Monday, December 28, 2009

Monday Morning QBing Ad Rates

Note: I started working on this post a few weeks back, almost immediately after I finished my exploration of the viewers/demos debate, but shelved it because for some reason it just didn't seem to have any real message, and the method is kinda shaky. I thought I might come back and spruce it up later, but I've decided to just go ahead and post the draft, and maybe someone will find it interesting.

Everybody's got an opinion on how any given fall season's crop of new TV shows will do. But the opinions that matter most are those agreed upon by networks and advertisers to sell TV spots. We're only about halfway through the new season, but I thought I'd do a little Monday morning quarterbacking and see how they've done so far.

My last series of posts used '08-'09 ratings and Ad Age's estimates for fall 2009 ad rates to illustrate that the A18-49 demo is, if not a direct cause, then a good indicator of ad rates. I stuck to shows in the same timeslot both seasons to try to limit speculation. This time it's all about speculation, using the price-per-demo averages I got from the last posts to try to estimate what they estimated about new shows. Since I'm only looking at scripted shows, I'm going to use the average price per demo point among scripted shows in the same timeslot, which was $41,165.29. Note that's a bit higher than the $40,207.62 per point across scripted and nonscripted, meaning scripted is a bit more valuable. (More on that in a future post, maybe.)

To get the speculated demo or "spec," I'm dividing the ad price by $41,165.29. "Demo" is the average of all originals this season (thru the week ending 12/13).

Price Spec Demo Diff
Eastwick 110558 2.7 1.8 -33%
Flashforward 175724 4.3 2.9 -31%
Brothers 54157 1.3 0.9 -30%
Hank 91167 2.2 1.6 -27%
Three Rivers 95633 2.3 1.8 -24%
Community 120000 2.9 2.3 -20%
The Forgotten 97527 2.4 1.9 -19%
Mercy 91172 2.2 1.9 -14%
The Middle 97812 2.4 2.1 -10%
The Good Wife 126882 3.1 2.9 -7%
Cleveland 158701 3.9 3.8 -1%
Trauma 75928 1.8 1.8 -1%
Glee 127350 3.1 3.3 +7%
Modern Fam 130388 3.2 3.6 +12%
AccOnPurpose 107817 2.6 3.0 +15%
Cougar Town 103314 2.5 3.2 +26%
V 110450 2.7 3.8 +42%
NCIS: LA 106508 2.6 3.8 +48%


Instead of trying to find a trend in the whole list, which is tough when there are shows being undersold by 40%+ and oversold by 30%+, I'm gonna break it down into a few points.

The 10pm hour
- The projections concerning Jay Leno are some of the most interesting things to me about all the ad rate stuff I've looked at. Let me first of all bring in Jay's ad rates.

Price Spec Demo Diff
Leno (Mon.) 53640 1.5 1.4 -2%
Leno (Tues.) 65678 1.8 2.1 +16%
Leno (Wed.) 62012 1.7 1.6 -6%
Leno (Thurs.) 57295 1.6 1.6 +2%
Leno (Fri.) 48803 1.3 1.3 0%


*- I'm not including Jay's premiere week, because it was much more inflated than the average premiere. For example, his Monday premiere drove the Monday average up more than 0.3 by itself! I think these are much more indicative of his "true strength."

First thing to be noted is that while a lot of people have made Jay out to be an absolutely colossal disappointment, if you go by the unscripted average per demo (which is what I applied to get the Spec) the advertisers pretty much saw it coming. He's right where they thought he'd be on most days, and throwing in his better-than-expected showing after The Biggest Loser and that huge premiere week (which, again, isn't included), they almost certainly aren't going to be all that upset.

But while the ad buyers did have Leno pretty well pegged, they messed up on everything else at 10pm. It seems they bought into the widely held belief that Leno's arrival would throw a bunch of 10pm NBC drama watchers into the mix, boosting the numbers for the other two nets. Of the three new 10pm dramas, they guessed too high on all of them. They probably would've been farther off on The Good Wife, but they underestimated its lead-in to such a great extent that it looks closer. Still, they had TGW building by about a half point in the demo on an NCIS spin-off that was airing directly out of the original. Pretty high praise, and the show hasn't lived up to that standard.

And that isn't the only evidence; Numb3rs fetches more dollars than the usually stronger Ghost Whisperer. CSI:NY, which almost always declined out of Criminal Minds, outdid it in the ads department. (You could say that L&O and SVU competition drove those speculations downward, I guess....) Private Practice was another returning show projected to do much better at 10pm. Maybe shares are a big piece of the puzzle, meaning there's a premium on shows in the lower-Households-Using-TV 10pm hour, but still, mistakes were made. Even looking at share, TGW hasn't been as strong as NCIS:LA and CSI:NY hasn't been as strong as CM.

ABC drama disappointments - It kinda ties in with the previous point about 10pm, since The Forgotten and Eastwick were part of that mix, but all the new dramas for fall on ABC have ugly differences. I don't really know where the bar is for make-goods, the payments made back to advertisers when something drastically underperforms, but I feel pretty confident that all three ABC fall dramas falling 19%+ below expectations means some of those are gonna come. The big speculation for The Forgotten largely resulted from them overpaying for the weakening DWTS Results show on Tuesday. They went way overboard on Flashforward, which they thought would average over a 4.0 (or maybe a bit lower with huge A18-34s) but only premiered with a 4.0 and has slowly but surely gone way down. That percentage difference looks bad now, and its season-long average is only going to get worse barring some kind of miracle.

The only good news on the hourlong front was the four-ep run of V, which vastly outdid expectations. The huge pilot ratings inflate that more than usual since there are only four data points, but even just going with the last two weeks (3.1) it's ahead of projections. Now we've got to see how it does after Lost this spring.

ABC Comedy Wednesday - I'd be scared to death to pay money to advertise on a night with five new shows, where there's such a huge spectrum of possible outcomes, but at least for the four comedies, they got the structure of the night fairly well. They were way off on Hank, but generally correct that it would be a night led by Modern Family. MF and Cougar Town both turned out stronger than even they expected, and The Middle a bit weaker (though with a stronger Hank lead-in they may have been almost exactly right), but outside of Hank they had a decent grasp on what would happen here.

Advertisers in the NCIS duo are happy - The NCIS mothership, one of the oldest-skewing shows on TV, fetched $133,304, which translated to about a 3.3 expected demo, down around 10% from last season's average. Instead, the show so far has been way up from last year, usually well into the 4's, and it's brought its spinoff with it. The ads on NCIS:LA went for about 80% of the mothership's on average, and retention-wise that's about right, but in retrospect, it still seems these shows should've gotten a lot more, even if they're really light on A18-34. (Which seems pretty likely considering how colossal their P2+ numbers are.) Spending less than $110k for an ad on NCIS:LA which often approaches 4 in the demo is quite a steal.

In general, as I said at the beginning, it's tough to come up with a sweeping statement about how they did, but with only a few exceptions (Flashforward, Eastwick, NCIS:LA being the really glaring ones) they seemed to have some general idea about how shows would do. It's just that a lot of the shows picked to struggle actually struggled much worse. Makes sense, because why would a network greenlight a show that they can't sell as anything more than a massive bomb?

Sunday, December 13, 2009

Peetooplus - What Sells on Friday?

One more urban legend for me to tackle: the notion that total viewers are somehow an important number to Friday advertisers because the demo "isn't home." I'm keeping this post short because there are very few data points, so it's hard to say anything for sure.

Demo Price Price/pt
Ghost Whisperer 2.5 78047 31218.8
Numb3rs 2.3 85007 36959.57
Dollhouse 1.5 56370 37580

Viewers Price Price/mil
Ghost Whisperer 10.3 78047 7577.379
Numb3rs 9.4 85007 9043.298
Dollhouse 3.7 56370 15235.14

These are the only three same-timeslot shows aside from 20/20, which I don't have demo/viewer averages on. All three are being "underpaid" by the standard per-demo used for all of primetime, but the per-demo prices are still fairly consistent. And again, anything you can say negatively about the demo goes double for viewers. The prices certainly have no connection to viewers, with Dollhouse getting 36% of GW's viewers on average but managing 72% of its ad dollars. A P2+ on Dollhouse costs over twice as much. If you want to bring skew into the equation, that helps a bit. Certainly CBS Friday is very light on A18-34. But Dollhouse is one of the youngest-skewing shows on TV and is still underpaid.

So there's something else that's hurting Friday beyond even the hurting ratings. There's a good chance that all these shows were speculated (and rightfully so) to decline this year, and that explains them being underpaid. GW was inflated by a big November/December arc that had dissipated by season's end, while Dollhouse bottomed out at a 1.0. Declining ratings suggest advertisers should be abandoning this night, and they are. I'd say they're overcompensating, but I think they just expected even more drops, and they were right.

That's about all I've got to say on total viewers. They don't seem to matter in general, they don't matter (at least positively) in terms of "skew," they don't matter to CBS, and they don't matter on Fridays. If any other old wives' tales about total viewers pop up, I may do another one of these. And I hope to have a few more posts in a little while using all this ad rate data I've compiled, but none of it will have anything to do with total viewers, so thus ends the Peetooplus discussion for now. At least as far as I'm concerned. ;-)

Peetooplus - CBS Cares... About Total Viewers?

CBS talks about total viewers a lot when they write their press releases. Combined perhaps with people getting hypnotized by all those "CBS Cares" commercials, this has led to two schools of thought on the Internet:

1) CBS cares about total viewers significantly more than other broadcast networks. They use the number as a major part of their network strategy out of respect for the people over 50 years old who watch their network. Press releases are good indicators of network strategy.
2) CBS is considerably stronger relative to other networks in total viewers than they are in most other measurements. Press releases have a tendency to play up the strengths of the network, no matter their importance in a network strategy.

So which one is true?

If you'll recall my scatter plots from an earlier post, I showed visually that ad rates are considerably better connected to the adults 18-49 demo than they are to total viewers. If CBS cares about the old folk, it should be different this time. Right?

Fewer data points, but the viewer one arguably looks a little better than the viewer one for all the networks, and it plays out that way in the correlation coefficient, where r = a stronger 0.71. But the demo correlation is better, much better than even the overall demo one was. r = 0.97. That's science experiment good.

CBS' average going rate for the demo is a little lower than broadcast in general, about $36,700 per demo point, but the majority of their shows are very close to that, with only demo-heavy 2.5 Men falling more than 15% away in either direction. Here's a complete list.

Demo Price Price/pt
2.5 Men 5.1 226635 44438.24
CSI 4.8 198647 41384.79
CSI: Miami 4 140065 35016.25
Survivor 4 152246 38061.5
NCIS 3.65 133304 36521.64
Criminal Minds 3.6 116553 32375.83
CSI: New York 3.4 127941 37629.71
Amazing Race 3.2 109736 34292.5
60 Minutes 2.7 93772 34730.37
Ghost Whisperer 2.5 78047 31218.8
Numb3rs 2.3 85007 36959.57
Gary Unmarried 2.2 79986 36357.27
Old Christine 2.1 80106 38145.71

Viewers Price Price/mil
CSI 18.1 198647 10974.97
NCIS 17.8 133304 7488.989
2.5 Men 14.9 226635 15210.4
Criminal Minds 14.4 116553 8093.958
CSI: Miami 14.1 140065 9933.688
60 Minutes 14.1 93772 6650.496
CSI: New York 13 127941 9841.615
Survivor 12.5 152246 12179.68
Amazing Race 10.5 109736 10451.05
Ghost Whisperer 10.3 78047 7577.379
Numb3rs 9.4 85007 9043.298
Old Christine 7.1 80106 11282.54
Gary Unmarried 7.1 79986 11265.63

The differences aren't as stark here, but still... they charge just 42% more per demo point on their priciest show (2.5 Men) than on their least pricey (GW). On the viewer chart, they charge more than 100% more per viewer to be on 2.5 Men than to be on 60 Minutes. Correlation-wise, advantage: the demo.

So maybe CBS "cares" about total viewers in the sense that they'd like to be able to get money for them, but they're still selling their ads based on the exact same number as everyone else. Some people say that CBS is more of an A25-54 network, and maybe that correlation's even better, but the A18-49 correlation is quite good, and much better than it is across all the networks combined. If they genuinely "care" about total viewers and use that number to make decisions about their network, they are throwing away money, and I offer to run the network. I'll even do it cheaper than the guys you have now, CBS.

Saturday, December 12, 2009

Peetooplus - Skewering "Skew"

In this day and age, when the demo seems to have generally "won" the "debate" in the minds (if not the hearts) of most people following TV ratings, there are basically three routes that a lowly spinster can take to address total viewers. And I've seen all of these routes taken many times by many people on the Internet.

1) Old-skew-is-better. AKA total viewers still matter on some level. "If advertisers can get some of the over-50 crowd watching their show along with their demo, they certainly won't mind that."

2) "A show's strength is purely what its demo is. Skew is meaningless."

3) Young-skew-is-better. "This show is likely to be canceled because it's old-skewing."

But how does it work out in the ad rates world? Unlike in the previous post, when I kind of expected the demo would hold up against viewers, I wasn't sure going in how this one would turn out.

Again, I'll try to look at this in a couple different ways, but it's using the same general data. The number I'll be using for "skew" is the percentage of total viewers that fall within the adults 18-49 demo. I derive that using numbers I've shown you in previous posts, plus assuming 132 million adults 18-49 total live in the US. (Nielsen's number for '09-'10. It may be slightly off the '08-'09 number, which is the year of all this data, but it's very, very close.) The average skew for a primetime show is about 44%.

We know from last post that on average, a demo point is worth about $40,000 per 30-second spot. So how do the oldest-skewing shows and the youngest-skewing shows do relative to that price? (That's what the "diff" number is, a percentage difference from the $40,207.62 average I got for my 34 shows)

These are the ten youngest-skewing scripted shows that returned to the same timeslot.

Demo Price/pt diff Skew
Family Guy 4 53687.5 +33.53% 68%
The Office 4.3 44473.49 +10.61% 68%
Simpsons 3.4 59388.24 +47.70% 63%
30 Rock 3.3 48386.06 +20.34% 63%
American Dad 2.9 47030.34 +16.97% 62%
Parks & Rec 2.4 49995.83 +24.34% 59%
Dollhouse 1.5 37580 -6.54% 54%
House 5.1 35940.78 -10.61% 50%
Private Practice 3.8 46171.05 +14.83% 50%
Grey's Anatomy 5.7 42186.32 +4.92% 50%

In other words, these are the ten scripted shows with the fewest viewers relative to their demo. And eight of them are "overpaid" relative to that approximately $40k/point average across all shows. The two others are not deeply in the red, and they could easily be down because of speculation; nobody really thought Dollhouse would average a 1.5 when it was below that for most of the end of its run, and House, even though that 5.1 average is strictly for airings within the Monday 8pm timeslot, was showing declines toward the end of the season as well. Family Guy and The Simpsons are by far the most overpaid, even though they're two of the three youngest-skewing shows on TV. Strike one for total viewers.

Here are the ten oldest-skewing scripted shows.

Demo Price/pt diff Skew
NCIS 3.65 36521.64 -9.17% 27%
Ghost Whisperer 2.5 31218.8 -22.36% 32%
Numb3rs 2.3 36959.57 -8.08% 32%
Criminal Minds 3.6 32375.83 -19.48% 33%
Castle 2.4 38625 -3.94% 34%
CSI: New York 3.4 37629.71 -6.41% 35%
CSI 4.8 41384.79 +2.93% 35%
CSI: Miami 4 35016.25 -12.91% 37%
Bones 2.7 39978.52 -0.57% 38%
Old Christine 2.1 38145.71 -5.13% 39%

Nine of ten are underpaid, and it's a real stunner to me that CSI is the lone one of these shows to get a good value, because its average is incredibly inflated by the much higher numbers when Petersen was still around, and I would think it'd have been speculated downward. Perhaps being in that really tough Thursday 9pm timeslot helped it. Most of these are not far from even, but they're mostly still down. Old skewers getting underpaid.

A couple other examples: let's take some shows with very similar demo but very different viewer numbers. I'll try to avoid Simpsons/Family Guy, which were major outliers even on the generally pretty consistent demo plot.

Demo Price Viewers Skew
Private Practice 3.8 175450 10 50%
NCIS 3.65 133304 17.8 27%

Private Practice is just a tick and a half higher than NCIS (4%) but its ads go for a whooping 32% more!

Demo Price Viewers Skew
CSI: New York 3.4 127941 13 35%
30 Rock 3.3 159674 6.9 63%

CSI:NY's originals were 3% higher on average but 30 Rock fetched about 25% more per ad despite having barely over half the viewers.

Now, this isn't to say that this is a very good correlation. When I got a coefficient for "Skew vs. diff," that number was only r = 0.67. Still stronger than ad rates against viewers, but not all that strong. So I'm not willing to come right out and say "advertisers pay more the younger you skew," because it's hard to conclude that from top to bottom. All I can say is that at the very top of the skew chart, the youngest-skewing shows seem to get a good value. Beyond the top 6 or so, it's pretty jumbled the rest of the way. But there aren't any good examples in the other direction on my chart, examples of a big viewer show hugely outdoing a little viewer show of the same demo. So I can say pretty confidently that total viewers don't even matter (in a good way) on some kind of secondary level. Given the three schools of thought I listed at the top of this post, the truth lies somewhere between 2 and 3.

This isn't to say that the programs are getting penalized for their high total viewer counts or rewarded for low ones. I think the young-skewing shows typically have appeal to advertisers that we can't see when we just see these two numbers. Usually a show with a small percentage of out-of-demo viewers is also likely to have more of its in-demo viewers be part of the premium A18-34 subset. Maybe they're more heavily DVRed and that factors in somehow. I'll leave the "why" to you. These posts aren't "why," they're just "what."

Tomorrow, a look at the notion that CBS sells total viewers.

Friday, December 11, 2009

Peetooplus - The Basics of Viewers/Demos

Sorry for not posting in the last few days, but I'm back with more on the "debate" about whether total viewers or the adults 18-49 demo is what matters in the analysis of TV ratings. In the last post I presented these two scatter plots, one plotting '08-'09 viewer averages (for originals) against fall '09 ad rates and one plotting demo averages against ad rates. Again, I'm only using shows that returned in the same timeslot, in an attempt to minimize the influence of speculation by the ad buyers.

As I remarked back then, it's easy to tell which one has a stronger correlation; the demo one. And the eye test is my best way of explaining it, because it's been several years now since my last statistics class. So while I can tell you that the correlation coefficient on the viewer chart is just r = 0.46 and the correlation on the demo chart is r = 0.90, and the strength of that r value has to do with how close you are to 1, I can't really say exactly how much better 0.90 is than 0.46. But it's a lot.

I'll try to make the point in a couple other ways. Still a lot of numbers, but hopefully not advanced statistics.

Demo Price
Grey's Anatomy 5.7 240462
Desperate Housewives 5.3 228851
2.5 Men 5.1 226635
House 5.1 183298
CSI 4.8 198647
DWTS (Monday) 4.7 178687
The Office 4.3 191236
CSI: Miami 4 140065
Survivor 4 152246
Bachelor 4 139500
Family Guy 4 214750
DWTS (Tuesday) 3.9 172570
Private Practice 3.8 175450
NCIS 3.65 133304
Criminal Minds 3.6 116553
The Biggest Loser 3.6 128295
Brothers & Sisters 3.5 140445
CSI: New York 3.4 127941
Simpsons 3.4 201920
Extreme Makeover: Home Edition 3.3 136743
30 Rock 3.3 159674
Amazing Race 3.2 109736
Celebrity Apprentice 3.2 110283
American Dad 2.9 136388
60 Minutes 2.7 93772
Bones 2.7 107942
Ghost Whisperer 2.5 78047
Castle 2.4 92700
Parks & Recreation 2.4 119990
Numb3rs 2.3 85007
AFHV 2.2 75893
Gary Unmarried 2.2 79986
Old Christine 2.1 80106
Dollhouse 1.5 56370

These are all the points on the demo scatter plot above. As you can see, almost all the shows at the top of the demo ranking are near the top of the price-per-spot list as well. (Family Guy and Simpsons being the most glaring exceptions.) For the most part, as the demo goes down, so does the price.

Now here's the viewer one.

DWTS (Monday) 19.6 178687
CSI 18.1 198647
NCIS 17.8 133304
DWTS (Tuesday) 16.2 172570
Desperate Housewives 15.4 228851
Grey's Anatomy 15.2 240462
2.5 Men 14.9 226635
Criminal Minds 14.4 116553
CSI: Miami 14.1 140065
60 Minutes 14.1 93772
House 13.4 183298
CSI: New York 13 127941
Survivor 12.5 152246
Bachelor 11.1 139500
Amazing Race 10.5 109736
Ghost Whisperer 10.3 78047
Private Practice 10 175450
Brothers & Sisters 10 140445
Extreme Makeover: Home Edition 10 136743
Bones 9.4 107942
Numb3rs 9.4 85007
Castle 9.3 92700
The Biggest Loser 8.9 128295
The Office 8.4 191236
Celebrity Apprentice 8.2 110283
AFHV 8.1 75893
Family Guy 7.8 214750
Simpsons 7.1 201920
Gary Unmarried 7.1 79986
Old Christine 7.1 80106
30 Rock 6.9 159674
American Dad 6.2 136388
Parks & Recreation 5.4 119990
Dollhouse 3.7 56370

Much worse correlation here. Out of 34 shows on the list, eight charge less than $100k for a 30-second spot. All eight of those were in the bottom ten of the demo list. Only four of the eight are in the bottom ten of the viewer list, and The Simpsons/Family Guy, two of the most expensive shows, are near the bottom of the list. NCIS, despite being the third most viewed show, is near the middle of the pack as far as price goes. Difficult to say those shows are being sold even remotely based on their total audience.

Now, here's another way of looking at this. I'll include price per demo point and price per million total viewers.

Demo Price/pt
Simpsons 3.4 59388.24
Family Guy 4 53687.5
Parks & Recreation 2.4 49995.83
30 Rock 3.3 48386.06
American Dad 2.9 47030.34
Private Practice 3.8 46171.05
The Office 4.3 44473.49
2.5 Men 5.1 44438.24
DWTS (Tuesday) 3.9 44248.72
Desperate Housewives 5.3 43179.43
Grey's Anatomy 5.7 42186.32
Extreme Makeover: Home Edition 3.3 41437.27
CSI 4.8 41384.79
Brothers & Sisters 3.5 40127.14
Bones 2.7 39978.52
Castle 2.4 38625
Old Christine 2.1 38145.71
Survivor 4 38061.5
DWTS (Monday) 4.7 38018.51
CSI: New York 3.4 37629.71
Dollhouse 1.5 37580
Numb3rs 2.3 36959.57
NCIS 3.65 36521.64
Gary Unmarried 2.2 36357.27
House 5.1 35940.78
The Biggest Loser 3.6 35637.5
CSI: Miami 4 35016.25
Bachelor 4 34875
60 Minutes 2.7 34730.37
AFHV 2.2 34496.82
Celebrity Apprentice 3.2 34463.44
Amazing Race 3.2 34292.5
Criminal Minds 3.6 32375.83
Ghost Whisperer 2.5 31218.8

Viewers Price/mil
Simpsons 7.1 28439.44
Family Guy 7.8 27532.05
30 Rock 6.9 23141.16
The Office 8.4 22766.19
Parks & Recreation 5.4 22220.37
American Dad 6.2 21998.06
Private Practice 10 17545
Grey's Anatomy 15.2 15819.87
Dollhouse 3.7 15235.14
2.5 Men 14.9 15210.4
Desperate Housewives 15.4 14860.45
The Biggest Loser 8.9 14415.17
Brothers & Sisters 10 14044.5
House 13.4 13678.96
Extreme Makeover: Home Edition 10 13674.3
Celebrity Apprentice 8.2 13449.15
Bachelor 11.1 12567.57
Survivor 12.5 12179.68
Bones 9.4 11483.19
Old Christine 7.1 11282.54
Gary Unmarried 7.1 11265.63
CSI 18.1 10974.97
DWTS (Tuesday) 16.2 10652.47
Amazing Race 10.5 10451.05
Castle 9.3 9967.742
CSI: Miami 14.1 9933.688
CSI: New York 13 9841.615
AFHV 8.1 9369.506
DWTS (Monday) 19.6 9116.684
Numb3rs 9.4 9043.298
Criminal Minds 14.4 8093.958
Ghost Whisperer 10.3 7577.379
NCIS 17.8 7488.989
60 Minutes 14.1 6650.496

I know this post is in severe danger of becoming too numbers-laden, but here's the point I'm trying to make. The Simpsons is the most expensive show per demo point and it charges not even twice what Ghost Whisperer, the cheapest show, charges. The average price per demo point is right around $40k , and 32 of 34 shows charge within 25% of the average demo price (between $30k and $50k) with only the animation tentpoles being outside of that range. 21 of 34 shows are in an even closer range of 12.5% (between $35k and $45k).

On the other hand, The Simpsons, also the most expensive show per-viewer, charges well over four times what cheapest-per-viewer 60 Minutes charges per viewer. If we're trying to say viewers sell advertising, the average price on this chart would be about $14k. Using the previous benchmarks, just 17 of 34 shows are within 25% of the viewer price and only 10 of 34 are within 12.5%. The viewer correlation with ad rates is much weaker.

I can't really think of any other ways to drill it home, so in conclusion: the demo matters a lot, and it matters a lot more than viewers. The correlation is not as exact as something you might hope for in an actual scientific study, due to other demographics and perhaps some speculation being factored in, but it's a heckuva lot closer than the viewer correlation, and anything you could say to diss the correlation between demo averages and ad rates would have to go at least double for the viewership chart. If there's one number that we as industry outsiders trolling around on the Internet can look at to find a connection with advertising money, it's adults 18-49.

But, sayeth the CBS lover, maybe it's viewers and demos! Well, if that's true, then demos are still much more important, and that's being generous. And considering some of the most "overpaid" shows relative to their demos are shows with very low viewership (especially The Simpsons and Family Guy, but to a lesser extent stuff like The Office and 30 Rock), it would seem the opposite is more likely to be true: that given two shows of about the same demo, having fewer viewers helps. I'll look at that question in an upcoming post about "skew." And I'll also look at the urban legend that "CBS sells total viewers." Not sure, with only three data points, if I'll have enough to tackle the "Total viewers sell on Friday" urban legend, but we'll see.

Monday, December 7, 2009

Peetooplus - The (Un)importance of Total Viewers

It isn't sweeps, there aren't any new shows coming up, so I'm going to take this blarg in a couple different directions during the holiday season. For starters, in honor of tonight's premiere of TNT drama Men of a Certain Age, I'm getting back to the very basics of TV ratings: total viewers (AKA P2+, which sounds like "peetooplus," hence the post title) vs. the adults 18-49 demo. Just how significant are "people of a certain age" in the world of advertising?

At some point early on in my interest in TV ratings, it became drilled into my head that adults 18-49 is the number that matters to the networks. Most sites that relay daily TV ratings give you both the total number of viewers and the number of adults 18-49, but for some reason it just seemed like people who seemed to be "in the know" were focusing on the demo, like Media Life Magazine, which trafficked almost exclusively in those numbers. But for awhile in the fledgling world of Internet analysis of TV ratings, VIEWERS VS. DEMO was a legitimate debate, mostly between the CBS watchers who watch old-skewing shows and the everything-else watchers who watch young-skewing shows. It seems that debate is not as big a deal at this point. Most people just go to the demo, but will sometimes bring total viewers into the equation. (If it suits their spin. Wink.) The heavily visited TVbythenumbers primarily uses A18-49, which has probably helped.

In this day and age everybody can tell you why the demo is the number that matters: usually you'll hear some combination of "they don't have brand loyalty" and "they don't watch as much TV, thus they're in high demand." A lot of people are very unhappy about those reasons, but they're the best ones we've got. But thinking back on the evolution of the debate, I realized something: I've never seen any of the stuff proven. I've just gone along with the explanations, they made enough sense, but I haven't seen any of it backed up mathematically. And that shouldn't be good enough for a numbers nerd like me.

So here I am, with the "debate" pretty much settled, giving you the most untimely post ever; I'm gonna show you that the demo matters a lot more than total viewers in determining the price of advertising. And for those who still don't "believe" the demo is all that important, consider this series of posts TV Ratings 101.

I'm going to take the rates paid for a 30 second spot in '09-'10 and compare them to the shows' total viewer and demo averages (in originals) for the previous season. (The complete list of ad rates is here.) It's not really this simple, because there are many other things that go into these ad rates, like other metrics (A18-34, upscale, maybe DVR numbers, etc.) and some degree of speculation. To try to limit the power of speculation, I stuck to shows that returned in the same timeslot.

I'm also going to look at a couple other urban legends surrounding total viewers, namely "Total viewers matter on a Friday, because the demo ain't home!" and "Total viewers matter to CBS, because they care about their audience!" I'll also take a look at another way to potentially (and perhaps inversely) apply total viewers to the ad rates world: the concept of "skew." I've gone by the creed that a show's strength is what its demo is, and that it doesn't matter if it has a low viewer count or if it "skews old." But is that true? Will advertisers pay more for a "younger skew" (potentially getting more adults 18-34) or an "older skew" (meaning they can throw in some old folks on the cheap)? Or is "skew" statistically insignificant and just another fabrication of Internet spinsters?

More to come, but here's a little tease. These are a couple scatter plots that graph viewer averages against ad rates and then demo averages against ad rates.
Suffice to say, the demo one (blue) is a lot closer to looking like a line; that is, ad rates seem to have a considerably more linear relationship with the demo than they do with total viewers. Even if you're desperate to cling to the importance of total viewers, you have to admit that on the top graph, the line is considerably thicker. And that's being generous, because there isn't much of a line at all. More like a blob.

© 2009-2022. All Rights Reserved.