[PDF] The Netflix Experience 20 sept. 2016 Asia is

2 ForMamá&Papá,Forencouragingmetobethecreatorofmyownlife.

3 Abstract.Thisprojectproposesauser-focusedapproachtostudythealgorithmlogicofon-demandapps,usingNetflixasacasestudy.ThemainresearchinterestistheperceptionthattheuserhasaboutthesuggestionandrecommendationlogicofNetflix.Inordertogathertheinformation,awalkthroughmethodonNetflixwasappliedaswellaspersonal,in-depththinkaloudinterviewswerecarriedout.Thesampleconsistedonasel ecti onofheavyusers,millennialsex-patslivinginSingaporeandworkinginthe creativein dustrytogetspecificinsightsontheirrelationshipwiththealgorithm.Toanalyzethegatheredmaterial,qualitativecontentanalysiswascarriedout.Thiskindofstudyisimportantwithintoday'scontemporarymediaenvironmenttohaveintegralapproachtouser sperceptionsinsteado fjustanalyticalfiguresandnumbers.Thetheoreticalcontextusedtoenlightsomeoftheconclusionsdiscussedonthisresearchwerebasedonthestudyofmediaineverydaylife,globalculturalindustrystudies,aswellas algorithmcultureandthesc ienceandtechnologystudies.Howalgorithmsareperceivedhavemajorrepercussionsnotonlyonon-demandapps,technologybusinessmodelsorentertainmentindustrybutalsoanintenseinfluenceonthewaypeopleconsumecontent.Re-thinkingtheuserasaco-producerofinformationandknowledge,consideringsomeoftheimplicationsthisphenomenonmighthaveonthecreativeindustryandhowthataffectsonourdailylifearesomeoftheissuesthisresearchelaboratedon.Itcanbesaidthattheselectedsampleappreciatesthesuggestionlogicsandithasmultiplefunctionalities:recommendation,curation,entertainment,companionshipandleisur e.NetflixOriginalsarevery wellvalidated;beingoneofthe mainattractionsoftheapp.Interface,functionalityandfeaturesarealsoitemsthatthesamplepositivelyhi ghlights.Theaccur acyperceptionofthealgorithmisgood, althoughlowwhencomparedtoothercountrieswherethesampleusedtheapp.Thesameappliestotheamountofcontentandtitlesavailable,beingtheselasttwo,issuesthatNetflixcouldimprove.Thisresearchwasconductedfor8months,fromOctober2016toMay2017,forSodertornUniversity-Stockholm,Sweden,with theguidancean dsuppor tofAssociateProfessorAnneKaun.Keywords:Netflix,On -Demand,Apps,Algorithms,ANT,GlobalCulturalIndu stry,MediaandEver ydayLife, UserExperience,Content,Entertainment, User,WalkthroughMethod,ThinkAloudMethod,ContentAnalysis.

4 TABLEOFCONTENTS:1.Introduction.................................................................................................................62.Theoreticalcontext,objective&researchquestions..............................................72.1PreviousResearch.............................................................................................72.2ObjectiveLimitationsoftheStudy.................................................................92.3ResearchMainandSupportQuestions.........................................................92.4LimitationsoftheStudy..................................................................................102.5.TheoreticalContext.........................................................................................112.5.1CurrentMediaLandscape:mediaineverydaylife&globalculturalindustry.............................................................................112.5.2AlgorithmCulture&ScienceandTechnologyStudies....................143.MaterialandMethods...............................................................................................203.1FieldofStudy:Netflix-Towardsadefinitionanddescription...................203.2SampleDefinition............................................................................................243.3Whattowatchnext:amethodologicalbriefintroduction........................263.4Theresearchprocess:gatheringinformation..............................................263.4.1Thinkaloudinterviews...............................................................263.4.2Thewalkthroughmethod.........................................................273.5Theresearchprocess:analyzinginformation...............................................303.5.1.Qualitativecontentanalysis.....................................................303.5.2.Categories...................................................................................313.6Limitationsofthemethods..............................................................................324.AnalysisandResults..................................................................................................334.1OfflineBehavior-Netflixexperience:where,whenwithwhom..............334.1.1Time.............................................................................................344.1.2Companion..................................................................................354.1.3Location......................................................................................364.2OnlineBehavior-Netflix'sUXandinterfaceanalysis...............................384.2.1Devices.......................................................................................,444.2.2Accounts&Profiles...................................................................464.2.3Recommendations....................................................................494.3PressPlay-Algorithmperception&userexperiences'highlights...........544.3.1Functionality...............................................................................544.3.2AlgorithmPerception................................................................584.3.3Content........................................................................................605.Conclusion&FinalDiscussion..................................................................................635.1Users'Perception..............................................................................................645.2Co-CreatorsofContent.....................................................................................665.3ReshapingtheCreativeProcess:CulturalCo-ProductionofKnowledge...676.Bibliography&ElectronicSources..........................................................................70

5 7.Appendix..................................................................................................................737.1Interviewsquestionnaire.............................................................................737.2Intervieweesprofiles....................................................................................757.3Interviewstranscriptions.............................................................................777.4Photographicsupportivematerial............................................................1287.5Tableofresults...........................................................................................1317.6Netflixtermsandconditions.....................................................................135

6 1.IntroductionIfonewatchesNetflix'sacquiredseriesBlackMirroronecancertainlyhaveaglimpseoftoday'smedialandscapeandpeople'sinteraction-andwhynot,reaction-withmediatechnology.Frommediagadgetsandw earablestomediaitself,BlackMirroronlydramatizeswhathasbeenhappeningduringthepast15years,whenprivateandpublicsphereboundarieshavebeenblurredduetotheusageofmediaasanenhancerordisruptivefactor,makingchangesinpeople´sbehavioranddecisions.BlackMirrorisaBritishsciencefiction TVseriesthatexaminesmo dernsociety,itsinter actionsand unexpectedconsequenceswithnewtechnologies.Hisveryowncreator,CharlieBrooker,explaineditduringitsfirstpremierontheoriginalChannel4Networkinterview,beforebeingboughtbyNetflix:"Overthelasttenyears,technologyhastransformedalmosteveryaspectofourlivesbeforewe'vehadtimetostopandquestionit.Ineveryhome,oneverydeskineverypalm-aplasmascreen,amonitor,asmartphone-ablackmirrorofour21stcenturyexistence.Ou rgriponrealityisshifting.W eworshipatthealtarsofGoogleandApple.Facebookalgorithmsknowusmoreintimatelythanourownparents.Wehaveaccesstoalltheinformationintheworld,butnobrainspacelefttoabsorbanythinglongerthana140-charactertweet.BlackMirrortapsintothecollectiveuneaseaboutourmodernworld1"(Channel4Editorial,2011).InBlackMirror,users'choicesaredramatizedbychips,artificialintelligenceandallsortsofdevices.Buttoday,theinvisiblethreatthatcontributestoandallowsthesepossibilities-thatatafirstglanceseemendless-aremainlyguidedbutcertainlynotultimatedefinedbythemagicofalgorithms.Asideoftheinherentattractivenessofitsstoryline;itishighlyprobablethatifoneisreadingthisandhasaNetflixaccount,oneprobablyknowsaboutBlackMirror.Beinganoriginal,itisoneofitsmorepushedseriesso,notonlyitwillbethereonthefirstsuggestedchoicestowhattowatchnext,butalsowillhaveavideotrailerplayingassoonasyouopentheapp.ItisindeedaninterestingparadoxthatNetflix,oneofthoseinternationalcompaniesthatcurrentlyisshapingnotonlythecontentbutmainlythewaypeopleconsumeentertainmentworldwide,isactuallyproducinganddistributingaseriesthatisadramaticcritiqueofthatalgorithmicpersonalizationprocess. 1 Charlie Brooker talks about Black Mirror, Channel4, (http://www.channel4.com/info/press/news/charlie-brooker-talks-about-black-mirror) 20/09/2011 searched on 14/01/2017

7 Thisprojecthas,asBlackMirror,alsoconsideredthenatureoftoday'smedialandscapeasafieldofstudy,particularlyNetflix,itsalgorithmicsystemanditsondemandlogic.Howisthealgorithm-andthepersonalizationpossibilitiesitallows-beingperceivedbyusers?Inotherwords,howdoestheNetflix'suserexperiencethesuggestionlogicofNetflix?Thesekindsofresearchquestionshavemajorrelevancefortoday'smedialandscape.Consideringauser-focusedapproach,thisresearchreformulatesitsproblemandstirsthedirecti ontowardstheuser;towardsthehumanexp eriencer atherthantheengineeringofalgorithms.Thisisawayto"re-humanize"thealgorith mculture,specificallyinNetflix 'scase.Thiscontributes todifferentperspectivesonthefie ldwherehardware,mediatechniquesandtechnologieshavebecomesospectacularandpowerfulthatsometimesitiseasytoforgetthatpeoplearebehindandinfrontofthem.Thereisalongtraditionofaudiences'studies,withRaymondWilliamsasaniconicfigure,butthistimearound,thedigitallandscapeinvitestheresearchertothinkofadifferentapproach.Nolongertheaudiencewillbejustareceptorbutageneratorofcontent.AsEdwardFinnclaims,"theideathatbigdataandmachinelearningsystemsmightgeneratespectacularresultsbutoffernonewhumanreadableinsightsintothesubjectathand"(Finn,2017,p.90)iswhatthisresearchwillattempttotackle.Itisofvitalimportancetoplacethisstudyinrelationtopreviousresearches.Someresearchhasbeenconductedregardingalgorithmtheoryanditsimplicationsonourdailylife:itsreach,itsinvisiblethreatoritsapocalypticmanipulationofinformation,fromaMarxismpointofview.Butverylittlehasbeendoneregardinghowalgorithmsareperceived,fromauser'spointofview,onaspecificarea."Thisapphasassembledasophisticatedalgorithmmode lfordescribing theculturalre lationshipsamongindividualfilmandtelevisionworks,amodelthatfullyembracesthegapbetweencomputationandculture"(Finn,2017,p.93).2.Theoreticalcontext,objective&researchquestions2.1.PreviousResearchWithinthemediaandaudiencestudiesresearchfields,Netf lixasanondemand,entertainment-streamingservice,isaconsiderablen ewfield of study.SearchingGoogleScholar,andsev eralresearchjournalssuc hasNew MediaandSociety,EuropeanJournalofCulturalStudiesandportalssuchasSAGEorJSTOR,fewtononeresultsappeared.

8 Themostinterestingandaccuratearticleregardingtheseissueswastheonewhichscrutinizedaboutthe"NetflixPrize"contestanditsproductionofalgorithmcultureconductedbyBlakeHallinanandTedStriphas.Thisarticlesummarizesandanalysesthe"NetflixPrize"project.Basically,itwasanonlinecontestofferingUS $1milliontowhoevercouldimprovethe accuracyofCinematch,thecompany 'sexistingmovierecommendationsystem,by10%(Hallinan&Striphas,201 6,p. 117).Bythen,Cinematchdidnotconsiderothervariablesasideraitingsstars:notactors,directors,genresnorperiods.SothemainissuewiththewayinwhichCinematchworkedwasthat"twouserswhogivethesamemovieafive-starraitingmightmeancompletelydifferentthings"(Finn,2017,pp.88,89).Therewerenoestablishedcriteriawhetherthatlike ordislikewas doneacco rdinglytothephotogr aphy,totheplot,totheperformanceoftheactors,tothesoundseffects.Noarticulationonwhyorwhattheuserwasevaluatin gwasprov ided.Thiswasoneofthe keymain findingsof thiscontest,andalthoughnoneofthesuggestionswereappliedattheend,Netflixdidcomeupwithasolutiontothismatterbyincorporating"taggers",meaningamoreprecise,indepth"human"characteristicsintheformofadjectivesordescriptions,wheremoreinformationwassuddenlyavailableandthesystemcouldperformatitsbest(Finn,2017,p.89).Someinteresting findingsofthisaforementionedresearch-besidesimprovingthecurrentrecommendationalgorithm-iswhattheauthorsdiscussastheway"hownewmeaningsandpracticescanin sinuatethe mselvesintolong-establishedroutines,transformingthelatterinwaysthatmaybejustreachingpopularawareness"(Hallinan&Strip has,2016,pp.118,119) .Theyquestionedwhatw asthedif ference, ifany,betweenhumandeterminationofwhattoconsumeandwatchandacomputersystemselectionofcuratedmoviesperapprovedandpopular"taste"(Hallinan&Striphas,2016,p.119).Theirmaingoalwasto"adddepthanddimension"totheirdefinitionofalgorithmcultureand,fromthere,startbuildingnewframeworksofculturalpractices(Hallinan&Striphas,2016,p.119).AlthoughtheyunderstoodtheconstrainsoftheNetflixPrizeproject-itbeingaworldwidepitchdrivenbyamoneyrewardtoimprovethecompany 'salgorithm-theyconsider editbeingalegitexample o feffort"toreinterpretwhatcultureis-howisevaluated,bywhomandtowhatends"(Hallinan&Striphas,2016,p.119).Hallinan&Striphas(2016)resultscontributetothealgorithmculturedefinitionandconceptsusedonthispaper.But,regardingtheuser'sperceptionsofthealgorithm,theirrecordslackofdepthandexamples,sincetheirmainvariableistheraitingsystembutnoteachuserpersonaljourneyandalgorithmicsuggestionexperience.

9 Ontheothe rhand,mo reresearchesan darticleswere foundthroughoutGoogleScholarregardingthecontentNetflixcreatesanddistributes.Netflix'soriginalcontentsuchasOrangeistheNewBlack,BlackMirror,HouseofCardsandMakingofaMurder,withtheirin nersubjectssuc hascrime,politic s,femalerepresentation,e tc.wereformallyandacademicallydiscussedonmultiplepaper s.Particularly,OrangeistheNewBlackarticles,focusingonitsfemalerepresentationandculturalconnotations,werewidely available,butalmostnon eonhowtheuser experien ceisperceived .Recently,onhisbookWhatAlgorithmWants(2017),EdFinnstudiedtheimplicationsofalgorithmsonthecreativeprocesswhilecraftingHouseofCardsandthepossibleoutcomesandconsequencesthosedecisionscanhaveonuser'sbehaviorandfurthercontentcreation.2.2.ObjectiveThepurposeofthisresearchistocriticallyexploretheondemand,algorithmbasedappNetflix,focusingonhowtheuserexperiencesthosesuggestionsandredefineswhatondemandreallymeans.Thestudyisanattempttounderstandtheusers'NetflixexperienceandcontrastitagainsttheexpectedusageimpliedbyNetflix'sinterface.ThiscontrastwillbedonethroughthewalkthroughmethodproposedbyLight,BurgessandDuguay(2016),analyzingthevision,theoperatingmodelandthegovernanceoftheappagainstuser'sjourneyexperiencesandinterviews,gatheredbythinkaloudtechniquestobefinallystudied,analyzedandstandardizedviaqualitativecontentanalysis.2.3.ResearchQuestionsMainquestion:- HowdotheNetflixusersexperiencethesuggestionlogicofNetflix?Supportquestions:- Doesthislogicexpandoralienatetheuserexperience?- DoesthislogicempowerNetflix'susers?- Whatstatusdoesoriginalcontenthaveinrelationtoothercontentwhenitcomestotherecommendationlogic?- WhatothervariablesoutsidethealgorithmcontributeframingtheNetflix'suserexperience?- Whathappenstotheuser´sexperiencewhenthealgorithmfails?- WhatareotherunexpectedpracticesoftheusageandexperienceofNetflix?Inordertoachievetheseobjectives,themostaccuratewayistotheoreticallysubdividefromabroadertoasingularperspectiveapproach.Firstly,startingfromthedefinition

10 andcontextualizationoftheproblemwithinthecurrentmedialandscape,tounderstandingthealgorithmcultureanditsimplications,tofinallyarrivetoaNetflix'sdescriptionandanalysisofitscharacteristicsforabettercomprehensionofeachuser'sexperienceattheempiricalresearch. 2.4.LimitationsoftheStudyReturningtoourmainpointofstudy,howdopeopledealandorcopewiththefactofalgorithmsbeingimplicitlyincorporatedinourdailylife?Aretheyindeedorareweconsciousaboutit?Howmuchdowefeeltheyinfluenceus?Isitforgoodorforbad?TakingNetflixasanexample,Iconsiderthatoneofthemostimportant-ifnotthemost-outcomeofthisprojectwillbetheconclusionsofhowNetflixusersexperiencethealgorithms,toenlightenpeople'sreactiontothison-demandlogic,torethinkandconsiderwhetherthislogicexpandsorlimitspeople'sexperience,payingparticularinterestonitsimplicatio nontoday ´stethered,digital entertainmentandcultu ralindustry.ThisstudyfocusesonNetflixbuthopefully theresultsandlearningscanbealsoconsideredforfurtheron-demandapps,whichaccordingtothecurrentstatusquo,areandwoulddefinitelybecomethestandardwayofconsumptioninwesternsocieties:fromgrocer ies,dinnerorservicesdeliveredtoyourd oor, toe-commerceor,asdisplayedonthiscase,entertainment.Itisimportanttounderstandhowthechosenplatformworks,soonecancomprehendthepossibleoutcomeswhenalgorithmsgowrongandtheexperiencegetsbroken.Themainfocuswouldbedoneontoeachconsumer'sjourneyandindividualexperience.Havingsaidthat,iti simportantto defineso meof thedelim itationsofthisstudy.Programming,codingandNetflix'sstruc turewillbeonlytaken intocon siderationthroughtheuser'seyesandnotthesoftwarestructure.Howitsalgorithmandcontentisprogrammed,codedandsuggestedissomethingthisresearchcaninferredonlightoftheusers'experiencesbutwillnotattempttoclearlydefinethemastheyactuallyaredesignedonandfortheapp.Regardlessofthisresearchfocusingandprioritizingtheuserexperienceabovethemereanalysisofalgorithms,exploringthesefromaconceptualperspectivewillcontributetoabetterunderstandingofNetflixandwillenlightenthemoretechnicalconceptssuchasapps,platform,software,algorithms,whichwillhelpinterpret,frameandmaybeenlightensomeofthereasonsbehindtheconsumerexperiences.

11 2.5.TheoreticalContext2.5.1.CurrentMediaLandscape:MediainEverydayLife&GlobalCulturalIndustryMuchhaschangedinthelast15yearsregardingeverydaylifeandthemedialandscape.Fromtransportationmeanstofood,mediaanditstechnologyfoundthewaytosqueezethemselvesinandtransformedpeople'shabitsandroutines.Inboth,amanifestbutalsoonamorelatentway,computationalprocesseshavebeenearningtheirwaythroughthe"sorting,classifyingandhierarchizingofpeople,places,objectsandideas"(Striphas,2015,p.395).Ononehand,algorithms,platformsandsoftwareevidentlyandbyourvoluntarywillcontributetotheshapingofthewayweconsume,communicateandconnect.While,onothereverydayissues-asimportantasthepreviousones,butthatmanytimesconsideredmundaneortrivialbysome-suchasentertainment,cultureandmediaconsumption,thesecomputationalprocessesactinasubtlerway,sometimeswithoutusevennoticing.Thiseverydaylifeshouldbeframedwithinabiggercultural,politicalandeconomicchange,whichcanbenamed"GlobalCultureIndustry"asLashandLury(2007)baptizedit."Culturalobjectsareeverywhere;asinformation,ascommunications,asbrandedproducts,asfinancialservices,asmediaproducts,astransportandleisureservices,culturalentitiesarenolongertheexceptionbuttherule"(Lash&Lury,2007,pp.5,6).Moreover,theseculturalobjectsnowneedtoberedefinedaswellastheculturalconceptitself.AccordingtoStriphas,bigdataandlargescalecomputationlogics-suchtheoneNetflixuses-altersthewayhumanthink,conduct,organize,practice,experienceandunderstandculture(Striphas,2015,p.396).Thesechangesinculturedefinitelyaffectourdailylifeandroutine.Netflixhadbecomeoneofthoseapps,featuredinmorethanonedevices(computers,tabletsandmobilephones)whichhelpedrestructureeverydayroutinesandtherefore,changedandinfluencedourculture.AsSaraPinkestablished,thismajorpresenseofNetflixinourdailylifecanbeawaytoexplore"theroleofmediainthemakingandexperiencingofenvironments,centeringontheirsaliencetodailyroutinesoftransitioninthe(...)mediasaturatedhousehold"(Pink&LederMackley,2013,p.677).Someofthechangesseeninmediaandineverydaylifearealsopresentwithintheconsumer'seconomyandtheyhavebecomequitesubstantial.Manovichexplainsthisinteractiverelationshipasatwo-waylink,betweentheuserandtheculturalproduct(apponthiscase).Theauthormentionsthatthisphenomenonoftenappliesto"born-digitalindustriesandmedia,suchassoftware,computergames,websites,andsocial

12 networks.Theirproductsareexplicitlydesignedtobecustomizedbyusers"(Manovich,2009,p.323).Notonlytheincorporationofnewtechnologyandthemediatizationofroutineshavehappened,butalsoevolved.Thisoccurredtotheconnectivityaspectofournewmediatizedculture.AccordingtovanDijck,theconnectivityistheelementthatwiresandconnectseverypieceofculturetogether,inaliteralandmetaphoricalway,inwhichtechnologiesandconnectedgadgets"shapeandareshapenotonlybyeconomicandlegalframes,butalsobyusersandcontent"(vanDijck,2012,p.1).Itisnolongeraboutlettingmessagesordatain,butalsocrafting,producingandspreadingthemoutasactiveprosumers.InNetflix,thisphenomenonisnotasliteral,buthavingitasoneofthemanyon-demandappsthatcustomizeeachuser'sprofilesaccordingtoeachuser'sinteractionsonlyreflectstheinfluenceandtowhatextentthismediaticculturalindustryhasreached.Howdidallthesehappen?Whatwastheturningpointofthephenomena?ItisofcommonagreementthatthemostrelevantshifthadhappenedsincetheintroductionanddevelopmentoftheInternet,butnottheWorldWideWebasweknowit,buttheWeb2.0.ThetermWeb2.0suggestsan"upgradedandupdatedversionoftheweb"beingdefinedas"(...)'collectiveintelligence'throughthedevelopmentofa'participatoryculture'"(Beer,2008,p.5-6).NotonlyparticipationhelpedshapedthisWeb2.0evolution,butalsoconnectivityandmulti-screenphenomena:"[...]Applicationsbecomeaccessiblefromanynetworkedinterfaceorportalratherthantheinformationbeingstoredonasingledevice.Here,informationmovesfromtheprivatedevicetothenetworkallowingittobeaccessedfromarangeofmobileanddesktopinterfacesatanytimeandfromanywhere"(Beer,2008,p.6). StriphashasagoodpointwhichcanbeusedtoclearlyexemplifyhowalltheseWeb2.0practiceshelpshapetheway"shopping,merchandizingandhostofothereverydayculturalactivities"because,duetothemedialandscapetransformationspreviouslymentioned,theynowbecame"datadrivenactivitiessubjecttomachinebasedinformationprocessing"(Striphas,2015,p.398).Thiscanbeseenasanintertwined,complicatedandreciprocalprocess:therawdatamininggeneratedbytheusers'activityistheonewhichcontributestothatWeb2.0,two-waystreet,interactiveformofcommerce,education,andoverallcommunication.

13 TheseissuesofprivateandpubliclifespheresaresignificanttothisNetflixresearch,because,withoutbeingnoticed,thisdynamicframedwithinthisnewhyperconnectedworld,createsanothersphere,avirtualone,wherethepublicandprivatespheresboundariesareaclickawayfromeachother,makingtheirlimitsblurryandconstantlychanging.Thissphereiswhereproductionanddistributionofmultimediacontentisbeingheld,andwheretheuserlivestheon-demandNetflixexperience.Thisproductionorusergeneratedcontentisreferringtotheselections,raitingsclassificationandeachparticularclicktheuserdoesontheplatform.Byinteractingwithit,wegeneratevaluableandinfiniteamountofdatathatwouldreturninanalgorithmrecommendationformbacktous.Animportantaspectwhichalsoshowcasesthat"tethered"self"(Turkle,2006,p.6)andtheblurryofboundariesisthatpersonalizationproductionhappensattwolevels:anautomatedone,wherealgorithms,marketinginterestsandpreviousconsumerbehaviorisbeingconsidered,anda"human"one,basedonuser'sagency,wherepersonalchoicesarealsodeterminedbypeers,friendsandareferencecommunity.AsJonessays,thesecustomized,individualchoicesofcontentare"notbasedoninvisibleinteractionswithmachines(...)weshouldnotbeblindtothefactthatisrealpeoplewhooccupythatspace,virtualorotherwise"(Jones,2002,p.3).Jonesregardshisanalysistothemusicalspherebutthesameissuehappenstothebroaderentertainmentscene. Jones(2002)andTurkle(2006)illustratethatcurrentlythereisnodoubtthatoureverydaylifehasbeenconditioned,modifiedandeventually-evolved-duetothisconnectivityprocessthe[global]culturalindustryandtheworlditselfhavebeenexperiencing.Thismeansthatourdailyroutines,includingourleisureoptions,havebeenaltered.Passingthroughpaperpagesisnowoutdated:scrollingisthenewblack:one'sKindle,one'sphone,one'stabletinordertoreadnews,one'sfavoritebookorourfriend'sstatus.Weread"feeds"meaningwearebeingfedbyallsortsofinformationdonebyinfinitivesources.Goingtothemoviesnowhadchangedby"Netflixandchill",thenewpickuplinepeopleuseinTinder,thedatingapp,todescribethefactofstayinglyinginbed,watchingsomeawesomeseriesandmaybesomethingmore.Aninterestinginterpretationofthisoverallstatus-quosituationisthatitpresentsourselvesalimbo-oxymoronwherewenegotiate,struggleandliveondailybasis:thehomogenizationoftheculturalindustry,whereeverythingisautomatedandequalversusourstrongurgetoindividualizeandpersonalizetheuserexperiencesthankstothepersonalizationtechniquesprovidedbythealgorithmculture.

14 2.5.2.AlgorithmCulture&ScienceandTechnologyStudiesConceptsWhatisitmeantbyalgorithmculture?TedStriphasdefineditastheshiftofdelegatingtheworkofcultureof"sorting,classifyingandhierarchizingpeople,places,objectsandideas"tocomputationalprocessesthateventuallymodifythewaywepractice,experienceandunderstandthem(Striphas,2015,p.395).TheauthortakesRaymondWilliams'(1958)definitioninordertounderstandacurrentviewonculture,definingit"notonlyasthegeneralbodyoftheartsbutalsoawholewayoflifematerial,intellectualandspiritual"(Striphas,2015,p.398).Hisdefinitionsaredonethroughoutahistoricalapproachonhoweachterminologyhavebeenmutating,hence,evolvingtotoday'scurrentusage.Hishistoricalapproachtoinformationhelpsitsunderstandingas"somethingthatemanatesfromsomeexternalsourcetooneself"(Striphas,2015,p.399).Thisupdatedversionisimportantsinceitconceivestheprocessofabstraction:informationcanbeconsideredthat"rawmaterial"thatiseithergivenorreceivedviasenses,speciallycomprehendthroughour"cognitivefaculties"(Striphas,2015,p.399).Informationisnowconsideredanindividualthingforitself,whichcanliveoutsideourbodies.Thismeantahugerevolutionsinceknowledgecanbestored;itnolongerdependsonourphysicalattitudesormemoryskills.Books,machines,cloudsandrecordsbecamemorerelevant.Anotherimportantissueisthat,afterWWIIonwards,machinestoredinformation"beginbeingseennotmerelyasausefulthingbutascustodianoforderliness"(Striphas,2015,p.400).Thisprocessisrelevantnowadaysforthisalgorithmicculturetheauthormentions,andparticularlytotheon-demandplatforms,sincehumansdonotholdtheexclusiverightsofcultural"producers,curatorsandinterpreters.Now,thankstooutsidestorage,informationcanbeproduced,storedandusedinaprogramedyetautonomousway.Also,itisofextremeimportancetoNetflix'salgorithmlogic.Turkle's(2006)virtualworldwithtastesandpreferencestheuserlikesanddislikesneedstobestoragesomewhere,andtheappallowsthisinformationnotonlytobethereandbeaccessedbyanydeviceasBeer(2008)established,butalsoandmainlyitallowstheinformationtogrowexponentiallyandbemodifiedinrealtimeaccordingtoeachuser'sselectionsandclicks.Thisisalsorelatedtorealtimesatisfactionandentertainment,whichisverydifferentfromwhatitusedtobe.Thisisonemajorshiftinourdailymediaconsumptionaffectingoneconomicalandproductionissues,asstatedonFinn's(2017)work.

15 Ontheotherhand,Striphas(2015)alsofindsinterestingtoreachadefinitionofcrowd.Whenhestartsanalyzingthisterm,hereferstoTheOxfordEnglishDictionarydefinition,understandingthisnounasasynonymof"mass,mob,multitudeandlargegatheringsofpeople,generallyinpublic,especiallyinurbansettings"whereindividualidentityisbeingblurredandusuallythesepeoplearegathered,engagingwithcertainobjectiveoraction(Striphas,2015,p.401).Then,hementionedMackay,andhisarticulationof"thinkinginherds"and"thepopularmind"(Mackay,2001[1841]).Mackay'sconceptionclearlydifferentiatesitselffromtheconceptofTheOxfordEnglishDictionary,designatingan"active,livingprocessalbeitoneinwhichanyindividualcontributionregistersdiffusely.Thenounform,"popularmind"largelyelidesthatprocess,positingsomeoverarchingthingreferringtoeveryoneingeneralandnooneinparticular"(Striphas,2015,p.401).Ifinformationcanbeproducedandstoredoutsideahumanbody,thenacrowdcanbeageneratorofinformation.MyunderstandingofwhythesetermsareimportanttoStriphasisbecause,althoughheclearlyspecifiesthatcrowdsareusuallydoneinurbansettings,Ibelievethattheanonymity,publicnessandcurrentavailabilityofInternetconnection,allowsthatthisinformation,-henceknowledge-generationanditscorrespondentstoragetohappenonline.Peoplegenerateinformation,willinglyandunwillingly-byleavingdigitaltracesofone'sonlineactivityduetocookiesandcaches-,andthetrends,socialnetworksandcrowdsourcingpossibilitiesarethosenew,digital,"publicspaces"wherepeoplegatherandthe"popularmind"isbeingsettled.Thisphenomenonispossibleduetoalgorithms,whichisthethirdandlasttermtheauthorexplains.HeexplainsthatoriginallythewordcomesfromtheoldGreekworldfornumberarithmos.Nowadays,thecontemporarydefinitionoftheworldrefersto"aformalprocessorsetofstepbystepproceduresoftenexpressedmathematically"(Striphas,2015,p.403-404)althoughheinsistsontheimportanceofthemisinterpretationoftheoriginalArabicconceptsforcalculationsuchasarithmeticandalgorisms,theArabicsystemofnumerationitself.So,whydidalgorithmsasatermbecomemorepopularintoday'sworld?Theauthorclaimsbacktheimportanceofinformationandculturetoexplainthisissue.OnhispaperTransmissionofInformationRalphHartley,aBellLaboratoriesengineer,explains,"inanygivencommunicationthesendermentallyselectsaparticularsymbol(...)astheselectionsproceed,moreandmorepossiblesymbolssequencesareeliminatedandwesaythattheinformationbecomesmoreprecise"(Hartley,1928,p.536).Thismeansthat,withincommunication'sentropy,wherechaosis,bydefinition,thecurrentstateofthatinformation,anyorderneedtobeengineered,andthesecreations,

16 transformationorselectionofsymbolsledtounderstandingandorder(Striphas,2015,p.405).ThisorderistheoneNetflixappliestostructureitself,providingacustomizedoffertoeachuseraccordingtotheinputsorsymbolstheusersubmitswithintheappnavigation,andofcourse,theculturalmediapracticestheydo.AccordingtoNetflix,thecompanydonotcrossdatawithotherappsorplatforms-atleastnotmorethanquantitativevaluessuchasemail,age,gender,anddevice-butofcoursethecompanynourishitselfbythetrends,culturalvaluesandeventsthatarecurrentlyrelevanttoitsaudienceforthecreationoforiginalcontentandcurationoftheirexternalcontent.Forexample,whenFidelCastropassedaway,thelaunchofNetflix'sOriginal,TheCubaLibreStory,aswellasotherColdWarandGuerrillaConflictcontentwaspushedorreleased.Thisblurrylinebetweenwhattheviewergenuinelywanttowatchandwhatispushedorsuggestedtosee-bythealgorithmorbyasocialevent-,changingandshapinghisorherhabitsonwhattowatchnextcanbeawayof"corruptpersonalization"asChristianSandvigcallsitandFinnexplainsit(Finn,2017,p.21)."Nowadays,algorithmshave"significantlytakenonwhat(...)hasbeenoneofculture'schiefresponsibilities,namelythetaskof'reassemblingthesocial'throughusinganarrayofanalyticaltoolstodiscoverstatisticalcorrelationswithinsprawlingcorpusesofdata,correlationsthatwouldappeartouniteotherwisedisparateanddispersedaggregatesofpeople"(Striphas,2015,p.406).Algorithmsshouldbeunderstoodas"'socio-technicalassemblages'joiningtogetherthehumanandnon-human,theculturalandcomputational"(Gillespie,2014,p.404-405).Buildinguponthatstatement,StriphasmentionsFlusser'sapproachthatalgorithmculture"istheautomationofculturaldecisionmakingprocesses,takingthelattersignificantlyoutofpeople'shands"(Flusser2011:117citedinStriphas,2015:408). ThealgorithmculturecannotemergewithoutthepossibilitiesoftheWorldWideWebanditsconnectivity.Thislastconceptisaclearexampleonhowthisresearchisrelevanttotoday'smedialandscape.Connectivitycannotbeunderstoodalonebyitstechnicaltermsandconnotations.Therelevanceofitshouldbecloselypairedwiththeinfluenceithasonpeople'sdailybehaviors.Inordertodothis,vanDijck's(2012)approachisaveryinterestingone,sincethreeconceptsofconnectivity-platform,protocolandinterface-aregoingtobepracticalwhileexplainingthealgorithmcultureweareimmersein.

17 Firstly,aplatformiseitherthematerialorimmaterialsupportforasocialactivitytohappen.Usually,thesesocialactivitiesareformattedintoprotocols,meaningtheexpectedorcorrectwaytosucceedand,thisphenomenonispresentedtothefinaluserwithafriendlylookorinterface(vanDijck,2012,p.4)."Anyplatform'sconnectivestructureismediatedbyprotocols:formaldescriptionsofdigitalmessageformatscomplementedbyrulesforregulatingthosemessagesinorbetweencomputingsystems"(vanDijck,2012,p.4).Protocolscanbe"technicalsetsofrules"thatworkindependentlyandindifferentlyfromitsveryowncontentbuttheycanalsoimproveandreframetheirusabilityandgoal,anddifferentfromitsoriginalprogrammingandintent,duetothewaytheirownersusethem(vanDijck,2012,p.4).Inthiscase,theplatformistheNetflixappinaTV,tablet,phoneorcomputerthatplaysentertainmentcontent,theprotocolsaretheprogrammedandformattedseries,moviesanddocumentariesavailabletostreamandthen,whentheusersdecidewhattowatchandtheirpreferencesstartgettingsetup,thenthoseprotocolsmutate.Theseprocessesareavailabletotheuserthankstoafriendlyinterfacewithwhichtheyinteract.ThearchitecturevanDijckhighlightsiswhattheregularuserisusuallyunawareofandwhatthesavvyuserhesitatesabout.Thiswouldbetheprogrammingtechniquebehindtheappsandtechnologyweuse.Forexample,Netflix'scopyrightslaws,whichdonotlettheuserchangeitsIPlocationformoreordifferentcontent.Ontheseekofhumanizingthealgorithmasamainobjectiveofthisresearch,itisextremelyimportanttokeepinmindwhyarewestillmentioningthesetechnicalsubjectsasrelevant.Platforms,protocolsandinterfacescontributetoillustratethismatter,particularlythelinkbetweentechnologicalandsocialaspects. Thealgorithmsareembodiedonthatfriendlyinterfacetheusersinteractwith,bytouchorclick.ThereisaninterestingapproachprovidedbytheScienceandTechnologySciencescalledtheActorNetworkTheory(ANT)."ANTforegroundsarelationalontologyaccordingtowhichsocioculturalandtechnicalprocessesaremutuallyshaping"(Light,Burgess,&Duguay,2016,p.6).Thismeansthattheinterfaceisthoughtanddesignedtakingtheuserintoconsiderationandtheuseritself,afterinteractingwithit,providesfurtherinputstotheapp,henceitsdevelopersanddesignerscanadjusttheinterfaceforbetterresultsaccordingtheactualuserusage.Thisisthekeyofsuccessforagoodandusefuluserexperience(UX)."ANTdifferentiatesbetweenintermediariesandmediators,whichcanadditionallybehumanornon-human.Intermediariespassunchangedmeaningalongthroughouta

18 networkofrelationswhilemediatorsaretransformative,theyalterthemeaningorcircumstanceswithinasystem"(Light,Burgess,&Duguay,2016,p.6).InNetflixcase,ahumanintermediarycanbeaNetflixusertalkingtoanother,recommendingaseriesortalkingaboutaspecificchapter,whileanon-humanonecanbethewelcomingtrailerwithNetflixoriginalcontentpresentineveryapp.Ontheotherhand,ahumanmediatorcanbetheusageoftheraitingoptionswhileanon-humanonecanbetheactualresultofthatraiting,alteringthealgorithmpersonalizedoffer."Socialpracticesareincreasinglymediatedbyplatformsthataffectpeople'sdailyinteractionsandreciprocalrelationships.Moreprecisely,platformsrunonaccountofcodedprotocolsthatappeartomediatepeople'ssocialactivities,whileinfactsteeringsocialtraffic"(vanDijck,2012,p.5).OnNetflix'sexample,certainconsiderationssuchaswatchingstyle,time,behavior,soloorgroupwatchingareinfactissuesthatrepresent"socialtraffic"vanDijckmentionsandwillbeconsideredasofflinebehaviorsandvariablesthataffecttheoverallexperience.Thesedefinitionswerenotmerelyinformativefortheforementionedaspectsbutalsonecessarytogetabetterunderstandingofthealgorithm'slogic.But,whatisanalgorithmexactly? "Analgorithmisanywell-definedcomputationalprocedurethattakessomevalue,orsetofvalues,asinputandproducessomevalue,orsetofvaluesasoutput.Analgorithmisasequenceofcomputationalstepsthattransformtheinputintotheoutput.Weshouldconsideralgorithms,likecomputerhardware,asatechnology"(Seaver,2014,p.1).Thisisenlighteningbecauseitprovidesclarificationontothemergingfactwhen"rigid,quantitativelogicofcomputationtanglesthefuzzy,qualitativelogicsofhumanlife"(Seaver,2014,p.2).Asstated,brands,communicationsandmediacannotbeseparatedfromeverydaylifeandtherepresentationoftheonandofflineself,sothesezerosandones,canactuallybeconsideredasatwentyfirstcenturyfingerprint,theusers'digitaltrace,whichreflectsontheirofflinepersona.SamethingappliestoNetflix."Giventhepersonalization[of]algorithms(...)allinteractionswiththesystemaretailoredtospecificuseraccounts"(Seaver,2014,p.5).AmorehumanapproachthatisinterestingandaccurateforthisresearchisZiewitz's(2011)streetexperiment,whereinsteadofconsideringanalgorithmasapowerfulentity"thatincreasinglytakescontrolofourlivesandgeneraterankingsthatforegroundsomebutbackgroundotheroptions"(Ziewitz,2011,p.5)onecanthinkaboutitas"somethingthatisbeingdoneinpractice,inmanyplaces,everyday,again"

19 (Ziewitz,2011,p.5).Algorithmsareusually"seemtobegivenandfinal",whenthecorrectwaytoconsideritis"farfromstatic,butconstantlyunderconstruction"(Ziewitz,2011,p.10). "Algorithmsplayanincreasinglyimportantroleinselectingwhatinformationisconsideredmostrelevanttous,acrucialfeatureofourparticipationinpubliclife.(...)Recommendationalgorithmsmapourpreferencesagainstothers,suggestingnewforgottenbitsofcultureforustoencounter.Algorithmsmanageoutinteractionsonsocialnetworkingsites,highlightingthenewsofonefriendwhileexcludinganother's.Algorithmsaredesignedtocalculatewhatis"hot"or"trending"or"mostdiscussed",andskimthecreamfromtheseeminglyboundlesschatterthat'sonoffer"(Gillespie,2012,p.1).Notonlyalgorithmsproviderawinformation,butalsotheyhavebecomethetoolsinordertogather,classify,understandandfindmoreandfurtherinformation.Theyarethecurrent"keylogic"thatruletheinformationandmediaflowinwhichwe,asproactiveusers,participate.Theyhavethe'powertoenableandassignmeaningfulnesstocontent',managinghowinformationisperceivedbyusers"(Gillespie,2012,p.1).Therearethreeveryimportantaspectsregardingalgorithms.Innoparticularorder,thefirstoneisrelatedtotherolealgorithmshave."Thesealgorithmsareproducingandcertifyingknowledge"(Gillespie,2012,p.2).Thesecondone,ishowalgorithmsarereadable,andthereforeusable,"onlycontrastedandincooperationwithdata"(Gillespie,2012,p.3-4).Lastly,itisimportanttostatethatisnotthealgorithmresultswhatmatters,iswhattheuserdoeswiththat(Gillespie,2012,p.4).Ifthealgorithmsresults,accordingtothedatabasepairedandthepreviousprogrammingarecorrectbutnotrelevantorupdatedfortheuser'spurposes,thenthealgorithms,andthereforetheapp,ismeanttofail.Onceagain,hereiswheretheinteractive,engagementandperformativeactionsfromusersareclearlyshowingthechange,fromapassiveintoanactiveaudiencebehavior,responsibleofreshapingmedia,andthereforeculture,forthelast15years.Theproductionofknowledge,itsvalidationandthetrendsettingaspectsaresomeoftheissuestobecontrastedonthefindingsattheresearch.

20 3.MaterialandMethodsInthissection,thereaderwillbeabletofindadefinitionanddescriptionofthefieldofstudyaswellasfortheselectedsampleusedforthisresearch.Anexhaustiveexplanationaboutthegatheringandtheanalyzingmethodswillbesubdividedandexplained.Firstly,thewalkthroughmethodaswellasthethinkaloudinterviewswillbedetailedasforgatheringtechniques,while,ontheotherhand,qualitativecontentanalysiswillbeusedandexplainedasthemethodappliedtostudyandinterprettheresults.3.1FieldofStudy:Netflix-TowardsadefinitionanddescriptionPeopletendtothinktheyknowwhatNetflixis;butdowereallyknowfromatechnicalpointofview?Forthispaperpurposes,IwilldefineNetflixasanapp."Softwareapplications(apps)arenowprevalentinthedigitalmediaenvironment.Theyarethesitewheresignificantsocioculturalandeconomictransformationsacrossmanydomainsfromhealthandrelationshipstoentertainmentandeverydayfinance"happen(Light,Burgess,&Duguay,2016,p.1).Itisdifficulttoreachafinaldefinitionofwhatanappis;itdependsonhowitisprogrammed,onitstechnicalpossibilities,itsobjectivesetc.AccordingtoTraction,aCanadianCRMconsultingandsoftwaredevelopmentfirm,anapplicationis"theactofputtingsomethingtoaspecialuseorpurpose"(TractiononDemand).Light,Burgess&Duguayalsodefinesitas"asubsetof[closedandcontrolledsystemslike]computerprograms-orplatforms-thatsolveparticular,oftensingularuserneeds,originallybusinessneeds"(Light,Burgess,&Duguay,2016,p.4-7).Hence,sinceboth,thesoftwareindustryandtheacademicfieldtendtorefertothemasplatformswithaparticularpurpose,forthisresearch,Iwillalsodefineanappasaplatform,invanDijck'ssense,withaspecifictask,onthisparticularcase,theonlineandondemandstreamingservice.Accordingtoitsofficialwebsite,Netflixisthemaintelevisionstreamingserviceworldwide,"leadingthepathofdigitalcontentsince1997".Netflixisavailableinmorethan190countriesandcurrentlyhasmorethan93millionusers,whoenjoyapproximately125millionsofTVshowsandmoviesperday(Netflix,2017).Netflixuserscanwatchwhatevertheywant,whenevertheywant,almostinanyscreenwithInternetaccesswithoutadvertisement.Netflixphenomenonis overwhelming:duringanygivendayof 2014,o nethirdofInternet'sdatadownloadduringpeakhourswereNetflixstreamingfiles."Bytheendof

21 2013,intheUSonly,thecompany's40millionsubscriberswatchedabillionhoursofcontenteachmonth"(Finn,2017,p.88).Dataisnolongerjustzerosandones,dataisrelevantonceitbecomesinterpretedandusedbyandforusers'behaviors,experiencesandchoices.Anapp'svisionclearlyestablishesitspurpose,targetuserandidealscenariosofusage.Italsomentionstheactivitiesissupposedtoperform,supportorenable.(Light,Burgess,&Duguay,2016,p.9)."Anapp'svisiontellsuserwhatitissupposedtodoandbyextension,implieshowitcanbeusedandbywhom"(Light,Burgess,&Duguay,2016,p.9).Itusuallygoesunnoticedbytheusersbuttothisresearchpurposeisimportantsinceitprovides"abaselineforidentifyinguserappropriation"(Light,Burgess,&Duguay,2016,p.9).Blogs,pressreleasesandpublicstatementsalsoconveytheapp'svisionandhelptowidespreaditamongthecommunityandnotonlyitssubscribedusers(Light,Burgess,&Duguay,2016,p.9).ThereisnoofficialNetflix'svisionpublishedneitherinsidetheappitselfnoronitsofficialwebsite.Theonlyinformationavailableunderthe"about"sectionistheoneusedforthedefinitionofwhatNetflixis.TheBalance.com,anonlineconsultancy,gatheredsomeofthemissionstatementsoftechnologycompanieslastyearandNetflixwasamongtheselectedones.Theyalsoconfirmthatthereisnoofficialpublishedmissionstatement,buttheymentionthatin2011,Netflix'sCEO,ReedHastingsexpressedhisvisiononhowheembodiestheupcomingfutureofthecompany.Hestatedtheywantedtobecomethebestglobalentertainmentdistributionservice,tobeabletolicenseentertainmentcontentaroundtheworld,tocreatemarketsaccessibletofilmmakersandtohelpcontentcreatorsaroundtheworldtofindaglobalaudience(Farfan,2016).ThewebsitealsostatesthatNetflixhaspublisheditscompanyvalues,whichcanprovideadditionalinformationandclarificationaboutthe"principleswhichguideditsemployeesintheirdailydecisionsandactivities.TheNetflixCompanyValueswerepublishedas:Judgment,Productivity,Creativity,Intelligence,Honesty,Communication,Selflessness,ReliabilityandPassion"(Farfan,2016).AccordingtoFacebook´sCOOSherylSandberg,Netflix'scompanyculturedocument"maywellbethemostimportantdocumentevertocomeoutoftheValley¨(2013)2.Theoperatingmodelusuallyinvolvestheapp's"businessstrategyandrevenuesources,whichindicateunderlyingpoliticalandeconomicinterests"(Light,Burgess,&Duguay,2016,p.10). 2 GregoryFerensteinforTechCrunch,31-01-2013:https://techcrunch.com/2013/01/31/read-what-facebooks-sandberg-calls-maybe-the-most-important-document-ever-to-come-out-of-the-valley/

22 AninterestingmodelstrategythatNetflixarticulatesandFinnanalyzesisthecurationfilterinsteadofcensorship."Facebook,Google,Netflixandtherestdonotoftenengageinovertcensorship,butratheralgorithmicallycuratethecontenttheywishustosee,aprocessmediascholarGanaeleLangloisterms'themanagementofdegreesofmeaningfulnessandtheattributionofculturalvalue'"(Finn,2017,p.111).OnNetflix'sOfficialBlog,onecanfindanicearticulationofitsbusinessobjectivecloselyrelatedtothesubjectofthisresearch:"Ourbusinessobjectiveistomaximizemembersatisfactionandmonth-to-monthsubscriptionretention,whichcorrelateswellwithmaximizingconsumptionofvideocontent.Wethereforeoptimizeouralgorithmstogivethehighestscorestotitlesthatamemberismostlikelytoplayandenjoy"(Netflix,2012).Notonlymoneyexchangemeansrevenueintheappworld,theusers'data,likes,dislikesandpreferencesaretoday'scurrency.Thesedatacanusuallybesoldtoadvertisersanddataminers(Light,Burgess,&Duguay,2016,p.10)althoughthisisnotthecase.Netflix,undertheirPrivacyStatement,clearlyestablishthattheinformationgatheredbythemorprovidedbytheuserwillbeusedwithinTheNetflixFamilyofCompanies,theservicesprovidersandsomeotherthirdpartiesbutonlytheonesrelatedtotheviewingexperience.Theusercanchoosetodiscloseinformationwithsocialplug-ins,whichareoperatedbythesocialnetworksthemselvesandregulatedundertheirownprivacystatements."Weuseinformationtoprovide,analyze,administer,enhanceandpersonalizeourservicesandmarketingefforts,toprocessyourregistration,yourordersandyourpayments,andtocommunicatewithyouontheseandothertopics"(Netflix,2017).Thisisanadvertisingfreeservice.Asanon-demandserviceandpay-per-viewphilosophy,itispartofitscoredefinition.TheNetflixcasemightbeparticularsincetheydonotselltheuser'sinformationtothirdparties,buttheykeepitanduseitthemselvesforthebettersuggestionandveryowncontentcreation.Theyguaranteetheprivacyandwellmaintenanceoftheinformation:"WheneverinthecourseofsharinginformationwetransferpersonalinformationtocountriesoutsideoftheEuropeanEconomicAreaandotherregionswithcomprehensivedataprotectionlaws,wewillensurethattheinformationistransferredinaccordancewiththisPrivacyStatementandaspermittedbytheapplicablelawsondataprotection"(Netflix,2017).Personally,Ibelieveisageniusstrategyforanoperatingmodel,sincealltherevenue-itscostandusers'data-stayswithinthecompany,inordertodevelopbetter

23 suggestionsandawiderscopeofcontent.Thisinformationwillallowtracingtrendsineachofitsregions,allowingNetflixtopartnerwithlocalproductioncompaniesinordertodeveloptherelevantcontentforeacharea.Thisshowstheglocalapproachthecompanyhas:providingaworldwideservice,selectingsomeoriginalcontentandthemostwatchedcontenttodistributearoundallregions,whileprovidingaspecial,locallyrelevantandappealingselection.Also,underNetflix'sTermsandConditions,theuserisbydefaultenlistedonperiodicallytestingandtrialsofnewcontentandprogrammingstrategiestobenefittheimprovementoftheoverallapp,speciallyitsrecommendationengine.Incasetheuserdoesnotwanttoparticipate,heorshewillneedtologintotheaccountinordertodeactivatethisoption.Forexample,differentraitingsystemsaswellasdifferentdistributiontechniques(weeklyreleases)arebeingconstantlytested."Netflixregularlymakeschangestotheservice,includingthecontentlibrary.Inaddition,wecontinuallytestvariousaspectsofourservice,includingourwebsite,userinterfaces,promotionalfeaturesandavailabilityofNetflixcontent.Youcanatanytimeopt-outoftestsbyvisitingthe"YourAccount"pageandchangingthe"Testparticipation"settings"(Netflix,2017).Duringthesubscriptionprocess,thenewuserhasthepossibilitytochoosethebestplanoptionamongthreealternativestheywouldprefertokeepiftheydecidetostayafterthefreetrialmonth.Netflixoffers3kindsofplans:basic,standardandpremium.Eachoneofthemclearlyestablishesitsconditions,featuresandprice.Differentpricingrangesareofferedaccordingtothecountrytheuserislocated,butitisusuallybetween$USD7to$USD10forthebasicplan,from$USD7to$USD13forthestandardplanandfrom$USD8to$USD18forthepremiumplan,allofthempermonth(OZBargainBlog,2015).ThecheapestplandealsarefoundinLatinAmerica(Argentina,Chile,Brazil)CanadaandtheUS,whilethemostexpensiveonesareinSwitzerland,NewZealand,NorwayandSweden.Itwouldbeinterestingtoseeifthepricerangesaffectstheamountofregisteredusersineachregion,butthatwouldbeforanotherresearch.Netflixcontemplatesitspricerangeaccordingthelivingcostofeachcountrytoitkeepsanaveragerelationandratiobetweenthecostsofitsplansandbetweenitspricingoffers.Afteranalyzingtheformalandtechnicaldefinitions,aswellastheNetflixexperienceitselfasaresearcherandafontuser,IwouldliketodefineNetflixasa"paid,internationalstreamingandentertainmentserviceapp,whichactsasaselfcontentandthirdpartycontentdistributoraswellasanonlineentertainmentcurator".

24 3.2.SampleDefinitionTheempiricalmaterialcollectionforthisresearchwillbeconductedinSingapore.SingaporeisoneofthelatestkeymarketswhereNetflixhaslaunched,togetherwithSouthKorea,HongKongandTaiwanonitsglobalexpansionplan(Loh,2016,N/A).ThecompanystartedtoexpanditselffirstlyonEnglishspeakingmarkets,suchasUKandAustralia.EuropeandLatinAmericanfollowedafterwards,wherewesternculturalgoodsweremorefrequentlyconsumed.AsiaisthelatestregionNetflixhasdecidedtoexpandin.Singaporelaunchjusthappenedayearago,makingthissampleanew,freshandattractiveonetoresearch.Itwillbeinterestingtoknowhowthestreamingandentertainmentexperiencehaschangedduetothisnewplayerinthemarket,withitsuniquealgorithmtechniqueandbusiness'model.Singaporeisgloballyrenownforbeingaculturalpot:notonlyitisacountrymadeoutaculturalblendfromHindi,MalayandChineseroots,butalsohomeofabigcommunityofex-pats.AccordingtotheDepartmentofStatisticsofSingapore,44%oftheentirepopulationisundertheworkpermitpassduring2016,meaningtheyarenotconsideredpermanentresidentsnorSingaporeancitizens(Statistics,2016).Therefore,Idecidedtotakethisfactintoanadvantageduetoacoupleofreasons:1)Itiseasiertofindaglobalsampleonaninternationalcitystate,wherethenumberofex-patsishighlyavailable,2)Itwouldmakethesampleandtheresultsmoreinterestingtocompare,sinceIwillbeabletograbsamplesfromLatinAmerican,UnitedStates,EuropeandAsiaitself3)Withoutbeinganintention,theexpatswillundoubtedlycomparedtheirpreviousstreamingservice,beingthisoneNetflixornot,andhowtheybehavedwiththosepreviouslymovingabroadtoSingapore.Lastly,thisondemandlogicisworldwidephenomenon,sohavinganinternationalsample,regardlessofitssize,makessense.AsstatedonitsAboutsection,Netflixisthe"worldwideleaderondigitalcontent"since1997(Netflix,2017).Ondifferentlevels,accordingtohowdevelopandwidespreadtheInternetconnectionandusageisspread,theon-demandconsumptionofproducts,especiallyentertainmentisbecomingaglobalnorm.Singapore,beingthepioneerontechhubsworldwide,withexcellentInternetconnectionislandwideandwithagovernmentemphasisonthedevelopmentofthecreativeindustryasadifferentiatorassetforthecountryandtheirprojectcalledSmartNation,seemsliketheperfectplacetodothisresearch.ThisiswhyIbelievethereisanaddedvalueforthisresearchingeneratinga"globalsample".Ofcourse,quantitywisewouldnotberepresentative,butatleastmighthelptogetanindepthandpluralcomprehensionofthephenomena.Candidatesfromthreedifferentcontinents(Asia,AmericasandEurope)willbeanalyzed.

25 Genderisalsoanothervariabletotakeintoconsideration.Isthisphenomenonaffectingequallymalesandfemalesalike?Oristhereagendermorevulnerabletoalgorithmsrecommendationsthattheother?Dotheyexperiencetherecommendationlogicequally?Therefore,sincenationalityandgenderwillbepartofmysample,agewillbecomemyfilterstrategy.Asavant-gardeleadersofchange,Iwouldchooseyouthtoperformthisanalysis.Youth,especiallymillennials,isthegenerationmorepronetochange:theirageandskillallowthemtobemoreflexibleandeasilyadaptable/learnersofnewexperiences.Asmanyofotherconceptsusedonthisresearch,theterm"millennials"alsodrawscontroversy.ThechosendefinitionforthismatterwillbetheoneprovidedbyHoweandStrauss,meaningthegenerationbornbetween1981and2001,ageneration"morenumerous,moreaffluent,bettereducated,techdrivenandmoreethicallydiverse(...)whichfocusesonteamwork,achievement,modestyandgoodconduct"(Howe&Strauss,2000,p.4).Finally,Ialsoconsideredtheusage.Forthissampleandforthemainpurposeoftheresearchquestions,theselectedNetflixusersshouldbeheavyusers.TheyshouldhaveaNetflixaccount-notnecessaryoneoftheirown-andwatchitatleast3timesaweek.Thiswouldprovideasavvyuserapproach,maybenotanexpertone,butatleastcertainkindofuserwhodealswiththeapponusualbasisandisfamiliarwithit. WherecouldIfindinternational,young,Netflix'sheavy-usersmillennials?Aspartoftherecruitmentstrategy,Idecidedtogoandsearchwithinthecreativeindustry,wheremarketers,creative,designersandcontentcreatorsarefound,andwoulddefinitivebeawareandfontofNetflix.Thisisbecause,asactivepro-sumers,theyhelpco-createandaremoretechsavvyduetotheirnaturalinclinationtowardstheindustry,thantheregularusers.Also,itisimportanttodefinetheunderstandingofculturalindustriesusedforthisresearch.Inthiscase,wouldbetheonediscussedbyFlew(2012),inwhichhefocusesontherelationshipbetweenthegoods,orsocall"creativegoods",theirproduction,demandandconsumption."Themorepopulartheculturalformsandpracticesare,themorelikelyitisthattheirproductionandconsumptionistiedupwithwidereconomicstructuresandrelations"(Flew,2012,p.3).So,ifoneconsidersNetflixasoneoftheculturalproductsonecanconcludethatthecreativeindustryconcernstherelationshipsbetweenthoseculturalformsandtheeconomicstructuresonwhichtheyareproduced,distributedandconsumed,whichinthiscasewillbecontentproducers,writers,productionhouses,directors,creatives,artdirectors,programmers,etc.(Flew,2012,p.3).

26 Havingallthesevariablesintoconsideration,onecanbrieflysumupthatthisstudywillfocusonhowtheNetflixalgorithmisperceivedbyusers,donethroughagenderpanelsampleofmillennial,ex-patusers,fromthreedifferentcontinentscurrentlylivinginSingaporeandworkinginthecreativeindustry.Table1:OverviewofthesampleAmorein-depthdescriptionofeachinterviewee'sprofilecanbefoundontheappendix.3.3.Whattowatchnext:amethodologicalbriefintroductionThisresearchcanbestructuredintotwoverydifferentstagesthatallowthereaderabettercomprehensionoftheprocess:firstly,onemustgathertheinformationandsecondly,oneshouldanalyzeandstandardizetheresults.Forbothstages,differentresearchmethodswillbeapplied.3.4.Theresearchprocess:gatheringinformationFortheconsumers'experiencesandjourneys,thinkaloudinterviewsandawalkthroughmethodwillbeused.Thechosenmethodsweretheonesthateasilyadaptedtotheaimsofthisresearch,whilegivinginterestingandrichresults.3.4.1.ThinkAloudInterviewsThethinkaloudinterview"isaresearchmethodinwhichparticipantsspeakaloudanywordsintheirmindastheycompleteatask"(Charters,2003,p.68).Inthiscase,therewasnotaspecifictaskbuttowatchNetflixandtoansweraquestionnaireabouttheirbehaviorwiththeappwhileinteractingwithit.Thethinkaloudinterviewsweredoneontheselectedsample,lastingfortyminuteseach,providingvaluableinformationoneachuser'sexperienceandjourneywhileusingtheapp.Thequestionnairecanbefoundontheannexofthisresearch.Questionswerespecificallycraftedinordertotackleimportantissuesthatwouldberelevantforthisanalysis,consideringoffline

27 behaviors,onlinebehaviorsaswellassocial,technical,algorithmicandmixedvariables,whichwillbediscussedontheanalysis.IpersonallyattendedtowheretheywatchedNetflixandexperienceditwiththem.Iaudiorecordedtheinterviews,tookfieldnotesandsomevisualsaidssuchaspictures,mainlyfromtheirsurroundingsanddevices.Manyoftheintervieweeshadapartnerwhometthesampleselectionandmadethewholeresearchmoreinteresting.Istillkepthalfandhalf:inatotalofeightinterviewees,Ikepttwocouplesandfourindividuals.Thisendedupbeingveryrichsincemanykeyfindingswereactuallygoingagainsttomypreviouspre-conceptionsornotionsofwhatIexpectorthoughtoncouples'behavior.Overall,themeetingswentfromaninformaltoformalapproach:firstlywetalkedaboutlifeandthenwenarrowitdowntothequestionnaireaswellaswatchingNetflix.Thedevicewasalwaysonasasourceofinformationandconsultation.Thefirstoneswerethelongestandtrickiestones,whiletheflowenhancedthroughouttheinterviews.AccordingtothecategoryselectionIhadpreviouslyestablished-whichwillbediscussedindepthonthefollowingchapter-,whatInamedas"offlinebehavior"wasfirstlyanalyzed:contextualinformationsuchaslocation,time,withwhomtheyexperienceit,aswellasstreamingplatformsalternatives.Secondly,Ifocusedonthe"onlinebehavior",withhardandsoftindicatorsalike:thedevices,theamountofaccountsandprofiles,theofferedcontentvs.theycontenttheylikeandtheonetheyactuallywatch,theirtastesandapp'sfunctionality,payingspecialattentiontoNetflix'soriginalcontentwhichwasdefinedasanindicatorofthealgorithm'slogic.3.4.2TheWalkthroughMethodOntheotherhand,inordertounderstandtheNetflixexperienceandcontrastitwiththeexpectedintentorusage,thistechniquewasapplied.Awalkthroughwasutilized,includingandemphasizingitsthreedifferenttechnicalstages:registrationandentry,everydayuseandsuspension,closureandleaving.Thewalkthroughmethodallowstoexaminetheappinterfaceandprovideabetterunderstandingoftheembeddedculturalreferencesshownonitthatguidesandshapesusers'experiences,mergingsymbolicandmaterialculturalobjects(Light,Burgess,&Duguay,2016,p.5-7).Awalkthroughmethodwouldprovidevitalinformationonhow

28 toaddresssomeofthequestionsfortheusers'interviewsandwillprovidehintsfortheprocessofstandardizeandanalyzethequantitativecontentanalysis."Thewalkthroughmethod(...)canbebuiltamoredetailedanalysisofanappintendedpurpose,embeddedculturalmeaningsandimpliedidealusersanduses.Thewalkthroughalsoservesasafoundationforfurtheruser-centeredresearchthatcanidentifyhowusersresistthesearrangementsandappropriateapptechnologyfortheirownpurposes"(Light,Burgess,&Duguay,2016,p.1). Thewalkthroughmethodallowstoexaminetheappinterfaceandprovideabetterunderstandingoftheembeddedculturalreferencesshownonitthatguidesandshapesusers'experiences(Light,Burgess,&Duguay,2016,p.5).So,inordertodoitproperly,notonlyastepbystepobservationanddocumentationoftheappwouldbenecessary,mergingspecifictheoreticalframeworkssuchasScienceandTechnologyStudiesandCulturalStudies-fromwhich,inthiscase,algorithmcultureandtechnicalconceptsanddefinitionswillbeselected-,andtheusers'journeysanalysisdrawnbythequalitativecontentanalysis.Whyisthisnecessary?Because,withbothapproaches,itwouldbeeasierto"identifyconnectionsbetweenthesecontextualelementsandtheapptechnicalinterface(...)andrecognize[its]discursiveandsymbolicrepresentations"(Light,Burgess,&Duguay,2016,p.2-4).Rogers(2013)developsaquestofdefiningadigitalmethodnotion,underliningtheurgeofusing"methodsofthemedium".Thismeansthatthestudyofsocietyandculturalphenomena'sshouldbedoneconsideringthefunctionsandpracticesofeverydaymediadigitaltechnologies,whichundoubtedlyshapeandco-createoursocioculturalbehaviorsandrepresentations,byprovidingdataandfeatures(Rogers,2013,p.3-19). Thethreetechnicalstagesthatthisresearchwillfocusonare: a-Registrationandentry:thisstageiswheretheusereithersignsinorlogsintoaccesstheapp.Duringthisprocess,theappusuallyestablishesandcommunicatesitsgovernance,waysofusage,termsandconditionsandaskstheusertoagreewiththeminordertoproceed(Light,Burgess,&Duguay,2016,p.12).b-Everydayuse:thisstageistheactualusageoftheapp.Thisincludestheanalysisofitsmenu,thenavigation,itsdifferentfeaturesandoptions:theinterfaceandhowitisperceived,thecolorselection,thebriefdescriptions,theminitrailers,etc.Allofthese

29 iconsandkeyfeatureswillbeanalyzedandcontextualizedsincetheycansymbolizedifferenttypesofculturalrepresentation(Light,Burgess,&Duguay,2016,p.14). c-Appssuspension,closureandleaving:whathappenswhentheuserdecidestoleavetheapp,deletetheaccountorinthiscase,cancelthesubscription?Whathappenswiththeuser'sinformationandpersonaldata?Howeasyitistodeletethedata?Isthedatareallydeleted?Alltheseconcernswillbeaddressedunderthisstage (Light,Burgess,&Duguay,2016,p.14-15). Aspartofthewalkthroughmethoditisimportanttostartfromthebeginning.Forthepurposeofthisresearch,Isimulatedsubscribinganewaccountandalsogeneratedanewprofileonmyalreadyexistingusertohaveafirstglanceofwhatanun-altered,virgininterfacewouldlooklikewhentheuserfirstapproachesNetflix.ThefirstcontacttheuserhaswithNetflixisviasomekindofhardware,meaninghowthisstreamingserviceisprovided.ItmaycomewithintheTVasanintegratedapportheusercandownloaditonanydevice:tablets,desktops,portablecomputersandsmartphones.Itworksinbothsystems:AndroidandIOS.ThisissuethatmightseemtrivialisinterestingsincethematerialitywhereNetflixisbeingplayedhelpstoidentify"physicalinteractionsencouragedbytheapp"(Light,Burgess,&Duguay,2016,p.6).Inthiscase:touch,scrollandclickaretheactivitiesthatcanbedone.So,asitsdefinitionexplains,this"ondemand"servicealsoprovidesan"ondemand"physicalexperienceaccordingtothedevicetheuserdecidestowatchit.Forthiswalkthrough,theuseddevicewouldbeaportablecomputer.Anotherimportantissuetoconsideristhatthewalkthroughmethodwasonlydoneonce:specificallymadeunderanewprofileforthepurposeofthisresearch.ThethinkaloudinterviewsincludedthesurfingandwatchingexperienceofNetflix,butthemainfocuswasputontheanswersandnotontheeachuser'sjourney.Fieldnotesandalouddiscussionhappenedregardingtheuserexperienceandsomeofthoseitemsthatwereanalyzedonthewalkthrough,butnotaspecificanddetailedwalkthroughwasdoneforeachoftheeightinterviewees.Thismayseemasadisadvantagetothisresearch,butactuallycontributestothevalidityandrelevanceoftheuserexperience:comparingtheactualusagetotheexpectedoneprovidesinsightfulandinterestingresults.ThemaincontributionofthismethodforthisresearchwasgettingfamiliarwiththeinterfaceandNetflixinfrastructure.Undoubtedly,theseitemswilldeterminetheuserexperienceandthereceptionofthealgorithmicsuggestion.

30 3.5.Theresearchprocess:analyzinginformation3.5.1Qualitative,theme-basedcontentanalysisThroughouttherichnessofthequalitativecontentanalysismethod,thisresearchwillhavethepossibilitytostandardizetheoutputsoftheinterviewees'responsesandwithdrawanalyticallearningsoutofthem.AsexplainedbyPhillipMayring,thistechniqueallowstheresearcher"todescribeproceduresofsystematictextanalysis,whichtrytopreservethestrengthofcontentanalysisincommunicationsciencetodevelopqualitativeprocedureswhicharemethodologicalcontrolled"(Mayring,2000,p.8).Also,thecontentanalysisprovidestheresearcheradeeperunderstandingofthecontent,since,asBeckerandLissmannexplained,differentlevelsofcontentcanbeanalyzedwiththistechnique:themanifestorprimarycontentasthemeandmainidea,butalsocontextualinformationaslatentcontentwhichisequallyrelevant,sinceitcanprovide-inthisparticularNetflixresearch-hintsofuttermeaningsanduserexperiences(Mayring,2000,p.2).Themergeofthisqualitativecontentanalysisonthelightofthewalkthroughmethodfindingswillbetheselectedequationforadeep,meaningfulanalysisonhowtheuserexperiencesthelogicofNetflixalgorithm.Accordingtothepre-establishedtheoreticalframes,thiscontentanalysiscanbeunderstoodundertwomajoranalyticalapproaches:thesocialaspectsandthemoretechnicalones.Eachoneofthemcorrelateswithanoffline(previoustotheapp'slogin)andonlinebehaviorwhich,joinedtogether,canindicateimportantvariablesthatcreatestheNetflixexperience,andtheinterpretation/experiencingofthealgorithmlogicasawhole.Onthelightoftheglobalculturalindustryandthepracticesofeverydaymedialife,thisresearchcanexplorevariablessuchaslocationwheretheuserexperiencetakesplace;time,howlongandhowoftentheexperiencelastandwithwhom,whethertheuserexperienceissharedornot.Otherissuessuchasdevicespreferencesandpractices-whichscreenisthemostselectedonefortheNetflixexperience-(thiswouldhavedirectcorrelationwiththewalkthroughanalysis,wherephysicalandmaterialityanalysiswillbedone)wouldbeanalyzedundertheAppGenerationtheoreticalframeaswellastheAlgorithmCultureandScienceandTechnologyStudies(STS).Lastly,contentselection,curationandrecommendation,aswellasandalgorithmfunctionalitywillbeaddressedwithAlgorithmsCultureandSTStoo.Animportantcue

31 ofNetflix'salgorithmsandessenseis,undoubtedlyitsoriginalcontent,whichrecommendations'willbedoneundertheSTStheoriesaswell.Content,suchasseries,documentariesorfilmsproducedandexclusivelydistributedbyNetflixiswhathelpsthemasabusinessstrategytomaintaintheircompetitiveadvantageontheon-demandstreamingservicesmarket.Theywerethefirstwhonotonlymadeavailableandaffordableanicecuratedamountofonlinecontent,butalsostartedgeneratingtheirveryown-withveryhighquality,aimingtocompetewithTVcableblockbusters-.So,Netflixoriginalcontentwillalsobeanalyzedasanimportantcueandinsighttounveilthehumanexperienceofthealgorithmiclogicbehindtheapp.3.5.2CategoriesOfflineandonlinebehaviorcategoriesarethetwomainobviousdistinctionsonecandrawfromthetheoreticalcontextused.Thesecategorieswillbesubdividedintovariables(social,technical,algorithmicandmix)thatwillprovidebetterandeasierthemestolookforsuchasLocation,Time,Devices,Account,ContentandOriginalsaswellasFunctionality.Onemaywonderwhythesecategories,speciallythosesoftvariables,whicharenotsorelatedtothelogicsuggestionbehindtheapp.But,onecanreachtotheconclusionthat"suddenly,culturalrealityleaksintothestatisticalcleanroomwherealgorithmscountsnothingbutusers,movies,raitingsandtimestamps"(Finn,2017,p.91),makingthemverymuchneededandinterestingassets.Table2:Pre-establishedanalyticalcategories,basedontheoereticalframes.Considerablevariablesandthemestolookfor.Underthecurrentestablishedconditionswherealgorithmsandtechnologyseemstobegoverningtheentirecommunicationsresearchfield,givingahumantwisttothemseemsmorethannecessary.Thiscanbeachievedbyathoughtfulcontrastofa

32 technicalwalkthroughmethodofNetflixagainstthinkaloudinterviewswithparticipantobservationtechniques.Theanalysisandconclusionswillbefilteredandanalyzedthroughaqualitativecontentanalysis,tofinallystandardizeandordertheresults.Tosumup,methodologicallyspeaking,inordertogathertheinformation,thinkaloudinterviewswereperformedtothedefinedsamplewhileanextensivewalkthroughwasdonewithanewprofileonNetflix,createdspecificallyforthisresearch.Ontheotherhand,thegatheredinformationwasanalyzedusingqualitative,theme-based,contentanalysiscategories,tofinallystandardizeandsummarizetheusers'Netflixsuggestionlogicexperience.3.7.LimitationsoftheMethodsItisimportanttoremindthereadertheapproachandintentofthisproject:Idevelopedanin-depth,qualitativeunderstandingoftheuserexperienceoftheNetflixalgorithm.Theselectedsampleisasmall,veryspecificone,whichcouldprovidethoroughinputs.Thiswastheselectedapproach,onethataddressedtheresearchquestionsinamoresuitablewayratherthantohaveaverybigsampleamount,withlittlespacefortheusertoexpressthemselvesindepth.Thesamplenumbercouldbebiggerifaquantitativeapproachwastakenintoconsideration,andprobablyforthatmatteranonlinequestionnairewouldbemoreefficientthanindepth,personalinterquotesdbs_dbs14.pdfusesText_20