Nielsen's mobile-TV challenge

Since the October release of Apple Computer's video iPod, engineers at a Nielsen Media Research laboratory outside Tampa Bay have been studying the device, looking for ways to measure what television shows are playing on the iPod's tiny tube.

The iPod, and the $1.99-per-episode television programming that Apple introduced with it, have brought home how much the TV business that Nielsen tracks is changing. Viewers once tied to their living room couches are becoming mobile, watching TV when and where they want.

Those changes mean the longtime ratings maven has to overcome major technology challenges if it wants to keep its customers--the TV networks and their advertisers--happy. To get a handle on things, Nielsen is expanding its tools to reach TiVos, cell phones and iPods. It's also grappling with a broad rethinking of what its viewership measurements mean.

"The problem with the way Nielsen has approached the problem is that their focus has been completely on the program," said Kate Sirkin, global research director for media buying giant Starcom MediaVest Group. "We want to know the truth about who's actually watching the ads we're putting out on television."

For an industry that pulls in an estimated $70 billion per year in advertising, it's hard to understate how important getting at that "truth" is. In the age of portable devices and TiVo ad-skipping, advertisers want to know if people are indeed watching the commercials, not just the shows. And if Nielsen can't help them do it, they may not be so willing to buy the pricey reports the ratings company sells.

Already, they're increasingly getting data directly from companies like TiVo, which can offer a look at exactly what viewers are watching.

That's hardly good news for Nielsen. But to understand how tricky these changes will be for the ratings giant, it's good to explain how the 70-year-old service has traditionally collected data, as well as the changes it's made in recent years.

The majority of Nielsen's local ratings are still produced the old-fashioned way: Randomly selected panels of viewers are given diaries to record their daily TV watching habits as accurately as possible. That's less accurate than electronic measurements, but small markets don't have enough advertising funding to support investment in expensive equipment, the company says.

In larger markets, Nielsen has long used tools that can sense what channel a TV is tuned to. And since the early 1990s, Nielsen has been developing a way to instead identify specific programs, no matter when or how they are being viewed.

To make this possible, Nielsen developed technology to insert a code into TV shows' audio tracks that makes a sound, inaudbile to human ears, about every two and half seconds. That sound carries information about which station the show came from and when it was broadcast.

For the last year and half, Nielsen has been distributing detectors that can "hear" this signal to some members of its audience panels. By comparing the time encoded in the signal with the actual clock time, the company can now tell whether those viewers are watching a live broadcast or a recording.

This system also has a backup that takes a full audio "fingerprint," or unique audio signature, of shows being watched on a home TV, and compares this to a central database of shows. This level of redundancy helps the ratings be even more accurate, the company says.

[Earl Benser]

Funny thing....

.... All the various ratings seem to say that show I consider to be
less than junk are the most watched. I'd sure love to think that
the rating systems are wrong. I'd hate to think that the rating
systems are right - the inferences are very unflattering to the
viewing population. Or, of course, I could just be weird in
disliking flatulent sitcoms, totally illogical 'mystery' shows, and
over-scripted 'reality' shows.

So do statistics lie????? Or do liars use statistics????? Or is the
whole network TV world really going to hell in a hand basket?????

[Mister C]

It would be funny

That is, if your assessments weren't so accurate. Unfortunately the current state of TV is just sad! A medium controlled by the marketing folks and reduced to the lowest common denominator.

And So It Goes: Adventures in Television by Linda Ellerbee, ISBN 0-425-10237-8 does a real nice job of explaining the system.

[TV James]

Not a new problem...

I've always maintained that the Neilsen number was too small, and probably skewed. How else do certain shows stay on the air? I'm using the flawed "no one I know watches this show" argument, but seriously, at some point there has to be some merit to it. If for nothing else than to say that the Neilsen uses too small a measurement, or needs to break it out more locally if, in fact, there is a part of this country that really does watch "Two and a Half Men" -- and I'm not counting people comatose on their couches drooling at the TV because they don't own remote controls.

Now that the cat is out of the bag, I don't see how Neilsen is going to stuff it back in. I know so many people drool over the "overnights" that there are people paid to read them early each morning, record them and make them available at non-published numbers at the networks so you can call in and hear them.

I imagine that the tide is really turning towards people who time-shift. We rarely watch things as they air, often not even the same night. We'll make an exception for reality shows, but we'll start late so that we "catch up" by the end of the program.

If I could, I would get the east coast affiliates because with our schedule, it's impossible to stay up until 11 pm and still be useful at work the next day.

[TV James]

Tracking iPods?

I don't have an iPod, but the shows are commercial free, aren't they?

If it's for measurement purposes, I would assume that Apple can already tell ABC how many times someone has watched the episode of Lost they purchased. And if TiVo can provide stats and these cable set-top boxes are getting more and more computer like, it seems like Neilsen is going to eventually become largely irrelevant unless they reinvent themselves as an aggregator purchasing the data from TiVO, Apple, Comcast, DISH, etc., and packaging it all up in a neat tidy bundle for the networks.

Perhaps it's time they get out of their arguably flaw data collection business.

[suznick]

"TV is basically an advertising medium"

I learned this in a 1964 Journalism class (20 years before the Macintosh). The iPod and similar devices will force TV to break from that business model. Let's see if TV can innovate as well as Apple (not the Beatles Apple Records) has done in the last 20 years. Maybe Jobs needs to run a network.

[Peter Halmos]

Compure ACTNow Audio Clip Detection

To detect specific video or audio clips you can also use the audio clip detection technology covered by the ACTNow SDK. It tells you exactly when and where a specific audio clip (jingle, song, noise, etc.) was played back or recorded respectively. An evaluation version can be found at www.compure.com/download.asp. However it is an SDK so it's rather a toolkit for programmers.