The chart below demonstrates the benefit of using Simulmedia Segmentsb” targeting in promotional plans, based on results of an internal study of recently launched programs. Using anonymous set top box data, we first examined promotion exposure by segment during the eight weeks prior to the showbs original air date. We then looked at the first episode of the program and examined average tune-in for two different groups: Overall audience exposed to the promotions, and Simulmedia Segmentsb” exposed to the promotions and previously designated as the most receptive to these programs.

Results show that Simulmediabs data-driven predictive models consistently have a higher response rate than existing promotion plans. Furthermore, the targeted and untargeted exposed groups watched the same amount of the given network in the preceding eight weeks, suggesting that the response discrepancy is, indeed, the result of our targeting rather than simply the result of differences in television consumption.

Targeted advertising works, and moreover, the benefits are widespread.B Advertisers clearly benefit, as they can efficiently target their desired audiences rather than wasting money on audiences that are unlikely to respond to the promotions.B Viewers benefit as well.B They are reminded about shows that they already like and are informed about new shows that they would probably enjoy.B Targeted advertising even helps the TV networks that carry the promotions, since viewers are less likely to change the channel during a commercial, and potentially not return to that network, if they are intrigued by the programs advertised in the promotions.

The key remaining question, particularly for advertisers, is how to most efficiently target the desired audiences.B Online advertisers can pinpoint specific individuals who just bought a dog, or a home, or just got married, and can advertise accordingly, spending money to reach only the users who are most likely to respond to the ads.B Television advertising is more difficult.B Target audiences can be located and reached, but only by also reaching viewers who are less likely to respond.B Everyonebviewers, advertisers, networksbloses when advertisements are shown to the wrong people.B Through continued data collection and analysis, we can work to reduce this waste and improve the situations of numerous parties in the television world.

The promotional response rate is calculated by using this formula:

Targeting low loyalty high attentiveness viewers is an ideal way to drive ratings for established programs. We define low loyalty viewers as ones that watch only one episode of a program in a season. As we previously determined, most shows have large numbers of such low loyalty viewers and their attention is for sale. Imposing the additional requirement of bhigh attentiveness,b we ensure that our target is more receptive to a particular program.

Our recent study examined whether exposure to promotion makes the highly attentive low loyalty segments more likely to tune in to subsequent episodes of a program.

We isolated high attentiveness low loyalty February 2009 audiences for ABCbs Lost and FOXbs 24.B Any viewer who saw at least thirty non-consecutive minutes of Lost on exactly one out of four possible occasions in February was included in the high attentiveness low loyalty audience for Lost. The same criteria were used to create a high attentiveness low loyalty audience for FOXbs 24.

Subsequently, we split the February high attentiveness low loyalty audience into two segments based on their exposure to promotion in the first week of March, prior to the first March episode. We then examined first March episode tune in of the promo-exposed and promo-unexposed highly attentive low-loyalty viewers and compared their rates. The results for both programs have confirmed our hypothesis that promotional exposure moves highly attentive-low loyalty viewers to tune in at higher rates.

Additionally, the graphs below illustrate that the high attentiveness low loyalty audience is a more effective target of promotion than the general set-top box audience. In fact, the high attentiveness low loyalty audience exposed to promotion is more than three times as likely to tune in to the program as the exposed general audience.

In our future analysis we plan to investigate whether exposure to promotion made the low loyalty viewers more loyal, and determine the persistence of the effect.