Tuesday, June 05, 2012

How accurate are Nielsen ratings?

No one really knows.

There are statistical models to support their validity. They’ve been doing it a long time and have revised and improved methods over the last fifty years. But still – one schlep in Clovis, New Mexico who hates Amy Poehler because she reminds him of a girl who once dumped him could represent 2,000,000 viewers.

With more and more people getting cable boxes or satellite boxes it’s easier to track just what is being watched. But those still don’t tell you WHO is watching. How many are in the room? What are their ages? How many left the room after the 15th vagina joke on 2 BROKE GIRLS (so five minutes in)?

And then there is the issue with shows on the DVR. How long do you give someone to watch a program before it no longer counts? Nielsen has a formula but is it valid or just arbitrary?   Last week I watched a sitcom episode that I had DVR’d and noticed there were Christmas commercials in it. I’m sure the sponsors could care less that I finally got around to it. And just because someone records a show doesn’t mean he’ll ever watch it.

And if you record a show and don’t watch it for weeks, is that your pattern or you were just on vacation?

How many viewers like a show but won’t watch it on the air, preferring to wait for the DVD to come out so they can watch the whole season at once? So you’re big fans of the show but Nielsen doesn’t know that.

But wait! There’s more!

Watching content on-line. That’s become even more popular because there’s that much more television being streamed and with inexpensive devices like AppleTV you can watch these shows on your big screen TV.  And don't forget about people now watching shows on their iAnythings.  The A.C. of A.C. Nielsen stands for “Aw crap!”

And yet, all programming and advertising decisions are based on the Nielsen numbers.  Yes, it's a crooked card game but it's the only game in town. 

Radio had a similar problem. Ratings were taken via diaries issued to listeners. On good faith, you were expected to keep a detailed log of what station you listened to, when you turned it on and when you turned it off. So essentially it charted the listening habits of diligent people.

And then a few years ago the People Meter was introduced. This is a little device that looks like a beeper. You just take it with you. It picks up and documents any radio signal you hear. How accurate is this an indicator? You could be in your car, stopped at a light for two minutes and the clown next to you has Eminem blasting. The People Meter hears it and registers that you -- a 78 year-old white lady from Beverly Hills --  listens to a hip hop station.

Still, it’s a far more accurate system than the diaries. And interestingly, the results when they started using People Meters were very different from the diary input. Big case in point: the oldies format.

Before People Meters, the oldies format was out of favor. Lots of station that had been playing oldies for decades saw their numbers dwindle and bailed on the format. Even WCBS-FM in New York, maybe the country’s number one oldies station, abandoned the format for the ill-advised “Jack” format.

Then People Meters came along and what do you know? Not only did listeners still listen to oldies stations, there were way more of them and listened way longer than anyone had ever thought. WCBS-FM is back to playing the Beatles. In some markets, the oldies station -- once thought passe -- is the number one station in town.  That's a big discrepancy between diaries and People Meters. 

Might the same be true of television? If an accurate measurement could be made, might we find that shows we thought no one was watching actually had decent numbers? How many series got cancelled that should have been given another chance? Is it possible that even fewer viewers watch WHITNEY than NBC thinks and they renewed a certified bomb?

This always reminds me of the chain gang in football. These are the guys who stand off on the sidelines with the yardage markers. When a ball is close to a first down the chain gang is summoned out to the field. Two guys run out, they place the markers down and rule. Is the nose of the ball over the line? We’re talking inches here. Entire games sometimes depend upon these calls. And two nudniks run out to the field, not even in a straight line, and their word is final. Huh? In this day and age of high technology they can’t install lasers that accurately tell you whether the football has crossed the first down plane?

And the NFL could install this hardware. At least TV has an excuse – they don’t have such sophisticated methods of calibration. Damn Steve Jobs for dying before he could solve this problem too!

In the meantime, the livelihood of thousands of people in the television and advertising industry still depend on the two nudniks zig-zagging their way to mid-field.


Rays profile said...

I forget who said it, but "if one Nielsen family shuts off the TV and goes to visit its grandmother, do a million other families shut off their TVs and go visit their grandmothers?"

Roger Owen Green said...

I believe that Nielsen counts DVR-TiVO-type recordings for a week after broadcast. (Which would void about 90% of what I record.)

Johnny Walker said...

I desperately want them to implement this system for TV. And what's wrong with People Meters for the job? Sounds like it would work just as well for TV!

JT Anthony said...

What's a better way? Statistically guessing at a particular target market based on what a sample size of that target watches?
It does seem to be the only racket in town, but the bottom line--it seems--is what advertisers are willing to pay for the ad space, no matter what system is used. It's easier for product companies to estimate its target market and then pour money into the shows they think resonate.
Just give Facebook and Goggle time and they will figure all of this stuff out...

Former Nielsen said...

Nielsen counts DVR-viewings as "full" if you watch it within 24 hours of the original broadcast. It counts as "half" if you watch it within a week. Later than that, they don't count it at all. (at least that's what they told me when I participated) They also ask you if you watch TV streaming on the internet or other devices before you start. At the time I did not and said so, so I don't know if they would've put one of their tracking thingies in my laptop if I'd said yes.

Blaze said...

I suddenly have vision. Like with the radio situation you described, some new, fresh, polling system suddenly gives a far more accurate graph of what the population is actually watching. It turns out that only these Neilsen Families watched "Jersey Shore" and "Duck Dynasty". The simmering resentment in the back of peoples' minds that they were the only ones who appreciated quality TV and everyone else must be mouth-breathing morons disappears! The years pass and the idiot shows are relegated to the fringe and quality of TV keeps rising as producers respond to the new information. With shows that are actually entertaining and intriguing, the populace is happier, a bit sharper mentally and more productive in their day. Life becomes so much better.

The Ames Family said...

Funny! I was just called last night by Nielsen and will be doing a week-long diary. We don't have cable, and since it's summer, hardly watch network TV, but hey! I'll take their $30 and fill in my streaming time.

I've always wondered how someone gets picked to be a Nielsen demo in the first place. I've heard it's really hard and there are year+ waits. Is that true?

Charles H. Bryan said...

When I was a kid, we got one of the blue Nielsen diaries (along with a fifty cents and a free golf pencil, as I recall). My mother actually lied on it to make our viewing habits look better. You're welcome, early 70s PBS!

spreng said...

I worked for Nielsen for a couple of years (2004-2006) and helped modify their software to handle time-shifting devices (DVRs, VCRs, etc.). Since those devices were all fed from the same antenna/cable that fed the TVs, it was simple to adapt the software to handle them. There was no online viewing available then, so I don't know how they handle that now.

Brian Fies said...

Charles H., funny, my family did those diaries a few times in the '70s and we did the same. Popular wisdom says that "Star Trek" exploded in popularity when it went into daily syndication. It didn't; that was just me.

As to the chain gang, as a high-school chain ganger I'd only point out that the back guy, while standing on the sideline, grabs the spot where the chain crosses the nearest 10-yard line with his fingertips (a 10-yard chain is always crossing some 10-yard line). It doesn't matter how the gang zig-zags onto the field--if he lays his marked spot on the same 10-yard line, the measurement is actually pretty accurate. The real uncertainty: was the ball position spotted accurately on the sideline in the first place?

Erich Eilenberger said...

If what Spreng says is true, then the company is obviously adapting. It is a flawed system in some ways, and obviously newer delivery methods may make tracking easier, but don't forget that many people (especially young people), have gotten rid of their cable subscriptions, making them once again more difficult to track.

Unfortunately, I have found that many people believe that television programming is somehow a right or end in itself when, as showrunner Noah Hawley said, it's just an Audience Delivery Device. Programming exists to sell air time. Advertisers want consumers to watch their commercials, so they have much, much less interest in time-shift viewing because consumers are far less likely to watch their commercials. Similarly, DVD viewing means nothing at all to them, since DVDs don't come with commercials. Even online viewing is less relevant because, while you actually have to watch commercials online and it's easy to track how many people have watched something, you are losing out on local ad buys. On top of which, it's easier for a YouTube video of a zit being popped to get online viewers than it is for an episode of Mad Men. So the competition is much more competitive and complicated.

Television is still based on a very old-fashioned model, which is why they still use an old-fashioned system for monitoring viewership.

Unknown said...

You hit on a very interesting point. Online tracking of viewing habits is very detailed. I'm an IT Project Manager and have worked at a couple of companies that create online rich media ads (those annoying pop up videos and shit). Via data gathered from your browsing, they can see where you came from, how much of an ad you've seen, where you're at, to some extent- who your are and where you go after that.
This is nothing shocking. It's just the world that we live in today. I would assume that with AT&T Uverse being IPTV, they are able to implement some of that technology to a greater degree.
These cable companies are probably gathering much more data than any of us realize. Especially with digital cable. Analog cable is a little more difficult because of its nature. However, a digital signal carries much more data and can be parsed more easily.
Frankly, I can't imagine that Nielsen's traditional way of collecting data will last much longer.

Brent said...

Ken, just for your edification, I've been a football referee for over 15 years. Started with the 8 year-olds and worked up to the high school level. There's a strict, set procedure in handling the chains to ensure accuracy. I won't go into it here because it would take FAR too long. It may not appear so to a casual observer, but we take VERY seriously the accuracy of the spot of the ball and measuring for a first down.

Okay, off the soap box. With respect to DVR use. Any scripted show is DVR'd so I can watch the one hour show in 42 minutes by skipping commercials. I control my schedule, not the TV. Long gone are the days of everyone in the family gathering around to watch a show on a specific day at a specific time.

Also, for me, anything that's on after about 9 PM gets DVR'd. Some folks have to get up at 4 AM and are in bed by 8 or 9. I'll have the Mariner game on the clock radio with the sleep timer running. Although the way they're scoring runs lately they aren't as much of a snooze fest anymore.

Mike said...

What I'm wondering is if there's any evidence to back up the idea that ads should be targeted to younger audiences?

I would think CBS would be interested in funding an answer.

Wendy M. Grossman said...

Ironically, they could get very good data on who actually watches what by looking at torrent numbers. But they're so busy suing torrenters that the data that might have been available to them is increasingly obfuscated. (Of course, there is the little detail that torrenters don't watch ads, but still - the torrent numbers tell you a lot about which shows are being underestimated by the ratings. GOSSIP GIRL, for example, was much more obviously a hit much earlier if you looked at the torrent numbers.


BigTed said...

Don't DVR companies like TiVo (or in my case, DirecTV) know exactly what you're watching at any given time, whether or not it's time-delayed? I thought they even knew when you paused over a Victoria's Secret commercial. It seems as if that could give the networks far more accurate information about how popular their shows are.

Mike said...

Great post - as an amateur TV geek and an amateur stats/demographic geek, this hits my sweet spot.

One point though:

Is it possible that even fewer viewers watch WHITNEY than NBC thinks and they renewed a certified bomb?

I'm not sure I read this right - but it seems your suggesting that there are people who wrote down in a Nielsen report that they watched Whitney even though they did not. As someone who has watched enough episodes of the show to know it improved greatly over the course of the season, let me assure you that nobody would possibly admit to watching it on purpose unless they actually did. If anything, I'd worry that critical darlings like Mad Men or Downton Abbey would be the ones who would benefit from this behavior -- TV that's supposedly good for you, so you don't feel so embarrassed.

It's interesting that somebody would bring up torrents as a way to track ratings because, if anything, I've found that they correspond to the Nielsen ratings -- the most popular viewed shows according to Nielsen are often the most popular downloaded, even on private sites where you'd think there would be some sort of 'niche favorites' skewing happening. That's probably not true if you look overall, but I'm still always surprised to see the popularity of things like Two and Half Men on online sites -- it's one thing to just not change the channel but another thing altogether to go through the effort of downloading it.

As much as I think the current system is flawed, I don't envy those who are trying to come up with a more accurate model.

Matt Tauber said...

I think the last time I saw the word "nudnik" was in an early '80s issue of Cracked Magazine.

Michael Hagerty said...


Neilsen has what they call the LPM (local people meter. It not only measures viewing directly, it does demographics, too...it's in the 56 largest markets currently and will eventually roll out to the entire country.

Basically, when they sign up a family, each member of the family is assigned a number. Whoever turns on the TV logs in on the remote. As members of the family enter the room to watch, they are supposed to log on (it's like a one-touch button). As they leave, they're supposed to log off.

Neilsen knows Dad is 49, Mom's 46, Sis is 20 and Junior's 18, so they've got the demographics.

Anyone who lives in a house with a family knows the strict adherence to the log-on/log-off protocol is going to last maybe a week.

Realistically, here's what happens:

Sis turns on the TV, logs in and finds a Gilmore Girls marathon on ABC Family.

Credit: 5 minutes of viewing Gilmore Girls from a 20 year old female.

Junior comes in, sees Gilmore Girls, says "this blows", grabs the remote without logging in and flips to "1000 Ways To Die" on SPIKE. Sis argues for a moment, rolls eyes and leaves the room without logging off.

Credit: 25 minutes of viewing 1000 Ways To Die from a 20 year old female.

Mom comes in says "over my dead body" just as Junior flips to "It's Always Sunny In Philadelphia." Junior leaves the room. Mom, the only conscientious member of the family, logs on but forgets to log Junior off (doesn't matter, he wasn't logged in anyway, it was Sis), and turns to HGTV.

Credit: 5 minutes of viewing Divine Design from a 46 year old female and 5 minutes of viewing from a 20 year old female).

Before Mom can settle in, the phone rings. She takes it in the other room. Dad walks in, grabs the remote, doesn't log in and flips to Top Gear on BBC America.

Credit: 55 minutes of viewing Top Geara from a 46 year old female and 55 minutes of viewing from a 20year old female.

When I was programming an independent station here in Phoenix a few years ago, I used to debate the LPM numbers with the Nielsen reps, who insisted they should be treated with the same credibility as stone tablets carried down from a mountaintop by a very hairy man.

"Why?" I'd ask.

"Because we explain to the Nielsen Families when we set them up that they must log in and log out."


"And...they do."

"Based on...?"

"Well, they said they would."

Frankly, adopting Arbitron's radio technology, giving each member of the family a device that is their own to wear that then senses what's on, is probably the only way around this. Logging in clearly doesn't work, and a camera in the room showing Nielsen who's actually watching might prove embarrasing when Cobie Smulders appears on-screen and the wife and kids are out of the.....

I may have said too much.

Unknown said...

Great article! I recently mentioned the Nielson ratings in a post about the future of TV shows online. The new world just has no place for this older than dirt system.

If anyone has any interest, the post can be found here - http://www.nathanhartman.net/2012/night-of-the-living-tv-show

Another great post Ken. One of my favorite blogs!

Movie Don said...

The solution seems pretty simple to me. Nielson needs to pay the cable companies for the usage information instead of the consumers, since the companies already know this information. That cuts out the obvious user 'lying to look better' scenarios.
Hollywood should demand that Nielson buck up for the info instead. Nielson could then hire phone people to call and ask about the demographics in the home. I'm sure a good number of folks would share the info to them for free. Wouldn't it be nice to know the ACTUAL numbers instead of the 'suggested by a formula' numbers?

Side note: I've never known anybody who has known anybody (to my knowledge) that actually had a box, so I'd suggest that our group of quirky watchers may not have been repressed in the marketplace at all. How many of these households 'happened' upon "Firefly" ? Not that many apparently. However, DVD and netflix numbers would suggest that it was in fact a well watched show in the most coveted demo.

Michael Hagerty said...

Movie Don: Most cable companies won't admit to being able to determine what any individual home is watching (other than things they can be billed for). Invasion of privacy issues, none of which are covered in your agreement with the cable company.

And even if they we-wrote the agreements, their existing technology can't tell them who's actually in the room watching. Sure, you can call the house and ask, but now you're back to people trying to remember who watched what a week or a month ago. And it's demographics that sells advertising, not raw numbers.

Really, Arbitron's got the right idea, using a wearable people meter. Nielsen just needs to license it from them and get channels and networks to start encoding their broadcasts so the devices recognize them.

Andrew said...

Who cares how many people are watching these shows on their DVRs or online? They're all fast forwarding through commercials (or if they're watching on Hulu, I'm sure the website can give advertisers an accurate view count).

Nielsen ratings are all about advertising and which shows should be allowed to charge more for commercials. If that show you love to watch is getting cancelled, maybe it's because you record it and then watch it Saturday afternoon instead of actually putting effort into supporting it.

Johnny Walker said...

I'm working my way through Cheers at the moment, enjoying all the old shows I'd either forgotten or missed the first time around. A few questions (and random observations) leap to mind about it:

Firstly, there was an episode where the cold open was the guys in the bar singing the funeral dirge for an empty keg. I thought to myself, "that's SO Ken Levine!", and lo! if it wasn't the first season three episode of Cheers written by Ken Levine and David Isaacs. (I presume AfterMASH had found its way to TV heaven by then.) No idea who was actually responsible for that scene, but it was a strange moment when your name actually came up :)

Other than that, Coach was still missing (Nicholas Colasanto was in hospital, according to Wikipedia) and I realised it must have been the episode where you and David had had no time to write Coach out of the script.

Which brings me to some questions: Nicholas Colasanto's departure from the show was a little odd. He was gone for a few episodes and then suddenly returned for one, "Cheerio Cheers". He looked pretty good in that to me, but then he disappeared again until the final episode of that season, where he was only featured in the cold open.

The strange thing about that cold open was how, when it ended the joke, there was no audience laughter. It was a surreal moment and I have to wonder if the laughter was removed deliberately? I don't know, it was just such an odd moment.

At the same time that Coach was missing, Diane was on holiday with Frasier, which meant that Shelley Long's part was reduced to two scenes an episode.

Was this because Long had outside commitments, or was just because they show was gearing up for shooting in Florence?

Finally, did the writers have any idea how they were going to tie up the cliff-hanger at the end of the season? Or did you just decide to worry about that next year?

Boy, Colasanto was great as Coach. I could barely imagine himself playing a different part.

Alexandra said...

I think the general consensus among people I work with is that the Nielsen system is antiquated and needs to be completely revamped in order to get an accurate look at who's watching what and why. The Nielsen family samples are just too small and too diverse to possibly represent what the millions of other families are doing, all across the country. It's a system I know a lot of people - especially people who work in television - would like to overhaul.

Craig Russell said...

Just to clarify Ken, The Arbitron People meters are only in the top 50 markets. From basically 51 to 284 still uses the diary method. So yes in New York, LA, Seattle and Baltimore they get the people meter. Casper, Wyoming and Grand Rapids, Michigan still have the diaries. Boo.

Gareth said...

I've done radio ratings surveys myself, through diaries. There's a basic problem that filling out a diary is boring and you don't get anything out of it. So people will agree to participate and then not fill the thing out, or not bother to be accurate. This is particularly bad with people under 25 who aren't supervised by parents. My worst result was probably the woman who left the radio on to entertain her cats, and left the house, but put the station down as "her" listening. Rest assured that didn't go into the final results.

Anonymous said...

Great call on the very inexact chain gangs used in football. The NFL has enough billions in it's coffers that they could spend a few million to have a laser system developed with sensors implanted on the filed and in the ball to ensure accurate placement down to the inch.

But they won't because it's yet another thing taken out of the hands of the on field officials and yet one more 'tradition' (insert roll eyes here) being killed off.

And speaking as someone who got paid to be on the chain gang for three years I can assure you that it was beyond inexact, and depending on how bored we were we would cheat for one or both teams.

Former Nielsen said...

Movie Don, it's not even a box anymore, they hardwire their devices inside your tvs (and dvrs, and dvd players, and xboxes, and...)

As for football, the guy behind hawkeye keeps saying he can do it for football or baseball. It's just the leagues rejecting it. I love the speed and exactness of the tennis challenge system. Push a button, it says "in" or "out" and zooms in as needed to show it. Done. No arguing. Within 3 mm I believe is the precision?

Anonymous said...


Next time, if you're going to use an example of someone whose heart was broken by a girl, and consequently embittered, cast someone with better aesthetic credentials then heavy-weight middle-aged over-dyed leprechaun Amy Poehler.

Trevor J. Cohen said...

Hey, Anonymous Poehler-Hater: you forgot to attach the photo to your comment that shows how you, by contrast, are God's gift to women. Get onto that, please.

McAlvie said...

It surprises me not at all that oldies stations are more popular than the suits realized. Think about who listens to radio now that we are in the age of the i-thingie. Those of us who grew up with radio still like to listen to the same stuff we did back when. Radio abandoned us for a long time, trying to grab the younger market; but the younger market was downloading their own music and not interested. This is why radio stations change format every 6 weeks these days. In the meantime, we oldies fans discovered online radio. And now we have our own ithingies and have figured out how to use fm transmitters to listen in our cars.

In the end, like you've been saying about tv scripts, good music lasts longer than the blonde chic du jour who is headed to rehab anyway.

Johnny Walker said...

Brilliant breakdown, Michael! Hilarious and informative!

Liggie said...

Our family got radio diaries for a week when I was a teen. However, I think I screwed up the demographics. I was into the youth-oriented pop of the day, but I logged the adult contemporary station my dad listened to when he drove us to school. Hence, that AC station got ranked in the top 10 for teens. I'm sure many teens uttered the "WTF?" of the day when they saw that result.

Alan C said...

I'm certain Quark and Holmes & Yoyo were more popular than the Nielsens suggested!

T.J. said...

Whether or not the Nielsens are accurate generally depends on whether your series just got picked up or canned.

al said...

Here's what I want to know:
Okay, my favorite show is coming on TV. It's ratings are slipping, so I do my part: I dvr it and watch every commercial (or at least let it play). But a few months later, I move figure Hulu is cheaper than cable and cancel my cable subscription: no more dvr. But my favorite show still comes on the air. Ratings are even lower, so I watch it live but since I haven't got a nielsen diary no one knows I watch it.
Well, support the cause, I figure, and go on and watch it again on hulu (it's a really good episode). Finally my view is counted! Not worth much, but better than nothing.

The show's getting threatened with cancellation. I really like it, so I decide... I'm going to watch it on Hulu AGAIN. When my girlfriend comes over, I get her to watch it too. In fact, I'm going to let it play when I walk out of the room, just for the extra rating. For the next week, anytime I leave the house or go to sleep I put the show on.
Is anyone counting these extra ratings? Have they figured, this IP address viewed the show once, no more count? Or since I'm still watching the ads, do I contribute more to the ratings?

The point is, if we want to save a show that's about to die... what's really more worthwhile? Tweeting and writing the network execs, or getting fan campaigns to convince people to have the show play online as much as possible (kind of like how everyone rallied to buy Firefly DVDs and convinced them to make a movie)?

al said...


Can't believe I did that.