追求相关性报告：一切皆因广告个性化

1、WHAT WE KNOW TODAY123PERSONALIZED ADS MAKE BETTER WEB EXPERIENCES+9% more enjoyable and +8% moreseamless experienceGENERALLY, THE MORE PERSONALIZED, THE BETTER THE WEB EXPERIENCEAds with the highest level of personalization lead to +6% more enjoyable experiences compared to ads with the lowest level

2、 of personalizationABOVE ALL, PEOPLE EXPECT RELEVANCE BASED ON THEM AND THEIR INTERESTSInterest based relevance is the top reportedexpectation, even above useful info2SOURCE: The Consumer Take On Data + AdPersonalization, Verizon, MAGNA, IPG Media Lab123B U T T H EQ U E S T I O NI S DOPERSONALIZED A

3、DSDRIVEMETRICS THATMATTERFORBRANDS?3THE SCOPEOF OUR TESTINGdemospast purchaselocationlife eventsearch termTYPES OF PERSONALIZATIONINDUSTRY VERTICALSAUDIENCE TYPESAutoRetailerFinanceDemo TargetsHVAs (High Value Audiences)DEVICES4Demo targeted based on age; Past purchase targeted based on reported pas

4、t product purchases made; Location targeted based on current location of user; Life events targeted based on reported major life events (anniversary, new baby, etc.) ; Search terms targeted based on reported terms searched for recentlyDEMOGRAPHICSInitial demographic and personalization questions use

5、d for ad personalization as part of larger batteryRANDOMIZATIONRespondents randomized to see personalized ad or standard ad within mocked up version of YouTube page. Personalization is determined based on previous survey questionsYOUTUBE EXPERIENCEParticipants visit YouTube testing page, where they

6、select and play video content based on their interestsSURVEYPost exposure survey to measure ad opinions, traditional branding metrics, and consumer perceptions of ad personalizationRECRUITMENTYouTube users recruited fromonline panelNationally representative sample for demo targetHVAs identified thro

7、ugh Kinesso database matchTotal n=6,603 (4,105 Demo,2,498 HVA)HOW WE DID IT01020304055PERSONALIZATION ISBENEFICIAL,BUTDOESNTGUARANTEEBRANDING67YES, PERSONALIZED ADS ACHIEVE RELEVANCE IN THE MOMENT= significant difference between standard ads and personalized ad at =90% confidence(Demo Targets ) Stan

8、dard Ads: n=1,133; Personalized n=1,151:43%39%48%48%43%52%RelevantSomething Im open toA good reminderAD PERCEPTIONSTotal Exposed - % Agree Standard Ads Personalized Ads = significant difference between standard ads and personalized ad at =90% confidence(Demo Target; Retailer Brand) With Brand in Tag

9、line - Standards Ads n=74; Personalized Ads n-=77 Without Brand in Tagline - Standards Ads n=323; Personalized Ads n=3088BUT PERSONALIZED ADS ARENT IMMUNE TO BRANDING ISSUESEnsure personalization doesnt take away from branding; consider including brand name in personalization text+8%WITH BRANDING IN

10、 PERSONALIZED MESSAGENO CHANGEWITHOUT BRANDING IN PERSONALIZED MESSAGEUNAIDED AD RECALLRetailer Brand - % Recall (Personalized Ads - Standard Ads)= significant difference between standard ads and personalized ad at =90% confidence(Demo Targets) Standard Ads n=586; Personalized Ads n=5299+7% AD PERCE

11、PTIONSThose Who Recall Ad - % AgreeStandard AdsPersonalized AdsSTRONG BRANDING+ PERSONALIZATION= AMPLIFIED BRAND STORY(+6)= significant difference between standard ads and personalized ad at =90% confidence(Demo Targets) Standard Ads n=586; Personalized Ads n=52910PERSONALIZED ADS FORGE MORE PERSONA

12、L CONNECTIONS,LEAVING PEOPLE OPENWANT TO HEAR MORE FROM BRAND (T2B)FELT CONNECTEDTO BRAND (TB)53% 21%26%47%TO HEARING MOREBRAND PERCEPTIONSThose Who Recall Ad - % AgreeStandard AdsPersonalized AdsULTIMATELY, A PERSONALIZED MESSAGE CAN LEAD TO +5% ON PURCHASE INTENT= significant difference between st

13、andard ads and personalized ad at =90% confidence(Demo Targets) Standard Ads n=586; Personalized Ads n=52911PURCHASE INTENTThose Who Recall Ad - % Agree55%60% STANDARD ADSPERSONALIZED ADSNOTALLTYPESOFPERSONALIZATIONARE CREATEDEQUALLY12BRAND PERCEPTIONSOVERALL FAVORABILITYBRAND PREFERENCEPURCHASE INT

14、ENTHIGHER LEVELS OF PERSONALIZATION MAY BE NEEDED TO SHIFT MORE DIFFICULT TO MOVE METRICS(Demo Targets) Standard Ads n=586; Personalized Ads n=52913BRAND METRICSAGELOCATIONLIFE EVENTSSEARCH TERMSPAST PURCHASEMore PersonalizationLess PersonalizationSUMMARY OF IMPACTThose Who Recall Ad Personalized Ad

15、s Outperform Standard AdsMiddle FunnelBottom Funnel= significant difference between standard ads and personalized ad at =90% confidence(Demo Targets, Location) Standard Ads n=105; Personalized Ads n=95 Note: Personalization example is mock-up only14TRACKING IS OBVIOUS WITH LOCATION MESSAGING, BUT IT

16、 DELIVERS BY BUILDING FAVORABILITY THROUGH RELEVANCETHE AD42%24%+14%+14%RelevantFeel like Im being trackedTHE BRANDNIGHT OUT IN AUSTIN?SMELL FRESH TONIGHTLOCATION PERSONALIZATIONThose Who Recall Ad - % Agree Standard Ads Personalized Ads+16%“Brand is Relevant” (T2B)+14%Overall Favorability (T2B)RUNN

17、ING OUT?KEEPSMELLINGFRESH= significant difference between standard ads and personalized ad at =90% confidence(Demo Targets, Past Purchase) Standard Ads n=110; Personalized Ads n=100Note: Personalization example is mock-up only15MESSAGING TAILORED TO PAST PURCHASES CONVEYS INNOVATION AND HIGHLIGHTS S

18、TORYTELLINGPAST PURCHASE PERSONALIZATIONThose Who Recall Ad - % Agree Standard Ads Personalized AdsTHE AD34%+12%48%+12%51%+12%Good reminderInnovativeTHE BRANDWould share/discuss49%61%“Has A UniqueStory To Tell”= significant difference between standard ads and personalized ad at =90% confidence(Demo

19、Targets, Search Terms) Standard Ads n=136; Personalized Ads n=122Note: Personalization example is mock-up only16SEARCH TERM- ENABLED MESSAGING OFFERS UTILITY AND BUILDS PREFERENCETHE ADTHE BRAND54%+12%61%+10%Good reminderProvided new info22%30%Brand PreferenceSEARCH TERM PERSONALIZATIONThose Who Rec

20、all Ad - % Agree Standard Ads Personalized AdsSCENTBEST COLOGNEWOODSYEARTHY SCENTPLEASANT SCENTSUBTLE COLOGNEDEODORANTHOT+16%+12%+12%+12%+12%= significant difference between standard ads and personalized ad at =90% confidence(Demo Targets, Life Events) Standard Ads n=126; Personalized Ads n=121Note:

21、 Personalization example is mock-up only17PERSONALIZING BASED ON LIFE EVENTS STRIKES AN EMOTIONAL CHORD, RESULTING IN PERSUASIONTHE ADTHE BRAND49%53%+11%+13%InspiringTold an interesting storyWant to hear more from (T2B)Respect (T2B)Overall Favorability (TB)Purchase Intent (TB)Recommend (T2B)LIFE EVE

22、NTS PERSONALIZATIONThose Who Recall Ad - % Agree Standard Ads Personalized AdsANNIVERSARY COMING UP?REMIND HER OF YOUR FIRST DATEINDUSTRYVERTICAL SHOULDINFORMPERSONALIZATIONSTRATEGY18HIGHEST ACCEPTABILITY TOWARDS LEAST PERSONAL PURCHASESQ: In your opinion, how acceptable is it for each type of brand

23、 to use your personal data to create personalized ads? (Demo Targets, Control Only) n=1,44019ACCEPTABILITY OF PERSONALIZED ADS BY VERTICALHighly Acceptable - % Agree (Scale: 8-10)30%Restaurants28%Food &Beverage28%Entertainment27%Retail27%Travel16%Pharma16%Auto15%Finance+_Personal+_AcceptabilityPERFO

24、RMANCE VARIES BY INDUSTRY THOSE WITH MORE PERSONAL PRODUCTS SHOULD BE PARTICULARLY THOUGHTFUL20IMPACT SUMMARY Personalized Ads Outperform Standard Ads Similar PerformanceRETAILERAUTOFINANCEAGELOCATIONLIFE EVENTSSEARCH TERMSPAST PURCHASENUANCED APPROACH IMPORTANTFOR SEARCH TERMS AND MAJOR LIFE EVENTS

25、Q: On a scale of 1 to 10, how protected do you think each of the following types of personal information should be? (Demo Targets, Control Only) n=1,44021PERCEIVED SENSITIVITY OF DATA TYPEIndexed to Average (100)109109989391Things Ive BoughtAgeCurrent LocationSearch TermsMajor Life EventsABOVEAVERAG

26、EMOBILE,ACOMPELLING CANVASFORPERSONALIZATION22PEOPLE ARE MOST ACCEPTING OF DATA USE FOR PERSONALIZATION ON MOBILEQ: What types of data do you find acceptable for advertisers to use to create personalized ads on device? (Demo Targets, Control Only) n=1,44023Differences In Acceptance Of Data Use For P

27、ersonalization By DeviceDelta (Mobile % Acceptable Desktop/Laptop % Acceptable)-4%+10%+11%+11%+11%+12%+13%Things Ive BoughtCurrent LocationDemosPredicted MoodMajor Life EventsSocial Media HabitsPersonal EmailsApps InstalledWhere You Go OftenTime ZoneEmails from BrandsEntertainment HabitsSearch Histo

28、ryDevice+6%+5%+4%+3%+2%0%-1%Brand I want to hear more from (T2B)Brand I trust (T2B) Overall opinion (TB) Search intent (T2B)MOBILE ONLYPERSONALIZED ADS WORK HARD ON MOBILE, WHERE PEOPLE ARE MOST COMFORTABLE WITH IT24IMPACT OF PERSONALIZATION BY DEVICEThose Who Recall AdPurchase intent (TB & T2B)PC +

29、 MOBILEBrand is relevant (T2B)(Demo Targets) PC (Personalized n=398 , Standard n=322); Mobile (Personalized n=231 , Standard n=264)COUPLINGPERSONALIZATIONW I T HPRECISION TARGETING25P E R S O N A L I Z A T I O NW O R K SW H E N T A R G E T I N GB R O A D D E M O S .BUT,WHATABOUT HIGHVALUEAUDIENCES (

30、HVAS)?26WHATARE HVAS?27WHO THEY AREHVAs are specific personas that the brand has identified as having more value than consumers reached via traditional strategies (e.g. demo). HVAs are heavily customized based on demographics, category status, and psychographics.HOW THEYRE CREATEDHVAs are created fr

31、om a big data set with audience behavior tracked at the consumer level. Modeling is used to predict conversions from behaviors.HVAs are developed ethically with a focus on people.B R A N DThose thatlove yogaThose that drive far to workThose that are veganThose that have a pet catBikes everydayThose

32、that cookPAST RESEARCH SHOWS HVAs NEED MORE CUSTOMIZED MESSAGINGSource: From Cookies to People, Magna, IPG Lab28HVAs can be particularly critical of ads not customized to them0%25%50%AD IS “RELEVANT”AD “PROVIDED NEW INFORMATION”AD “ADDRESSED A CONCERN I HAD”AD PERCEPTIONS - % AGREE BASED ON PAST RES

33、EARCHHVADemoHVADemoHVADemo2933%PERSONALIZED ADS28%STANDARD ADSPERSONALIZATION MAKES ADS MORE RELEVANT FOR HVA AUDIENCESAD IS “SOMETHING IM OPEN TO RIGHT NOW”HVAs - % AgreeStandard AdsPersonalized Ads= significant difference between standard ads and personalized ad at =90% confidence(HVA Targets for

34、Pharmacy Retailer & Auto) Standard Ads n=371; Personalized Ads n=25630ULTIMATELY, HVAs RESPOND ESPECIALLY WELL TO PERSONALIZED ADSAdding precision targeting to precision messaging leads to best results= significant difference between standard ads and personalized ad at =90% confidence (Demo Targets

35、for Pharmacy Retailer & Auto) Standard Ads n=586; Personalized Ads n=529 (HVA Targets for Pharmacy Retailer & Auto) Standard Ads n=371; Personalized Ads n=256+11%+5%+8%+4%+2%0%Search Intent (T2B)Purchase Intent (TB)Brand PreferenceIMPACT OF PERSONALIZATION BY TARGETING TYPEThose Who Recall Ad Delta

36、(Personalized Ads Standard Ads)Demo TargetsHVA TargetsBUT, WHAT ABOUT PERSONALIZING MESSAGES BASED ON PERSONAS?demospast purchaselocationlife eventsearch termMessages Customized Based On:PERSONA-BASED PERSONALIZATIONSTANDARD PERSONALIZATION31Demo targeted based on age; Past purchase targeted based o

37、n reported past product purchases made; Location targeted based on current location of user; Life events targeted based on reported major life events (anniversary, new baby, etc.) ; Search terms targeted based on reported terms searched for recentlyMessages Customized Based On:Cooks at homeYogaHave adogBikes dailyVeganHVApersonas32BECAUSE EACH HVA IS UNIQUE, MAXIMUM IMPACT IS ACH