MRC Social Measurement Guidelines

MRCSocialMediaMeasurementGuidelinesNovember2,2015version1.0-Final

KeySponsoringAssociations:AmericanAssociationofAdvertisingAgencies(4A’s)InteractiveAdvertisingBureau(IAB)WordofMouthMarketingAssociation(WOMMA)

SocialMediaMeasurementGuidelinesv1.0FINAL November2,2015

2

TableofContents

Overview ....................................................................................................... 51.1 GoalofGuidelines .................................................................................. 5

1.2 DevelopmentProcess ............................................................................. 5

1.3 DevelopmentTeam ................................................................................ 6

1.4 ScopeandApplicability ........................................................................... 7

1.5 RevisionProcess ..................................................................................... 7

1.6 Definitions ............................................................................................. 8

2 GeneralMeasurement ............................................................................ 112.1 Authorship ........................................................................................... 12

2.2 Engagement ........................................................................................ 13

2.2.1 Interaction ...................................................................................................... 142.2.2 ContentRedistribution: .................................................................................... 152.2.3 AdvocacyandInfluence: ................................................................................... 16

2.3 Reach .................................................................................................. 18

2.4 CostMetrics ........................................................................................ 19

2.5 DemographicsandTargeting ................................................................ 19

3 SocialMediaCoverageandBasisforProjection ........................................ 193.1 Internet/MobileInternetUsers–Universe .............................................. 19

3.2 CompatibilityorMeasurementLimitations ............................................. 20

3.3 TrackingSocialMediaExposuresOutsideofApplications ......................... 20

3.4 SocialMediaApplications ..................................................................... 21

3.5 UsageRestrictions,IfAny ...................................................................... 24

4 SocialMediaActivityMeasurementSpecifics ........................................... 244.1 TrackingofUsersAccessingContent–TechnicalDetails .......................... 24


3

4.1.1 ClientSide(andViewable): ............................................................................... 254.1.2 TrackingMethod ............................................................................................. 254.1.3 ProjectionorOtherEstimationMethods,IncludingInferences ............................. 254.1.4 MetricsandKeyAttributes ............................................................................... 264.1.5 DifferencesbyPlatform .................................................................................... 27

4.2 UserGeneratedContent(UGC)Analysis–TechnicalDetails ..................... 27

4.2.1 TrackingMethod ............................................................................................. 274.2.2 ThreadedContent,OverlappingContentandDownstreamEditing ....................... 284.2.3 ExcludingNon-RelevantInformation ................................................................. 29

4.3 Video .................................................................................................. 30

4.4 FrequencyCapping ............................................................................... 31

5 EnhancingSocialMediaMeasurementAccuracy ....................................... 315.1 CacheBusting ...................................................................................... 31

5.2 FiltrationforKnownInvalidActivity ....................................................... 31

5.3 SegregationofPre-FetchandPre-RenderActivity .................................... 32

5.4 Auto-InitiatedContent,OtherNon-UserIntendedContent ....................... 32

5.5 Applications–On-Linevs.Off-LineActivity ............................................. 32

5.6 InactivityConsiderationsandLimits ....................................................... 33

6 GeneralReportingParameters ................................................................. 336.1 GeneralParameters ............................................................................. 33

6.2 Time ................................................................................................... 33

6.3 Location .............................................................................................. 34

6.3.1 TraditionalGeo-LocationMethods .................................................................... 346.3.2 ApplicationLocationServices ............................................................................ 346.3.3 HybridandEmergingLocationTechniques ......................................................... 35

6.4 SegregatingNon-LikeAdContentorActivity ........................................... 35

7 DisclosureGuidance ................................................................................ 357.1 General ............................................................................................... 35

7.2 DataOwnershipDisclosures .................................................................. 37


4

8 AuditingGuidelines ................................................................................. 378.1 General ............................................................................................... 37

8.2 U.S.CertificationRecommendation ........................................................ 38

9 WhoWeAreandContact ........................................................................ 39


5

Overview

ThisdocumentpresentsaguidelineforthemeasurementofcertaintypesofSocialMediaactivityconductedbyconsumerswhousemedia.Forpurposesofthisdocument,SocialMediaisdefinedascreated,sharedorexchangedinformation,ideas,contentandopinionsinvirtualcommunitiesandonlinenetworks,includingconversationsaboutorwithabrand.ThedocumentwaspreparedfortheuseandbenefitofthemediaIndustry,especiallythoseconstituentsthatanalyzebehaviororuser-generated/postedinformation,measureadvertisingandthosethatattempttomonetizeevents(whetherbuyerorseller)inaSocialMediaenvironment.

Furthermore,forpurposesofthisdocument,aSocialMediaPlatformisdefinedasamediavehicle,site(asawholeorinpart),app,widgetorothermediamechanismthathasasocialorientation;specifically,thatcapturesusercommentaryandfacilitatessharingofinformationamongusersofadefinednetwork–i.e.,representingachannelforsharinginformation,opinionsorexperiencesfromusersofthatplatform–orallowscross-usercollaboration

1.1 GoalofGuidelines

Thisdocumentestablishesadetailedsetofmethods,definitionsandcommonpracticesforentitiesthatmeasureSocialMediaactivity.Guidelinesareintendedtoestablishanddocumentgoodpracticesofmeasurement;improvepracticesanddisclosuresusedbypractitioners;andalsoprovideeducationtousersofSocialMediadatafromallsegmentsoftheIndustry.ThisdocumentalsoestablishesarecommendationandabenchmarkforauditprocesseswherebythepracticesanddisclosuresofSocialMediameasurementorganizationscanbevalidatedbythirdparties.

1.2 DevelopmentProcess

TheguidelinescontainedinthisdocumentoriginatedfromaprojectledbytheMediaRatingCouncil(MRC)withsponsorshipandsupportbytheAmericanAssociationofAdvertisingAgencies(4A’s),InteractiveAdvertisingBureau(IAB)andWordofMouthMarketingAssociation(WOMMA),withtheparticipationofalargegroupofSocialMediavendors,publishers,measurementorganizations,ad-servers,ad-networksandotherinterestedorganizations.Theseguidelineswerealsoreviewedbymajorbuyer-sidetradeorganizations(ANA)andtheirconstituentsandthereafterprovidedtothepublicthroughaformalperiodofpubliccommentpriortoformaladoption.


6

Thefinalguidelineisexpectedtobepublishedandavailableonthewebsitesofthe4A’s,IABandWOMMA,andwillbere-assessedperiodicallytoensureitremainsapplicableovertime.

1.3 DevelopmentTeam

Associations–PrimarySponsors/FacilitatorsinadditiontotheMediaRatingCouncil(MRC)AmericanAssociationofAdvertisingAgencies(4A’s)InteractiveAdvertisingBureau(IAB)WordofMouthMarketingAssociation(WOMMA)ParticipatingOrganizations

140Proof Ernst&Young(EY) PricewaterhouseCoopers(PwC)360i ESPN PubmaticAddThis eyeReturnMarketing PurchAdMaster Facebook QuintlyAdsNative GeddesAnalytics SapientNitroAdvertisingResearchFoundation(ARF) GfK SASInstituteAllen&Gerritsen HavasWorldwide ScrippsNetworkInteractiveAllianceforAuditedMedia(AAM) IDGCommunications Shareablee,Inc.Allrecipes.com IMServices ShareThisAmericanExpress IPRMeasurementCommission SocialbakersAssociationofAmericanAdvertisers(ANA) JEGI SonyPicturesTelevisionBigValleyMarketing KantarMedia SpredfastBPA KellerFay TheDailyNewsBrandwatch KPMG TheNewYorkTimesCompanyBuzzfeed LinkedIn TimeInc.Centro ListenFirstMedia TrackMavenChartbeat MansuetoVentures TravelersCMP.LY Mashable TwitterConclaveonSocialMediaMeasurement MECGlobal UnmetricConverseon Merkle UnivisionCrowdtap MomentumWorld UnrulyMediaCrosbyMarketing Nativo U.S.BankC-SuiteCommunications NeoReach VideoAmpCyberAlert NBCUniversal VMLDeloitte Nielsen WWEDoubleVerify Ogilvy Yahoo!Eleven OnIdeas ZócaloGroupEnergyBBDO Ernst&Young(EY) ZuumEngagementLabs PainePublishing


7

1.4 ScopeandApplicability

TheseguidelinesareintendedtocoverthemethodsusedformeasurementofcertaintypesofSocialMediaactivityincluding:(1)tracking/countingusersaccessingSocialMediaorrelevantcontentwithinSocialMedia,(2)InteractionorEngagementwithSocialMediacontent,and(3)consumerlisteningandconsumersentiment.Activitygeneratedondevicesofalltypes–personalcomputers,mobiledevices,tablets,etc.–areconsideredherein.ThisdocumentisprincipallyapplicabletoSocialMediaPlatformsorUserGeneratedContent(UGC)collectors(thatproduceorenablemeasurementoftheirplatformactivity),thirdpartySocialMediameasurersoraggregators,Internetandmobilecompanies,ad-servingorganizationsandothergeneratorsofSocialMediaincidenceinformationandisintendedasaguidetoacceptedpractice.Thesites,applicationsandplatformsforwhichthisdocumentisintendedgenerallyincludethepresenceofsomeformofadvertisingorbrandedcontentorusercommentary(UGC)abouteither.Inthecontextofthisdocument,UGCgenerallyreferstoorganic(definedhereasholisticallysourcedcontentcreatedbyusersandnotpaid)contentgeneratedbypeersorusersofSocialMediaPlatforms.OwnedSocialMediacontentandPaid/SponsoredUGCisspecificallyaddressedseparatelywithinthisdocument.Thisdocumentexcludes“top-down”sources(professionalInfluencers)suchasbloggerswithtopicinfluence.WhileblogsareaviableandvalidaspectofSocialMediaandtheseguidelinescanandshouldbeappliedtoblogmeasurement,uniqueaspectsofblogssuchascontentsourcingandsubscriptionswerenotconsideredherein.SocialMediaInfluencers,asdefinedlaterinthisdocument,areincludedherein.Additionally,marketers,advertisingplannersandbuyerscanusethisdocumenttoassistindetermininghowaccuratelymeasurementparametersareexecuted.Measurementtechniquesbasedonreal-timedatacollectionaswellasperiodic/batchdatacollectionandanalysesareconsideredherein.

1.5 RevisionProcess

TheMRCintendstoestablishanongoingprocesstorefineandmaintaintheseguidelinesinthefuturedrivenbyinformationgatheredfromauditactivitiesaswellasindustryandparticipantfeedback.Weintendforthistotakeplaceatleastannually.


8

1.6 Definitions

Authorship:Theoriginalpieceofcontentor,originalactionunrelatedtooriginalcontent(exposuremetricsthatmeasuregeneraluseofSocialMediaPlatforms,withoutengagement).Auto-InitiatedActivity:Activitynotrequiringdirectuserinvolvementincludingauto-playcontentanduser-configuredautomatedactivity.Auto-Initiatedeventsshouldbesegregatedanddescribedforreportingpurposes.BrandAdvocacy:ActivitywherebyausercreatesfavorableUGCaboutabrandorproduct,andthenpassesonpositivemessagesaboutthebrandtootherusers(suchasinarecommendation).ClientSideCounting/UserInitiatedEvents:Countingbasedoneventsoriginatingfromtheuser’sbrowserorothermediaoperatedbytheuserwhichareclosesttotheactualSocialMediaeventbeingmeasured(forexample,countinganeventupontheentryofUGCmeetingrelevantcriteria),whicharefilteredtoexcludeknowninvalidactivity.ContentRedistribution:IntentionalanddirectinteractionwithintentofredistributionorpromotionofcontentoradvertisingbyindividualusersofSocialMediaPlatformstootherusersofSocialMediaplatforms,eitherindividually,toasubsetofusers(friends,followers,etc.)oracrosstheuserbase.Conversion:Thepointatwhicharecipientofanadvertisementperformsadesiredactionsuchasaclick,openingofafileoremail,fillingoutaform,navigatingthroughtoapageormakingapurchase.EarnedMedia:User(consumer)generatedactivitybasedonaccessing,watching,listening,readingorotherwiserespondingtomediaorabrand/productexperience.Engagement:Aspectrumofconsumeradvertisingactivitiesandexperiences(interactionsandinterest)—cognitive,emotional,andphysical.Engagementassumesactiveparticipation,butmayalsodescribeacognitiveoremotionalconnection(suchasinSocialListeningandMonitoringmetricsdescribedbelow),inadditiontoaphysicalaction.Imputation:DatarelevanttoaSocialMediameasuredeventthatdidnotoriginatefromauser,butwasinsteadinferredbasedonotherinformationassociatedwiththateventortheuserorrelatedusers.Imputeddataoriginatesfromthemeasurementorganization,nottheuser.Imputationcanbeareasonableandacceptedstatisticalprocessbuttheassociatedinferencesanddataprocessesshouldbeexaminedforempiricalsupport.Influencers:Userswiththeabilitytoreachotherusersoraffectotherusers’thinkinginasocialonlinecommunity,generallydefinedbythenumberofdirectlyconnectedusersorthesizeofthenetworkofusersassociatedwithanaccount,userorprofile.Aperson,grouporentitywith


9

theabilitytoreachandaffectanotherpersonorgroupofpeople’sthinkingandbehaviorduetoattributessuchasexperience,expertise,reputationandsocialfootprint.Interaction:Interactionmetricsincludemeasurementofindividualuserinteraction(generallyquantifiablephysicalormechanicalactionswithcontent)withoriginalcontentoradvertisingonSocialMediaPlatforms.SuchinteractionsmaybeInternal(visibleonlytotheplatformormeasurementorganization),orPublic(visibletooneormoreusersoftheplatform,therebyextendingvisibility).MetricsUnit(Word,Sentence,Post,etc.):Themeasurementbaseforaconsumerlisteningevent.Thisbaseshouldbespecifiedinmeasurementreports.Opinion:Avieworjudgmentformedaboutsomething,notnecessarilybasedonfactorknowledge.Organic:Holisticallysourcedcontentcreatedbyusersandnotpaidorincentivizedbyabrandorothermarketingorganization.OrganicReach:Totaluniquecountofusersexposedtoorganic(unpaid)contentthatincludesonlythoseactualusersexposedtocontentinitsoriginalformat(withinageneralfeed,ticker,page,etc.).OwnedMedia:AmediachannelthatisOwned,andbyimplicationcontrolled,byamediaoutletorbrand;forexample,awebsite,page,account,station(radioortelevision)orapp.OwnedMediashouldbedefinedbytheverticalorcontextofthemeasurement.PaidMedia:Abrandorothermarketingorganizationpaystouseanownedmediaoutlet–ingeneral,advertising,messagingorbrandsponsorshipthatappearsonownedmedia.MayalsobereferredtoasSponsoredMediaornativeadvertisingcontent.PaidReach:Totaluniquecountofusersexposedtopaidcontentoradvertisingthatincludesonlythoseactualusersexposedtocontentinitsoriginalformat(withinageneralfeed,ticker,page,etc.).PotentialReach:ProjectedReach(totalcountofuniqueusers)basedonacountofasubsetofplatformusersoranetworksuchasfriends,followersorfans.PotentialReachshouldbelimitedtouseasaplanningmetriconly(notacurrencymeasurement),however,properlydisclosedandsupportedestimatedactualReachmaybeusedasanevaluationmetricwhereappropriate.Scraping:HarvestingordataextractionwherebysoftwaretechniquesareusedtoextractHTML,non-APIinformationandcontentfrompages,platformsorotherpropertieswherepermissibleintermsandconditions.


10

Sentiment:Avieworattitudetowardasituation,experience,event,productorbrand;afeeling,senseoremotionexpressedbyaSocialMediauser.SentimentisacomponentofOpinionandBrandAdvocacy.Sentimentisthefeelingtheauthoristryingtoconvey,oftenmeasuredthroughcontextsurroundingthecharacterizationofanobject.SocialMedia:Created,sharedorexchangedinformation,ideas,contentandopinionsinvirtualcommunitiesandonlinenetworks,includingconversationsaboutorwithabrand.SocialMediaPlatform:Amediavehicle,site(asawholeorinpart),app,widgetorothermediamechanismthathasasocialorientation;specifically,thatcapturesusercommentaryandfacilitatessharingofinformationamongusersofadefinednetwork–i.e.,representingachannelforsharinginformation,opinionsorexperiencesfromusersofthatplatform–orallowscross-usercollaboration.SponsoredMedia:Mediainfluencedbyabrandthroughpaymentorotherincentive,butwherethemessageisnotcontrolledasdirectlyasabrandcreatedadvertisement.Targeting:Technicaloptionsofferedtoadvertiserstoputaspecificadplacementandadtypeinfrontofadefinedaudience.TargetingoptionswithintheSocialMediaenvironmentcanincludelocation,demographics,interestsandbehaviorslikepurchaseactivity.TheoreticalReach:ModeledReach(totalcountofuniqueusers)basedonagrossmultiplierofusers(suchasprojectingpost-addeliverybasedonagrossmultiplierrepresentingsomeestimateofpotentialexposure).TheoreticalReachshouldgenerallybeavoided,particularlyasareportingmetric.ThreadedMessage:Asetofcommunications,entriesoreventsthatarerelatedtoacommonpieceoforiginalcontent.UserGeneratedContent(UGC):Contentthatisentered,copied-to,posted(orotherwisecreated)byusersofaSocialMediaPlatformforsharingwithothersonthatPlatform.Inthecontextofthisdocument,UGCgenerallyreferstoorganic(holisticallysourcedandnotpaidorincentivizedbyabrandorothermarketingorganization)contentgeneratedbypeersorusersofSocialMediaPlatforms.Paid/SponsoredUGCisincluded,butspecificallyaddressedseparatelywithinthisdocument.Thisdocumentexcludes“top-down”sources(professionalinfluencers)suchasbloggerswithtopicinfluence.WhileblogsareaviableandvalidaspectofSocialMediaandtheseguidelinescanandshouldbeappliedtoblogmeasurement,uniqueaspectsofblogssuchascontentsourcingandsubscriptionswerenotconsideredherein.Viral/EarnedReach:TotaluniquecountofactualusersexposedtocontentoradvertisingthatincludesOrganic(Earned)and/orPaid,butalsosubsequentexposureandamplificationresultingfromcontentredistribution(Viralor“virality”;sharing,re-tweeting,etc.).


11

2 GeneralMeasurementSocialMediaAudienceActivitySocialMediaAudienceActivitygenerallyconsistsofaccessingorinteractingwithanytypeofcontentand/oradvertising(orotheruser),useractionsrelatedtocontentandcountsofcertaintypesorincidencesofUGConSocialMediaPlatforms.SocialMediaAudienceActivitymayoccurthroughoneormoremechanismssuchasabrowserorabrowserequivalentoranapplicationprogram(an“app”),withinaSocialMediaPlatformandshouldbefilteredtoremoveinvalidactivity(furtherdiscussedbelowandasdefinedbytheMRC’sInvalidTrafficandFiltrationGuidelinesAddendum).CapturingSocialMediaAudienceActivity,orassociatedspecificevents,isthefoundationformeasurementofSocialMedia.Inorderforauser-actiontobeconsideredpartofSocialMediaAudienceActivityitshouldbetrackable/measurable(publiclyorprivately),fallwithinthecampaigntimeperiod(forcampaignleveloradvertisingcounting)ordefinedmeasurementperiodandincludeallvalidtraffic.Examplesoftrackable/measurableuser-actionmayincludeinteraction(clicks,shares,re-tweets,likes,favorites,etc.)withSocialMediacontent,applicationdownloadsorsocialsharesfromabrandwebsitetoSocialMediaPlatforms.BeyondageneraldefinitionofSocialMediaaudienceactivityanarrayofmetricsareproducedwhichresultorarederivedfromthisactivity.Whiletheseguidelinesdonotseektodefineandstandardizeallpossiblemetrics,thebroadSocialMediametriccategoriesarepresentedbelowwithexamplesofavailablemetricswithineachcategory.SomemetricsarespecifictoaSocialMediaPlatform;however,theseguidelinesseektoaddressSocialMediaactivitybroadlyandgenerallytopreserverelevanceoftheirapplicability.Whereapplicableandubiquitous,standarddefinitionsarepresentedbelow.Subsequentsectionspresentguidelineswithregardtomeasurementanddisclosure,whichisintendedtoapplyacrossthesemetricsandcategories.Thespecificassignmentofmetricstocategoriesbelowisnotmeanttobedefinitiveascertainmetricsmayrepresentoneormorecategories,dependingonthemeasuredoruserentity.PointInTimeMeasurementWiththeexceptionofmetricsdesignedtomeasureuseractivitylongitudinally(asdefinedbelow,e.g.duration,completion,etc.)manySocialMediametricsinvolvesmeasurementofcountsatapointintime.SocialMediaisdynamicandfluidandcertainactivitydiscussedfurtherbelowmayeitherbesubsequentlydeletedor“undone”aftermeasurementofthem(suchasun-liking,un-following,etc.).Measurersshoulddevelopconsistentmethodologiesforcountingsuchmetricsatoneormorepointsintimeandfullydisclosetherulesapplied,thetimeperiodofmeasurementandthelimitationsofsuchtechniquesinrelatedreportingandmethodologicaldisclosures.


12

Shillingvs.OwnedContentandPaid/SponsoredUGCReportingShillingGoodfaithattemptsshouldbemadebythemeasurementorganizationtoidentifyandsegregate“shilling”,orpaid/incentivizedbrowsingoractivitywiththeexpressintenttoinflatemeasurement,withinSocialMediaactivityandUGC.Suchinstancesofshillingmaynotbereadilyidentifiableandmaybeintentionallyobfuscated(suchasinthecaseoftheuseofsoftwaretomasktheidentityofcontentoriginators).IfameasurerisdirectlyawareofshillingwithinUGCcontent,theseinstancesshouldberemovedwhereknown.OwnedContentorPaid/SponsoredContentUseractivitygeneratedfromexposuretoandengagementwithOwnedContentorPaid/SponsoredUGC(suchasPaidBrandAdvocacyorInfluencers)isvalid.Suchactivityshouldbedisclosedandsegregatedinreportingfromactivityassociatedwithbrandedcontent(adsoutsideofOwnedchannels)asthenatureofthecontent,thedisproportionateinvolvementofconsumersandtheimpactofthemeasurementmaydiffer.ThisincludesactivitygeneratedfromSponsored/PaidBrandAdvocacyorInfluencersandtheactivityofOwnedchannelsandassociateduseraccounts(definedbelow).However,theinitial(seeAuthorship;definedbelow)OwnedContentorPaid/SponsoredUGCactivityshouldnotbecountedinreportedmeasurement.However,subsequentexposuretoorengagementwiththisinitialactivitybyotherplatformusersshouldbereported,butsegregatedfromorganicmeasurement.DistinguishingIn-Feed,NativeorPaid/SponsoredContentPaid/SponsoredcontentappearingonaSocialMediaPlatform(oftenreferredtoasNativeorIn-Feed)shouldbeclearlylabeledinamannerthatallowsareasonableconsumertodistinguishbetweenwhatispaidadvertisingandwhatisUserGeneratedContent.Furthermore,inorderforameasurertosupportidentificationofPaid/SponsoredContentasrequiredabove,SocialMediaPlatformsmustspecifycontentassuchviapubliclyavailableAPIs.

2.1 Authorship

Theoriginalpieceofcontentororiginalactionunrelatedtooriginalcontent(exposuremetricsthatmeasuregeneraluseofSocialMediaPlatforms,withoutengagement)asmeasuredbyaplatform,publisherorthirdpartysuchas:

Completion(pageandauto-playvideo)CountsPageVisitsPosts1


13

Shares(originalAuthorshipsuchasslideshares)1TimeSpentMetrics(includingauto-playvideoduration)1Tweets1 ViewableImpressions

1:Thesemetricsmayalsobeusedasmeasuresofengagementwhenpresent.

SubsequentEngagementandReachmetricsaredirectlyrelatedtoandderivedfromoriginalAuthorship.Authorshipmetricsshouldserveascurrency,baselinemetricsforSocialMediameasurement.

TheMRCandIABhavepublishedcertainguidelinesapplicabletoimpressionmeasurementacrossdigitalmediaincludingdisplay,video,richmediaandmobile(webandin-app).Thisdocumentisnotintendedtosupersedethoseguidelines.Instead,SocialMediameasurersshouldapplythoseguidelinesasapplicablebasedonadvertisingplacementtype.Furthermore,theMRChaspublishedguidelinesrelatedtoviewableimpressions,whichlikewise,shouldbeappliedaccordingly.Itisexpectedthatspecificusecasesrelatedtoimpressionandviewableimpressioncountingsuchasmobileviewability,in-feed,in-streamandnativeadvertisingmeasurementwillbeaddressedasupdatesoraddendatothosepre-existingguidelines.ItissuggestedthatViewableImpressionsserveasabaselinecurrencymetricformeasuringExposureonSocialMediaPlatforms.Finally,varioususerattributesanddataheuristicsmaybeusedtofurtherclassifyexposuremetricsasactivevs.passive(hovers,scrolls,dwelltime,etc.).Suchmethodsshouldbeempiricallysupportableandanyuserattributesorheuristicsusedshouldbeabletobedemonstratedtoserveasaproxyorindicatorforactivityorattention.

2.2 Engagement

Generally,perIABdefinitionreferencedbelow,aspectrumofconsumeractivitiesandexperiences(interactionsandinterest)—cognitive,emotional,andphysical.Engagementassumesactiveparticipation,butmayalsodescribeacognitiveoremotionalconnection(suchasinSocialListeningandMonitoringmetricsdescribedbelow),inadditiontoaphysicalaction.Thisdocumentbreaksengagementintothefollowingcategories:(1)Interaction,(2)ContentRedistribution,and(3)AdvocacyandInfluence.Measurersareencouragedtoreferenceandadopttheapplicablegeneralmetricdefinitions(includingEngagement)publishedbytheIABincollaborationwiththe4AsandtheANA;DefiningandMeasuringDigitalAdEngagementinaCross-PlatformWorld:

http://www.iab.net/media/file/Ad_Engagement_Spectrum2014_FINAL2-5-2014-EB.PDF



14

2.2.1 Interaction

Interactionmetricsincludemeasurementofindividualuserinteraction(generallyquantifiablephysicalormechanicalactionswithcontent)withoriginalcontentoradvertisingonSocialMediaPlatforms.SuchinteractionsmaybeInternal(visibleonlytotheplatformormeasurementorganization),orPublic(visibletooneormoreusersoftheplatform,therebyextendingvisibility)andinclude:Internal:Clicks(guidelinesexist:ClickMeasurementGuidelines)Completion(postclickorinitiation)ConversionDirectMessagesDownloadsDuration(postclickorinitiation)InstallsSaves(includingvideo“watchlater”functionality)SwipesTimeSpentMetrics(excludingauto-playvideodurationwithoutphysicalengagement)TapsOtheractionstaken(votes,redemptions,etc.) Public:+1sCheckInsComments(CountsorUGC)FavoritesFollowsHeartsHashtags1LikesMentions1PostPinsRatingsReplies(CountsorUGC)SubscribesTagging1ThumbsUp1PublicInteractionmetricssuchasHashTags,MentionsandTaggingwithintenttore-distributecontent(wheremeasurablequantitativelyorempirically)mayalsobecategorizedasContentRedistributionmetrics.

http://www.iab.net/media/file/click-measurement-guidelines2009.pdf


15

2.2.2 ContentRedistribution:

IntentionalanddirectinteractionwithintentofredistributionorpromotionofcontentoradvertisingbyindividualusersofSocialMediaPlatformstootherusersofSocialMediaplatforms,eitherindividually,toasubsetofusers(friends,followers,etc.)oracrosstheuserbase.

ContentredistributionbynatureisaformofPublicengagement.TheextentthatPublicContentRedistributionmetricsarecompiledthatrepresentonetoonevisibility(limitedtotwousers)shouldbeclearlydefined,labeledandsegregatedfrommorewidelyPublic(visibletoanetworkofusersorallusers)metrics.Wherepossible,theintentoftheinteraction(onetoone,ormorewidelyPublic)shouldbegauged(supportedempiricallywithconsiderationtoprivacyrestrictionsoftheplatform)andfactoredintothemeasurementmethodology.

Theseinclude:

Emailing1Embeds2HashtagsLinking1MentionsRe-blogsRe-pinsRe-postsRe-tweetsRe-vinesSharesTagging

1GenerallyconsideredforemailingorlinkingfromwithinaSocialMediaPlatformthatcanbedirectlymeasured(asopposedtocopyingorsharingoflinksoutsideoftheplatform[alsoreferredtoas“darksocial”activity]wheremeasurementmaybelimited).Suchmetricsshouldbeaccompaniedbycleardefinitionanddescriptionofcapabilitiesandlimitations.2Embeds,orcopyingthesourcecodeoforiginalcontent(Authorship)andincorporatingorembeddingitintositesorplatformsotherthantheplatformfromwhichtheyoriginated,isaformofContentRedistribution.However,asthisactiongenerallyinvolvessitesorplatformsthatmaybeexternalfromtheoriginalplatform,measurementofactivityassociatedwithembedsoutsideoftheoriginalSocialMedia


16

Platformmayrequirespecialdisclosureandreportingsuchasdiscretereporting(atthesiteorplatformlevel)andsegmentationfromAuthorshipordirectEngagement.Theabovemetricsmayalsobereportedbasedondepth(iterativeorcascading)andalternateplatformsourceordestination(contentredistributionbetweendifferentplatforms)withappropriatedisclosureanddefinition.SeeSection4.2.2ofthisdocumentforfurtherguidanceonthreadedoroverlappingcontentanddownstreamediting.

2.2.3 AdvocacyandInfluence:

BrandAdvocacyisactivitywherebyausercreatesfavorablecontentaboutabrandorproduct,andthenpassesonpositivemessagesaboutthebrandtootherusers(suchasinarecommendation).DefinitionofAdvocacyshouldbeconsistentandempiricallysupportedbasedoncontextofthemediabeingmeasured(vertical)andtheactivitiesgenerated.Influencersareuserswiththeabilitytoreachotherusersoraffectotherusers’thinkinginasocialonlinecommunity,generallydefinedbythenumberofdirectlyconnectedusersorthesizeofthenetworkofusersassociatedwithanaccount,userorprofile.DefinitionofInfluenceshouldbeconsistentandcouldbeempiricallysupportedbasedonsomequantitativemeasure,usuallythetoppercentileofReach,connectedusers(friends,followers,etc.)oractivity(interactionorengagementwithcontent).Influencersmayalsobeaperson,grouporentitywiththeabilitytoreachandaffectanotherpersonorgroupofpeople’sthinkingandbehaviorduetoattributessuchasexperience,expertise,reputationandsocialfootprint.Asaresult,determinationofInfluencemayalsobebasedonqualitativefactorswherebyabsolutethresholdsaremoredifficulttodefine.Furtherthequalityofwhatcontentoruser(s)arebeinginfluencedmayimpacttheusefulnessofInfluencemetrics.ThesefactorsshouldbeconsideredwhendefiningandmeasuringInfluenceandassociatedassumptionsorlimitationsofthismeasurementshouldbedisclosed.BothAdvocacyandInfluencecanbespecifictoamediatype,brandorplatform,andcanbeexplicitorimplicitbutinallcases,thedefinitionandbasisshouldbeclearlydisclosedwithreportedmetricsandwithininmethodologicaldescriptions.LikeallotherSocialMediametricsandasdiscussedabove,measurersshouldseektoincludeonlyOrganicorEarnedactivityassociatedwithAdvocacyandInfluence.Authorshipofcontent(initialactivitygeneratedaspartofOwnedchannelsorPaid/Sponsoredsitesorusers)shouldnotbeincludedinreportedmetrics(subsequent


17

EngagementwiththisAuthorshipisvalidandreportable).However,exposuretoorengagementwiththisactivityshouldbereportedonasegregatedbasis. SocialListeningandMonitoring:SocialMediametricsaimedatmeasuringqualitativeaspectsofactivitysuchas:

IntentOpinionSemantics(usermeaningandrelevance)Sentiment(mayhavequantitativeaspects)

BrandTrackingSocialMediametricsaimedatmeasuringbrandperformanceandhealth,suchas:

BrandHealthMetricsLoyaltyResonance

WhilemuchoftheguidancecontainedwithinthisdocumentmaybeappliedacrossSocialMediameasurementingeneral(includingSocialListeningandMonitoringaswellasBrandTracking),duetothequalitativenatureofmanyofthesemetrics,thisdocumentisnotintendedtostandardizeormandatedefinitionofengagementwithregardtoqualitativemeasurement.Additionally,someSocialListeningandMonitoringmetrics(suchasBrandAwarenessandRecall)maybederivedfromofflinesurveys(responsesurveys)asopposedtodigitalcensusmeasurement.Suchofflinesurveysshouldbesubjecttocertainqualityparametersincluding,butnotlimitedto,theMRCMinimumStandards(specificallywithregardtosamplerepresentation,reductionofbiasandsampleperformanceandresponserateaswellasempiricalsupportofanyestimation,dataadjustmentandmodeling).However,measurersandusersofSocialListeningandMonitoringaswellasBrandTrackingdataareencouragedtoconsiderthedisclosureguidancecontainedinSection7.0andthroughoutthisdocumenttopromotetransparencyandeducationaroundthemeaninganduseofsuchmeasures.Forfurtherguidanceonengagement,seeDefiningandMeasuringDigitalAdEngagementinaCross-PlatformWorld,adocumentauthoredbytheIABincollaborationwiththe4A’sandtheANArelatedtoMakingMeasurementMakeSense(3MS).SocialListeningandMonitoringorBrandTrackingisgenerallynotprojectabletotheopinionofallusers.Thelimitationsandunderlyingassumptionsofsuchdatashouldbe




18

clearlydisclosed.PrimaryResearchmaybedonetoaccompanytheSocialListeningandMonitoringorBrandTrackingtooffsetsomeoftheselimitations.

2.3 Reach

Reachisdefinedhereasthetotalunique(deduplicated)countofusersexposedtoorganic(unpaid)contentoradvertising.SocialMediaReachthatincludesonlythoseactualusersexposedtocontentinitsoriginalformat(withinageneralfeed,ticker,page,etc.)isoftenreferredtoasOrganicReach.Totaluniquecountofusersexposedtopaidcontentoradvertisingthatincludesonlythoseactualusersexposedtocontentinitsoriginalformat(withinageneralfeed,ticker,page,etc.)isoftenreferredtoasPaidReach.

SocialMediaReachthatincludesOrganicand/orPaid,butalsosubsequentexposureandamplificationresultingfromcontentredistribution(sharing,re-tweeting,etc.)isoftenreferredtoasEarnedorViralReach.

CertainSocialMediameasurersmodel,imputeorprojectPotentialReachbasedonthenetworkofthoseactuallyexposedtocontent.ProjectedReachbasedonacountofasubsetofplatformusersoranetworksuchasfriends,followersorfans,isreferredtoasPotentialReachthroughoutthisdocument.Ifnecessary,PotentialReachshouldbelimitedtouseasaplanningmetriconly(notacurrencymeasurement).MeasurersofPotentialReachmayutilizealgorithmstodivideorsegmentanetworkofusersintothosemorelikelytobeexposedtocontent.Whenempiricallysupported,thesedivisionsofPotentialReachmaybeutilized,butshouldbeclearlydefinedandlabeled.ProperlydisclosedandsupportedestimatedactualReachmaybeusedasanevaluationmetricwhereappropriate.

ModeledReachbasedonagrossmultiplierofusersisoftenreferredtoasTheoreticalReach(suchasprojectingpost-addeliverybasedonagrossmultiplierrepresentingsomeestimateofpotentialexposure).TheoreticalReachshouldbeavoidedasitmayresultinunreliable,inconsistentandinaccurateestimates.Again,MRCandIABguidelinesexistwithregardtoAudienceReach.Furthermore,theMRCintendstoestablishdigitalaudience-basedcurrencyguidelines.WhileSocialMediameasurersshouldapplythoseguidelines,thisdocumentprovidesfurtherguidancewithregardtoReachmeasurementacrossSocialMediaPlatforms(seeSection4.1.3ofthisdocument).Validcomparisonsofdifferentplatform-specificReachmetricsmaybeperformedwherethesemetricsarewithinthesameReachcategoryandrepresentthesamelevelofuseractivityorengagement.Suchvalidcomparisonsmaybeusedforreportingrelativeperformanceacrossplatforms.


19

2.4 CostMetrics

PaidmediapresentwithinSocialMediaPlatformsmayalsoallowpost-level(afterconsumptionorengagement)metricsderivedfromthecostofadvertisingplacementascomparedtootherSocialMediametrics(exposure,engagement,Reach,etc.),suchas: Conversions Costper(CPM,CPC,CostperEngagementetc.)

2.5 DemographicsandTargeting

Theabovemetricsmaybereportedonthebasisofdemographiccategoryforgeneralreportingortargeting,whichcanbeobtainedorassignedusinganumberoftechniques(registrationdata,panelorsamplebasedmeasurement,acombinationofdirectmeasurementtechniquesandmodelingorattribution,etc.).

ThebasisusedtomeasureandreportdemographicdatashouldbefullydisclosedaspartofaDescriptionofMethodologyanddemographiccategoriesshouldbeconsistentandclearlydisclosed.Relevantstandards(e.g.,MRCMinimumStandardsandDataIntegrationGuidelinesaswellasIABAudienceReach)shouldbeconsideredwhendesigningandevaluatingthedemographictechniquesused.

Limitationsofthedemographictechniquesandapplicablequalityorperformancemetrics(forsampleorpanelapproaches)relatedtothemshouldbefullydisclosedtousersofthedata.Forregistrationdata,collectionandupdatepoliciesshouldbefullydisclosedincludingifconductedthroughuseofathirdpartydataenrichmentprovider.Furthermore,organizationsutilizingregistrationinformationforlocationarestronglyencouragedtodevelopstronglogicchecksandinspectiontechniquesandconsiderexistingdatacollectionguidance(e.g.,CRE,etc.).

3 SocialMediaCoverageandBasisforProjection

3.1 Internet/MobileInternetUsers–Universe

ItiscriticalthattheimpactofreportedSocialMediameasurementsbeappropriatelysized.Sizingisdependentonplatformpenetration,datacollectionmethodandthenatureofrepresentationassociatedwiththatmethod.InmanycasesSocialMediameasurementsshouldonlybereportedasgrossactivity,essentiallyrepresentingthenumberofoccurrencesofacertainconditionthatwereencounteredwithintheplatform–forexample,projectionisnotappropriatewhenonlyasubsetofdataisreceivedorcollectedfromaplatformacrossthereportingperiod.


20

However,insomecaseswhendatacollectionmethodsjustifywidersizing,measurementscanbeprojectedorindexedtototalplatformusersorinternet/mobile/tabletusers(asapplicable)ortoevenawiderbase,suchastotalcountrypopulation.

Furthermore,mediatypeorcategorysizingshouldbeconsistentandclearlydefinedwithinplatformmeasurementwhenutilizingthisasabasisforindexingorprojection.Sizingandrelativemeasuresmayalsobereportedonaweightedorscaledbasisasrelatedtotheobjectiveormeasurement(intendedaudienceorcontext),withclearandobjectivedefinition.

Ineithercasethebasisforprojectionshouldbespecifiedinreports(orifnoprojectionexists),andthesupportforthisprojectionmethodshouldbespecifiedinthedescriptionofmethodologyofthemeasurer.Forcross-mediacomparisons,thisguidelinerequiresatotalpopulation,ortotaladultpopulation(dependingonSocialMediaPlatform)projection,andnotaplatformcoverageuniverse.Additionally,validcomparisonsofdifferentplatform-specificmetricsmaybeperformedwherethesemetricsarewithinthesamecategory(discussedinSection2ofthisdocument;e.g.LikestoFavoritesto+1s,etc.)andrepresentthesamelevelofuseractivityorengagement.Suchvalidcomparisonsmaybeusedforreportingrelativeperformanceacrossplatforms.

3.2 CompatibilityorMeasurementLimitations

IncaseswherethemeasurerandtheirmeasurementtechniquesortheSocialMediaPlatform’sprovideddata(adirectorperiodicdatastreamprovidedbytheSocialMediaPlatform)islimitedtoonlyasub-setofpotentialusers’activity(forexample,notinclusiveofcertaintypesofdevices,notinclusiveofexternalactivitysuchasemailingorcopyinglinksoutsideoftheSocialMediaPlatform(alsoknownas“darksocial”)ornotinclusiveofcertaintypesofoperatingsystems,browsersorappversions,theselimitationsshouldbefullyandprominentlydisclosed,ifmaterial(typically,greaterthan5%inaggregate).

Activitycapturedfrompartialdatastreams,forexample,datastreamsnotrepresentingtheentireuser-baseoftheplatform,shouldbedisclosedandappropriatelysized.

3.3 TrackingSocialMediaExposuresOutsideofApplications

ForSocialMediathatisaccessedoutsideofanapplication(regardlessofuserdevice),ormorespecifically,whentrackingisconductedthroughabrowseroremail,itiscriticaltocaptureinformationrelatedtotheSocialMediaPlatformforreportingpurposes.ThiswillassistinassigningSocialMediaactivitybyplatformandfacilitatethecombiningofmeasuresacrossbrowseroremailaccessedmediawithothermethodssuchasapps(whichmaynaturallyindicatetheplatform)orgames.Additionally,datashouldbecollectedwherepossiblethat


21

allowsthemeasurertodeduplicatemeasurementordetermine“uniques”(seeSection4.1.2forfurtherdetail).Gross,unduplicatedmetricsresultingfromloggedoutactivityorthatoutsideofapplicationsshouldbefullydisclaimed.

AsdiscussedinSection2.2.3,someSocialListeningandMonitoringmetricsmaybederivedfromofflinesurveys(responsesurveys)asopposedtodigitalcensusmeasurement.Suchofflinesurveysshouldbesubjecttocertainqualityparametersincluding,butnotlimitedto,theMRCMinimumStandards(specificallywithregardtosamplerepresentation,reductionofbiasandsampleperformanceandresponserateaswellasempiricalsupportofanyestimation,dataadjustmentandmodeling).

3.4 SocialMediaApplications

SocialMediaapplications(apps)comeinnumerousforms,butessentiallyrepresentclosedlogicalenvironmentswithspecifiedcontrolsandfunctionalitywheretheusercanaccessandgenerateSocialMediacontent.

TheIAB,MRCandtheMMA(MobileMarketingAssociation),haveproducedaguidelineforthemeasurementofadvertisinginanapplicationenvironment.Whilethisguidancegenerallyappliestoadimpressions,theSocialMediameasurerthatconductsbusinessinapplicationenvironmentsshouldconsidertheprovisionscontainedinthisguideline.Thefollowingarekeyaspectstoconsider(adaptedfromtheIn-AppGuidelinefortheSocialMediaenvironment):

PhysicalandLogicalAppEnvironmentPhysicalattributesoftheadvertisement,SocialMediaeventorUGCspaceanalyzed,andtheplacementofthesewithintheapplicationshouldbedisclosed;thisdisclosurecanbemadeaspartofinitiatingthemeasurementproductwiththemediaoutlet,advertisingagencyorbuyer.Additionally,anyvariableexposureattributessuchasadsizeoftheadvertisingduringtheapplicationsessionortheUGCopportunitiesinvolvedshouldbedescribed.Measurementeventsshouldnotbecountedifoccludedorotherwiseexposure,sharingordistributionisimpairedinasignificantmanner.EventbasedSocialMediainstancesshouldbedescribedintermsofattributesandtriggercriteria.TheSocialMediameasurementorganizationworkingwithapps(includinggames)shouldhavesufficientcontrolsorassurancetodeterminethat:

• Theapplicationwasdownloaded,openedandinitializedasdesignedonthatClient

UserpriortothemeasuredSession.• Theapplicationitselfwasfunctioningasintendedduringthesession.SocialMedia

activityassociatedwith“faulted”conditions(situationsoffunctionalityissueswiththeapplication,errorsornon-workingconditions)shouldbetrackedandsegregatedfromactivityinfullyfunctioningenvironments.


22

MeasurementConditionswithintheAppChangesinpageorientationduringapplicationusecanbetrackedasanativeorSDKfunction,however,thisgenerallydoesnotaltertheparametersforcountingSocialMediaactivity,excepttotheextentitmaychangethesizeofUGCorrelevanttext(forexample,whichmaycauseviewableparameterstobemoreorlessdifficulttoachieve).

Instanceswhereusersareencounteringapplication-loadingscreensaregenerallynotcountedasapplicationactivity,noristheloadingdurationconsideredin-appduration.IncertaininstancestheseloadingscreenscancontainadvertisingmessagesorSocialMediarelevantcontent.Ingeneral,theseimpressionsoranyothermeasuredactivityassociatedwithloadingscreens(e.g.,duration,mouseactivity,etc.)shouldbesegregatedforreportingpurposesfromactualapplicationuseandadvertisingconsumedduringtheapplication’suserexperience.

MeasurementOrientationswithanAppApplicationtransactionrecords,whichcontainevidenceofSocialMediaexposureorrelevantSocialMediaevents,canbederivedandtransmittedtotheapplicationmeasurementorganization:(1)onareal-timebasisduringapplicationexecution,(2)inbatchedgroupsthataretransmittedperiodically(inwholeorinpart)duringanon-lineapplicationSessionor,(3)firststoredduringoff-lineapplicationuseandlatertransmittedduringasubsequenton-lineSession(notnecessarilyassociatedwiththesameapplication)oftheapplicableClient-User.Insituationswheretheconnectionspeedoftheclient-usercanimpactcountingeffectiveness,scrapingeffectivenessorthecountedactivityitself,theapplicationdeveloperorSDKdeveloper,ifapplicable,shouldmakereasonableeffortstoensurecountingisaccurate.Editinganderrorhandlingrulesshouldbedevelopedtodetect,segregateandreportcountingsituationswithsuspectaccuracyrelatedtoconnectionspeed.

ControlswithinAppsOriginatedbytheMeasurer(orRelied-OnbytheMeasurer)Thegeneralinternalcontrolspresentattheorganizationthatdevelopsorapprovestheapplicationforreleaseareacriticalcomponentoftheoveralloperatingeffectivenessofmeasurementassociatedwiththeapplication.Thesecontrolsdonothavetonecessarilyresideattheoriginalapplicationdevelopment/programmingfacility;however,certainlevelsofqualitycontrolshouldbepresentatsomestageofapplicationrollout.Ultimately,itistheresponsibilityoftheSocialMediameasurementorganizationtoensurethatpropertestingandreleaseprocessesarefollowedandthatcontrolleddevelopmentprocesseswereemployedinbuildingtheoriginalapplication.

Ingeneral,theSocialMediameasurementorganizationthatreliesonappsshouldhavesufficientcontrolstoensure:


23

• Developmentofandchangestoapplicationsareauthorized,testedandapprovedpriortobeingrolledoutforUserdownload(release).Segregationofversionsshouldbemaintainedwheremeasurementfunctionalityhasbeenchanged.

• AccesstoapplicationsoftwareassociatedwithadvertisingorUGCscraping/capture,storageofads,adorUGCplacementandservingfunctionalityisrestrictedtoauthorizedpersonnel(nonUser)andprograms.Usersshouldnothavetheabilitytoalterappfunctionalityinthisarea.

• User-setparametersthatrelatetomeasuredeventsaredocumented,recordedandincludedindatatransmissionsbacktothemeasurementorganizationifchanged.

• Theapplicationisdocumented,andmeasurementassociatedfunctionalityisdocumented.

• Onlyauthorizedformattedcontentisacceptedasinputbytheapplication,regardlessofwhetherthatcontentisenteredreal-timeorstoredforlateruse.

• Anycalculationsordataaccumulationprocesseswithintheapplicationhavebeentestedforefficacy.

• Datatransmissionsfromapplications(whetherreal-timeorbatched)arecomplete,accurateandprotectedfrommodification.

• Errorsandmeasured-eventrelateddatarejectedforqualitypurposesislogged,evidencesupportingtheerrorisretainedanderrorsarefollowedupontocorrectpotentialcasesofsystematicorrecurringissues.

InSDKorientedmeasurementenvironments,theapplicationdeveloperorsellershouldhavesufficientconfidencethattheabovecontrolsaremaintainedfortheSDKfunctionality.Developmentofthisconfidencecanencompassaperiodicreviewand/ortestingconductedbytheapplicationdeveloper,inwhichcasetheapplicationdeveloperthenistakingresponsibilityforthecontrolsattheSDKdeveloper.AnotherapproachisfortheSDKdevelopertobeauditedbyathirdpartywithsomeformofobservableassuranceprovidedsuchascertification,accreditationorathird-partyCPAattestation.Inthislattercase,iftheapplicationdeveloperislookingtobecomecertifiedoraccredited,theauditingorganizationcanbuildacaseforrelyingoftheSDKprovider’sassurance(dependingonconditionsofthatassurance).

ProvidedDataStreamsversusIn-AppMeasurementInsomecases,measurementorganizationswillreceivedatadirectlyfromtheSocialMediaPlatformineitherareal-timeorbatchenvironmentsuchasviaanAPIorfirehosedata.Thisnecessitatesadifferentanalysisworkflowandlevelofintegration(lesserintegration)fromthosemeasurementfunctionsthatareintegratedwithorintotheapplicationitself–collectingdatadirectly–whicharedescribedabove.InsituationswheredataisreceiveddirectlyfromtheSocialMediaPlatform,othercontrolsapply.Specificallythemeasurementorganizationshouldhavecontrolsthatensure:

• Anunderstandingoftheprocessesusedtocreatethetransferreddata.


24

• Anunderstandingofthemeaningofthedatafieldstransferred.• Completenessofthedatatransferred,andcontinuoustime-periodsfromtransferto

transfer.• Datachecksforlogicalconsistencyandappropriatefieldcontents.• Back-upandrecoveryprocessesfortransferreddata.• Anunderstandingofthecompletenessofthedatatransferred,andifincompleteor

asubsetthenatureofthemissingdata.• Forreal-timecollectionprocesses,processesthatensuretransfercompletenessin

theeventoflostconnections.Furthermore,inorderforameasurertosupportidentificationofPaid/SponsoredContentasrequiredbythisdocument,SocialMediaPlatformsmustspecifycontentassuchviapubliclyavailableAPIs.MobileViewabilityAsdiscussedabove,theMRChaspublishedguidelinesrelatedtodesktopviewableimpressions,whichshouldbeappliedaccordingly.Itisexpectedthatspecificusecasesrelatedtoimpressionandviewableimpressioncountingsuchasmobileviewability,in-feed,in-streamandnativeadvertisingmeasurementwillbeaddressedasupdatesoraddendatothosepre-existingguidelines.However,thedesktopviewabilityguidelinesdonotdirectlyaddressmobileviewability(weborin-app).TheMRCintendstopublishguidelinesspecifictoviewabilitymeasurementinmobileenvironments.Likewise,itisexpectedthatSocialMediameasurersconsidertheseguidelinesforviewabilitymeasurementofSocialMediainamobileenvironment.

3.5 UsageRestrictions,IfAny

UsagerestrictionswithSocialMediadatathatimpactmeasurementorgroupstobemeasuredshouldbefullydisclosedandifpossible,quantified.Arelevantexampleofthiswouldbeagerestrictionsorotherlimitationsforparticipationbyminorpersons(orotherdemographicinformation/originofuser)suchasCOPPArestrictions.

4 SocialMediaActivityMeasurementSpecifics

4.1 TrackingofUsersAccessingContent–TechnicalDetails

Example:QuantifyingusersaccessingorotherwisepotentiallyinfluencedbySocialMediaUGC,postingsorotherevents.


25

4.1.1 ClientSide(andViewable):

AllusertrackingassociatedwithSocialMediameasurementshouldbemeasuredbasedonclientsideactivityoractualusereventsfromuserdevices,filteredtoexcludeknowninvalidactivityandinsofarasapplicableviewableeventsonlyshouldbecounted.

4.1.2 TrackingMethod

ThespecifictrackingmethodandcountingassetsusedbytheSocialMediameasurementorganizationshouldbefullydisclosedinthedescriptionofmethodologyandmayinclude,butarenotlimitedto:

Ad/AssetTagsCommunicationBasedMeasurement(Redirects)OnlineorOfflineSurveysPageImplementationsPlayerPlug-InsScripts

Specificmachinebasedtrackingofusers–suchasuniquebrowsersorcookies–shouldbespecifiedanddistinguishedfromactualuserbasedtracking(whichrequiresmoreadvancedadjustments,registration,etc.).

De-duplicationofusertracking(into“uniques”)forreachmeasurementsisdesirable–butisrecognizedtobeadifficultmetrictocaptureaccurately.Ifameasureislabeled“unique”de-duplicationisrequired.Itiscriticalthatthoroughdescriptionsofhow“users”areactuallydeterminedandhowde-duplicationduringmeasurementperiodsisaccomplishedareprovidedtousersofuniqueusermeasurements.WerefertheuseroftheseguidelinestopriorIABguidanceonAudienceReachMeasurementswithregardtothehierarchyofuniquemetrics(cookies,browsers,etc.).Trackingmethodsshouldbesubjecttointernaltestingandvalidationuponoriginalimplementationaswellasperiodicinternalverificationtesting.Trackingmethodsanddisclosureswillbethesubjectofintenseproceduralverificationifexternalcertificationissoughtbythemeasurementorganization.

4.1.3 ProjectionorOtherEstimationMethods,IncludingInferences

Projection,imputationorotherdataadjustmentsmadetotrackingmeasurementsbytheSocialMediameasurementorganizationshallbefullydescribed,basedonsystematicandreasonableproceduresandshouldbesupportedbyunderlyingempiricalstudies.Asageneralrule,activityshouldnotbeattributedtoaSocialMediauserwithoutdirectevidenceofsuchactivitysuchasassumptionsofactivityorfans/followersorTheoreticalReachmultipliers.SocialMediameasurersandresearchersare


26

encouragedtopublicallycirculaterelevantresearchsupportinganyassumptionsinthiscontext,whichcanbecompliedandlinkedtoaspartoffutureiterationsofthisdocument.

Readersoraccesstocontentshouldnotbeassumedbasedonmereusersofaplatform;actualtransactionalevidenceofcontentconsumptionshouldbeusedasabasisformeasurement.Inferencesbasedonbrandspecificsubsetsofthetotaluserpopulation(fansorfollowers)shouldnotbeappliedtothetotalpopulation.Suchinferencesshouldalsobedisclosed.BaselinemetricsutilizedforOrganic,PaidandEarnedReachshouldbeclearlydisclosedandconsistentacrossmediacategoriesandtypeswithinaplatform.Ifnecessary,PotentialReachcouldbeusedasaplanningmetric(notacurrencymeasurement),butshouldbelimitedtoplanningcircumstancesonlyandthebasisforsuchmeasurementaswellastheinherentlimitationsshouldbefullydisclosed.MeasurersofPotentialReachmayutilizealgorithmstodivideorsegmentanetworkofusersintothosemorelikelytobeexposedtocontent.Whenempiricallysupported,thesedivisionsofPotentialReachmaybeutilized,butshouldbeclearlydefinedandlabeled.However,properlydisclosedandsupportedestimatedactualReachmaybeusedasanevaluationmetricwhereappropriate.TheoreticalReachingeneralshouldbeavoidedformediaplanningorbuying.Ifpresentfordirectionalanalysis(non-transactionalandnon-currency),thebasisforprojectionusedinTheoreticalReachshouldbefullydisclosedanddisclaimedaswellassupportedbasedonsystematicandreasonableproceduresandunderlyingempiricalstudies.Furthermore,underlyingmodelassumptionsshouldbefullydescribed.GrossmultipliersusedforderivingTheoreticalReachwithoutempiricalsupportshouldnotbeusedforanypurpose.SeeSection4.4ofthisdocumentforguidancewithregardtofrequencycappingthatmayimpactReachmetrics.

4.1.4 MetricsandKeyAttributes

MetricsanddefinitionsofattributesofkeymetricsshallbedisclosedbytheSocialMediameasurementorganization.Importantmetricdefinitionsshouldappearonorwithineasyaccessfromthereportscontainingthemeasurementdata.Measuredunits,forexample,viewableimpressionsorviewsofUGCorotheractions(forexample,“likes”)shouldbedefinedandshouldconformtocommonindustryaccepteddefinitionstothefullestextentpossible.


27

4.1.5 DifferencesbyPlatform

AnydifferencesinprocessesormeasurementmethodologybySocialMediaplatformshallbedisclosed.ThisisarelevantissuewhenestimatesorcomparisonsarereportedacrossSocialMediaplatformsorwhenthemeasurercombinesplatforms.

4.2 UserGeneratedContent(UGC)Analysis–TechnicalDetails

Example:SortingthroughoranalyzingUGC,generallyusingmachine-basedprocessesandlearning,tofindbrandmentionsorothereventsconsideredmeaningful,andseekingtodeterminethecontextandsentimentofthecontentofthepost.Also,determiningcontextandmeaningofsubsequentthreadedpostings.

Inadditiontotheguidelinesinsection4.1above(whichshouldbeappliedtoSocialMediameasurementbasedonUGC),thefollowingadditionalconsiderationsshouldbeconsidered:

4.2.1 TrackingMethod

ThespecificcollectionmethodandwordorcontentrecognitionprocessesusedbytheSocialMediameasurementorganizationshouldbefullydisclosedinthedescriptionofmethodologyandmayinclude,butarenotlimitedto:

APIDirectMeasurementMachineLearningOfflineExportPartnerDataScraping(wherepermissiblebytermsandconditions)Third-PartyTracking

Specificmachinebasedprocesses(andlearningfunctions)forgatheringofUGCwordsorkeywords,numberofmentionsoreventsshouldbedisclosed.Datacollectionmethodsshouldbesubjecttointernaltestingandvalidationuponoriginalimplementationaswellasperiodicinternalverificationtesting.Thistestingshouldincludethevalidityofthecontextandsentimentmeasures.Datacollectionmethodsanddisclosureswillbethesubjectofintenseproceduralverificationifexternalcertificationissoughtbythemeasurementorganization.Thecontentsourceshouldbeclearlydefinedandscopedinthemeasurementservice’sdisclosures,forexample:

• Whattypeofcontentiscapturedfromwhatplatform• Usercoverageformeasuredplatforms


28

• Collectionmethod(direct,API,offlineexport,partner,third-party,scraping,etc.)

• APItypewhereapplicable(public,closedclient,platformorfirehose)• Depthofcontentcollection(scraping)applied• Storagemethods,timing• Natureofmachine(orhuman)analysisprocessesemployed• Timeperiodsanalyzed• Editsordataadjustmentsapplied• Natureofspecialhandlingofambiguouscases• Errorcorrectionandreissueprocedures• Sentimentaccuracy

4.2.2 ThreadedContent,OverlappingContentandDownstreamEditing

Engagementmetricsmayalsobereportedbasedondepth(iterativeorcascading)andalternateplatformsourceordestination(contentredistributionbetweendifferentplatforms)withappropriatedisclosureanddefinition.Toolsforanalyzingandaccumulatingthreadedcontentshallbedisclosed,andthesetoolswillbesubjectedtoextensiveinternalqualitycontroltestingbythemeasurementorganization–especiallywhenlinkagesofthreadedcontentarenotclear(forexample,situationswheresequentialfollow-onlistsarenotused).Duplicateoroverlappingcontentprocessesshouldexistthatminimize/eliminatedoubleormultiplecountingofsingleevents.ChangesmadetoUGCbyoriginalpostersorothersshouldbedistinguishedwherepossiblebythemeasurementorganization.Defaultordering(mostrecent,mostrelevant,etc.)ofUGCinthreadsorfeedsasconfiguredbyaplatformormeasurer(asopposedtouser-configuredordering)shouldbedisclosed.Additionally,algorithmsemployedbySocialMediaPlatformstoeitherotherthecontentusersseeorsuggestcontentshouldbeempiricallysupportedandtheuseofsuchtechniquesshouldbedisclosedaspartofmeasurementandreporting.Changestothisorderingortheunderlyingalgorithmsshouldbedisclosedinadvancefortransparencyandtrendingpurposeswherepossible.Finally,thepresenceandprevalenceofuser-configuredorderingmayimpactmeasurementandshouldbestudiedandconsideredbymeasurersofplatformswherepresent.


29

4.2.3 ExcludingNon-RelevantInformation

Themeasurementorganizationmayexcludenon-relevantcontent(basedonkeywords,subjectmatterorothercontextualmeasures)frommeasuredevents.Contextualrelevancemaybesubjectiveandvarybetweenplatforms.Aswithanydataadjustmenttechniques,anyprocessestodeterminetherelevanceofandexcludeorincludecontentformeasurementshouldbefullydescribed,basedonsystematicandreasonableproceduresandshouldbesupportedbyunderlyingempiricalstudies.Measurersareencouragedtodeveloprobustdictionariesortaxonomiesthatareconsistentlyappliedforthepurposeofestablishingcontextualrelevance.Alternativelyoradditionally,machinelearningoralgorithmicapproachesarepermissibleifsupportedbyempiricalanalysis,consistentlyappliedandperiodicallyupdated.SimilartoAdVerificationguidance,SocialMediaUGCcontextualrelevanceparametersshouldbedevelopedthroughconsultationbetweentheplatformormeasurerandthecustomer.ContextualrelevancecriteriashouldbemaintainedbyeachmeasurerandclearlydisclosedtocustomersTimingofUGCThetimingofthecontextualrelevancedeterminationorunderlyingdataassumptionsshouldbedisclosedrelativetothetimingoftheUGCanalyzed(“freshness”)andameasurershouldhavestateddatastalenesspoliciesempiricallysupported.ItisstronglyencouragedthatmeasurersanalyzethedegreeofuseractivityassociatedwithtypesofUGClongitudinallyasabasisforfreshnessorstalenesspolicies.Atimestampisrecommendedtoinformusersabouttheageoftheservice’smostrecentassessment.Empiricallysupportedreportingtimecut-offeditrulesarestronglyencouraged,withfulldisclosure,especiallywithregardtomeasurementofSocialMediaactivityrelatedtoaspecificadvertisingcampaign.Reportingcut-offrulesshouldconsiderthenatureoftheposts(e.g.,Sponsoredevergreenposts)andcampaigntimebounds(whereapplicable)aswellasiftheactivitybeingmeasuredrelatestooriginalAuthorshiporre-postingoforiginalAuthorship.Measurersmayutilizecustomizabletimecut-offreportingorUGCagesegmentationwhereclearlydefinedandreportedwithtransparentmethodology.Insituationswhereactivityismeasuredrelatedtore-Authorshipoforiginalcontentandcut-offrulesareappliedseparately,reportingshouldbesegregatedfromthatofactivityrelatedtooriginalAuthorship.


30

4.3 Video

MeasuringAccessestoSocialMediaEventData

WhenmeasuringaccesstoSocialMediaonvideo,orUserGeneratedVideo,therequirementsassociatedwithconventionalvideomeasurementapply.Videoaccessshouldbecountedusingaclient-sidetechniqueonlywhenthespecificSocialMediacontentisencountered(ifitisnotrepresentedintheentirevideoexecution),intheviewableareaofthescreen,postcaching.

Videoaccesscanbemeasuredusingseveraltypesofmetrics:

• VideoAdImpression(served)–Thismetricrepresentsthecountofthefirstframeof

thevideoad,postbufferingifapplicable.Thisisnolongerconsideredacurrencymetric.

• VideoAdImpression(Viewable)–Generally50%ormoreofpixelsinviewforatleast2continuousseconds(aswellasotherrequirementsoftheviewabilityguidelines)isrequiredforvideoinadesktopenvironment.ThisguidancemaybereplacedorsupersededbyadditionalguidancepromulgatedbytheMRCwithregardtomobileandstream/feedenvironments.

• Variousmetricscanbeusedtomeasureprogressthrougheitherthevideoortheapplicationcontent(completion,quartiles,percentage,retention,view-throughrate,costpertimemetrics,etc.),whichcanactasasurrogatetouserinterestorengagementatabasiclevel.

• Variousmetricscanbeusedforuserinteractionswiththevideoortheapplicationcontent,whichagaincanactasasurrogatetouserinterestorengagement–forexample,swipes,taps,movementorhovers,etc.

AnalyzingUGCVideoAnalyzingandunderstandingthecontentandsentimentwithinuser-generatedvideoishighlyrelevanttothisguideline.Severalgeneralprinciples,someofwhicharedescribedinsection3ofthisdocument,apply:

• Verificationoftheuser-generated(Earned)statusofthevideo,asdistinguishedfrom

Paidmedia(seeSection2.0forfurtherguidanceonthesegregationinreportingofOwnedContentandPaid/Sponsoredmetrics).

• Evaluatingtheentirestringofcontent,ratherthanasample.Ifsamplingisemployed,specificparametersshouldbedisclosed.

• Ensuringthatusergeneratedvideocontentisnotanalyzedmorethanonceforameasuredmetric–thiscanbecomplexgivenduplicationand/ormultiplecopiesorpartialcopiesthatcanresidewithinplatforms(seesection3.3).

• Applyingaconsistent,transparent(disclosed)setofoverallqualificationcriteria,searchtermsandcontextevaluationrules.


31

• Applyingmanual-interventionwhenmachinelearneddecisionrulesencounterambiguitiesorwheresignificantdoubtexistsastotruecontext.

4.4 FrequencyCapping

Incertainsituations,socialplatformsmayenforcefrequencycapping,oralimitationofthenumberofadsthatmaybeservedeithertothetotaluserpopulationorasubsetofusers(suchasusersoutsidethenetworkofaspecificbrand).Suchaparametermayeitherbeenforcedbydefaultoftheplatformorasconfiguredbyabrandorotherentityservingadsontheplatform.SuchlimitationsofexposureorReachshouldbefullydisclosedtousersoftheplatforminthecontextofreportedmetricswherepossible.

5 EnhancingSocialMediaMeasurementAccuracy

Note:MostofthissectionisapplicabletotrackingorcountingSocialMediauserexposures,ratherthanUGCcontentanalysis.

5.1 CacheBusting

Incertainsituations,CacheBustingtechniquesshouldbeemployedbySocialMediameasurementorganizationstominimizeundercountingofSocialMediaactivity,ifthatactivityisrelatedtoimpressions,web-pagecallsoruseraccesses.CacheBustingtechniquesareappliedbysitesorad-servingentitiesandconsistofHTTPheadercontrolsand/orrandomnumberassignmentsappliedtotransactionalinformationtomakethemuniquefromothereventssotheyarenotsuppressedforcountingpurposes.

5.2 FiltrationforKnownInvalidActivity

AllmetricssubjecttoauditbyMRCorcertificationauditorsareexpectedtocomplywiththeMRC’sInvalidTrafficandFiltrationGuidelinesAddendum.ThisincludesSocialMediametrics,whichshouldbefilteredforknownGeneralInvalidTrafficasrequiredbythoseguidelines.Furthermore,likeothermetrics,applicationofSophisticatedInvalidTrafficdetectionprocessesarestronglyencouragedforSocialMediametrics.SuchSophisticatedtechniquesmaybefurtherenabledinaSocialMediadatacollectionenvironmentandplatformsareencouragedtoextendthesetechniquestorobustaccountanduservalidationprocedures.

However,whileSocialMediaPlatformsorUGCcollectorsmayhavetheabilitytoutilizegranulardatatoconductGeneralandSophisticatedInvalidTrafficdetection,third-partyaggregatorsorusersofAPIandfirehosedatamaybelimitedintheirvisibilityintogranulardataandthereforelimitedintheirabilitytofilter.InsituationswheresuchpartiesutilizedatafromupstreamplatformsorUGCcollectorswithaccreditedfiltrationtechniques,thisshouldbeconfirmedand


32

disclosedbytheaggregatorwithnofurtherrequirementforadditionalfiltration.Howeverwheresuchaccreditedupstreamfiltrationtechniquesareeithernotpresentorunknown,suchpartiesshouldmakereasonableeffortstoeitherobtainplatformdatathatallowsthemtocomplywiththeseguidelinesorperformadditionaldownstreamanalyticalprocedurestoserveasaproxyformoregranularanalysis.EffortsshouldbemadetoidentifyandexcludeSpamfromAuthorshipandEngagementtotheextentknownbythemeasurementorganization.

5.3 SegregationofPre-FetchandPre-RenderActivity

Eventsthatoccurbecauseofnon-user-initiatedpre-fetchorpre-rendertechniquesshouldbesegregatedforSocialMediameasurementreportingpurposesandappropriatelydescribedsotheusersofmeasurementdataareinformedofspecificorigins.AspartoftheMRC’sefforttomodernizeexistingdigitalmeasurementguidelines,wecontinuetopushformeasurementattheclient(user)sideandaslateinthedeliveryprocessaspossible.ItisimportantthatSocialMediameasurement,particularlythatofmobileactivity,considerthedelivery,loadingandrendering(withconsiderationofAdBlocking)ofcontentasrequiredpriortoinitiatingmeasurement.

5.4 Auto-InitiatedContent,OtherNon-UserIntendedContent

Valid,measuredSocialMediaeventscanincludethosewheretheunderlyingactivitywaseitheruserorautoinitiated,howeveranauto-initiated(notrequiringdirectuserinvolvementincludingauto-playcontentanduser-configuredautomatedactivity)eventbasedonsocialeventsshouldbesegregatedanddescribedforreportingpurposes.

Additionally,ifameasurerhasknowledgethatausereventoccurredinanunintendedmanner,throughtricking,misdirectingormaskingtruelocations,theseeventsshouldnotbecountedinreportedevents.

5.5 Applications–On-Linevs.Off-LineActivity

Inanapplication-basedplatformpertainingtoSocialMedia,activitycanbeconductedon-lineoroff-line(off-lineactivityisposted,updatedand/ormeasuredupontheuserre-engagingon-line).Alloftheseeventscanbelegitimatelymeasuredbuttheexistenceandnatureofoff-lineeventsshouldbedescribedinreportsofSocialMediameasurement.Ingeneral,itisdesirabletodisclosethepercentageofmeasuredeventsthatoccurredoff-line.Thepassageoftimecanmakeoff-lineeventslessvaluable,soexcessiveamountsofdelayinre-engagingon-line(forexample,thosewhereprioron-lineexposurewaspriortothereportingperiod)shouldbequantifiedanddisclosed.


33

5.6 InactivityConsiderationsandLimits

WerefertheuserofthisguidelinetopriorIABguidanceonAudienceReachMeasurementswheretheterm“Session”isdefined.ThisguidanceindicatesthedeterminationofasinglecontinuoussetofactivityattributabletoabrowserorusercaninvolvemakingassumptionsaboutuseractivityandissubjecttomeasurementlimitationssuchasdynamicIPaddressingandcookiedeletion.Measurementorganizationsshouldconsiderthesedifficultieswhencreatingrulesfordeterminationofsessions/visits.SocialMediabehaviormaysupportSessiondeterminationanddefinitiondifferentthangeneralInternetbehaviorandmeasurersareencouragedtostudyandsupportanyalternatedefinitions.

Inmeasuringsessions,itisgenerallynecessarytoterminatesessionsifcertainperiodsofinactivityareencounteredforauser.IftheSocialMediameasurementreliesonidentifyingusers,uniqueusersorsessions,webelievethisguidanceishighlyrelevantandrequired.

6 GeneralReportingParameters

6.1 GeneralParameters

Generalreportingparameters(dayparts(andmoregranulartime),weekparts,timezones,etc.)provideforconsistencyandcomparability.Theseshouldbebasedonthelogicalapplicationofinformationabouttheusagepatternsofthemedium.

InordertoprovideformorestandardizationinSocialMediameasurementreporting,thefollowinggeneralreportingparametersarerecommended.Notethattheseareonlyseveralofthepossiblereportingparametersthatmaybeused.Ifparametersinadditiontothesearereported,similarrulesshouldbedefinedandapplied.

6.2 Time

Day—12:00midnightto12:00midnight

Dayparts—SocialMediausagepatternsneedfurtheranalysistodeterminetheusefulnessofestablishingeffectiveandlogicalstandardizedreportingdayparts.Weencouragesuchanalysistodeterminetheneedforstandardizationofthismeasurementparameter.

TotheextentthatSocialMediameasurementisspecifictoamediavertical(e.g.,TV),measurersareencouragedtoconformtoexistingandstandardizeddayparts,especiallywithregardtocross-mediacomparisonsorGRPs.However,itislikelythatmedia-agnosticmeasurementwillneedtobefurtherstudiedtodeterminetrafficandusagepatterns.SocialMediaspecificdaypartsshouldbesupportedbyempiricaltrafficanalysis.Customdaypartsshouldbefullydisclosed.


34

TimeZone—Fulldisclosureofthetimezoneusedtoproducethemeasurementreportisrequired.Itispreferable,althoughnotacurrentcompliancerequirement,forcertifiedpublishersoradserverstohavetheabilitytoproducemeasurementreportsinaconsistenttimezonesobuyerscanassessactivityacrossmeasurementorganizations.ForUS-basedreportsitisrecommendedthatreportsbeavailableonthebasisoftheEasternTimezone,fornonUS-basedreportsthisisrecommendedtobeGMT.

Week—MondaythroughSunday

Weekparts—M-F,M-Sun,Sat,Sun,Sat-Sun

Month—Threereportingmethods:(1)TVBroadcastmonthdefinition.Inthisdefinition,theMonthbeginsontheMondayoftheweekcontainingthefirstfullweekendofthemonth,(2)4-weekperiods–(13peryear)consistentwithmediaplanningforothermedia,or(3)acalendarmonth.Forfinancialreportingpurposes,amonthisdefinedasacalendarmonth.

6.3 Location

6.3.1 TraditionalGeo-LocationMethods

Ifinformationaboutthegeographiclocationoftheusersiscollectedusingtraditionalgeo-locationmethods(forinstance,usingregistrationinformationorIP-baseddatafromathird-partyvendortodeterminethelocationoftheuser),andthisinformationisusedinreporting,anylimitationstothemethodsusedshouldbedisclosed.Ambiguitiesinwirelessoperatorroutingshouldbeaccountedforinlocationdeterminationandestimatedthroughprocessesderivedfromcarrier/adservercooperation.

Forregistrationdata,collectionandupdatepoliciesshouldbefullydisclosed.Furthermore,organizationsutilizingregistrationinformationforlocationarestronglyencouragedtodevelopstronglogicchecksandinspectiontechniquesandconsiderexistingdatacollectionguidance(e.g.,CRE,etc.).

6.3.2 ApplicationLocationServices

Ifinformationaboutthegeographiclocationoftheusersiscollectedusingapplicationlocationservices(suchasGPS-enableduserlocationdetermination,devicefingerprintingoranotherapplication-basedmeans),andthisinformationisusedinreporting,themethodusedandanyknownlimitationstoitshouldbedisclosed.


35

Totheextentthatplatformsormeasurersutilizebothtraditionalgeo-locationmethodsandapplicationlocationservices,discrepanciesbetweentheseapproachesshouldbereconciledwithsignificantconflictsaddressed.

6.3.3 HybridandEmergingLocationTechniques

Certainhybridandemergingtechniquesforlocationassignmentanddetectionareeithercurrentlypresentorbeingdeveloped,includingadvancedtechniquesinvolvinglocationmappingandproximatepointsofinterest.SuchtechniquesshouldbedisclosedindetailalongwithinherentlimitationsandcoveragewhenutilizedbySocialMediameasurersorplatforms.

6.4 SegregatingNon-LikeAdContentorActivity

Forreportingpurposes,measurementsshouldbesegregatedbythevarioustypesofads,UGCoreventsincludedinthecampaign.Forinstance,countsshouldbereportedseparatelyforadswithinthecampaignofdifferentsizesorfunctionalities,differentbrands,differentbrandsub-components,etc.

7 DisclosureGuidance

7.1 General

SocialMediameasurementorganizationsshouldfullydisclosetheirSocialMediaactivityrecordingprocesstobuyersandotherusersofthemeasurementdata.Anorganization’smethodologyforaccumulatingSocialMediaadvertisingmeasurementsshouldbefullydescribedtousersofthedata.Specifically,thenatureofmeasurements,methodsofsamplingused(ifapplicable),datacollectionmethodsemployed,dataeditingproceduresorothertypesofdataadjustmentorprojection,calculationexplanations,reportingstandards(ifapplicable),reliabilityofresults(ifapplicable)andlimitationsofthedatashouldbeincludedinthedisclosure.

SeeSection4.2.1forfurtherguidance.PlatformsandmeasurementorganizationsshoulddisclosethesedetailstousersofthedataviaaDescriptionofMethodology,updatedatleastonceannually.

Thefollowingpresentsexamplesofthetypesofinformationdisclosed.


36

NatureofSocialMediaMeasurements•NameandTypeofApplicationIncludedintheMeasurement•NameofMeasurementReport•TypeofMeasurementsReported

o TimePeriodsIncludedo DaysIncludedo BasisforMeasurement(MeasurementUnit)o GeographicAreaso SignificantSub-GroupingsofDatao Demographiccategories

•FormatsofReportedData•SpecialPromotionsImpactingMeasurements•NatureofAuditingAppliedandDirectionstoAccesstoAuditReport•Sampling/ProjectionsUsed

o SamplingMethodsUsedo ExplanationofProjectionMethods

DataCollectionMethodsEmployed

•MethodofDataCollectiono LoggingMethodo LoggingFrequencyo LoggingCapturePoint

•TypesofDataCollectedo ContentsofLogFiles

•ContactswithUsers(ifapplicableandallowableconsideringprivacy)•ResearchonAccuracyofBasicData

o CookieParticipationPercentages(ifapplicable)o LatencyEstimates

•RateofResponse(ifapplicable)

EditingorDataAdjustmentProcedures•CheckingRecordsforCompleteness•ConsistencyChecks•AccuracyChecks•RulesforHandlingInconsistencies•CircumstancesforDiscardingData•HandlingofPartialDataRecords

o AscriptionProcedures

ComputationofReportedResults•DescriptionofHowEstimatesareCalculated

o Illustrationsaredesirableo FrequencycapsenforcedimpactingadorReachmetrics

•WeightingTechniques(ifapplicable)


37

•VerificationorQualityControlChecksinDataProcessingOperations•Pre-ReleaseQualityControls•ReprocessingorErrorCorrectionRules

ReportingStandards(ifapplicable)

•RequirementsforInclusioninReports,BasedonMinimumActivityLevels

ReliabilityofResults

•SamplingError(ifapplicable)

Dataretentionrules•Maintainingsufficientdataorprocessesthatallowforaudittrail

LimitationsonDataUse•Non-samplingError•ErrorsorUnusualConditionsNotedinReportingPeriod•LimitationsofMeasurements

7.2 DataOwnershipDisclosures

UsersandmeasurersofSocialMediadatashouldclearlydiscloseanyapplicabledataownershiprequirementsorpoliciesaspartofbothcontractualarrangementsandmethodologicaldisclosures.SocialMediameasurersorplatformsshouldpromoteformaltransparencysurroundingdataderivedfromuseractivityonmeasuredplatformsandanyrelateduserprivacypoliciesviadisclosurewithinaDescriptionofMethodology.UsersofSocialMediadataandclientsofSocialMediaPlatformsormeasurers(includingadvertisers,agenciesandpublishers)areencouragedtoincorporatetermsandclarificationregardingdataownershipintocontractualarrangements,termsofserviceortransactionstipulationsinadvance.

8 AuditingGuidelines

8.1 General

ThirdpartyindependentauditingisencouragedforallSocialMediameasurementsusedinthebuyingandsellingprocess(orwhenpresentedascomparabletothosemeasurementsthatare).Thisauditingisrecommendedtoincludebothcountingmethodsandprocessing/controlsasfollows:


38

1.CountingMethods:Independentverificationofactivityforadefinedperiod.Countingmethodproceduresgenerallyincludeabasicprocessreviewandriskanalysistounderstandthemeasurementmethods,analyticalreview,transactionauthentication,validationoffiltrationproceduresandmeasurementrecalculations.

2.Processes/Controls:Examinationoftheinternalcontrolssurroundingallphasesofthemeasurementprocess.Processauditingincludesexaminationoftheadequacyofappliedfiltrationtechniques.Althoughauditreportscanbeissuedasinfrequentlyasonceperyear,someaudittestingshouldextendtomorethanoneperiodduringtheyeartoassureinternalcontrolsaremaintained.Auditreportsshouldclearlystatetheperiodscoveredbytheunderlyingaudittestingandtheperiodcoveredbytheresultingcertification.

8.2 U.S.CertificationRecommendation

AllSocialMediameasurementapplicationsusedinthebuyingandsellingprocessarerecommendedtobecertifiedascompliantwiththeseguidelines,atminimumannually.Thisrecommendationisstronglysupportedbythe4Asandothermembersofthebuyingcommunity,forconsiderationofmeasurementsas“currency.”

SpecialAuditingGuidanceforsoftwareprovidersthatmarketSocialMediameasurementsoftwaretoSocialMediaPlatformsorpublishersforuseonthepublisher’sITinfrastructure(i.e.,“outsourced”)shouldconsiderthefollowingadditionalguidance:

1.Thestandardizedmeasurementsoftwareshouldbecertifiedonaone-timebasisatthemeasurementorganization,andthiscertificationisappliedtoeachcustomer.Thiscentralizedcertificationisrequiredatminimumannually.

2.Eachcustomer’sinfrastructure(andanymodificationsthatcustomerhasmadetothemeasurementsoftware,ifany)shouldbeindividuallyauditedtoassurecontinuedfunctioningofthesoftwareandthepresenceofappropriateinternalcontrols.Processesperformedinthecentralizedcertificationapplicabletotheoutsourcedsoftwarearegenerallynotre-performed.Theassessmentofcustomerinternalcontrols(andmodificationsmadetooutsourcedsoftware,ifany)isalsorecommendedtobeatminimumanannualprocedure.Thesecertificationproceduresareonlynecessaryforoutsourceclientswhowishtopresenttheirmeasurementsforusebybuyers.

SpecialAuditingGuidanceforAdvertisingAgenciesorOtherBuyingOrganizations

IfbuyingorganizationsmodifyorotherwisemanipulatemeasurementsfromcertifiedSocialMediameasurementorganizationsuponreceipt,auditingoftheseactivitiesshouldbeconsidered.


39

InadditiontoMRCanditscongressionalsupportedcertificationprocessforthebroadcastindustry,thereareanumberofothercertifiersandtypesandlevelsofcertificationavailabletoadservingorganizations.

9 WhoWeAreandContact

AbouttheMediaRatingCouncil(MRC)

TheMediaRatingCouncilisanon-profitindustryassociationestablishedin1963comprisedofleadingtelevision,radio,printanddigitalmediacompanies,aswellasadvertisers,advertisingagenciesandtradeassociations,whosegoalistoensuremeasurementservicesthatarevalid,reliableandeffective.MeasurementservicesdesiringMRCaccreditationarerequiredtodisclosetotheircustomersallmethodologicalaspectsoftheirservice;complywiththeMRCMinimumStandardsforMediaRatingResearchaswellasotherapplicableindustrymeasurementguidelines;andsubmittoMRC-designedauditstoauthenticateandilluminatetheirprocedures.Inaddition,theMRCmembershipactivelypursuesresearchissuestheyconsiderprioritiesinanefforttoimprovethequalityofresearchinthemarketplace.Currentlyapproximately110researchproductsareauditedbytheMRC.AdditionalinformationaboutMRCcanbefoundatwww.mediaratingcouncil.org.

AbouttheAmericanAssociationofAdvertisingAgencies(4A’s)

The4A'sisthecatalystforbringingtogethertherightpeopleintherightplaceattherighttimetoaddresstheadvertisingindustry'smostcriticalbusinessissues.Theyprovideleadership,advocacy,guidanceandcommunitytotheirmembersandtheindustryatlarge,withproprietaryaccesstothepeople,informationandtoolsneededtomakesmartermanagementdecisions.Theirmissionistohelpagenciesbecomemoresuccessful.

The4A'sSocialMediaCommittee,anagency-ledseniorexecutivecommitteeofthe4A's,wasformedin2013,andisdedicatedtodefininganddrivingbestpracticesforsocialmediamarketingandadvertising.Theirgoalistoactasaresourceandthought-leaderforthesocialmediapracticeinsideagenciesandinspire,recognizeanddevelopfuturetalent.

AbouttheInteractiveAdvertisingBureau(IABU.S.)

TheInteractiveAdvertisingBureau(IAB)empowersthemediaandmarketingindustriestothriveinthedigitaleconomy.Itiscomprisedofmorethan650leadingmediaandtechnologycompaniesthatareresponsibleforselling,delivering,andoptimizingdigitaladvertisingormarketingcampaigns.Together,theyaccountfor86percentofonlineadvertisingintheUnitedStates.Workingwithitsmembercompanies,theIABevaluatesandrecommendsstandardsandpracticesandfieldscriticalresearchoninteractiveadvertising.Theorganizationiscommittedto

http://www.mediaratingcouncil.org


40

professionaldevelopment,elevatingtheknowledge,skills,andexpertiseofindividualsacrossthedigitalmarketingindustry.TheIABalsoeducatesmarketers,agencies,mediacompaniesandthewiderbusinesscommunityaboutthevalueofinteractiveadvertising.Foundedin1996,theIABisheadquarteredinNewYorkCity.Formoreinformation,pleasevisitwww.iab.net.

AbouttheWordofMouthMarketingAssociation(WOMMA)

TheWordofMouthMarketingAssociation(WOMMA)istheofficialtradeassociationforthewordofmouthandsocialmediamarketingindustries.Foundedin2004,WOMMAistheleaderinethicalwordofmouthmarketingpracticesandoffersonlineandofflineeducation,professionaldevelopmentopportunities,andknowledgesharingwithtopindustrymarketers.WOMMA’smembershipismadeupofthemostinnovativecompaniescommittedtoadvancingthewordofmouthmarketingindustrythroughadvocacy,education,andethics.Formoreinformation,visitwww.WOMMA.org

ContactInformationForquestionsrelatedtothecontentofthisguideline,pleasecontact:

MRC:RonPinelliJr.,VPDigitalResearchandStandards212-972-0300rpinelli@mediaratingcouncil.org4A’s:JulesShapiro,[email protected]:SusanBorst,Director,[email protected]

WOMMA:PeterStorck,BoardMember&[email protected]

http://www.iab.net

http://www.womma.org/

http://www.womma.org/

mailto:[email protected]




MRC Social Measurement Guidelines

Documents

fina november2

final november2