Invidi says its system will allow a company -- say Procter & Gamble (PG
) -- to buy 30 seconds of ad time on a cable system and be certain that an ad for Secret antiperspirant goes, for example, to an audience that is 95% female, while an ad for P&G's Fixodent denture adhesive goes to viewers 55 and older.
BEHAVIORAL APPROACH. Invidi is confident that its software can anonymously profile households based on the "clickstream" of the remote control. An Invidi-armed digital cable box monitors the TV viewer by his or her remote-control choices and makes assumptions after about 20 to 25 clicks about the TV watcher's gender with about 95% accuracy. When age and pastimes are part of the targeting as well, accuracy drops to 75%.
Because Invidi is compatible with digital cable set-top boxes, it's the first system that can be rolled out on a nearly national scale that can target and profile consumers based on the programming they watch. Other systems exist that target viewers based on Zip code, for example, but Invidi's is the first to do it according to TV-watching behavior.
Time Warner (TWX
) will first test the system in its laboratories in Colorado. Then, probably by yearend, it expects to test the system in a metropolitan city market. Comcast (CMCSA
) plans a trial in 2006 in both a lab setting and households.
Invidi Technologies CEO and President David Downey recently spoke with BusinessWeek Marketing Editor David Kiley about the way the system works and the potential it has for catching on nationally. Edited excerpts of their conversation follow:
Q: Your system seems very "Big Brotherish." How can people be sure their TV-viewing patterns won't be monitored in a way that violates privacy?
A: The data is collected as bits in a digital stream and is never matched to the address of the house or the serial number of the cable box. We have been very careful about that, because, naturally, we don't want to set off any alarm bells with consumers. The cable companies want this technology, but not at the expense of people canceling their cable subscriptions.
Q: Does this work like Internet cookies [small files that Web sites place on a visitor's computer to track how often you visit a site, among other things]?
A: No. Internet cookies sit on your computer hard drive and wait for you to come back. That's not how this works. We monitor [a randomly chosen] 5% of a cable system's audience all the time. When a commercial break is coming up, the system looks at the 5% -- and the 5% changes all the time -- and their set-top boxes "vote" in a way automatically for which one of maybe five or six ads gets downloaded to that TV.
The cable system might have a Revlon ad, which will get sent not only to that portion of the 5% we are monitoring who match Revlon's desired profile, say women 18-34, but to all the households in the system where we detect a woman of the same profile is most likely watching [based on data collected and stored in our database and data stored in the cable set-top box in a given household. By collecting and storing the data from the changing 5% we are always monitoring, the 5% we are linked to at any one time accurately triggers the right ad to the whole cable system.]
A Home Depot (HD
) ad in the same break might be directed by the same method to go to the households in which a man 34-50 is watching.
Q: So, if a remote control has gone from MTV to Buffy The Vampire Slayer to Real World, etc., your system figures a woman in that age bracket is watching?
A: That's oversimplified, but you have the right idea.
Q: Will people have a chance to opt out of this if they want?
A: You'd have to ask the cable companies who are our clients. The hope is that consumers will actually want to opt in. If cable companies can earn more ad revenue, cable bills may stay low or even go lower. And we think that if ads are better targeted to be more relevant, and fewer ads waste the time of fewer consumers, then everybody wins.
Q: What is the biggest benefit of this system?
A: For advertisers, it's being able to send ads to people who would have an interest in seeing that particular ad. The biggest reason consumers skip ads when they have the chance is that too few of the people seeing a particular ad care about the product. If I don't have a dog, why do I want to see a pet food commercial?
If I am a man, I don't care about seeing a perfume ad. I skip that. And when consumers get trained to skip most advertising, they miss ads that are really relevant to them. Advertisers are very concerned about the level at which consumers are ignoring and avoiding their messages.