putjohncusack

MarketingBulletin,2014,25,Article1ExpertsversusAudience'sOpinionsattheMovies:EvidencefromtheNorthAmericanBox-Office1LeeYoongHonTheimpactofcritics'moviereviewshaserodedinrecenttimeswhileusers'moviereviewshavegainedsignificanceintermsofinfluencingmovie-goers'decisions.
Whilesomemovie-goerscontinuetotrustmoviecritics,othersrelymoreontheirpeersorfriendsforinformationregardingamovie'squality.
SocialnetworkingplatformslikeTwitterandFacebookcontinuetodrivethelatter'sinfluenceandunsurprisingly,users'reviewsarenowincreasinglybeingrecognisedasasignificantfactorinmovieconsumptiondecisions.
UsingNorthAmericanbox-officedata,thispapercomparestheeffectsofon-lineuserreviewsagainstcritics'reviewsonthesuccessofamovieintermsofitsopeningweekend,postopeningweekendandoveralltotalcumulativebox-officegrossing.
Thepaperfoundsignificanteffectsforbothcritics'andusers'reviews,buttheusers'reviewshadstrongereffectsinallthreemeasuresofmoviesuccess.
Theresultsshowthatusers(moviegoers)notonlyhaveagreaterinfluenceonattendancethancritics,butalsoprovidemoreaccuratepredictionsofamovie'sbox-officesuccess.
Keywords:Motionpictureindustry,consumerbehaviour,box-officereceipts,moviecritics'opinions,moviegoers'opinions,socialnetworksIntroductionMotionpicturebox-officereturnsarenotoriouslydifficulttopredict.
Overtheyears,academicandmarketplaceresearchhasfoundvariousfactorsthataccountforrevenue,butthefindingsareinconsistentatbest.
Someofthesefactorsincludethenumberofscreensreleased,starpower,productionandmarketingbudget(seeDeVany&Walls,1999),critics'reviews(seeEliashberg&Shugan,1997;King,2007andGemser&Oostrum,2006),timeofrelease,andmoviegoers'word-of-mouth.
Thelastfactorhasbeendeemedbymanyasincreasinglyinfluentialasthewebcommunity,whomostlyreliedontheinternetnetworkingplatformsforsocialandinformationexchange,getslarger.
GiventheexplosivepopularityofsocialnetworkslikeTwitterandFacebook,theimpactofthe"word-of-mouth"processonmoviesaleshasincreasedandmanyintheindustryareincreasinglytakingnoteofthis.
Infact,suchelectronicword-of-mouthismoreimmediate,thus,insomecases,mayevenmakeorbreakafilmwithin24hours.
RichardCorliss,amoviecriticwithTimeMagazinecallsFridaythenewweekendgiventhepoweroftheinstantmessagingorthe"TwitterEffect"asitispopularlyknown.
Asopposedtotheirpeers'opinions,moviegoersalsoreadcritics'reviewsbeforetheybuymovietickets,butsuchbeliefhasbeenquestionedbymanyresearchersalthoughit1Thispaperwaspresentedatthe2010OxfordBusiness&Economicsconference(OBEC)atOxfordUniversity.
Page1of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1remainssignificanttosomeextent.
King(2007)foundpositivecorrelationbetweencritics'reviewandamovie'sboxofficeperformancetakingintoaccountofonly"bigger"movies,i.
e.
thosethatwerereleasedonmorethan1000screens.
Healsofoundthatproductionbudget,thenumberofopeningscreensandthecritics'reportaccountedformorethanhalfofthevarianceinhissampleofmoviesreleasedinNorthAmericain2003.
TheImpactofUsers'CommentsviaWord-of-MouthManyresearchstudiesdoneonthemotionpictureindustryappeartofindcriticsplayinganinsignificantroleinmoviebox-officereturnsand,surprisingly,recordedsomenegativecorrelationintheirrelationship.
SomehaveattributedthisphenomenontothegrowinginfluenceofsocialnetworkingtoolslikeTwitterandFacebook,whichhavebeenincreasinglyimportantforexchanginginformationandcomments.
Giventheprevailingimportanceofthesetechnologiesthataredominatinghowcommunitiesinteractandsocialise,thepowerofword-of-mouthinfluenceonmovieattendancemayhavegrownmuchstrongerthanithaseverbeen.
However,theadventoftechnologieswouldalsofacilitatealowersearchcostforcritics'reportsonmovies.
Further,suchtechnologicalevolutionwouldalsoensureamorematureandsophisticatedmovieaudienceaswell.
Inaddition,thesourceofinformationformovieconsumptionconsiderationsmayalsoincludetheatricaltrailersandtelevisionadvertisements.
Thus,itisnotsurprisingtoseemoviestudiosinvestinghugesumstopropupthemovie'strailers,withsomeofthese"teasertrailers"appearingevenuptoayearbeforethemovie'srelease.
Thesemarketingploysinmanyinstancesserveastoolstocreateword-of-mouthawarenessandcuriositythatultimatelycarriesamovie'ssuccess.
Sometimesitisnotthemarketingbudgetbutthecreativeapproachthatleadstosuccess;forinstance,thecleverapproachusedbystudiosincreatingthehypesurroundingthelow-budgetsemi-documentaryhorrorpicture,TheBlairWitchProjectultimatelyledtophenomenalbox-officereturnsforthemoviebackin1999,underliningthepoweroftheinternet(seeHowell,2014).
SocialfeedslikeTwitterandthevariousothermessageblogsprovideup-to-dateinformationofthemoviespremieringand,assuch,itisnowonderthattheseusers'commentshaveplayedaprevalentroleinthesuccessorfailureofmovies.
Forinstance,itiswidelybelievedthattheinterestgeneratedthroughthesesocialnetworksledtothesuccessofsomemoviesliketheQuentinTarantino-helmedwarmovie,IngloriousBasterds,whichmanagedtopullinabetter-than-expectedmid-AugustweekendhaulofUS$$38,054,6762.
Anotherexampleofthisphenomenonwasthe2009blockbusterNewMoon(sequeltothe2008hit,Twilight),whichregisteredthethirdhighestweekendgrossofUS$140.
7millionandevenshatteredtherecordforasingle-dayearning,achievingthefeatthroughFridayscreeningswhichcumulatedtoUS$72.
7million3.
Whiletheexistingfanbaseisa2Despitebeingcritically-acclaimed,noneofthefourTarantino-directedmoviesafter1994'sPulpFictionmadeanythingcloseto$US100milliondomestically.
Asof15November2009,IngloriousBasterdshasmade$120,135,237intheNorthAmericamarket(source:www.
boxofficemojo.
com).
3Figuresquotedfromwww.
boxofficemojo.
comPage2of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1significantfactor,itisalsoclearthatthepoweroftheTwittereffecthadastronghandinthiseventastheon-linesocialnetworkcommunitieshavebeenbuildingonitshypewaybeforeitwasevenreleased.
Dataprovidedbywww.
fizziolo.
gy,awebsitethattabulatesthenumberofsocialconversationsonmanyissuesincludingmovies,placedNewMoonintheTop10sinceearlyOctober2009whenthesitebeganoperating.
BythesecondweekofNovember2009,itnotonlytoppedthechartbutalsoachieveditwithhighesteverscorerecordedatthetime.
Unsurprisingly,themovie,whichopenedon20thofNovember2009,shatteredseveralrecordsincludingthebiggestopeningdaygross(US$72.
7million),thebiggestmidnightscreeninggross(US$26.
3million).
Italsoachievedthedistinctionofbeingthethirdbiggestweekendgrossing(US$140.
7million)atthetime.
ThehorrormovieParanormalActivityalsodisplayedpatternssimilartothoseseeninNewMoonasnetworkingwebsitescontributedtoanenormousinterestamongmoviegoers,especiallytheTwittercrowd.
ItwasstillinthetopthreeintheFizziologychartfortheweekending17November2009eventhoughitwasreleasedonthe25thofSeptember.
Infact,theFizziologywebsitewasstillreportingawaveoffansgoingtoseethemovieagain(seewww.
fizziolo.
gy)anditislikelythattheso-calledpositiveon-lineword-of-mouthhadcontributedtothesuccessofthemovie.
ThisfilmhasseenasensationalreturnfromasmallbudgetofUS$15,000andwasoriginallyslatedonlyfora160screen-showing(themoviewouldgoontogross$107,854,596attheNorthAmericanbox-office).
However,theimpactofsocialmediacanalsobeadverse,asseeninthedisastrousoutingofthesummer2009releasedSashaCohencomedyBruno,whichwaspannedbymanyoftheusersthroughthemanysocialnetworkwebsites.
Afterasolidopeningweekendrelease(ittookinsomeUS$30millioninitsfirstweekendofrelease,adecenthaulbysummerreleases'standards)thefilmsubsequentlyfailedbadlyinthefollowingweeks.
Admittedly,thecritics'reportsonthismovieweremostlynegativebutitwasthesensationaldropinitsearningsthatunderlinedthespeedofword-ofmouthinanelectronicera.
4Overtheyears,therehavebeenmanysummermovieswhichhadpoorcritics'reviewsbutmaderespectableearningsespeciallywhengivenasignificantboostintheopeningweekend'shaul,asituationthatclearlydoesnotresonantwiththecaseofBruno's.
Holbrook(1999)foundsignificantdifferencesbetweenpopulartastesandexperts'ratingsinhisstudyusingHomeBox-officeTVchannelviewersurveys(representinguserreviews)andcriticalguidebooks(representingexpertjudgments).
Ifcriticsandusersdifferintheiropinions(astheyarelikelyto,sinceordinarymoviegoerstendtofavourentertainmentthatismorereadilyaccessibleandlessdemandingwhilecriticsareinsteaddrawntomorechallengingandintellectuallystimulatingartworks(seeHolbrook,1999;Bourdieu(1984)),thenwhichofthetwogroupsarelikelytoplayagreaterroleinamoviegoer'sconsumptiondecision4Bruno,whichopenedin2756screensacrossNorthAmericamadeUS$$30,619,130initsfirstweekendbutfell72.
8%and65.
9%respectivelyinitssecondandthirdweekend.
ItwentontotakeuponlyUS$$60,054,530overitsentirerunattheNorthAmericabox-officeandiswidelyregardedbymanyintheindustryasaphenomenathatwasduetothenegativebuzzgeneratedbyTwitter.
SeeCorliss(2009).
Page3of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1However,giventhatevenindustryplayersarenowbeginningtokeepakeeneyeonthephenomenaofusers'commentsviavarioussocialnetworkingwebsites,thesuggestionthatmoviegoersseektheirpeers'approvalbeforeconsumptioncannotbetakenlightly.
AccordingtothepresidentofmarketingatDreamWorksStudios,ChristineBirch,marketerstodayareincreasinglypayingmoreattentiontomoviegoers'commentsintheinternet.
SuchdevelopmentsarenothardtobelieveespeciallygiventhatabiggerproportionofTwitterusers'populationhavegreatermovieconsumptionhabitscomparedtothegeneralpopulation(EcommerceJournal,2009).
LiteratureReviewEchoingthefamousquotebylegendaryscreenwriterWilliamGoldberg,"Withduerespect,nobodyknowsanything",DeVanyandWalls'1999paperalsoacknowledgedthegreatextentofriskintheindustryandfoundnoindividualcausalfactorforamovie'ssuccess.
5Theyhighlightedthecomplexityofmovieasproductsandthecascadingofinformationamongfilmgoersduringtheirshortcourseofrunevolvesalongtoomanypathsthatmakeitimpossibletoattributethetruedeterminingfactorsofamovie'sbox-officereturns.
ConcludingthattherearenosuccessformulasinHollywoodandassertingthenotionoftheaudiencebeingtheultimatedeterminantofamovie'sreturn,theynonethelessconcludedthatamovie'ssuccessisalsodowntoitsquality.
Elsewhere,EliashbergandShugan(1997)foundthatcriticalreviewssignificantlycorrelatewiththelatebox-officerevenueandcumulativeearningsthusdisputingtheconventionalwisdomofcriticsbeingastrongdeterminantonlyintheearlystageofamovie'srun.
Usingweeklybox-officedata,theirpapersuggestedthatcriticsmaybemoreofforecastersofamovie'searningsuccessratherthanamotivating(orotherwise)factortoamovie'saudienceattendance.
Hence,theargumentputforwardbythemwasmoreoftheroleofcriticsratherthantheinfluenceofcritics,i.
e.
,moreofa"predictioneffect"ratherthanan"influenceeffect".
Meanwhile,King(2007)foundnocorrelationbetweencriticalratingsandboxofficegrossingusingreleasesintheNorthAmericain2003.
However,biggerreleases(whichsawopeningsofexceeding1000screens)exhibitedpositiverelationshipsbetweencriticalreviewsandmovierevenues.
Insuchbiggerreleases,productionbudgetandopeningscreennumbersalsopositivelyaffectedthemovies'box-officereturns.
King'sfindingssupportedthoseofEliashbergandShugan(1997)whichfoundgreatercorrelationbetweencritics'opinionstolaterearningsratherthaninitialearnings.
Thisissurprisingbecausegeneralindustrywisdomwouldexpectcritics'reportstohaveagreaterinfluenceonmovierevenueintheearlylife-cycleofthemovie(seeBurzynskiandBayer,1977)as5The"nobodyknowsprinciple"isalsousedtorefertothefilmindustrywhichischaracterizedbysymmetricalignorancebyRichardCaves.
Infact,thehighlevelsofdemanduncertaintywhichmakesitdifficulttoanticipatetheaudience'stastes(i.
e.
audiencesthemselvesmaynotknowwhattheywantuntiltheyseeit)coupledwiththehighdegreesofproductuncertainty(i.
e.
theproducerthemselvescannotguaranteeexactlywhattheaudiencewillgetuntilproductioniscomplete)characterizetheessenceofthefilmindustry.
Formore,seeCaves(2000).
Page4of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1informationfromword-of-mouthislesspotentatthestartofthemovie'sinitialreleaseperiod.
Holbrook(1999),usingpopularitypollsconductedbyHomeBoxOffice(HBO)intheirGuidetoMoviesonVideoCassetteandCableTVin1989,foundthepopularappealandexpertjudgmentssharingsimilartastesalthoughalbeitaweakcorrelation(r=0.
25)hencedisputingtheBourdieu-ledargument(seeBordieu,1984)thatprofessedpopulartastestendtogravitatetowardsmoviesthatareaccessiblebyvirtueoftheirgreaterrealisminrepresentingfamiliarsettingswhileexpertjudgmentsaremoreinfavourofmoviesthataremoreabstract,complexandbasingonlessconventionalvalues.
GemserandOostrum(2007),usingdatafromtheDutchfilmindustry,foundthatconsumersofarthousemoviesareledbyfilmreviewswhenmakingtheirparticularchoiceofmovies;consumersofmainstream-typemovies,incontrast,seekothersourcesofinformationnamelyword-of-mouth,moviestars,and/oradvertising.
Theyfoundthatfilmcritics'takeonan"influencer"roleinthecaseofarthousemoviesbuttheytendtotakea"predictor's"roleinthecaseofmainstreammovies.
Finally,antecedentstudiesoncritics'roleinbox-officereturnsappearedanythingbutconclusive.
SomelikeEliashbergandShugan(1997)donotsupportthenotionofcritics'reporthavinganymotivatingaffectsonfilmattendancealthoughKatzandLazarfeld(1955)interpretedtheirresultsascriticsbeinginfluentialinshapingmovieaudiences'interestsinattendingmovies.
King(2007)meanwhilefoundnocorrelationbetweencriticalratingsandbox-officegrossing.
Nonetheless,someofthestudiesthatusedcumulativebox-officeearningsinsteadofearningsatdifferentpointsinthelife-cycleofthemoviefailtoaccountfortheeffectsofword-of-mouthintheiranalysis.
Intermsoftrackingthecorrelationbetweenusers'on-linesearcheffortsandmoviesuccess,PanaliganandChen's(2013)researchpaperforGooglerevealthatmovie-relatedsearchesonGooglewentupby56%intheperiodof2011-12.
Theyalsofoundsimilartrendpatternswhencomparingbox-officeindexandsearchrelatedindexthusarguingthataconnectioncanbeestablishedbetweensearchactivityandboxoffice.
Moreinterestingly,theyalsofoundthatmoviesearchvolumesevendayspriortoamovie'sreleaseisastrongpredictorofweekendboxofficereceipts.
Similarly,Goeletal(2012)arguedthatsearchqueryvolumecanbeusedtoforecastopeningboxofficegrossing,firstmonthsalesofvideogamesandweeklyrankingofsongs.
Theirstatisticalfindingsonsearchqueryvolumeyieldedthestrongestpredictivepowersinthecaseofopeningweekendgrossingformoviesfollowedbyvideogamessalesandsubsequentlymusicrankings.
Page5of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Basedontheaboveliterature,thisstudyaddressedthefollowingnullhypotheses:H01:Moviecriticsandusershavedivergenttastes.
H02:Critics'opinionsmattermorethanusers'opinionsinthedeterminationofbox-officeoutcomes.
H03:Critics'opinionsaregreaterthanusers'opinionsattheimmediateandearlierstageofthemovie'srun,i.
e.
criticshavebigger"influencer"effect.
H04:Critics'opinionsaregreaterthanusers'opinionsatthelaterstageofthemovie'srun,i.
e.
criticshavebigger"predictor"effect.
MethodThispaperinvestigateswhetherexperts'opinionsorpeers'opinionshavethegreaterinfluenceonaconsumer'schoiceofmovies.
Inordertotestthis,thepaperusesratingsbyon-lineusersandcritics'reports.
TheaverageratingsforeachmoviegivenbythesetwogroupsweresourcedfromtheYahoo!
Movieswebsite.
6Thiswebsitewaschosenbecausetheon-linecontributionsweresubstantialforeachmovie,upto30,000to40,000usersforsomemovies(morecomprehensivethanmostotherwebsites).
However,thenumberofcritics'reportssourcedinthiswebsiteisfewer(inquantity)thansomeotherwebsiteslikewww.
metacritic.
comorwww.
rottentomatoes.
com,bothofwhichalsotabulatethefinaloverallcritics'scoresusingmoremeticulousandstringentcalculations.
Inanyevent,themovetousethesamesourceforinformationbothintermsofcritics'andusers'scoresistoensureaconsistentsourceofinformationsincethebox-officerevenuesdata(fromYahoo!
Movies)arealsoprovidedbythewebsitewww.
boxofficemojo.
com.
Further,datafromthissitearealsousedinthecollectionofthe"scores"pertainingtothepopularityofstars(bothdirectorsandactors),whichisrepresentedinthe"starpower"variable.
ToascertainthereliabilityoftheYahoo!
critics'scores,acorrelationtestwasconductedbetweenthecritics'scoresfromthetwosources(Yahoo!
andMetacritic).
Therewasastrongandpositivecorrelation(r=0.
92;significantat1%)confirmingtheirsimilaritiesandthus,thereliabilityoftheYahoo!
critics'scores.
Thisisbecausethemethodoftabulationofcritics'ratingsinwww.
metacritic.
comissignificantlymoremeticulousandthenumberofcritic'sreviewsbeingsourcedisalsohigherthantheYahoo!
site.
ThecorrelationpatternbetweenYahoo!
critics'scoresandMetacritic'sscoresforthe243moviesinthesampleisshowninFigure1.
6CorrelationtestbetweenYahoouserreviewsandMetacriticuserreviewssawsignificantrelationship(at1%)thuslendingsupporttothevalidityofusingtheYahoousers'datatorepresentusers'opinions.
Page6of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Figure1.
Yahoocritics'movieratingsscoresvsMetacritics'movieratings'scoresInordertotestthesehypotheses,thepresentpaperemploysastatisticalregressionanddescriptiveanalysisinvolvingvariableswithLikertscalescoresthatarerecalculatedtoa100-pointsystemscore.
Forinstance,tomeasureamovie'scriticalvalue,thepaperusestheaveragescoresofmoviecriticsonthemovieasprovidedintheYahoo!
Movieswebsite.
Mostofthemoviesreviewedbythiswebsiteusedcontributionsfromagroupofmoviecriticsandthecritics'reportcardcomesintheformofgradingrangingfromAtoF(note:includingthe+and–categoriesineachofthem,e.
g.
A+orA-).
Giventhatthereare13categories,thescoresarethenrecalculatedtoa100pointssystem,withA+being100whileFtakingonafigureof7.
6(i.
e.
basedonascaleof7.
7foreachcategory).
Suchanapproachissimilartotheproceduredonebythefilmwebsitewww.
metacritic.
com,whichmakesuchcalculationsbasedonthescalerangeoffourtoonestarratingthatiseventuallycalculatedandprocessedaccordingly,forexample100beingfourstars,88being3.
5starsandsoon.
7Meanwhile,consumers'ratingsofthemoviesareused(tabulationssimilartotheoneusedforthecritics)astheiropinionsonthemovieareassumedtobesharedamongtheirpeersthroughelectronicword-of-mouth.
ThispresumablytakesplaceviasocialnetworkingforumslikeTwitter,Facebook,Yahoo!
orFriendster.
Themeasurementofthescoresfromtheconsumers'perspectiveismeasuredbytheaveragescoresobtainedthroughtheYahoo!
Moviessite,asthecommentsthroughTwitterorFacebookabout7However,Metacritic'scalculationsalsoinvolvedamorecomplicatedassessmentbasedoncertainweightsbeingassignedtoeachparticularreviewer(inadditiontoothertypesofassessmentsforsomereviewers,e.
g.
basedonatenpointscaleasopposedtothefour-starscaling)inanunpublishedsubjectivemethodology.
SeeKing(2007)formore.
01020304050607080901000102030405060708090Yahoocritics'scoreMetacritics'scorePage7of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1moviesarenotavailable.
8ThehugequantityofmovieuserresponsesintheYahoo!
Moviessiteprovidesagoodsamplerepresentingtheonlinenetworkingcommunity'sopinionsofthemovies.
Inessence,Yahoo!
hasvarioussocialnetworkingservicesinmanyoftheirproductssuchasMyWeb,Yahoo!
Personals,Flickr,Yahoo!
Buzz.
Inaddition,italsoprovidesinternetcommunicationservicessuchasYahoo!
MailandYahoo!
Messenger.
Thesefactorsjustifytheselectionofitsusers'reviewsasaproxyrepresentativeoftheon-linemoviegoercommunity.
Themoviesexaminedare2008and2009releasesofessentially"popular"movies.
Thispaperselectedasampleofmorepopularmoviessinceitsmainobjectiveistotestthestrengthofusers'comments,i.
e.
,theword-of-mouth(albeitelectronically)versuscritics'views.
Inclusionofmovieswithfewerscreenreleasesorart-housemovies(usuallylimitedintheirreleases)maybesignificantforcritics'reviewsinputsbutgeneralaudiencesarelikelytogivethemamiss.
Inanycase,moviesthatarereleasedonlesserscreensalsosufferfromlesserword-of-mouthpotentialsincetherearesmalleraudiencesbeingcapturedintheseselected(limited)locations.
Furthermore,opinionsbetweencriticsandthegeneralaudiencearealsolikelytodivergeespeciallyinthecaseofart-housepiecesorforeignmoviesordocumentariesascriticsaremorelikelytoappreciateculturaldiversityasopposedtothemasses(seeKing,2007andBourdieu,1984).
Negativereviewsinthecaseofart-housemoviesarenotnecessarilyharmfulforthistypeofmovie;quitethecontraryinfact,asreportedbyGemserandOostrum(2007)ontheimpactofart-housemoviereviews.
Theyfoundthatitisbettertohavenegativereviewsascomparedtohavingnoreviewsatallasitisthesize(ofthereviewarticle)andnumberofreviewsthataredirectlyinfluencingconsumerdecisions.
Inlightofthesefactors,thepaperfocusesonmoviesofmorepopularappealtoallandassuch,thisresentanalysisusesthetop150moviesintheyearof2008and2009.
9Thisfairlyrecentpoolofmoviesalsogivesamoreaccuratesamplerepresentation,giventhaton-linesocialnetworkingeffectisarecentphenomenon.
Finally,theuseofmoviesfrom2008and2009alsoavoidstheproblemsofmakingpriceadjustments,astheaverageticketpricesforbothyearsarethesameespeciallysincepricesofmovieadmissionschangesovertimeasaresultofeconomicandtechnologicalreasons.
10Meanwhile,thesampleofmoviesselectedinthesetwoyearsarethemorepopularones;i.
e.
,thetop150moviesofeachyearalthoughinthecaseof2009,thetop8Commentsandconversationsonmovies(negativeorpositive)canbefoundinawebsitesourcecalledwww.
fizziolo.
gy,thefirstandonlycharttomeasuretheworldofentertainmentusingtheworldofonlineconversation.
Althoughitattemptstomeasuretheworldofentertainmentusingtheworldofonlineconversation,therearestilldatalimitationsasitscharttracksonlytop10discussions,thusnotallmoviesmaybesubjectedtoconversationsandassuch,nomeasurementpointswouldbeavailableforthesemovies.
Inaddition,sinceitisverynew(startedonlyasof8October2009),therearelimitationsofhistoricaldataaswell.
9Thesampleofmoviesfor2009excludesthosereleasedafter1October2009.
Thefiguresobtainedintheyear'stop150movieswereasthosetabulatedasof15November2009atwww.
boxofficemojo.
com10Sourcetakenformthewebsitewww.
boxofficemojom.
com(retrievedDecember2009).
Theaveragepricesofmovieticketin2008and2009arethesameatUS$7.
18.
Page8of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1150asof15November2009.
11ThoseinthelistbutreleasedafterOctober2009wereomittedduetothefactthattheirtheaterrunsmaystillbeinasignificantstageatthetimeofthisstudy.
Nonetheless,sincethebox-officerevenuesformoviesaremostsignificantduringthefirstfewweeksofthemovie'srelease,atimeframeofsixweeks(inthecaseoftheomissionreasoninginthe2009sample)wouldprobablyensuremoviesinthesamplehavingtakenmostoftheirbox-officereturnsbythen.
12Inanyevent,mostmoviecontractrunscallforaminimumof4to8weeks.
Infact,thepaperbyDeVanyandWalls(1999),foundabout35,19and12%ofallbox-officerevenuesareearnedinthefirst,secondandthirdweekofafilm'sreleaserespectively.
ResultsCorrelationPatternsThefindingsinthedescriptivestatisticalanalysisrevealedseveralinterestingresults.
Firstly,thereappearedtobeapositiverelationshipbetweentheusers'scoresandcritics'scores,albeitaweak(r=0.
413)butstatisticallysignificantone.
Thissuggeststhatthegeneralmoviegoer'stastesmaynotbetoodistinctfromthecritics'opinions,afindingincontrasttoBourdieu(1984)whoarguedotherwise,i.
e.
,thatordinaryconsumerspreferredentertainmentthatwasmorereadilyaccessiblewhereascriticstendedtogravitatetowardsworksofhighercomplexity.
Nonetheless,asthepresentpaper'ssampleof243moviesarethetopmoviesfrom2008–2009,theconvergencebetweenusers'andcritics'tastesmaybegreaterinthecaseofpopularmovies.
AsimilarfindingwasalsoreportedbyHolbrook(1999)whoalsofoundaweakbutsignificanttendency(r=0.
25)forpopularappealandexpertjudgmentstoreflectsharedtastesbetweenconsumersandcritics.
Thefindingsalsosupportmanyofthesuggestions(seeCrane,1992;Gans,1974,1999)thatculturaltastesintheU.
S.
arenotalignedtosocialstatus,incontrasttoothercountriessuchasFrance.
Figure2showsthecorrelationbetweenYahoo!
usersandYahoo!
critics'movieratingsscoresofmovies.
ThefindingsrejectthenullhypothesisH01,whichstatesthatthereissignificantdifferencebetweencritics'andusers'tastes.
Thecorrelationwithbox-officegrossingseemstobestrongerinthecaseofusers'scoresascomparedtocritics'scoresinallthedifferentcasesofbox-officeachievements.
Thegeneralperceptionaboutthemotionpictureindustryisthatcritics'scoreshaveagreatercorrelationwithweekendbox-officerevenuebecausetheyareavailablebeforeamovie'spremiere.
11Atthetimeofthewritingofthispaper.
12About65-70%ofallmotionpicturesearntheirmaximumbox-officerevenueintheirfirstweekofrelease(DeVany&Walls,1999).
Page9of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article101020304050607080901000102030405060708090Critics'ScoreUsers'ScoreFigure2.
Yahoocritics'movieratingsscoresandYahoousers'ratingsscoresGiventhatamovie'spriorqualityisnoteasytojudge,onewouldexpectmoviegoerstocheckoutthecritics'reportsbeforemakingtheirdecisions.
Boththeconventionalwisdomandpastliteraturespointtoapositivecorrelationbetweenweekendgrossingandcritics'scores,withthelatteralsohavingastrongerinfluencethanusers'reviewsduringtheearlyrunofthemoviesinceotherinformationlikeword-of-mouthislessavailable(seeBurzynski&Bayer,1977).
Inaddition,withlimitedsignalingpropertiesexceptforthepromotionalorteasertrailers,onewouldexpectmoviegoerstocheckoutthecritics'reportsfirst.
However,thepresentresultsfoundthecorrelationbetweenthecritics'reviewsandvariousbox-officeearningindicatorstobeweak,althoughstillpositivelyrelated.
Infact,thecorrelationisweakestinthecaseofopeningweekendgrossing,thuscontradictingtheconventionalwisdomofcriticshavingthebiggestimpactintheearlylife-cycleofthemovie.
Theweakfindingsforcritics'reviewsinthepresentanalysisofferssomesupportforCaves(2000),whoarguesthat"studiosanticipatingbadreviewswouldhaveratchetupthepromotionsbeforethemovie'sreleaseto"drown-out"thebadcriticalreviews".
Incontrast,userreviewswerefoundinthepresentanalysistohaveastrongercorrelationinallthebox-officeearningindicators,i.
e.
,theopeningweekendgrossing,thetotalcumulativegrossing,thetotalcumulativegrossingminustheopeningweekend'srevenueandtheaveragegrossing(seeTable1).
13Table1.
Correlationbetweenuser'sreviewsandcritics'reviewswithseveralbox-13Acrudecalculationinvolvingthetotalcumulativegrossingdividedwiththemaximumnumberofscreensduringthemovie'stheatricalrunlife-cycle.
Themovetoincludethisistocontrolforthenumberofscreeningsthatisverylikelytoimpactamovie'searningcapacity.
Page10of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1officegrossingachievementsTotalCumulativegrossingWeekendgrossingTotalgrossingminusopeningweekendAveragegrossingUsers'reviews0.
328**0.
196**0.
368**0.
430**Critics'reviews0.
171**0.
0500.
214**0.
321**averagegrossingiscalculatedbydividingthetotalcumulativegrossingwiththefilm'shighestscreencountlevels.
**significanceatp=0.
01.
Thepresentfindingssuggestthatmoviegoersmayrelymoreontheirfellowmoviegoersforopinionsorinformationpertainingtoamovie.
Inthecaseoftheopeningweekendgrossing,theusers'word-of-mouthmaybemuchfastergiventoday'stelecommunicationstechnology(e.
g.
,on-linesocialnetworkingwebsites,textingandsoon)hencemakingusers'commentsmuchstrongerinimpactingamovie'sopeningcomparedtocritics'reports.
Althoughtherearealsomanyotherfactorsinvolved,thefactis,intoday'sworldtheword-of-mouthprocessismuchfasterandup-to-datethanever.
Infact,sincetheusers'reviewsalsohadstrongercorrelationwiththepost-weekendmoviereceipts,itappearsthatmoviegoers'peers'reviewsmaybeastronger"influencer"comparedtocritics.
Inaddition,astheusers'reviewscoresarealsostrongerincorrelationwiththetotaloverallgrossingandaveragegrossing,itappearsthattheyarealsoabetter"predictor"thanthecritics'opinionsaswell.
Perhapsthestudiosshouldnotworryaboutuncooperativecriticsorevenfearnegativereviewsasthereislittle"predictory"and"influencing"effectonmoviegoers.
TheexperienceofmovieslikeTransformers:RevengeoftheFallen,theTwilightsagas,X-MenOrigins:Wolverineandsoon(allreleasedin2009)hasindicatedthatcritics'reviewshadlittleornoeffectonthesemovies'box-officeperformance.
Infact,ParamountdidnotevenscreenthemovieG.
I.
Joeforcriticstoreviewbutoptedinsteadto"lettheaudiencedefinethefilm"(Pomerantz,2009).
Inanyevent,thephenomenaofcriticslosingtheirinfluenceatthebox-officeisnotsomethingrecentandindustryreportspre-Twittererahavepreviouslyindicatedthechangingtrend,forexampleareportbyCopernicus,aglobalmarketingstrategyandresearchfirmin2000putaresoundingnototheanswertowhethercriticsmatterinbox-officereturns(http://www.
copernicusmarketing.
com/about/press11.
shtml).
Table2presentsthecorrelationbetweentheusers'andcritics'reviewsagainsttotal,weekend,andafterweekendtotalbox-officegrossinginaccordancetothreegroups;(1)"big"sizemovies,i.
e.
,moviesthatareraninmorethan3000screens,(2)"medium-size"group,i.
e.
,moviesthatraninbetween1000–3000screensandfinally,(3)"small-size"group,i.
e.
,moviesthathadlowerthan1000screens.
Unsurprisingly,thecorrelationisstrongestinthecaseofthosemoviesingroup1.
Page11of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Table2.
Correlationbetweenuser'sreviewsandcritics'reviewswithseveralbox-officegrossingachievementsbasedonbig,mediumandsmallreleasesmoviesamplesMovieSamplethatopenedinReviewsTotalCumulativegrossingWeekendgrossingTotalgrossingminusopeningweekendGroup1>3000screens(n=95)Users'reviewsCritics'reviews0.
557**0.
501**0.
418**0.
421**0.
585**0.
509**Group2Between1000-2999screens(n=120)Users'reviewsCritics'reviews0.
278**0.
250**-0.
029-0.
1280.
339**0.
342**Group3Lessthan1000screens(n=28)Users'reviewsCritics'reviews-0.
285-0.
076-0.
320-0.
319*-0.
2080.
110**significanceatp=0.
01,*significanceatp=0.
05.
Inthecaseofamoviethatisgivenasubstantiallywiderelease(Group1),thepublicitysurroundingthemoviewouldbemoreextensivebothintermsofcreatingcuriosityandhype.
Thisfactorislikelytoberesponsibleinpushinguptheinterestofmoviegoersinsearchingformoreinformation(e.
g.
critics'reports,users'reviews,askingfriendswhohaveseenitandsoon)aboutthemovieandeventuallyleadingtomovieattendance(orotherwise).
ThecorrelationscoresinTable2arealsoconsistentwiththecorrelationanalysiswhentakingthesampleoverall(seeTable1);i.
e.
,userreviewshavehighercorrelationtothevariousbox-officeindicatorsespeciallyingroupsofmoviesthatarereleasedandraninmorethan1000screens.
Fromtheresults(Table2'sreportingonthecorrelationscoreswhichmostlyindicatestrongercorrelationinthecaseofusers'reviewscomparedtocritics'reviewswiththevariouscategoriesofbox-officerevenues),itappearsthatusers'reviewsarebothastronger"predictor"and"influencer"(indicatingthatitistheword-of-mouthandnottheexperts'opinionsthatmatter)factorinmoviesthatraninmorethan1000screens.
Nonetheless,forsmallerreleases(thosereleasedonlessthan1000screens),thecorrelationswerenotsignificantandweremostlynegative.
Thismaybeduetothemuchsmallersamplesizebutitisalsopossiblethatsmallerreleaseswouldhaveamuchlowerword-of-mouthcapacityaslesserpeoplewouldhaveseenitimmediatelyafteritspremiere.
Page12of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Theotherinterestingresultinthecaseofthesmallerreleasesisthatthenegativecorrelationsaffectboththeusers'andcritics'reviewssuggestingthattheirtastesmaynotbethatthatdistinctlydifferenteveninthecaseofsmallermovies,i.
e.
,thoselessmainstreamHollywoodblockbusterfares.
Althoughthescoresareinsignificant,nonethelessthepatterndoesoffersomesupportinrejectingthenullhypothesisH01thusdisputingtheBourdieu-ledreasoningofculturalhierarchyamongmovieaudiences,i.
e.
theexpertsarelikelytofavourmoreintellectuallystimulatingfarewhiletherestofthemassesflocktomoviesofhighercommercialcontent.
RegressionAstherearevariousfactorsthatcanaffectamovie'sattendancebesidesusers'andexperts'opinions,severalregressionmodelswereperformed.
Inthecaseofregression,boththelinearandlogarithmicmodelsareused.
Intermsoftheotherfactors,theyarelikelytobe:thetimingofrelease(e.
g.
summerandChristmasseasonmoviesarelikelytobringinmorecrowdasaresultoftheholidayseasons);the"celebrity"or"star"factor(i.
e.
,majorbox-officedrawslikeTomCruiseorBradPitthavehugeappeals);thetotalnumberofscreensreleased(asamovie'srevenueishighestinitsfirstweek,thenumberofscreensreleasedisparamountinmanycasestoensurebox-officesuccessalthoughpositiveword-of-mouthmayalsoleadtotheincreaseinthenumberofscreensplayedsubsequentlyaftertheinitialreleasescreenquantityandhencetherevenuesaswell);theproductionbudget14(abiggerbudgetwillincreasetheprobabilitythata"star"isinvolvedorbiggerspecialeffectsandpresumablywillalsoincreasethelikelihoodthatstudioswillputinmoremoneyintothemarketingofthemovieaswell);continuationfrompastwork(i.
e.
,whetherthemovieisasequelorprequel–anymoviethatisbuildingfromsuchpriorworkwillhavetheadvantageofhavingacapturedandloyalaudience),andfinallythemovie'sratingcategory(i.
e.
,whetheritisratedasG,PG13,RandPG).
1514Althoughmarketingcostwillalsobeastrongfactor,itisnotincludedasthedataisdifficulttocollectandunreliableatbest.
Thesamegoestoproductionbudgetbutthewebsitewww.
boxofficemojo.
comprovidedproductionbudgetfor165moviesinoursampleof211(downsizefromtheearliersizeof243moviesaresultofnon-availabilityofuser/critics'ratings).
Therestofthemovies'productionbudgetsweresourcedfromvariouson-linemoviedatawebsites.
Thesourcesincluded(a)http://www.
aceshowbiz.
com,(b)http://movies.
about.
com/od/bruno/a/behind-the-scenes.
htm,(c)www.
wikipedia.
org.
(d)http://www.
the-numbers.
com/movies/2009/STPLY.
php,andfinally(e)http://coffeeandcelluloid.
com/2008/05/07/dinner-and-confessions-of-a-shopaholic.
15Astherearetwoandfourcategoriesrespectivelyinthe"continuationfrompastwork"and"movierating"variables,thereareoneandthreedummiesrepresentingthemrespectivelyintheregression.
Inthecaseof"continuationfrompastwork",thedummywilltakethevalueof1ifthemovieisasequel/prequelwhilethecontrolcategoryinthecaseofthemovieratingisPG.
Page13of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Overall,theindependentvariablesinthemodelare:theusers'ratings,thecritics'ratings,themovie'sproductionbudget,twodummyvariablestoaccountforthetimingofamovie'srelease(i.
e.
,summerseason,Christmasholidaysandothers16),themovie'sproductionbudget,thetotalnumberofscreens,continuationfrompastwork,themovie'srating,andavariabletoaccountfortheinclusion(ornon)ofacrowd-pullingstarpowerindividual(i.
e.
,apopularactor/actress/directorwhocan"guarantee"amovie'sopeningoroverallperformance).
Inthecaseofthe"star"variable,someauthorsusedvariouslistsofpowerfulandpopularmoviestarendorsedbycertainpublicationfortheircodedcalculation;forinstanceDeVanyandWalls(1999)optedtousethelistoftop100mostpowerfulpeopleinHollywoodbyPremieremagazineorJamesUlmer'slistofAandA+peoplewhileHolbrook(1999)codeddummyvariablestorepresent"stars"usingthenumberofmoviesdirected(inthecaseofdirectors)ornumberofmoviesappeared(foractors/actresses)asacriteria.
Thepresentpaper,however,optedtousestarpowermeasurementasdictatedthroughtheon-linehits/visitsofpopularcelebritypagesprovidedinthewww.
boxofficemojo.
comwebsite.
Thiswebsiteprovidestheannualtop100mostpopularcelebritiesintermsofthepageslinkingtothembeingviewed.
Thetablerankseachofthecelebritiesintermsofthepercentageoftotalmoviepagesviewedandprovidestherespectivesaturationscoresforeachcelebrityintermsofpercentageoverthe100.
Themostpopularcelebritywillhavethehighestpercentagescoreandtheirshareofthe100%aregiveninthesaturationscores,e.
g.
RussellCrowetopsin2008withashareof5.
86%whileAngelinaJolierankedfirstayearlaterwithashareof6.
06%.
Eachmoviewhichhasaleadingactororactress(orproducerordirector)inthetop100isassignedthepercentagescorebasedontheirrespectivesaturationscores.
Ifamoviehasmorethanonepersonmakingthetop100list,thehighestscoreamongthepoolof"stars'intheparticularmoviewillbetheoneselected.
Onlythescoreofone"star"(i.
e.
the"star"withthehighestscore)isselectedforeachmovie.
Intermsofthemultipleregressionmodels,threewereperformedforlinearandlogarithmicrespectively,using;16ThesummerseasonmoviesreferstomoviesreleasedbetweenmaytothefirstweekofSeptemberwhiletheChristmasholidaysseasonistakenasbetweenthelastweekofNovember(takingintoaccounttheThanksgivingholidays)throughtothefirstweekofJanuarythefollowingyear.
Page14of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1totalgrossing,weekendgrossingandtotalgrossingaftertheopeningweekendasdependentvariablesrespectivelyfromthelistof209movies.
17TheresultsforthelinearmodelarepresentedinTable3.
Table3.
DeterminantsofTotalGrossing,WeekendGrossingandTotalGrossingafterOpeningweekend:LinearModelTotalGrossingofMoviesWeekendGrossingTotalGrossingafterOpeningweekendUsers'reviews0.
233***0.
125***0.
267***Critics'reviews0.
169***0.
091*0.
194***ProductionBudget0.
345***0.
417***0.
300***TotalReleaseScreens0.
447***0.
393***0.
450***StarAppeal0.
0600.
077*0.
050Sequel0.
0540.
136***0.
018Summer0.
0690.
0210.
085*Christmas-0.
028-0.
123***0.
013G-rating0.
0480.
087**0.
030PG130.
126**0.
203***0.
088R-rating0.
0970.
140**0.
075R20.
7140.
7250.
672AdjustedR20.
6980.
7090.
654StandardisedCoefficients***p<.
01**p<.
05*p<.
10Thefindingsinthelinearregression(seeTable3)revealedthatboththeusers'andcritics'opinionsarepositivelyrelatedtoafilm'sbox-officeearnings.
Besidesconfirmingthesimilarityintastesbetweenexpertsandaudiences,thebiggercoefficients,(thepaperreportsthestandardisedcoefficientsinsteadoftheunstandardised17Thefinalsamplesizeisadjustedfromaninitialamountof243moviesduetonon-availabilityofdataforsomeofthevariables.
Page15of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1onesastheaimisnotaboutmakingpredictionspersebutrather,tryingtoseewhichofthevariablesthathaveaneffectonmovies'success18)inthecaseoftheusers'reviewsinallthreeregressions,indicatethefollowing:firstly,users'opinionsseemedtobestrongerinthecaseoftheweekendgrossingthusunderminingtheconventionalwisdomofcriticswieldingastrongerinfluenceintheearlylifecycleofafilm'sbox-officeandsecondly,itappearsthatusers'opinionsexertgreaterinfluenceinthelaterstagesofthemovie'sperformancethustheproponentsofword-of-mouthpowerappearedtobejustifiedaswell.
19Finally,thestrongercoefficientoftheusers'reviewsasopposedtothecritics'alsoindicateusers'reviewsbeingastrongerpredictorofamovie'soverallsuccessaswell.
Thefirstargumentisrationalisedalongthelinesthattoday'sword-of-mouthprocessgainsmomentumfasterthanitdidpreviouslyasaresultoftechnologicalchangeswhichareabletofacilitateinstantinformationexchangeviavariouson-linesocialnetworkingwebsitesorspecificwebsiteswhichprovideusers'feedback.
Assuch,theimpactofword-of-mouthcanbefeltimmediatelyafterthemovie'spremieredate.
TheTwittereffectasclaimedbymanyindustrypeople,islikelytobeoneofthereasonsresponsibleformanyfilms'suddenchangeinbox-officereceiptswithindays,bothnegativelyandpositively.
Inthecaseofthesecondpoint,itsuggestsaudiencestodayrelymoreontheopinionsofothermoviegoersratherthantheopinionofexperts.
Ifthisistrue,studiosshouldbelessconcernedaboutshowingtheirupcomingreleasestocritics(orspendingtimetryingtoinfluencethem)giventheirlimitedinfluence.
Finally,thethirdfindingsuggestthataudiencesarethebestpredictorofamovie'ssuccess,consistentwiththeargumentbyDeVanyandWalls(1999)whoprofessedthatitistheaudiencethatmakesamovieahit.
Assuch,theusers(audience)havebothstrongerinfluenceandpredictiveeffectsthancriticsbasedontheresults.
TheresultsthusrejectthenullhypothesisH02,asusers'reviewsarefoundtomattermorethancritics'reviewsinthedeterminationofbox-officeoutcomes.
AstheresultsfromtheTable3alsoshowedthatusers'reviewshaveabiggerstandardisedcoefficients(comparedtocritics')inallthethreebox-officeoutcomes,thepaperfindssufficientevidencetorejectboththeremainingnullhypothesesH03andH04–i.
e.
,users'opinionsaregreaterthancriticsatboththeearlierandlaterstagesofthemovie'srunthusconfirmingthatitisbothan"influencer"and"predictor"ofbox-office.
Otherthantheusers'andexperts'opinions,themultipleregressions(linear)alsoconfirmedtheotherdeterminantsofbox-officerevenue,namelytheproductionbudgets,andthetotalnumberofscreensreleased.
Allthreelinearregressionsprovidedtheexpectedpositiveandsignificantcoefficients.
However,starappealisonlysignificant(albeitweak)inthecaseofopeningweekendgrossing.
18Moviesuccessdefinedbythemovie'stotalgrossing,totalgrossingafteropeningweekendandweekendgrossing.
19Toallayfearsofpotentialmulticollinearitybetweenusers'andcritics'scores,allthreelinearregressionswererepeatedwithusersremovedandthenincludedandrepeatedtheexercisewithcritics.
Nosignificantlymajordifferenceswereobservedinthecaseoftheircoefficientsandrespectivesigns.
Theyremainedsignificantinallcases.
Page16of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Onepossibleexplanationforthisisthatmanystudiostodayareoptingtouseless"bigger"namesevenintheirblockbusterfeaturesandprefertoinvestmuchoftheproductionbudgetinexpensiveCGI(Computer-generatedimagery)effects(e.
g.
,"event"moviesliketheblockbusterdisasterflick2012featurednoA-liststarsotherthanthemoderatelyfamousJohnCusackintheleadrole).
Theotherpossibilityisthatstarpowerfactormaybestronglylinkedtotheproductionbudgetfactoranywaysincebigspendingmovieshaveahigherpropensitytofeaturebignames,hencethehugecostbeinganindicatorofbiggerpay-chequesofthestars.
Asforthetimingofrelease,thedummyvariableforsummer(relativetotheomittedcategoryofnon-holidayperiod)isnotsignificantforbothtotalandopeningweekendgrossing20whilethedummyvariableforChristmas(relativetotheomittedcategorynon-holidayperiod)isonlysignificantinthecaseofweekendgrossingandevenhavenegativecoefficientsinatleasttwooftheregressions.
Thisindicatesthatthetimingofreleasebearslessersignificanceintoday'smarket.
AsreportedintheAsianWallStreetJournal(2009),studiosarebeginningtochangetheirways(insteadofrelyingonholidayseasons)onthereasoningthatahitcanhappenanytimeoftheyeargiventherightmarketingstrategyandrightmovie.
21GenerallybothsummerandChristmasreleasesshouldperformbetterinthebox-officeduetotheextendedholidayperiodsandalsothefactthatthestudiosthemselvestendtoreservethebiggermoviesforsummerorChristmasrelease.
Meanwhile,inthecaseofmovieratings,onlythedummyofPG13issignificant(ascomparedtothecontrolgroupofPG-ratings)inthecaseoftotalgrossingthusindicatingthatmoviesratedPG13havehigherlifetimegrossingcomparedtomoviestaggedasPG.
However,allthreeratingcategoriesarepositiveandsignificantinthecaseofopeningweekendgrossing.
Finally,"priorwork"issignificantonlyinthecaseofopeningweekendgrossingindicatingthatsequels/prequels'captiveaudiencecanensuregoodbox-officeopenings.
Toprovidegreaterrobustnessinourusers'reviewsversuscritics'reviewsanalysis,ahierarchicalregressionwasperformed.
Table4providesthethreedifferentregressionblocksfortotalgrossing,startingfromblock1whichconsistsofalltheindependentvariablesexceptusers'andcritics'reviews.
Next,theusers'reviewsvariablewasaddedinthesecondblockregressionandsubsequently,critics'reviewsaswell,inthethird.
TheresultsinTable4showthatusers'reviewsadd8.
4%totheexplainedvariancewhilecritics'reviewsonlyaddedabout1.
8%more.
ThisfindingrejectsthenullhypothesisH02,asusers'ratingsdrivemovieconsumptionmorethancritics'reviews.
20However,asexpected,thesummerseasondummyvariablehadtheexpectedpositivecoefficientandwassignificantinthecasesofthetotalboxofficeandthetotalboxofficeaftertheopeningweekendasstudiosusuallyreservedtheirbiggestblockbustersforthesummerrelease.
21The4thinstallmentoftheFast&theFuriousfranchise(2009)hadthebiggestApril(anon-holidayseason)openingever.
In2010,AliceinWonderlandopenedtoUS$116minanearlyMarchweekendwhileClashoftheTitanstook$US61minanearlyAprilweekendindicatingthatmoviesmaynothavetorelyonholidayseasonstoenjoyhugeweekendopenings.
Page17of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Table4.
DeterminantsofTotalGrossing:HierarchicalModelFirstBlockSecondBlockThirdblockCritics'reviews0.
169***Users'reviews0.
299***0.
233***ProductionBudget0.
458***0.
409***0.
345***TotalReleaseScreens0.
324***0.
378***0.
447***StarAppeal0.
092*0.
0670.
060Sequel0.
0470.
0360.
054Summer0.
085*0.
076*0.
069Christmas0.
001-0.
025-0.
028G-rating0.
0550.
0560.
048PG130.
0830.
119*0.
126**R-rating0.
0730.
131**0.
097R20.
6120.
6960.
714ChangeinR20.
0840.
018StandardisedCoefficients***p<.
01**p<.
05*p<.
10Meanwhile,regressionsusingalogarithmicmodelarealsoperformedtoallowforthepossibilityofnon-linearrelationships(e.
g.
theeffectsoftotalopeningscreensortheproductionbudgetonamovie'sgrossingcouldbeexponential)aswell.
TheresultsarereportedinTable5.
FromthestandardisedcoefficientsinTable5,itappearsthatthereisverylittletoseparatetheusers'reviewsandcritics'reviews(althoughinthecaseofweekendgrossing,thenatureofthecorrelationsisnegativeforboth).
22Giventhatusers'reviewshadsignificantlystrongerunstandardisedcoefficients23,thefindingsthussupportthehypothesisofusershavingnotonlyastronger"prediction"effect,butalsobigger"influencer"effectaswell.
22Thenegativecomparisonheretakesonanabsolutesense,i.
e.
-0.
3039inthecaseofusers'beingbiggerthanthecritics'coefficientof-0.
1664.
Inthissense,theusers'ratingshaveabiggernegativeimpactonmovierevenuecomparedtocritics'reviews.
23Inthecasesoftotalgrossing,weekendgrossingandtotalgrossingafterweekend,theunstandardisedcoefficientsforusersare1.
211,-0.
304and1.
552whileforcritics,0.
768,-0.
166and0.
936.
Page18of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Asexpected,bothtotalscreensandproductionbudgetaresignificantandpositivelyrelatedinallthreemodels.
Thestarappealfactorissignificant(albeitweakat10%level)intotalgrossingandtotalgrossingafterweekendwhilethetimingofreleasesharedsimilarweakresultswiththelinearmodelthusfurtherlendingsupporttotheclaimsandobservationsofchangingstrategiesamongststudiosasfarastimingofreleaseisconcerned.
Table5.
DeterminantsofTotalGrossing,WeekendGrossingandTotalGrossingafterOpeningweekend:LogarithmicModelTotalGrossingofMoviesWeekendGrossingTotalGrossingafterOpeningweekendUsers'reviews0.
182***-0.
0270.
231***Critics'reviews0.
200***-0.
0260.
242***ProductionBudget0.
110**0.
233***0.
104*TotalReleaseScreens0.
688***0.
611***0.
612***StarAppeal0.
074*0.
0340.
079*Sequel0.
116***0.
086**0.
092**Summer0.
045-0.
0110.
067Christmas0.
041-0.
1810.
087*G-rating0.
0660.
0630.
061PG130.
0850.
106*0.
062R-rating0.
009-0.
24-0.
022R20.
6970.
7310.
638AdjustedR20.
6800.
7160.
618StandardisedCoefficients***p<.
01**p<.
05*p<.
10Comparedtothelinearmodel,movieratingsalsorecordedevenweakerfindings.
Meanwhile,continuationfrompriorworkregisteredsignificantfindingsinallthreecases.
Assequelsandprequelshaveacapturedandloyalaudience,thisfindingisinlinewiththeconventionalassumptionsonthisfactor.
Overall,asfarasthedebateofusersversuscritics'opinion,findingsfromthelogarithmicmodelsalsorejectthenullhypothesesofH02,H03andH04.
Basedontheoverallfindings,theseresultssuggestthatintoday'smotionpicturemarkets,itistheconsumerswhowillultimatelydictatethesuccessofamoviebothintermsofitsopeninganditsfinaltallypostcinematicrun.
Page19of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1ConclusionsWhilecriticshavealwaysplayedpivotalrolesintheartsandthemediasector(i.
e.
,movies,music,theaterplaysandsoon),theirinfluenceonthesesectorshavebeenincreasinglydilutedasaresultofmanyfactors.
Consumers,intheirquestforacquiringinformationpertainingtothequalityofsuchgoods,havevarioussourcesthatcouldprovidethenecessarysignalsofquality.
Inthecaseofthemovieindustry,oneofthesignalsofqualityoffilmsistheexperts'reviewswhichareusuallyprovidedinthemediabeforeamovie'srelease.
Infact,moststudiosscreentheirlatestmoviestocriticsbeforethemovies'premiereintheaters.
Thestrongercorrelationbetweenusers'opinionswithmovieperformanceascomparedtocritics'opinionsrevealedinthispaperisnotentirelysurprisinggiventhechangesinthemarketandsocialenvironmentandalsobasedonthefindingsfrompaststudies.
King's2007paperhadfoundalackofcorrelationbetweencriticalratingsandbox-officeearningswhileDeVanyandWallls'1999studyalsodidnotprovidesupportfortheroleofcriticsindictatingamovie'schancesofbeinga"hit".
Word-of-mouthamongconsumershaveinrecenttimesbecomeincreasinglymoreinfluential,especiallygiventhefactthattechnologicaladvancementhasacceleratedtheprocessandalsosignificantlyexpandedsocialnetworklinksthroughon-linefeedslikeTwitterandFacebook.
Thesesocialnetworkwebsitesarepossiblythereasonthathavecontributedtoundermininganddrowningouttheinfluenceofexperts'reportsastheyprovidetheplatformforstrong(andfast)word-of-mouth.
Thepresentpaperfoundusers'opinionstomattermorethancritics'judgmentsinbothinfluencingmovieattendanceandpredictingamovie'ssuccess.
24Thisisespeciallytrueforbigreleases,i.
e.
moviesthatarereleasedinmorethan1000screens,presumablybecausetheword-of-mouthismuchgreaterasaresultofmorepeoplehavingseenitimmediatelyafteritspremiere.
Inaddition,theon-lineinterestgeneratedthroughtheon-linesocialcommunityislikelytogetabiggerboostinbiggermoviessincetheinterestandcuriositylevelsaremuchhigherinthesemoviesastheyarelikelytobetherecipientsofmuchbiggermarketingcampaigns.
Asthepresentsampleincludedonlythetopmoviesfrom2008and2009,itisnotcertainwhethertheeffectsofcriticsgetdrownedoutbyusers'commentsinthecaseoflimitedreleasestypeofmoviesaswell,i.
e.
documentaries,foreignfilmsandsoon.
Fromthelimitednumberofsmallermoviesinthispaper'ssample,theimpactofbothcriticsandusersopinionshavehadnosignificantimpact.
24Lionsgate,thestudiobehindtherecentsummermovie"Killers"didnotscreenthemovieforcriticspriortothemovie'spremiereandinsteadworkonastrategythatallowtheviewerstoassessthefilm,presumablythroughwritingaboutitonTwitterorFacebook(Lemire,2010).
Page20of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Finally,theotherinterestingfindingsfromthispaperarethepossiblechangingstrategiesofstudiosintermsoftheuseof"starpower"andthe"timingofrelease",inbothcasessuggestingadivergencefromtheusualpracticesofthepastsbymoviestudios.
Inconclusion,moviestudiosmayneedtopaymoreheedtousers'opinionsratherthanspendingtoomuchtimeworryingaboutcritics'assessments.
Inthissense,thereactionsandcommentsfromthosepre-releasescreeningstoselectedaudiencesmayprovetobeverycrucialinprovidingthenecessaryinputsforfurthereditsandcuts.
Finally,itmayalsopaytodownplaythesignificanceofcritics'reactionsfromtheirscreenings(presumablybeforethemovie'srelease)andstudioexecutivesshoulddevotefewerresourcestomakingcontactswithcriticsorattemptingtoinfluencetheirviews.
ReferencesBourdieu,P.
(1984).
Distinction.
Cambridge,MA:HarvardUniversityPress.
Burzynski,M.
&D.
Bayer(1977).
Theeffectofpositiveandnegativepriorinformationonmotionpictureappreciation.
JournalofSocialPsychology,101,215-218.
Caves,R.
E.
(2000).
CreativeIndustries:Contractsbetweenartandcommerce.
Cambridge,MA:HarvardUniversityPress.
Corliss,R.
(2009).
Box-officeweekend:Brunoaone-daywonder.
http://www.
time.
com/time/printout/0,8816,1910059,00.
html,Accessed24November2009.
Copernicus(2000,March28).
Moviecriticreviewshavenoimpactonboxofficesuccess,saysnewCopernicusstudy.
Accessed24November2009)http://www.
copernicusmarketing.
com/about/press11.
shtml.
Crane,D.
(1992).
HighCultureversusPopularCultureRevisited.
InLamontandFournier,eds.
,(1992),CultivatingDifferences:SymbolicBoundariesandtheMakingofInequality,vii-xii,Chicago:UniversityofChicagoPresspp.
58-74,.
DeVany,A.
&W.
D.
Walls(1999).
Uncertaintyinthemovieindustry:DoesstarpowerreducetheterroroftheboxofficeJournalofCulturalEconomics,23,285–318.
EcommerceJournal(2009,October15)FacebookandTwittergreatlyinfluenceboxofficesalesofmoviemakingindustry,http://www.
ecommerce-journal.
com/news/24641_facebook-and-twitter-greatly-influence-box-office-sales-moviemaking-industrydrgn=1,Accessed14November2009.
Eliashberg,J.
,&Shugan,S.
M.
(1997).
Moviecritics:InfluencersorpredictorsJournalofMarketing,61(April1997),68-78.
Gans,H.
J.
(1974),PopularCultureandHighCulture,NewYork:Basic.
Gans,H.
J.
(1992),Preface,InLamontandFournier,eds.
(1992)CultivatingDifferences:SymbolicBoundariesandtheMakingofInequality,vii-xii,Chicago:UniversityofChicagoPress.
Page21of22http://marketing-bulletin.
massey.
ac.
nzMarketingBulletin,2014,25,Article1Gemser,G.
&Oostrum,M.
(2007),Theimpactoffilmreviewsonthebox-officeperformanceofarthouseversusmainstreammotionpictures.
JournalofCulturalEconomics,31,43-63.
Gray,B.
(2009).
NewMoonshattersopeningdayrecord.
(http://boxofficemojo.
com/news/id=2626&p=.
html,Accessed10December2009.
Gray,B.
(2009).
WeekendReport:NewMoonopenswithdeafeningowl,http://boxofficemojo.
com/news/id=2627&p=.
html,Accessed10December2009)Goes,S.
,Hofman,J.
M.
,Lahale,S.
,Pennock,D.
M.
&Watts,D.
J.
(2010).
PredictingconsumerbehaviorwithWebsearch.
ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica(PNAS),107(41)pp.
17486-17490d.
o.
i.
:10.
1073/pnas.
1005962107.
Accessed25October2014).
Holbrook,M.
B.
(1999).
Popularappealversusexpertjudgmentsofmotionpictures.
TheJournalofConsumerResearch,26(2),144-155.
Howell,P.
(2014).
HowtheTheBlairWitchProjectbecameablockbuster.
TheHamiltonSpectator,http://www.
thespec.
com/whatson-story/4617028-how-the-the-blair-witch-project-became-a-blockbuster/,accessed23November2014).
Katz,E.
&Lazarsfeld,P.
(1955).
PersonalInfluence.
Part2,Chapter13,Glencoe,IL:TheFreePress.
King,T.
(2007).
DoesfilmcriticismaffectboxofficeearningsEvidencefrommoviesreleasedintheU.
S.
in2003.
JournalofCulturalEconomics,31(3).
Lemire,C.
(2010).
Killers'reviewsnotalloweduntilopeningday.
http://www.
huffingtonpost.
com/2010/05/27/killers-reviews-not-allow_n_592206.
html,Accessed03March2011)Panaligan,R.
&Chen,A.
(2013).
QuantifyingmoviemagicwithGoogleSearch.
WhitePaperatGoogleThink(http://www.
google.
com.
au/think/research-studies/quantifying-movie-magic.
html,accessedon25October2014)Pomerantz,D.
(2009).
MovieCriticsvs.
TheAudience:DoesfilmcriticismstillplayaroleinaTwitterworldhttp://www.
forbes.
com/2009/08/27/movies-twitter-matt-atchity-business-entertainment-tomatoes.
html,Accessed14November2009)Schuker,L.
(2009).
Latestblockbusterdefiestheseason.
TheWallStreetJournal,April62009,http://online.
wsj.
com/article/SB123895426861490419.
html,Accessed10December2009)Walls,D.
W.
(2005).
Modelingmoviesuccesswhen'NobodyKnowsAnything':Conditionalstable-distributionanalysisoffilmreturns.
JournalofCulturalEconomics,29:177-190.
LeeYoongHonisaSeniorUniversityTeachingFellowattheUniversityofNottingham,Malaysiacampus.
Page22of22http://marketing-bulletin.
massey.
ac.
nz