July 16, 2024

Measure It, and They Will Come

The technology for measuring TV and radio audiences is about to change in important ways, according to a long and interesting article, in yesterday’s New York Times Magazine, by Jon Gertner. This will have implications for websites, online media, and public life as well.

Standard audience-measurement technology, as used in the past by Nielsen and Arbitron, paid a few consumers to keep diaries of which TV and radio stations they watched and listened to, and when. Newer technology, such as Nielsen’s “people meters”, actually connect to TVs and measure when they are on and which channel they are tuned to; family members are asked to press buttons saying when they start and stop watching. People meter results were surprisingly different than diary results, perhaps because people wrote in their diaries the shows they planned to watch, or the shows they liked, or the shows they thought others would want them to be watching, rather than the shows they really did watch.

The hot new thing in audience measurement involves putting quiet watermarks (i.e., distinctive audio markers) in the background of shows that are broadcast, and then paying consumers to wear beeper-like devices that record the watermarks they hear. A key advantage of this technology, from the audience monitor’s viewpoint, is that it records what the person hears whereever they go. For example, current Nielsen ratings for TV only measure what people see on their own television at home. Anything seen or heard in a public place, or on the Internet, doesn’t factor into the ratings. That is going to change.

Another use of the new technology puts a distinctive watermark in each advertisement, and then record which ads people hear. When this happens – and it seems inevitable that it will – advertisers will be willing to pay more for audio ads in public places and on the Net, because they’ll be able to measure the effect of those ads. Audio ads will no longer be coupled to radio and TV stations, but will be deliverable by anybody who has people nearby. This will mean, inevitably, that we’ll hear more audio ads in public places and on the Net. That’ll be annoying.

Worse yet, by measuring what people actually hear, the technologies will strengthen advertisers’ incentives to deliver ads in ways that defeat the standard measures we use to skip or avoid them. No longer will advertisers measure attempts to deliver audio ads; now they’ll measure success in delivering sound waves to our ears. So we’ll hear more and more audio ads in captive-audience situations like elevators, taxicabs, and doctors’ waiting rooms. Won’t that be nice?


  1. This technology has been developed and is in use in Switzerland for some years now – being a small country, there was not even a reason to use watermarking, as there are not that many programs anyway that have to be tracked. It’s called Radiocontrol (www.radiocontrol.ch). People that participate in the studies usually wear a watch that gets sound snippets around them, that are then analysed. According to the leader of the project, this seems to work remarkably well – as long as no stations are not reusing content of another …

  2. The article downplayed “advanced” cable boxes and DVRs that watch what you watch. TiVo has made quite a public splash with this kind of information (notably with statistics from the Superbowl “wardrobe malfunction”). Presumably, second-generation CableCards (which support two-way communication) can also watch what you watch. Ignoring privacy issues for the moment, these sorts of “bulk” statistics give fantastic data on what’s playing on TV sets, particularly for less popular shows that Nielson-style sampling might miss, but less data about who’s actually doing the watching.

    The Portable People Meters, assuming they actually work, are clearly attractive for bridging this gap. An amusing thought to ponder is what else they could do if they could get rid of the need for the audio watermarks and could just do straight-up audio recognition. Do people spend more or less time in a bar if the volume is above 90dB? What songs make people buy more expensive food at a restaurant? I imagine they’re salivating at the measurement opportunities.

  3. I R A Darth Aggie says

    Anything seen or heard in a public place, or on the Internet, doesn’t factor into the ratings. That is going to change.

    Oh, joy! Even more incredibly difficult to interpret correctly sets of numbers.

    Scenario: I’m in a sports bar that has 20 TV’s. All of them are on. 18 of them are muted, and two are not. Now my little observing thingie may pick up two seperate sets of watermarks during the course of an evening.
    Q: Which program am I watching?
    A: Neither. I’m watching one (or more) of the 18 muted programs.
    And yes, I’ve gotten much better at reading and deciphering closed captioning…and, yes, I also switch TVs when non-interesting ads come on. Interesting ads mean “they’re funny”. 🙂 Conversely, perhaps neither watermark is detected. I’m now listed as “having watched or seen nothing” in that time slot.

    Another scenario: I’m out and about, and some punk^Wyoung person is driving by blaring out their favorite tunes, and a watermark is detected. Am I now in the demographic that enjoys gangsta rap?
    So far as I can tell, all this is going to do take muddy water and make it sufficiently murky as to be mostly useless.

  4. Ray: This is what advertisers call the “illusion of invincibility”. Advertisement works, else no one would pay for it.

  5. Yes, good article. Unfortunately for them, they wouldn’t get much from me. When an ad comes on the radio, I switch to another channel. I watch TV with one eye while on the computer. Usually I record TV and my VCR auto fast forwards through the ads. If I’m watching live, I hit the mute button until the show comes back on. When dumped on by advertising in a public space, I tune it out. My ad filter blocks most net ads. Wonder how many people are similar to me?

    Advertising is a fool’s game. Advertisers have buyers convinced with bogus statistics that advertising is valuable. Buyers want to believe that they are spending their money wisely and are afraid to find out they might be wrong. Which is why Nielsen has fought so hard for so many years to not implement technology that would show who is really watching what. If producers/advertisers found out the real story, the advertising industry would implode.

    Has anyone with an IQ higher than 90 even been influenced to buy a car by the pretty (and expensive) ads they see on TV? Me, I’m still waiting for the pretty babes to cluster around when I open a beer like happens on TV [lol].

  6. I think that such watermarks in the long run could be the bane of the advertising industry. Consider: if you know exactly what ads a consumer hears, and can correlate it with what that consumer buys, then you can start weeding out ineffective advertising. Which, for my money, is almost all the advertising out there.

  7. Fascinating article. I would certainly be concerned about privacy, carrying a device that listens to everything I hear, especially if it knows where I am via GPS.

    I wonder if there is any way to tell advertisers how much contempt we have for them? I’ve stopped watching television altogether. The advertising itself is bad enough, but the evolution of televised content to be a vector for advertising, by bypassing critical thought and higher brain functions, has made said television content simply not worth watching.

  8. Grant Gould says

    Oh yes indeed. Extract the watermark from your favorite ad (ideally one that you are getting paid for…), and blast it over a crowd from time to time. Wow. That could be quite amusing, and profitable as well.

    I wonder how much you could get paid as, eg, a movie theatre operator to surreptitiously insert ad-watermarks into a film’s audio track. You’d be nearly certain to hit at least one beeper-wearer for a first-run theatre in a major market city.

  9. It would also be fun to follow a beeper-wearer around and just blast your favorite water-marked at at them to dilute their marketing reality.

  10. Someone is going to build a pair of noise-cancellation headphones that turn on when they detect one of these watermarks and make a bajillion dollars, at least until the attorneys break his kneecaps.

  11. Cypherpunk says

    Actuallly it will be great to have captive-audience advertising in private places like doctor’s offices and other businesses. It will be a perfect opportunity to provide feedback to the advertisers about how much this is costing the public, thereby internalizing the external costs.

    Imagine how much doctors or supermarket managers will enjoy having their patients and customers berate them for forcing them to listen to constant ads, and threaten to go to some other provider! This will force every doctor, every business to consider the costs as well as the benefits of the advertising. Unlike TV, they don’t HAVE to do it to stay alive. It will only happen if the benefits outweigh the costs, forcing advertisers to pay enough to compensate for those costs.

    The result is that advertisers will bear the full social costs that they impose, which will lead to an economically and socially optimal level of advertising. It will be far better than the present system.

  12. On the other hand (assuming fraud isn’t easy, and anyone can in theory detect and verify a watermark), such watermarks also provide an essentially-incontrovertible record of who is paying to have a particular piece of audio sound in a particular place, thus making enforcement of any rules on public and semi-public audio fairly straightforward. (It will be interesting to see how the venues and the broad/narrowcasters divvy up the pie and whether auditors get any part of it.)

    I wonder whether ASCAP is already salivating over this.

  13. So we’ll hear more and more audio ads in captive-audience situations like elevators, taxicabs, and doctors’ waiting rooms. Won’t that be nice?

    You forgot restaurant bathrooms. Gotta love the forced advertising in front of the urinal.