHappyMotoring,,silkscreenonvintageroadmap,"x"SpotlightSpotlightonBigDataBigData:TheManagementRevolutionExploitingvastnewflowsofinformationcanradicallyimproveyourcompany’sperformanceButfirstyou’llhavetochangeyourdecisionmakingculturebyAndrewMcAfeeandErikBrynjolfssonoctober HarvardBusinessReview SpotlightonBigDatatheanalyticsthatwereusedinthepastWecanmeasureandthereforemanagemorepreciselythaneverbeforeWecanmakebetterpredictionsandsmarterdecisionsWecantargetmoreeffectiveinterventions,andcandosoinareasthatsofarhavebeendominatedbygutandintuitionratherthanbydataandrigorAsthetoolsandphilosophiesofbigdataspread,theywillchangelongstandingideasaboutthevalueofexperience,thenatureofexpertise,andthepracticeofmanagementSmartleadersacrossindustrieswillseeusingbigdataforwhatitis:amanagementrevolutionButaswithanyothermajorchangeinbusiness,thechallengesofbecomingabigdata–enabledorganizationcanbeenormousandrequirehandsonorinsomecaseshandsoffleadershipNevertheless,it’satransitionthatexecutivesneedtoengagewithtodaywhat’snewhereBusinessexecutivessometimesaskus,“Isn’t‘bigdata’justanotherwayofsaying‘analytics’”It’struethatthey’rerelated:Thebigdatamovement,likeanalyticsbeforeit,seekstogleanintelligencefromdataandtranslatethatintobusinessadvantageHowever,therearethreekeydifferences:VolumeAsof,aboutexabytesofdataarecreatedeachday,andthatnumberisdoublingeverymonthsorsoMoredatacrosstheinterneteverysecondthanwerestoredintheentireinternetjustyearsagoThisgivescompaniesanopportunitytoworkwithmanypetabyesofdatainasingledatasetandnotjustfromtheinternetForinstance,itisestimatedthatWalmartcollectsmorethanpetabytesofdataeveryhourfromitscustomertransactionsApetabyteisonequadrillionbytes,ortheequivalentofaboutmillionfilingcabinets’worthoftextAnexabyteis,timesthatamount,oronebilliongigabytes“Youcan’tmanagewhatyoudon’tmeasure”There’smuchwisdominthatsaying,whichhasbeenattributedtobothWEdwardsDemingandPeterDrucker,anditexplainswhytherecentexplosionofdigitaldataissoimportantSimplyput,becauseofbigdata,managerscanmeasure,andhenceknow,radicallymoreabouttheirbusinesses,anddirectlytranslatethatknowledgeintoimproveddecisionmakingandperformanceConsiderretailingBooksellersinphysicalstorescouldalwaystrackwhichbookssoldandwhichdidnotIftheyhadaloyaltyprogram,theycouldtiesomeofthosepurchasestoindividualcustomersAndthatwasaboutitOnceshoppingmovedonline,though,theunderstandingofcustomersincreaseddramaticallyOnlineretailerscouldtracknotonlywhatcustomersbought,butalsowhatelsetheylookedathowtheynavigatedthroughthesitehowmuchtheywereinfluencedbypromotions,reviews,andpagelayoutsandsimilaritiesacrossindividualsandgroupsBeforelong,theydevelopedalgorithmstopredictwhatbooksindividualcustomerswouldliketoreadnextalgorithmsthatperformedbettereverytimethecustomerrespondedtoorignoredarecommendationTraditionalretailerssimplycouldn’taccessthiskindofinformation,letaloneactonitinatimelymannerIt’snowonderthatAmazonhasputsomanybrickandmortarbookstoresoutofbusinessThefamiliarityoftheAmazonstoryalmostmasksitspowerWeexpectcompaniesthatwereborndigitaltoaccomplishthingsthatbusinessexecutivescouldonlydreamofagenerationagoButinfacttheuseofbigdatahasthepotentialtotransformtraditionalbusinessesaswellItmayofferthemevengreateropportunitiesforcompetitiveadvantage(onlinebusinesseshavealwaysknownthattheywerecompetingonhowwelltheyunderstoodtheirdata)Aswe’lldiscussinmoredetail,thebigdataofthisrevolutionisfarmorepowerfulthan HarvardBusinessReview octoberCopyRigHtHaRvaRDBusinesssCHoolpuBlisHingCoRpoRationallRigHtsReseRveDVelocityFormanyapplications,thespeedofdatacreationisevenmoreimportantthanthevolumeRealtimeornearlyrealtimeinformationmakesitpossibleforacompanytobemuchmoreagilethanitscompetitorsForinstance,ourcolleagueAlex“Sandy”PentlandandhisgroupattheMITMediaLabusedlocationdatafrommobilephonestoinferhowmanypeoplewereinMacy’sparkinglotsonBlackFridaythestartoftheChristmasshoppingseasonintheUnitedStatesThismadeitpossibletoestimatetheretailer’ssalesonthatcriticaldayevenbeforeMacy’sitselfhadrecordedthosesalesRapidinsightslikethatcanprovideanobviouscompetitiveadvantagetoWallStreetanalystsandMainStreetmanagersVarietyBigdatatakestheformofmessages,updates,andimagespostedtosocialnetworksreadingsfromsensorsGPSsignalsfromcellphones,andmoreManyofthemostimportantsourcesofbigdataarerelativelynewThehugeamountsofinformationfromsocialnetworks,forexample,areonlyasoldasthenetworksthemselvesFacebookwaslaunchedin,TwitterinThesameholdsforsmartphonesandtheothermobiledevicesthatnowprovideenormousstreamsofdatatiedtopeople,activities,andlocationsBecausethesedevicesareubiquitous,it’seasytoforgetthattheiPhonewasunveiledonlyfiveyearsago,andtheiPadinThusthestructureddatabasesthatstoredmostcorporateinformationuntilrecentlyareillsuitedtostoringandprocessingbigdataAtthesametime,thesteadilydecliningcostsofalltheelementsofcomputingstorage,memory,processing,bandwidth,andsoonmeanthatpreviouslyexpensivedataintensiveapproachesarequicklybecomingeconomicalAsmoreandmorebusinessactivityisdigitized,newsourcesofinformationandevercheaperequipmentcombinetobringusintoanewera:oneinwhichlargeamountsofdigitalinformationexistonvirtuallyanytopicofinteresttoabusinessMobilephones,onlineshopping,socialnetworks,electroniccommunication,GPS,andinstrumentedmachineryallproducetorrentsofdataasabyproductoftheirordinaryoperationsEachofusisnowawalkingdatageneratorThedataavailableareoftenunstructurednotorganizedinadatabaseandunwieldy,butthere’sahugeamountofsignalinthenoise,simplywaitingtobereleasedAnalyticsbroughtrigoroustechniquestodecisionmakingbigdataisatoncesimplerandmorepowerfulAsGoogle’sdirectorofresearch,PeterNorvig,putsit:“Wedon’thavebetteralgorithmsWejusthavemoredata”howDataDrivenCompaniesperformThesecondquestionskepticsmightposeisthis:“Where’stheevidencethatusingbigdataintelligentlywillimprovebusinessperformance”ThebusinesspressisrifewithanecdotesandcasestudiesthatsupposedlydemonstratethevalueofbeingdatadrivenButthetruth,werealizedrecently,isthatnobodywastacklingthatquestionrigorouslyToaddressthisembarrassinggap,weledateamattheMITCenterforDigitalBusiness,workinginpartnershipwithMcKinsey’sbusinesstechnologyofficeandwithourcolleagueLorinHittatWhartonandtheMITdoctoralstudentHeekyungKimWesetouttotestthehypothesisthatdatadrivencompanieswouldbebetterperformersWeconductedstructuredinterviewswithexecutivesatpublicNorthAmericancompaniesabouttheirorganizationalandtechnologymanagementpractices,andgatheredperformancedatafromtheirannualreportsandindependentsourcesNoteveryonewasembracingdatadrivendecisionmakingInfact,wefoundabroadspectrumofattitudesandapproachesineveryindustryButacrossalltheanalysesweconducted,onerelationideainBriefDatadrivendecisionsarebetterdecisionsit’sassimpleasthatusingbigdataenablesmanagerstodecideonthebasisofevidenceratherthanintuitionForthatreasonithasthepotentialtorevolutionizemanagementCompaniesthatwereborndigital,suchasgoogleandamazon,arealreadymastersofbigdataButthepotentialtogaincompetitiveadvantagefromitmaybeevengreaterforothercompaniesthemanagerialchallenges,however,areveryrealseniordecisionmakershavetoembraceevidencebaseddecisionmakingtheircompaniesneedtohirescientistswhocanfindpatternsindataandtranslatethemintousefulbusinessinformationandwholeorganizationsneedtoredefinetheirunderstandingof“judgment”pHotogRapHy:MaRkRanDalloctober HarvardBusinessReview ForarTiClereprinTsCallor,orvisiThBrorgSpotlightonBigDatashipstoodout:Themorecompaniescharacterizedthemselvesasdatadriven,thebettertheyperformedonobjectivemeasuresoffinancialandoperationalresultsInparticular,companiesinthetopthirdoftheirindustryintheuseofdatadrivendecisionmakingwere,onaverage,moreproductiveandmoreprofitablethantheircompetitorsThisperformancedifferenceremainedrobustafteraccountingforthecontributionsoflabor,capital,purchasedservices,andtraditionalITinvestmentItwasstatisticallysignificantandeconomicallyimportantandwasreflectedinmeasurableincreasesinstockmarketvaluationsSohowaremanagersusingbigdataLet’slookindetailattwocompaniesthatarefarfromSiliconValleyupstartsOneusesbigdatatocreatenewbusinesses,theothertodrivemoresalesimprovedairlineEtasMinutesmatterinairportsSodoesaccurateinformationaboutflightarrivaltimes:Ifaplanelandsbeforethegroundstaffisreadyforit,thepassengersandcrewareeffectivelytrapped,andifitshowsuplaterthanexpected,thestaffsitsidle,drivingupcostsSowhenamajorUSairlinelearnedfromaninternalstudythataboutoftheflightsintoitsmajorhubhadatleastaminutegapbetweentheestimatedtimeofarrivalandtheactualarrivaltimeandhadagapofatleastfiveminutesitdecidedtotakeactionAtthetime,theairlinewasrelyingontheaviationindustry’slongstandingpracticeofusingtheETAsprovidedbypilotsThepilotsmadetheseestimatesduringtheirfinalapproachtotheairport,whentheyhadmanyotherdemandsontheirtimeandattentionInsearchofabettersolution,theairlineturnedtoPASSURAerospace,aproviderofdecisionsupporttechnologiesfortheaviationindustryInPASSURbeganofferingitsownarrivalestimatesasaservicecalledRightETAItcalculatedthesetimesbycombiningpubliclyavailabledataaboutweather,flightschedules,andotherfactorswithproprietarydatathecompanyitselfcollected,includingfeedsfromanetworkofpassiveradarstationsithadinstallednearairportstogatherdataabouteveryplaneinthelocalskyPASSURstartedwithjustafewoftheseinstallations,butbyithadmorethanEverysecondsitcollectsawiderangeofinformationabouteveryplanethatit“sees”ThisyieldsahugeandconstantfloodofdigitaldataWhat’smore,thecompanykeepsallthedataithasgatheredovertime,soithasanimmensebodyofmultidimensionalinformationspanningmorethanadecadeThisallowssophisticatedanalysisandpatternmatchingRightETAessentiallyworksbyaskingitself“WhathappenedalltheprevioustimesaplaneapproachedthisairportundertheseconditionsWhendiditactuallyland”AfterswitchingtoRightETA,theairlinevirtuallyeliminatedgapsbetweenestimatedandactualarrivaltimesPASSURbelievesthatenablinganairlinetoknowwhenitsplanesaregoingtolandandplanaccordinglyisworthseveralmilliondollarsayearateachairportIt’sasimpleformula:Usingbigdataleadstobetterpredictions,andbetterpredictionsyieldbetterdecisionsSpeedier,MorepersonalizedpromotionsAcoupleofyearsago,SearsHoldingscametotheconclusionthatitneededtogenerategreatervaluefromthehugeamountsofcustomer,product,andpromotiondataitcollectedfromitsSears,Craftsman,andLands’EndbrandsObviously,itwouldbevaluabletocombineandmakeuseofallthesedatatotailorpromotionsandotherofferingstocustomers,andtopersonalizetheofferstotakeadvantageoflocalconditionsValuable,butdifficult:Searsrequiredabout HarvardBusinessReview octoberexpertisefromSurprisingSourcesoftensomeonecomingfromoutsideanindustrycanspotabetterwaytousebigdatathananinsider,justbecausesomanynew,unexpectedsourcesofdataareavailableoneofus,erik,demonstratedthisinresearchheconductedwithlynnWu,nowanassistantprofessoratWhartontheyusedpubliclyavailablewebsearchdatatopredicthousingpricechangesinmetropolitanareasacrosstheunitedstatestheyhadnospecialknowledgeofthehousingmarketwhentheybegantheirstudy,buttheyreasonedthatvirtuallyrealtimesearchdatawouldenablegoodneartermforecastsaboutthehousingmarketandtheywererightinfact,theirpredictionprovedmoreaccuratethantheofficialonefromthenationalassociationofRealtors,whichhaddevelopedafarmorecomplexmodelbutreliedonrelativelyslowchanginghistoricaldatathisishardlytheonlycaseinwhichsimplemodelsandbigdatatrumpmoreelaborateanalyticsapproachesResearchersattheJohnsHopkinsschoolofMedicine,forexample,foundthattheycouldusedatafromgoogleFlutrends(afree,publiclyavailableaggregatorofrelevantsearchterms)topredictsurgesinflurelatedemergencyroomvisitsaweekbeforewarningscamefromtheCentersforDiseaseControlsimilarly,twitterupdateswereasaccurateasofficialreportsattrackingthespreadofcholerainHaitiaftertheJanuaryearthquaketheywerealsotwoweeksearliereightweekstogeneratepersonalizedpromotions,atwhichpointmanyofthemwerenolongeroptimalforthecompanyIttooksolongmainlybecausethedatarequiredfortheselargescaleanalyseswerebothvoluminousandhighlyfragmentedhousedinmanydatabasesand“datawarehouses”maintainedbythevariousbrandsInsearchofafaster,cheaperwaytodoitsanalyticwork,SearsHoldingsturnedtothetechnologiesandpracticesofbigdataAsoneofitsfirststeps,itsetupaHadoopclusterThisissimplyagroupofinexpensivecommodityserverswhoseactivitiesarecoordinatedbyanemergingsoftwareframeworkcalledHadoop(namedafteratoyelephantinthehouseholdofDougCutting,oneofitsdevelopers)SearsstartedusingtheclustertostoreincomingdatafromallitsbrandsandtoholddatafromexistingdatawarehousesItthenconductedanalysesontheclusterdirectly,avoidingthetimeconsumingcomplexitiesofpullingdatafromvarioussourcesandcombiningthemsothattheycanbeanalyzedThischangeallowedthecompanytobemuchfasterandmoreprecisewithitspromotionsAccordingtothecompany’sCTO,PhilShelley,thetimeneededtogenerateacomprehensivesetofpromotionsdroppedfromeightweekstoone,andisstilldroppingAndthesepromotionsareofhigherquality,becausethey’remoretimely,moregranular,andmorepersonalizedSears’sHadoopclusterstoresandprocessesseveralpetabytesofdataatafractionofthecostofacomparablestandarddatawarehouseShelleysayshe’ssurprisedathoweasyithasbeentotransitionfromoldtonewapproachestodatamanagementandhighperformanceanalyticsBecauseskillsandknowledgerelatedtonewdatatechnologiesweresorarein,whenSearsstartedthetransition,itcontractedsomeoftheworktoacompanycalledClouderaButovertimeitsoldguardofITandanalyticsprofessionalshavebecomecomfortablewiththenewtoolsandapproachesThePASSURandSearsHoldingexamplesillustratethepowerofbigdata,whichallowsmoreaccuratepredictions,betterdecisions,andpreciseinterventions,andcanenablethesethingsatseeminglylimitlessscaleWe’veseenbigdatausedinsupplychainmanagementtounderstandwhyacarmaker’sdefectratesinthefieldsuddenlyincreased,incustomerservicetocontinuallyscanandinterveneinthehealthcarepracticesofmillionsofpeople,inplanningandforecastingtobetteranticipateonlinesalesonthebasisofadatasetofproductcharacteristics,andsoonWe’veseensimilarpayoffsinmanyotherindustriesandfunctions,fromfinancetomarketingtohotelsandgaming,andfromhumanresourcemanagementtomachinerepairOurstatisticalanalysistellsusthatwhatwe’reseeingisnotjustafewflashyexamplesbutamorefundamentaltransformationoftheeconomyWe’vebecomeconvincedthatalmostnosphereofbusinessactivitywillremainuntouchedbythismovementanewCultureofDecisionMakingThetechnicalchallengesofusingbigdataareveryrealButthemanagerialchallengesareevengreaterstartingwiththeroleoftheseniorexecutiveteamMutingtheHiPPOsOneofthemostcriticalaspectsofbigdataisitsimpactonhowdecisionsaremadeandwhogetstomakethemWhendataarescarce,expensivetoobtain,ornotavailableindigitalform,itmakessensetoletwellplacedpeoplemakedecisions,whichtheydoonthebasisofexperiencethey’vebuiltupandpatternsandrelationshipsthey’veobservedandinternalized“Intuition”isthelabelgiventothisstyleofinferenceanddecisionmakingPeoplestatetheiropinionsaboutwhatthefutureholdswhat’sgoingtohappen,howwellsomethingwillwork,andsoonandthenplanaccordingly(See“TheTrueMeasuresofSuccess,”byMichaelJMauboussin,inthisissue)Forparticularlyimportantdecisions,thesepeoplearetypicallyhighupintheorganization,orthey’reexpensiveoutsidersbroughtinbecauseoftheirexpertiseandtrackrecordsManyinthebigdatacommunitymaintainthatcompaniesoftenmakemostoftheirimportantdecisionsbyrelyingon“HiPPO”thehighestpaidperson’sopinionTobesure,anumberofseniorexecutivesaregenuinelydatadrivenandwillingtooverridetheirownintuitionwhenthedatadon’tagreewithitButwebelievethatthroughoutthebusinessworldtoday,october HarvardBusinessReview Bigdata’spowerdoesnoterasetheneedforvisionorhumaninsightForarTiClereprinTsCallor,orvisiThBrorgSpotlightonBigDatapeoplerelytoomuchonexperienceandintuitionandnotenoughondataForourresearchweconstructedapointcompositescalethatcapturedtheoverallextenttowhichacompanywasdatadrivenFullyofourrespondentsratedtheircompaniesatorbelowonthisscaleNewrolesExecutivesinterestedinleadingabigdatatransitioncanstartwithtwosimpletechniquesFirst,theycangetinthehabitof

职业精品

(汽车)产品营销策划书范文.doc

HH牙膏营销方案策划书.doc

加班管理人力资源考勤管理系统方案.doc

物品采购管理制度-正式.doc

用户评论

0/200
    暂无评论
上传我的资料

精彩专题

相关资料换一换

  • Big Data-The Man…

  • MacUser_-_12_Oct…

  • Dynatom October-…

  • Weekly Pboy - 22…

  • Mag_EDN2012Octob…

  • Big Data_The Man…

  • Big Data_The Man…

  • The October Revo…

资料评价:

/ 0
所需积分:2 立即下载

意见
反馈

返回
顶部