How Data Mining Can Help Advertisers Hit Their Targets

Knowledge at Wharton

Ask about Knowledge at Wharton

Episode Description

New research from Wharton senior fellow Shawndra Hill looks at how “second screening” impacts consumer response to TV ads. More: http://knlg.net/2kz8i01 📂 Marketing 🏷️ Research

Read the full transcriptHide the full transcript
  1. we're here today with Wharton senior fellow Chandra Hill to talk about her new research which focuses on TV ads
  2. online search and the connections between them Chandra thank you for being here today
  3. thank you for having me first of all could you give us a short description of what you looked at in this research
  4. absolutely so what we aim to achieve is to find new ways to measure tv-out effectiveness so let me take a step back
  5. and kind of talk a little bit about how people typically measure effectiveness large brand advertisers will usually ask
  6. another company to survey consumers and ask them questions like did you see the ad would you like to recommend the
  7. product that was advertised to your friends how did you feel about the ad so questions about their attitudes they
  8. might also look at sales data and correlate that with the amount of spent that they've made what we hope to do is
  9. look at more granular data that reveals itself in the searches people post on large search engines and so what we're
  10. hoping to do or have done is link TV ad data at an aggregate level where they you know can tell us precisely which
  11. television show what time which locations an ad was shown and then we look at search data around that TV ad
  12. before and after to see whether there was an impact on the search behavior and what we're actually trying to do is look
  13. at the ability to coordinate advertising efforts so not just on television but also on digital platforms like sponsored
  14. search so what we do is combine data from TV ads so the location of those ads so where they were shown not just the
  15. location in terms of geography but also which shows they were shown in and then we link that to the search data not just
  16. the searches but also conditioned on somebody making a search did they click on a sponsored search ad or not and we
  17. combine data from all of these sources to make causal claims about the impact of TV ads on digital behaviors towards
  18. measuring the effectiveness of TV ads so this features kind of capitalizes on a phenomenon that's going on then that
  19. last couple years called second screaming in that no longer do we watch TV and just look at the TV but often
  20. we're sitting on the couch looking at the TV and then also scrolling through our phones the whole
  21. time so tell me like when you looked at this a little more closely what were some of the key takeaways that you found
  22. um so first of all that's a great observation and I should probably take a step back and tell you the research
  23. questions that we were interested in so the number one research question we're interested in is just that like how does
  24. behaviors in response to TV ad manifest themselves via these second screens and what we found was that the response that
  25. we were seeing so we do in fact see that there's an increase in search behavior after a TV ad is shown but that's
  26. manifesting itself primarily on smartphones so the smaller device the more likely someone is to respond
  27. directly after a TV ad digitally and we also were interested in because we have very granular level search data not just
  28. in you know whether people are searching more but as I mentioned this interaction
  29. with the sponsored search ads and then finally we wanted to look at how the TV ads impacted different users in various
  30. ways so for instance we're interested in heterogeneous effects on demographics so
  31. age and gender so do certain genders respond differently to a particular creative that's shown in a particular
  32. television show similarly we looked at device and that's how we were able to discern that the
  33. response was coming primarily from the mobile phone so what were some of the findings that were most surprising to
  34. you I know one thing that I kind of stood out to me is this idea that you found that really when people are when
  35. this increase in searching on your phone is going on it's basically only amounts
  36. to about three minutes that's right so you hit the nail on the head in terms of that three surprising I mean well the
  37. surprising findings so there were two that I think are obvious in hindsight but we didn't necessarily anticipate so
  38. the first one was one that we already talked about by disaggregating the data and looking at different cohorts of
  39. people searching from smart phones versus tablets versus PCs we were able to see that the significant effect in
  40. terms of the bounce and searches after TV ad was happening only on mobile phone right so that's the first thing that was
  41. surprising to us although in hindsight it makes sense right if you're sitting in front of the television you're not
  42. going to bring your desktop - you know probably watch a television right and then the second one was because we're
  43. looking at very fine-grained windows so really for the first time are we doing this sort of granular response to TV ads
  44. minute by minute we were able to see the dynamic change in how people search after TV ad and as you mentioned we
  45. found that really we're seeing it either in the first second or third minute after the TV on and at first were like
  46. wow we expected this thing to sort of slow down but maybe tell off and the reason is we suspect is that TV ad
  47. segments are almost always exactly three minutes and so people are probably switching their attention back to the
  48. television show after the TV and TV ads are aired so if I am an advertiser you're looking at you have this three
  49. minute window people are on their phones they're looking at these ads what can i
  50. what can I do with this information how can I use this and maybe synchronize first of all are people even
  51. synchronizing now with in this way like trying to make sure if someone sees an ad on TV they may also see it on mobile
  52. and if not what can I do to kind of capitalize on this information that you find right so the implications of our
  53. work I think are many right so the first one is that because we're finding that the search response to television ads is
  54. manifesting itself primarily on mobile phones and from prior research not ours we know that people are more likely to
  55. click on the first ad only on a mobile phone when compared to say a PC or desktop and that's primarily because of
  56. the footprint right so you only see the first ad so what that suggests is if people really are moving to mobile phone
  57. when they're watching television that if you're an advertiser and you really want
  58. to keep their attention you should spend the money to make sure you're the first
  59. ad that shows up for the advertiser so that's the first one but then I think the work has even broader implications
  60. so because we can see who is responding right so let's just say let's take two examples let's say you have only one ad
  61. creative like one TV ad one commercial and you want to know sort of for this let's
  62. say it's a new product who's responding you can launch that TV on and basically you know look at the response in the way
  63. that we have and see which types of customers are responding and we're like which geographies are responding and
  64. that can help you sort of optimize your other your other advertising efforts to do more here or less there depending
  65. upon what you find the other example I wanted to point out is if instead you have many ad creatives you don't know in
  66. advance like maybe you've done some focus test and you know which one small groups like but you don't know in
  67. advance what the broader audience will respond best to you can launch all four of those or however many ads and see who
  68. is responding the best and then you know adjust how you present those ads over time so what this approach allows for is
  69. to do near real time optimization of ads with very aggregate level data now you asked a question of like what are people
  70. doing now right so for the most part people are measuring advertising effectiveness in the ways that I
  71. mentioned when we first started so asking you know people via surveys did they see the ad or looking at sales data
  72. but the future is quite different in that now there are solutions for TV networks and and even sort of solutions
  73. that sit outside of TV networks that allow people to buy advertisement programmatically so right now we're not
  74. all the way there so there's gonna be programmatic buys that more advertisers will do as well as something called
  75. addressable TV where people can actually advertise to individual households that
  76. they know information right so if you could do that then looking at this aggregate level data kind of is
  77. unnecessary but until we get there this way is a good way to begin to optimize campaigns this sort of is a beginning of
  78. saying that my Pretty Little Liars crowd is maybe a little different than my scandal crowd which is maybe different
  79. than my Monday Night Football exactly now speaking of that what were there some interesting things that you found
  80. in terms of demographic differences and how people reacted to this we did and that was one thing that we thought would
  81. be interested to advertisers so because you can see who's responding with respect to
  82. we looked at really just age and gender but still that's enough to give an advertiser insights we could match the
  83. demographics of the TV show for instance if you look at sporting events those tend to skew male and then ask when an
  84. advertisement is shown in a sporting event who what audience members are most likely to respond and we found perhaps
  85. obviously in hindsight that when a TV ad is shown in a sporting event men are much more likely to respond to it then
  86. when an ad is shown in a sporting event women really don't respond more than they would otherwise so the idea is
  87. maybe that your ads the the audience you might want to go after with these ads are the people that are already watching
  88. anyway which I assume they know but then also that it transfers over to online searches as well that's right and so and
  89. you can just check right so two things right so one you can check that the people that you're targeting are
  90. actually the ones responding so that's kind of like a validity check that your strategy is a good one
  91. but then in addition to that if you have two types of shows that let's just say men because we use that example are
  92. likely to watch you can compare and see like which type of show when an ad is placed in it are men most likely to
  93. respond to that ad because there could be all kinds of things going on perhaps in some shows people are more engaged
  94. with the show and are less likely to turn away from the commercials for instance or you know get up and maybe
  95. they're a longer show and they get up and do other things so you can look at the match between the type of show and
  96. the audience that's responding and it has two implications now just this research play into also the idea that
  97. more and more people are maybe turning away from broadcast TV and going more towards streaming for example because I
  98. mean I know when I watch Hulu for example because I have the kind where you do get ads with it is that it's
  99. asking me do I want the experience of this or do I want here or do I want to learn about do I want to travel video
  100. about California or do I want something about a cleaner I mean can this also be applied to other like be on broadcast TV
  101. so it can be applied to other advertising strategies where you have a specific time stamp associated
  102. with the event so that could be you know sort of placing a billboard in a particular location and then taking it
  103. away it could be a you have a radio advertisement and it has a certain time so the type of methodology that we
  104. actually advocate for is one that allows us to tease out the causality between an
  105. event that has a specific time and behavior that happens you know after that event by comparing it not only to
  106. what happened before but also to some control group that we come up with but so any event based advertising it would
  107. work but to answer your question about Hulu and let's say Netflix those those solutions for sort of media consumption
  108. or a little bit different in that they know who you are right like they have your information so what they can do is
  109. closer to the addressable television example that I mentioned earlier where people can sort of now advertise
  110. directly to individual household so companies like Hulu and Netflix have the ability already to do sort of one-to-one
  111. advertising and they can use your your behavior either on their own site or by matching their data with third parties
  112. to target to you directly now what is their guess what sets this research apart from other research
  113. that's been conducted on this topic so there are a few things that set it apart right so the one thing is the
  114. granularity of the data so because of the scale of the data we were able to look at minute by minute response by
  115. different locations so that's one thing the second thing is that for these searchers we also have as I mentioned
  116. demographics very crude high level demographics but we were able to then look at these heterogeneous treatment
  117. effects at four demographics and then also by device type which no one's done before and then finally this combination
  118. between not just looking at search response but looking at the clicks so conditioned on a search looking at the
  119. sponsored ad clicks is something that's also novel so looking at how firms might
  120. begin to coordinate their advertising efforts something that hasn't yet been done
  121. before when looking at response to TV ads and what's next for this research I know you've done a lot with social TV in
  122. the past couple years but where were you gonna go with this next so there are a number of sort of obvious extensions so
  123. we want to well and we are have already started to bring in just organic search so we focused in the first paper on
  124. sponsored search and we can look at well you know when an organic search response
  125. actually is for the brands or not like does that make a difference in their likelihood to click after an ad and
  126. we're also looking at other types of digital responses not just search so we're looking at clicks on webpages
  127. associated with the brand and also looking at where people are coming in from when they make those clicks asking
  128. questions around which specific advertising platforms might be most effective right after a TV ad campaign
  129. and now after sort of those things that we're already working on what we plan to
  130. look at are assigning people to different categories so instead of thinking about it instead of assigning
  131. them to demographics like male or female we assign them to a place on the purchase funnel so are they ready to buy
  132. are they just seeking information are they doing comparison shopping and we want to know whether the TV ad is more
  133. or less effective depending upon where people are in that purchase funnel other things that we'd like to do but we'll
  134. you know need some convincing for a partner or we have used observational data techniques and we feel pretty
  135. strongly that our method for teasing out the causality the relation the causal relationship between the TV ads and the
  136. search response is pretty solid however what we'd really like to do is run an experiment while TV ads are running to
  137. make sure that what we're finding for this with the sponsored search results are it's really true that in fact there
  138. is the impact on sponsored search results that we're seeing and I guess the pie in the sky you know kind of
  139. future work would be to actually run experiments using addressable TV right so our work I think will last
  140. for quite a while because although addressable TV exists today they're not that many advertisers that have adopted
  141. right away but what we want to see again is whether these different combinations
  142. of advertising lead to sort of more or less spend more or less clicks more or less search for information and in using
  143. addressable TV solutions in combination with experiment on sponsored research you can get precisely at that answer
  144. Chandra thanks so much for being with us today thanks Rito you [Music]

Brief Summary

Chandra Hill discusses her research on TV ads' impact on online search behavior and the importance of mobile synchronization for advertisers.

Key points

  • Three-Minute Search Spike. After a TV ad, search behavior spikes for about three minutes, primarily on smartphones.
  • Impact of TV Ads on Digital Behavior. Chandra Hill reveals how TV ads primarily drive search behavior on mobile devices.
  • Demographic Insights on Ad Responses. Research shows that men respond more to ads during sporting events than women.

Tension

45/ 100Some tension

Related episodes