By Erwin Ephron

Research makes for bad conversation. It’s too fussy for casual talk and can leave the wrong impression. Have you been following the countless discussions on TV commercial ratings?

Since the people meter was introduced in 1987, TV ratings have been based on viewing during the average minute of a program. But advertisers also want to know how many people watch their messages. This has led Nielsen Media Research to propose commercial-minute ratings, which has led to lots of industry discussion. What is it all about?

The New York Times thinks Nielsen is going to report how many people see commercials (“A Question of Eyeballs,” Oct. 18, 2006), but that’s wrong. Nielsen can’t tell us how many people see commercials. They can only track whether people change channels or turn the set off or leave the room.

If Nielsen can’t measure how many people see commercials, what can they measure?

Nielsen records the tuning behavior of panel members who register themselves as “watching TV” by signing in to the set meter. This usually happens when they first enter the room or turn the set on. After that they are periodically, but not frequently, asked to confirm that they are viewing. But that query is not timed to coincide with commercials, so the viewing state of the Nielsen sample when commercials appear is uncertain. The question, “Did you see that commercial?” never comes up in any form.

What Else Does Nielsen Do?

The set meter also keeps a log of channels tuned. Nielsen integrates the viewer timelines with the schedules of programs carried on the channels tuned. A commercial-minute audience is the average number of people in their sample who are still registered as watching TV, during the minutes carrying commercials in the programs on the channels tuned, projected to the U.S. population. My apology if this sounds complicated. It is.

The resulting commercial-minute data that have been made available show losses in audience of about 2 to 10 percent during commercials compared to programs—a small difference. But the big hole in this small difference is the assumption that watching TV is a constant state, unaffected by what is on the screen. Researchers, who read the fine print, qualify a Nielsen commercial exposure as “an opportunity to see” a commercial. And given the opportunity, it’s obvious the probability is a lot less than one. So the Nielsen commercial-minute audience is an overstatement of people seeing commercials.

The 800-Pound Gorilla

It’s curious that in our discussions of commercial ratings, no one brings up clutter. MindShare reports there is close to a minute of nonprogram time for every three minutes of content in prime time. That is the average. On more cluttered channels, the ratio can approach one for two.

A new TNS analysis covering 46 broadcast and cable networks shows that if you turn the set on at random between 7 p.m. and 11 p.m., there is a 25 percent chance that there will be a commercial on the screen at that exact second. That’s the average. The high is 46 percent.

A companion piece of research, this from Ball State University, investigates the effect these many commercials have on viewing. Trained observers recorded the behavior of people watching prime-time television. They found that 58 percent of all commercial-pod viewing was interrupted by channel switching or an activity other than viewing (talking, reading, eating, leaving the room, etc).

The dynamic here is critical. Half of the interruptions were for channel-switching. But since commercial breaks are often synchronized across channels, a large fraction of those who changed channels would be counted by Nielsen as viewing commercials on the channels switched to.

The difficulty of avoiding commercials by changing channels is familiar to any viewer. Pick a given minute. TNS data show that on average at 7:09 p.m., 41.3 percent of the ad-supported networks will be showing commercials. At 8:09 it will be 23.5 percent. At 10:09 it will be 29.3 percent. These switch-in viewers are less than receptive.

Too Big a Promise

Commercial-minute audience is three words. But our public discussions have focused on the first two: whether it should be the commercial-second-weighted-average-minute, or the average for the break, or the specific minute when a brand’s message appears…And whether that minute should be live-only or include delays.

Far less attention is paid to the audience piece of the measurement. And that is unfortunate. Commercial audience does not tell us who sees commercials, but many advertisers think it does and assume that they will finally be getting full value. They won’t. It’s like New York’s proposal to remove trans-fat from food. Make it sound too important and people will assume we’ve solved the getting fat problem. No, we haven’t done that either.