Solving TV ad problems on- and offline
June 16, 2009 — Abigail Hamilton
Image courtesy of AdWeek
According to Ad Week, Ace Metrix is introducing a near-real-time TV commercial testing platform. That’s all fine and all, but the news reminded me of how more pressing advertising problems remain unsolved. Online advertising โ including online TV advertising โ is failing to monetize as well as people need it to.
With computers delivering rich user information to ad servers and quanitifiable (not estimated!) viewership and demographic information, online advertising should cost advertisers a lot to deliver to a few people โ because advertisers can find the right people. Their target audience, not a random crowd. Online TV is driving this formula in the right direction, because viewers are very intolerant of too much in-stream advertising.
Viewers are happy to “pay” for their TV by watching ads, but they “snap” and become hostile if the ad-for-content trade is imbalanced. After all, one of the reasons they are turning to the computer for TV is that primetime TV is now 41% commercials.
Online, TV viewers can be targeted to see ads that have a much, much better chance of being relevant to them. Traditional TV, not so much: How many Viagra, WeedEater, etc. ads and promos for shows I will not watch must I endure to get my TV content?
I’ll be interested to see if online ad rates go up, and their effectiveness goes up with it, due to sharp targeting and fewer competing/offputting ads.
Meanwhile, I’m pondering whether Ace Metrics is using the same old small-sample, unscientific approach, or whether its verbatim opinions (self-reporting) add something user behavior analysis misses?
The system evaluates dozens of variables across an array of demographic and geographic targets, but also asks panelists to provide their personal impression of spots. According to Goldman, more than 70 percent of panelists answer that open-ended question in detail, providing customers with a library of “verbatim” qualitative material to help assess ads.
The responses are crunched by proprietary algorithms that are used to generate an overall “Ace” score for each spot, as well as separate scores for individual variables.
Thoughts?!

