Tải bản đầy đủ
Chapter 2. Business Motivations and Drivers for Big Data Adoption

Chapter 2. Business Motivations and Drivers for Big Data Adoption

Tải bản đầy đủ

learningthatthisisnotsufficientinordertoexecutetheirbusinessmodelsina
marketplacethatmoreresemblesanecologicalsystem.Assuch,organizationsneedto
consumedatafromtheoutsidetosensedirectlythefactorsthatinfluencetheir
profitability.Theuseofsuchexternaldatamostoftenresultsin“BigData”datasets.
ThischapterexploresthebusinessmotivationsanddriversbehindtheadoptionofBig
Datasolutionsandtechnologies.TheadoptionofBigDatarepresentstheconfluenceof
severalforcestoinclude:marketplacedynamics,anappreciationandformalismof
BusinessArchitecture(BA),therealizationthatabusiness’abilitytodelivervalueis
directlytiedtoBusinessProcessManagement(BPM),innovationinInformationand
CommunicationsTechnology(ICT)andfinallytheInternetofEverything(IoE).Eachof
thesetopicswillbeexploredinturn.

MarketplaceDynamics
Therehasbeenafundamentalshiftinthewaybusinessesviewthemselvesandthe
marketplace.Inthepast15years,twolargestockmarketcorrectionshavetakenplace—
thefirstwasthedot-combubbleburstin2000,andthesecondwastheglobalrecession
thatbeganin2008.Ineachcase,businessesentrenchedandworkedtoimprovetheir
efficiencyandeffectivenesstostabilizetheirprofitabilitybyreducingcosts.Thisofcourse
isnormal.Whencustomersarescarce,cost-cuttingoftenensuestomaintainthecorporate
bottomline.Inthisenvironment,companiesconducttransformationprojectstoimprove
theircorporateprocessestoachievesavings.
DavenportandPrusakhaveprovidedgenerally-acceptedworkingdefinitions
ofdata,informationandknowledgeintheirbookWorkingKnowledge.
AccordingtoDavenportandPrusak,“[d]ataisasetofdiscrete,objective
factsaboutevents.”Inabusinesssense,theseeventsareactivitiesthatoccur
withinanorganization’sbusinessprocessesandinformationsystems—they
representthegeneration,modificationandcompletionofworkassociated
withbusinessentities;forexample,orders,shipments,notificationsand
customeraddressupdates.Theseeventsareareflectionofreal-worldactivity
thatisrepresentedwithintherelationaldatastoresofcorporateinformation
systems.DavenportandPrusakfurtherdefineinformationas“datathatmakes
adifference.”Itisdatathathasbeencontextualizedtoprovide
communication;itdeliversamessageandinformsthereceiver—whetheritbe
ahumanorsystem.Informationisthenenrichedviaexperienceandinsightin
thegenerationofknowledge.Theauthorsstatethat“[k]nowledgeisafluid
mixofframedexperience,values,contextualinformationandexpertinsight
thatprovidesaframeworkforevaluatingandincorporatingnewexperiences
andinformation.”
Astheglobaleconomiesbegantoemergefromrecession,companiesbegantofocus
outward,lookingtofindnewcustomersandkeepexistingcustomersfromdefectingto
marketplacecompetitors.Thiswasaccomplishedbyofferingnewproductsandservices
anddeliveringincreasedvaluepropositionstocustomers.Itisaverydifferentmarket
cycletotheonethatfocusesoncost-cutting,foritisnotabouttransformationbutinstead

innovation.Innovationbringshopetoacompanythatitwillfindnewwaystoachievea
competitiveadvantageinthemarketplaceandaconsequentincreaseintoplinerevenue.
Theglobaleconomycanexperienceperiodsofuncertaintyduetovariousfactors.We
generallyacceptthattheeconomiesofthemajordevelopedcountriesintheworldarenow
inextricablyintertwined;inotherwords,theyformasystemofsystems.Likewise,the
world’sbusinessesareshiftingtheirperspectiveabouttheiridentityandindependenceas
theyrecognizethattheyarealsointertwinedinintricateproductandservicenetworks.
Forthisreason,companiesneedtoexpandtheirBusinessIntelligenceactivitiesbeyond
retrospectivereflectiononinternalinformationextractedfromtheircorporateinformation
systems.Theyneedtoopenthemselvestoexternaldatasourcesasameansofsensingthe
marketplaceandtheirpositionwithinit.Recognizingthatexternaldatabringsadditional
contexttotheirinternaldataallowsacorporationtomoveuptheanalyticvaluechainfrom
hindsighttoinsightwithgreaterease.Withappropriatetooling,whichoftensupports
sophisticatedsimulationcapabilities,acompanycandevelopanalyticresultsthatprovide
foresight.Inthiscase,thetoolingassistsinbridgingthegapbetweenknowledgeand
wisdomaswellasprovidesadvisoryanalyticresults.ThisisthepowerofBigData—
enrichingcorporateperspectivebeyondintrospection,fromwhichabusinesscanonly
inferinformationaboutmarketplacesentiment,tosensingthemarketplaceitself.
ThetransitionfromhindsighttoforesightcanbeunderstoodthroughthelensoftheDIKW
pyramiddepictedinFigure2.1.Notethatinthisfigure,atthetopofthetriangle,wisdom
isshownasanoutlinetoindicatethatitexistsbutisnottypicallygeneratedviaICT
systems.Instead,knowledgeworkersprovidetheinsightandexperiencetoframethe
availableknowledgesothatitcanbeintegratedtoformwisdom.Wisdomgenerationby
technologicalmeansquicklydevolvesintoaphilosophicaldiscussionthatisnotwithinthe
scopeofthisbook.Withinbusinessenvironments,technologyisusedtosupport
knowledgemanagement,andpersonnelareresponsibleforapplyingtheircompetencyand
wisdomtoactaccordingly.

Figure2.1TheDIKWpyramidshowshowdatacanbeenrichedwithcontexttocreate
information,informationcanbesuppliedwithmeaningtocreateknowledgeand
knowledgecanbeintegratedtoformwisdom.

BusinessArchitecture
Withinthepastdecade,therehasbeenarealizationthattoooftenacorporation’s
enterprisearchitectureissimplyamyopicviewofitstechnologyarchitecture.Inaneffort
towrestpowerfromthestrongholdofIT,businessarchitecturehasemergedasa
complementarydiscipline.Inthefuture,thegoalisthatenterprisearchitecturewillpresent
abalancedviewbetweenbusinessandtechnologyarchitectures.Businessarchitecture
providesameansofblueprintingorconcretelyexpressingthedesignofthebusiness.A
businessarchitecturehelpsanorganizationalignitsstrategicvisionwithitsunderlying
execution,whethertheybetechnicalresourcesorhumancapital.Thus,abusiness
architectureincludeslinkagesfromabstractconceptslikebusinessmission,vision,
strategyandgoalstomoreconcreteoneslikebusinessservices,organizationalstructure,
keyperformanceindicatorsandapplicationservices.
Theselinkagesareimportantbecausetheyprovideguidanceastohowtoalignthe
businessanditsinformationtechnology.Itisanacceptedviewthatabusinessoperatesas
alayeredsystem—thetoplayeristhestrategiclayeroccupiedbyC-levelexecutivesand
advisorygroups;themiddlelayeristhetacticalormanageriallayerthatseekstosteerthe
organizationinalignmentwiththestrategy;andthebottomlayeristheoperationslayer
whereabusinessexecutesitscoreprocessesanddeliversvaluetoitscustomers.These
threelayersoftenexhibitadegreeofindependencefromoneanother,buteachlayer’s
goalsandobjectivesareinfluencedbyandoftendefinedbythelayerabove,inother
wordstop-down.Fromamonitoringperspective,communicationflowsupstream,or

bottom-upviathecollectionofmetrics.Businessactivitymonitoringattheoperations
layergeneratesPerformanceIndicators(PIs)andmetrics,forbothservicesandprocesses.
TheyareaggregatedtocreateKeyPerformanceIndicators(KPIs)usedatthetactical
layer.TheseKPIscanbealignedwithCriticalSuccessFactors(CSFs)atthestrategic
layer,whichinturnhelpmeasureprogressbeingmadetowardtheachievementof
strategicgoalsandobjectives.
BigDatahastiestobusinessarchitectureateachoftheorganizationallayers,asdepicted
inFigure2.2.BigDataenhancesvalueasitprovidesadditionalcontextthroughthe
integrationofexternalperspectivestohelpconvertdataintoinformationandprovide
meaningtogenerateknowledgefrominformation.Forinstance,attheoperationallevel,
metricsaregeneratedthatsimplyreportonwhatishappeninginthebusiness.Inessence,
weareconvertingdatathroughbusinessconceptsandcontexttogenerateinformation.At
themanageriallevel,thisinformationcanbeexaminedthroughthelensofcorporate
performancetoanswerquestionsregardinghowthebusinessisperforming.Inother
words,givemeaningtotheinformation.Thisinformationmaybefurtherenrichedto
answerquestionsregardingwhythebusinessisperformingatthelevelitis.Whenarmed
withthisknowledge,thestrategiclayercanprovidefurtherinsighttohelpanswer
questionsofwhichstrategyneedstochangeorbeadoptedinordertocorrectorenhance
theperformance.

Figure2.2TheDIKWpyramidillustratesalignmentwithStrategic,Tacticaland
Operationalcorporatelevels.
Aswithanylayeredsystem,thelayersdonotallchangeatthesamespeed.Inthecaseofa
businessenterprise,thestrategiclayeristheslowestmovinglayer,andtheoperational
layeristhefastestmovinglayer.Theslowermovinglayersprovidestabilityanddirection

tothefastermovinglayers.Intraditionalorganizationalhierarchies,themanagementlayer
isresponsiblefordirectingtheoperationallayerinalignmentwiththestrategycreatedby
theexecutiveteam.Becauseofthisvariationinregardtospeedofchange,itispossibleto
envisionthethreelayersasbeingresponsibleforstrategyexecution,businessexecution
andprocessexecutionrespectively.Eachoftheselayersreliesupondifferentmetricsand
measures,presentedthroughdifferentvisualizationandreportingfunctions.Forexample,
thestrategylayermayrelyuponbalancedscorecards,themanagementlayeruponan
interactivevisualizationofKPIsandcorporateperformanceandtheoperationallayeron
visualizationsofexecutingbusinessprocessesandtheirstatuses.
Figure2.3,avariantofadiagramproducedbyJoeGollnerinhisblogpost“TheAnatomy
ofKnowledge,”showshowanorganizationcanrelateandalignitsorganizationallayers
bycreatingavirtuouscycleviaafeedbackloop.Ontherightsideofthefigure,the
strategiclayerdrivesresponseviatheapplicationofjudgmentbymakingdecisions
regardingcorporatestrategy,policy,goalsandobjectivesthatarecommunicatedas
constraintstothetacticallayer.Thetacticallayerinturnleveragesthisknowledgeto
generateprioritiesandactionsthatconformtocorporatedirection.Theseactionsadjustthe
executionofbusinessattheoperationallayer.Thisinturnshouldgeneratemeasureable
changeintheexperienceofinternalstakeholdersandexternalcustomersastheydeliver
andconsumebusinessservices.Thischange,orresult,shouldsurfaceandbevisibleinthe
dataintheformofchangedPIsthatarethenaggregatedintoKPIs.RecallthatKPIsare
metricsthatcanbeassociatedwithcriticalsuccessfactorsthatinformtheexecutiveteam
astowhetherornottheirstrategiesareworking.Overtime,thestrategicandmanagement
layersinjectionofjudgmentandactionintotheloopwillservetorefinethedeliveryof
businessservices.

Figure2.3Thecreationofavirtuouscycletoalignanorganizationacrosslayersviaa
feedbackloop.

BusinessProcessManagement
Businessesdelivervaluetocustomersandotherstakeholdersviatheexecutionoftheir
businessprocesses.Abusinessprocessisadescriptionofhowworkisperformedinan
organization.Itdescribesallwork-relatedactivitiesandtheirrelationships,alignedwith
theorganizationalactorsandresourcesresponsibleforconductingthem.Therelationships
betweenactivitiesmaybetemporal;forexample,activityAisexecutedbeforeactivityB.
Therelationshipscanalsodescribewhethertheexecutionofactivitiesisconditional,
basedupontheoutputsorconditionsgeneratedbyotheractivitiesorbysensingevents
generatedoutsideofthebusinessprocessitself.
Businessprocessmanagementappliesprocessexcellencetechniquestoimprovecorporate
execution.BusinessProcessManagementSystems(BPMS)providesoftwaredevelopersa
modeldrivenplatformthatisbecomingtheBusinessApplicationDevelopment
Environment(BADE)ofchoice.Abusinessapplicationneedsto:mediatebetween
humansandothertechnology-hostedresources,executeinalignmentwithcorporate
policiesandensurethefairdistributionofworktoemployees.AsaBADE,modelsofa
businessprocessarejoinedwith:modelsoforganizationalrolesandstructure,business
entitiesandtheirrelationships,businessrulesandtheuser-interface.Thedevelopment
environmentintegratesthesemodelstogethertocreateabusinessapplicationthatmanages
screenflowandworkflowandprovidesworkloadmanagement.Thisisaccomplishedinan
executionenvironmentthatenforcescorporatepolicyandsecurityandprovidesstate
managementforlong-runningbusinessprocesses.Thestateofanindividualprocess,orall
processes,canbeinterrogatedviaBusinessActivityMonitoring(BAM)andvisualized.
WhenBPMiscombinedwithBPMSsthatareintelligent,processescanbeexecutedina
goal-drivenmanner.Goalsareconnectedtoprocessfragmentsthataredynamically
chosenandassembledatrun-timeinalignmentwiththeevaluationofthegoals.Whenthe
combinationofBigDataanalyticresultsandgoal-drivenbehaviorareusedtogether,
processexecutioncanbecomeadaptivetothemarketplaceandresponsiveto
environmentalconditions.Asasimpleexample,acustomercontactprocesshasprocess
fragmentsthatenablecommunicationwithcustomersviaavoicecall,email,textmessage
andtraditionalpostalmail.Inthebeginning,thechoiceofthesecontactmethodsis
unweighted,andtheyarechosenatrandom.However,behind-the-scenesanalysisisbeing
donetomeasuretheeffectivenessofthecontactmethodviastatisticalanalysisof
customerresponsiveness.
Theresultsofthisanalysisaretiedtoagoalresponsibleforselectingthecontactmethod,
andwhenaclearpreferenceisdetermined,theweightingischangedtofavorthecontact
methodthatachievesthebestresponse.Amoredetailedanalysiscouldleveragecustomer
clustering,whichwouldassignindividualcustomerstogroupswhereoneofthecluster
dimensionsisthecontactmethod.Inthiscase,customerscanbecontactedwitheven
greaterrefinement,whichprovidesapathwaytoone-to-onetargetedmarketing.

InformationandCommunicationsTechnology
ThissectionexaminesthefollowingICTdevelopmentsthathaveacceleratedthepaceof
BigDataadoptioninbusinesses:

•dataanalyticsanddatascience
•digitization
•affordabletechnologyandcommodityhardware
•socialmedia
•hyper-connectedcommunitiesanddevices
•cloudcomputing

DataAnalyticsandDataScience
Enterprisesarecollecting,procuring,storing,curatingandprocessingincreasingquantities
ofdata.Thisisoccurringinanefforttofindnewinsightsthatcandrivemoreefficientand
effectiveoperations,providemanagementtheabilitytosteerthebusinessproactivelyand
allowtheC-suitetobetterformulateandassesstheirstrategicinitiatives.Ultimately,
enterprisesarelookingfornewwaystogainacompetitiveedge.Thustheneedfor
techniquesandtechnologiesthatcanextractmeaningfulinformationandinsightshas
increased.Computationalapproaches,statisticaltechniquesanddatawarehousinghave
advancedtothepointwheretheyhavemerged,eachbringingtheirspecifictechniquesand
toolsthatallowtheperformanceofBigDataanalysis.Thematurityofthesefieldsof
practiceinspiredandenabledmuchofthecorefunctionalityexpectedfromcontemporary
BigDatasolutions,environmentsandplatforms.

Digitization
Formanybusinesses,digitalmediumshavereplacedphysicalmediumsasthedefacto
communicationsanddeliverymechanism.Theuseofdigitalartifactssavesbothtimeand
costasdistributionissupportedbythevastpre-existinginfrastructureoftheInternet.As
consumersconnecttoabusinessthroughtheirinteractionwiththesedigitalsubstitutes,it
leadstoanopportunitytocollectfurther“secondary”data;forexample,requestinga
customertoprovidefeedback,completeasurvey,orsimplyprovidingahooktodisplaya
relevantadvertisementandtrackingitsclick-throughrate.Collectingsecondarydatacan
beimportantforbusinessesbecauseminingthisdatacanallowforcustomizedmarketing,
automatedrecommendationsandthedevelopmentofoptimizedproductfeatures.Figure
2.4providesavisualrepresentationofexamplesofdigitization.

Figure2.4Examplesofdigitizationincludeonlinebanking,on-demandtelevisionand
streamingvideo.

AffordableTechnologyandCommodityHardware
Technologycapableofstoringandprocessinglargequantitiesofdiversedatahasbecome
increasinglyaffordable.Additionally,BigDatasolutionsoftenleverageopen-source
softwarethatexecutesoncommodityhardware,furtherreducingcosts.Thecombination
ofcommodityhardwareandopensourcesoftwarehasvirtuallyeliminatedtheadvantage
thatlargeenterprisesusedtoholdbybeingabletooutspendtheirsmallercompetitorsdue
tothelargersizeoftheirITbudgets.Technologynolongerdeliverscompetitive
advantage.Instead,itsimplybecomestheplatformuponwhichthebusinessexecutes.
Fromabusinessstandpoint,utilizationofaffordabletechnologyandcommodityhardware
togenerateanalyticresultsthatcanfurtheroptimizetheexecutionofitsbusiness
processesisthepathtocompetitiveadvantage.
TheuseofcommodityhardwaremakestheadoptionofBigDatasolutionsaccessibleto
businesseswithoutlargecapitalinvestments.Figure2.5providesanexampleoftheprice
declineassociatedwithdatastoragepricesoverthepast20years.

Figure2.5Datastoragepriceshavedroppeddramaticallyfrommorethan$10,000to
lessthan$0.10perGBoverthedecades.

SocialMedia
Theemergenceofsocialmediahasempoweredcustomerstoprovidefeedbackinnearrealtimeviaopenandpublicmediums.Thisshifthasforcedbusinessestoconsider
customerfeedbackontheirserviceandproductofferingsintheirstrategicplanning.Asa
result,businessesarestoringincreasingamountsofdataoncustomerinteractionswithin
theircustomerrelationshipmanagementsystems(CRM)andfromharvestingcustomer
reviews,complaintsandpraisefromsocialmediasites.ThisinformationfeedsBigData
analysisalgorithmsthatsurfacethevoiceofthecustomerinanattempttoprovidebetter
levelsofservice,increasesales,enabletargetedmarketingandevencreatenewproducts
andservices.Businesseshaverealizedthatbrandingactivityisnolongercompletely
managedbyinternalmarketingactivities.Instead,productbrandsandcorporatereputation
areco-createdbythecompanyanditscustomers.Forthisreason,businessesare
increasinglyinterestedinincorporatingpubliclyavailabledatasetsfromsocialmediaand
otherexternaldatasources.

Hyper-ConnectedCommunitiesandDevices
ThebroadeningcoverageoftheInternetandtheproliferationofcellularandWi-Fi
networkshasenabledmorepeopleandtheirdevicestobecontinuouslyactiveinvirtual
communities.CoupledwiththeproliferationofInternetconnectedsensors,the
underpinningsoftheInternetofThings(IoT),avastcollectionofsmartInternetconnecteddevices,isbeingformed.AsshowninFigure2.6,thisinturnhasresultedina
massiveincreaseinthenumberofavailabledatastreams.Whilesomestreamsarepublic,
otherstreamsarechanneleddirectlytocorporationsforanalysis.Asanexample,the
performance-basedmanagementcontractsassociatedwithheavyequipmentusedinthe
miningindustryincentivizetheoptimalperformanceofpreventiveandpredictive
maintenanceinanefforttoreducetheneedandavoidthedowntimeassociatedwith
unplannedcorrectivemaintenance.Thisrequiresdetailedanalysisofsensorreadings
emittedbytheequipmentfortheearlydetectionofissuesthatcanberesolvedviathe
proactiveschedulingofmaintenanceactivities.

Figure2.6Hyper-connectedcommunitiesanddevicesincludetelevision,mobile
computing,RFIDs,refrigerators,GPSdevices,mobiledevicesandsmartmeters.

CloudComputing
Cloudcomputingadvancementshaveledtothecreationofenvironmentsthatarecapable
ofprovidinghighlyscalable,on-demandITresourcesthatcanbeleasedviapay-as-you-go
models.Businesseshavetheopportunitytoleveragetheinfrastructure,storageand
processingcapabilitiesprovidedbytheseenvironmentsinordertobuild-outscalableBig
Datasolutionsthatcancarryoutlarge-scaleprocessingtasks.Althoughtraditionally
thoughtofasoff-premiseenvironmentstypicallydepictedwithacloudsymbol,businesses
arealsoleveragingcloudmanagementsoftwaretocreateonpremisecloudstomore
effectivelyutilizetheirexistinginfrastructureviavirtualization.Ineithercase,theability
ofacloudtodynamicallyscalebaseduponloadallowsforthecreationofresilient
analyticenvironmentsthatmaximizeefficientutilizationofICTresources.
Figure2.7displaysanexampleofhowacloudenvironmentcanbeleveragedforits

scalingcapabilitiestoperformBigDataprocessingtasks.Thefactthatoff-premisecloudbasedITresourcescanbeleaseddramaticallyreducestherequiredup-frontinvestmentof
BigDataprojects.

Figure2.7Thecloudcanbeusedtocompleteon-demanddataanalysisattheendof
eachmonthorenablethescalingoutofsystemswithanincreaseinload.
Itmakessenseforenterprisesalreadyusingcloudcomputingtoreusethecloudfortheir
BigDatainitiativesbecause:
•personnelalreadypossessestherequiredcloudcomputingskills
•theinputdataalreadyexistsinthecloud
Migratingtothecloudislogicalforenterprisesplanningtorunanalyticsondatasetsthat
areavailableviadatamarkets,asmanydatamarketsmaketheirdatasetsavailableina
cloudenvironment,suchasAmazonS3.
Inshort,cloudcomputingcanprovidethreeessentialingredientsrequiredforaBigData
solution:externaldatasets,scalableprocessingcapabilitiesandvastamountsofstorage.