SAN JUAN, Puerto Rico\u2014To TiVo\u2019s Lisa Lutz, it\u2019s not that people aren\u2019t asking the right questions about creating audience segments and targeting them across platforms and devices. \u201cIt\u2019s just that there\u2019s so many questions and so many problems to be solved, we\u2019re just not there yet,\u201d says the VP of Product Management.<\/p>\n
\u201cI think it\u2019s a wildfire right now,\u201d Lutz adds in this interview at last month\u2019s Beet Retreat 2018<\/a>.<\/p>\n She sees more agencies wanting to use TiVo\u2019s MVPD and box viewing data to create targets for viewers of connected-TV and non-linear devices \u201cand export that out as part of a segment.\u201d<\/p>\n It could be in combination with data about, say, politics or location. Based on viewing behavior, ads for a new horror movie could be targeted to people who are known to be theatre-goers and horror aficionados.<\/p>\n Then there is the concept of exposed and unexposed ads and how to achieve desired frequency and measure it. TiVo<\/a> can help to determine how to retarget people who have been exposed to specific ads or help to find \u201cthose viewers who I wanted to target but are light TV viewers and I didn\u2019t actually find them on TV,\u201d Lutz explains.<\/p>\n TiVo\u2019s second-by-second viewing data also provide insights into how to retain audiences and how effective efforts are to improve the commercial experience by adjusting pod lengths. While there\u2019s planning, measurement and attribution, \u201cthey\u2019re all very much in parallel. If I plan, I create my target, I will most likely model that target today because I want to get the scale that I need to find those viewers on platforms.<\/p>\n How does one determine whether a particular target is effective?<\/p>\n \u201cThe only way to know that is to start looking at attribution. During the course of that campaign, of that model segment who converted and look at how those converted viewers score against that model target. Was that model effective?<\/p>\n \u201cThere are so many different questions and measurements and evaluations that need to happen that are just iterative across a process,\u201d Lutz<\/a> says. \u201cI don\u2019t think you can look at things in a silo.\u201d<\/p>\n