Measuring Google TV Advertising and Privacy

Google’s advertising model goes beyond the Web to places like televison. Yesterday, the Official Google Blog ran a post on TV advertising through Google, Tuning in to TV data, which told us that they are gauging interest in ads shown on TV by whether or not viewers change channels during commercials. A video featuring Google’s Dan Zigmond discusses how television ads might retain audiences:

I wonder about the approach, personally. When you’re watching TV and a commercial comes on, do you change the channel to see what else is on? Do you get up and grab a snack, or run a brief errand? Or, do you pay as much attention to the commercials as you do the show that surrounds them? If you do stay in front of the screen and pay attention to the advertising, do you change the channel if you don’t like an ad, or do you suffer through it knowing that it will be gone very soon?

I recalled seeing a patent application about Google television ads from last year that I didn’t write about, and decided that it might be worth returning to. One of the listed inventors on the patent application is Dan Zigmond, who narrated the Google video. The patent application is:

Forecasting TV Impressions
Invented by Jason Bayer, Greg Hecht, and Daniel J. Zigmond
Assigned to Google Inc.
US Patent Application 20080263578
Published October 23, 2008
Filed: March 28, 2007


A computer-implemented method for forecasting television impressions comprises receiving information relating to previous television (TV) impressions at a time slot on a TV channel, information relating to one or more programs shown at the time slot on the TV channel and predicting a future TV impression at the time slot on the TV channel. The future TV impression is based on at least one of the information related to previous TV impressions and the information related to programs shown at the time slot on the TV channel.

It makes sense for an advertiser to be able to tell how many viewers have historically watched certain television shows before they decide to place an advertisement on the show, and the patent filing details how it might predict the numbers of viewers by looking at past views, or impressions, for particular shows at specific timeslots on different channels. That prediction may consider other factors, such as whether or not the show to be presented is a rerun, or a special event, or what kind of content might be shown on other channels at the same time.

The patent application also describes an “advertisement decision engine” which might be used to help advertisers decide when to show their ads based upon those predictions concerning how many viewers might watch certain television programming.

Some other details from the patent filing:

1. Demographic information about past audience members of particular shows might be available to advertisers if the telecast provider makes that information available, such as “zip code, phone prefix, occupations, average income, and the like.” This information might be combined with other information about viewers taken from sources such as Experian to learn more about households where certain shows were viewed.

2. Advertisers can access information about content being shown on other channels from the telecast provider, such as whether or not the Oscars or the Superbowl is being broadcast on another channel, or viewing metrics from external sources such as Nielsen Media Research, Tribune Media Services, and others.

3. Viewing habits and behavioral patterns for previous shows and viewers might be accessible, including information about viewer’s “buying habits, hobbies, interests, and the like.”

4. Actual views of television programming can be collected by the telecast provider, to enable this system to “compare[s] the predicted and actual impressions and uses any variations to improve future predictions of impressions by the prediction system.” This information might be collected by monitoring set top boxes and recording instruments, such as digital video recorders.

5. Information about other shows being offered at specific times might be gathered from external sources such as electronic programming guides. This information might include additional information related to the shows being offered such as “genre, category, advisory information about the content, and the like.”


The patent filing discusses impressions for television shows that viewers might watch, and how those impressions might be used to help advertisers decide where to advertise.

The Official Google Blog from yesterday tells us that Google doesn’t stop at looking at impressions for the shows themselves:

Each week, Google analyzes data from millions of anonymized set-top boxes (STBs) to see which channels they were tuned to second by second. This data is provided by our partner, EchoStar. We’re then able to use tuning metrics to provide our advertisers with next-day reports of how many televisions showed their ads nationwide and how the audience responded with their remotes.

How might you define the success of advertising on television? One signal that you might look at is how many viewers started watching a commercial and changed the channel before the commercial had finished playing. Does that tell you whether or not the commercial was a good one or a bad one? It’s possible that some percentage of the audience wanted to see what else was available on other channels. Another percentage may have run to the refrigerator for a snack, or when off on another errand.

I’ve written about how Google might look at social interaction with television broadcasts in the future, and how they could use a microphone from your household to monitor ambient noise in your home, which would be transformed from actual audio to a visual representation of those sounds – which can’t be transformed back into audio – keeping any conversations private. A method like that might help the search engine know whether or not viewers are still in front of their TV screens when a commercial comes on, or have left the room. That post is: Google Radio and TV Personalization, Ratings, and Advertising Patent Applications.

In the future, will Google be watching what you’re watching on TV, listening in to see if you’re still in front of the screen when ads come on, noting whether or not you’re channel surfing when those ads are broadcast? Actually, much of that (except for the listening in part) is happening now for viewers using certain telecast services.

The Official Google Blog post concludes:

Through our analysis of tuning data from millions of set-top boxes, we’re getting closer to matching the right ads to the right television audience. It takes a lot of processing power to make sense of the enormous amount of data, but the insights to be gleaned are very powerful. Not only are we able to offer advertisers better measurement and more accountability for their TV campaigns, our goal is to also create a better viewing experience for TV audiences by showing viewers what they want to see.

Being able to measure how many people are paying attention to advertising on television might be helpful to advertisers, but it comes with a cost. As viewers, and as consumers, do we really want to be scrutinized so closely by Google or our television programming providers or advertisers?

Do we want a record of our viewing habits maintained in databases next to information about our purchasing habits, our hobbies and interests, and our household income? Do we want microphones listening in to what we are doing while we are watching TV, even if the people recording assure us that they transform audio into visual representations of that sound that can’t be retransformed into the conversations that we might be holding?


Author: Bill Slawski

Share This Post On