Google's Move to TV - Black Box or no?

Google has entered into television advertising with a new partnership with Echostar. The partnership includes the creation of an automated platform for ads running on EchoStar DISH Networks' 125 networks. Google will gain access to DISH ad inventory from across all channels and dayparts; Google's platform will then allow it to sell that inventory and provide measurement on those buys.

The buying of DISH Network ads will work much like AdWords and use a web-based auction system. The real-time reporting allows advertisers to see how their ads performed on a second-by-second basis and adjust creative or daypart scheduling accordingly. Advertisers can target by age or demographic data. Reporting data will be pulled anonymously from the four million DISH boxes currently in use. The program will, according to Google's Eric Schmidt, allow for the ads delivered to be more relevant and therefore more valuable to the viewer and advertisers, because they can be targeted more closely.

On the surface, this sounds like good news for advertisers. After all, commercial data on a second-by-second basis is something advertisers have long been searching for, yet even Nielsen's new commercial minutes ratings do not provide such detailed measurement.

But relevance and the ability to measure viewership alone are not enough. Advertisers will not be acting in their own best interests when they purchase television time via the AdWords ad management platform. Online, advertisers can look at the quality score assigned by Google, along with keyword bid, ad copy and landing page, to try to determine how the minimum price bid was determined and how they can improve how often their ad is served.

But on television, a "black box" process could hamstring advertisers, because the same tools for improvement do not exist.

There's also the problem of testing. Online, Google can be used to good advantage because advertisers can test different combinations of keyword, ad copy and landing page to see which drives the best ROI.

With television, testing is obviously far more cumbersome and costly. So while Google may be able to offer better measurement than the broadcast and cable networks themselves, what exactly are they offering to measure? Without a control ad against which to run a test, the measurement is far less useful. An advertiser may find that 120,000 people watched an ad. But, did they take an action as a result of the ad? Would a different ad from the advertiser have worked better?

As consumers lean more and more toward creating and/or choosing their own media for consumption when and where they want it (via linear television, time-shifted television with Tivo, via online sources like YouTube or broadcast network sites, via mobile devices, etc.) ads that aren't relevant haven't got a chance.

But I would love to see Google TV move beyond the black-box automation that it uses online.  If they do so, I believe that advertisers will be well-served.