Google number-crunching to find good TV ads

Google believes the TV remote control is the next mouse, at least when it comes to measuring the fickle behavior of TV viewers and Web surfers regarding advertising.

Google TV Ads has been a formal project for a little over a year, and the company released new data Thursday showing how it's trying to help television advertisers quantify the impact of a good TV ad versus a bad one. Google has been quietly measuring the viewing habits of Dish Network customers, including how likely they are to watch all the way through a compelling, relevant ad or click away in disgust from the 10,000th airing of "Mini Sirloin Burgers."

This is a logical next step for Google's advertising brawn: television advertising metrics, and how the company can apply its expertise in Internet advertising to the small screen. The metric they focused on Thursday is called "% Initial Audience Retained," or the percentage of people who watch a given ad all the way through.

It may seem kind of obvious, but Google crunched the numbers to find out that past "stickiness," or %IAR, can predict future stickiness. That is, if over a certain period of time the numbers showed that more people were likely to stick around for an entire ad, future audiences will also be likely to stick around for that ad.

One thing this research doesn't seem to address is ad saturation. People may have initially responded to Toyota's "Saved By Zero," Subway's "Five-Dollar Footlong," or the ubiquitous (on the West Coast, at least) "Mini Sirloin Burgers" from Jack in the Box, but after you've seen that ad for the 50th time in the last three hours, why stick around?

But it seems the data gathered could be used to determine just that: when a successful TV ad campaign has reached the point of diminishing returns, or even outright backlash. Toyota's incessant campaign during last year's NFL season drove viewers up the wall, and might have hurt the company's image among a powerful demographic.