Nielsen ratings, TiVo, and ‘Arrested Development’

How might changes in the Nielsen ratings system help shows like “Arrested Development”? And how do the Nielsens work, anyway? Here’s how.

Some background. The Nielsen Media Research company has about 9,000 “Nielsen families” whose TV habits are monitored from week to week. Nielsen tries to get an accurate cross-section of different types of households so that from those 9,000 they can extrapolate what the entire nation is watching. For example, since African-Americans make up 12 percent of the population, they make up 12 percent of the Nielsen families.

So when they say “Lost” was viewed by 10 million households last week, they don’t actually know that. What they know is that “Lost” was viewed by 900 Nielsen households, and Nielsen extrapolated that to guess roughly how many American households watched it.

Basically, the formula works like this: There are approximately 109,000,000 households with TV in America. Since there are 9,000 Nielsen families, that means each of those families is representing 12,111 other American households — 109,000,000 divided by 9,000. So if 900 Nielsen families watch something, they take that to mean 10,899,900 households in America watched it — 900 x 12,111. Got it?

Of course, they often report not the number of households, but the number of VIEWERS. That requires even more approximating. Nielsen knows how many people live in each of its 9,000 households — but it DOESN’T know exactly how many people live in each of America’s 109,000,000 households. All it has are averages. So while reports of households might be fairly accurate (assuming Nielsen’s sample group is a representative sample), reports of total viewers are always going to be somewhat guess-y.

Nielsen claims that its sampling is a pretty darned accurate reflection of what America is watching. Every February, May, July and November — “sweeps” months — it augments its regular 9,000 families by sending out additional surveys to thousands of other households, in all 210 of the nation’s TV markets. This is to get a more detailed, more accurate sample — and Nielsen says it usually resembles its regular 9,000-family sampling very, very closely. How closely, they won’t say.

In theory, of course, a show could be watched by all 9,000 Nielsen households, and NO other households in all of America. Nielsen would report it as a phenomenal hit, even though it was actually only watched by 9,000 families (which makes it actually a huge flop). Statistically, it’s possible, but of course it’s astronomically unlikely. (Maybe if a network aired a sitcom ABOUT Nielsen families, one where ONLY Nielsen families would get the jokes….)

But what about the reverse? What if there were a show that was watched by a lot of average Americans, but for some reason NOT by Nielsen families?

UPN had some trouble with that in early 2005 when all of a sudden its shows’ ratings dropped. Why had viewers suddenly abandoned them? Turns out they hadn’t. Nielsen had merely changed its reporting methods in some major cities, and the new methods were under-representing African-Americans — UPN’s key demographic. To its credit, Nielsen took measures to improve the situation. I do think Nielsen WANTS to be accurate.

Yet until recently, a significant and fast-growing portion of American society — almost 7 percent of all TV households — was not represented by the 9,000 Nielsen families. Not represented AT ALL — as in, if you belonged to this group, Nielsen wouldn’t let you be one of its families.

The group? People with TiVo or other Digital Video Recorders (DVRs).

Nielsen’s information-gathering device, the PeopleMeter, was not equipped to handle DVRs, it said. People with DVRs were still included in the “sweeps” surveys — the forms Nielsen mailed out had a column to indicate that you had watched the show on DVR rather than “live” — but if you had a DVR, you could not be one of those 9,000 families who comprise the bulk of Nielsen’s research. Nielsen wouldn’t let you.

This is kind of a big deal. Nielsen strives to get an accurate sampling of American households, yet leaves out 7.4 million. (That’s how many have DVRs as of August 2005 — and the number is rising fast.) Added to Nielsen’s notorious lack of representation for TV viewing in non-home settings — at bars, in dorm rooms, etc. — we start to see a lot of cracks in Nielsen’s system.

Because here’s the thing: Nielsen and the networks should be paying particular attention to DVR viewers, not ignoring them. Not only are DVRs the way of the future — that 7 percent of American households is expected to rise to 35 percent within the next four years — but DVR users are also the people who CARE about TV. They’re the ones who are passionate about it, who choose their shows intentionally, who don’t watch something just because it happens to be on when they’re flipping channels.

They’re loyal, too. A non-DVR fan of “Desperate Housewives” might miss the occasional episode due to forgetfulness or being out of the house, but a DVR user never will, because the DVR records it each week automatically. Isn’t that the audience you’d want to cater to?

Nielsen has finally alleviated the problem and has done whatever was necessary to make the PeopleMeters DVR-compatible. As of this week (Dec. 26), 60 of its 9,000 families are DVR households. Nielsen will increase the number 100 per month until the percentage of Nielsen families matches the percentage of DVR households nationwide. That means it will be another six months before Nielsen is accurately reflecting DVR usage. (Seven percent of 9,000 is 630.)

It will be interesting to see how this affects the ratings. Are there shows that are watched more by DVR viewers than regular viewers? I suspect we will see these changes:

– Ratings for “The War at Home” will decline. The Fox sitcom, a wretched exercise in forced “outrageousness,” currently has a cushy slot between “The Simpsons” and “Family Guy.” I suspect many of the people who watch it do so simply because they like the shows before and after it, and there’s no particular reason to change the channel (or turn the TV off) for that half-hour in between. DVR users, however, don’t watch shows accidentally. They record what they want to see, period — which in this case probably means “The Simpsons” and “Family Guy,” but NOT “The War at Home.”

– The lead-in effect in general will be minimized. Many, many shows remain “popular” (that is to say, they are watched by a lot of people) simply because they come on immediately after shows that are genuinely popular. Fox used this to its advantage with “House,” which struggled on its own, then became a hit when it followed “American Idol.” After gaining an audience in that time slot, the show was sent out on its own — and did just fine, thanks to all those people discovering it and becoming fans when they saw it after “American Idol.” Had Fox not taken advantage of people’s apathy — they don’t change the channel or turn off the TV when their show ends; they just sit there and look at whatever comes on next — “House” would have continued to struggle.

This phenomenon will fall off once DVR usage is accurately reported, and especially as it becomes more prominent over the next few years. Like I said, DVR people generally don’t watch shows accidentally. They record them and watch them later. At the very least, they start watching a show half-way through, so they can skip the commercials.

– Ratings for Letterman and Leno will even out. At present, Letterman’s ratings tend to go up when CBS has a strong prime-time lineup leading into it, apparently because people watch the 10 p.m. show, then keep it on CBS for their local news, then keep it on for Letterman. Leno receives a similar boost when NBC is strong. Again, this is not generally how DVR people watch TV. I don’t expect Letterman to start matching Leno — Leno has a lot of people who watch him regardless of what they were watching earlier in the evening — but I do expect the gap to lessen somewhat, being no longer influenced as heavily by lead-ins.

– Ratings for some middle-of-the-road shows will go down. I’m thinking of CBS’ Monday night sitcom lineup here, particularly the 9 p.m. hour: “Two and a Half Men” and “Out of Practice.” Both are highly rated, regularly winning their time slot and charting among the top 20 shows of the week. But does anyone LOVE those shows? Do people look forward to them? Does anyone have their DVR set to record them each week? I suspect not. These aren’t “appointment” shows. They air on a night when people are very likely to be home, at a time when there’s not much else on. Plus, they’re on right before “CSI: Miami,” which is a much more “legitimately” popular show (as in, people actually anticipate it). Some viewers, with nothing else to do, will turn to a station early just to make sure they don’t forget to watch their show. Honestly, I suspect DVR households don’t watch middling shows like that nearly as much as other people.

A show I DO think DVR people watch in greater percentages than non-DVR people is “Arrested Development.” The show is densely packed with jokes, subtle references, and even fleeting visual gags. You almost HAVE to have a DVR (or at least have recorded the show with your VCR) to fully appreciate it.

It’s also, frankly, a show that is more likely to be watched regularly by people who are willing to commit to a TV show. It requires a viewer’s careful attention, and it requires faithful viewing, each and every week. (It also rewards it immeasurably with brilliantly funny writing and performances, but that’s another article.) What kind of viewer is most likely to give a show that kind of attention? The kind of viewer who is also passionate enough about TV to have a DVR.

I don’t mean to say that ALL DVR users are passionate, selective viewers, nor that ALL people without DVRs will indiscriminately watch anything that happens to land in front of their eyes. But there is certainly a trend, particularly in the attention DVR users will pay to a show.

It might be too late for “Arrested Development,” since Fox has all but canceled it. But other shows with devoted cult followings might find their ratings improving once the Nielsen ratings start accurately reflecting DVR usage. And maybe some of those crappy shows — the ones no one cares about, but that they sometimes watch anyway — maybe those will go away. Here’s hopin’.