Monday, October 25, 2010

CW demo: ...or is it?

Way back in my first post on the so-called "CW Demo" of women age 18 to 34, I basically decided that I would start evaluating CW shows based on women 18-34 ratings due to exactly one comparison: Gossip Girl getting about 80% more ad dollars than Supernatural despite it having almost the exact same adults 18-49 rating.

That's not a very solid foundation, but that discrepancy hasn't changed in 2010-11 ad rates: Gossip Girl got 62% more money per 30 second spot than Supernatural ($47,248 vs. $29,100) even though Supernatural actually had over 10% more A18-49 (1.22 vs. 1.09). In this year's case we do have to figure in some speculation, since Supernatural was headed to Friday night, where Smallville took a big year-to-year hit last season. But even if you give Supernatural a 1.00 demo (about what Smallville got last year), that still means Gossip Girl's getting about 50% more bucks per A18-49. That means adults 18-49 is an invalid number for looking at the CDub, right?!


That happened to be a fairly cherry-picked example. Plotting all eight of the CDub's returning shows based on last year's total viewers, adults 18-49, and women 18-34 vs. their ad dollars, here are the charts if you're interested:

Correlation coefficents are Viewers r = 0.80, Adults 18-49 r = 0.88, W18-34 r = 0.80. (remember, the closer to 1 the better)

None of these are particularly good coefficients, but the advantage would seem to go preliminarily to "the demo" rather than what I've named "the CW demo."

Throw out speculation-influenced Supernatural (the main outlier beneath the viewer and A18-49 "lines") and it's Viewers r = 0.92, A18-49 r = a very strong 0.98, W18-34 r = 0.76. In this case, the correlation for adults 18-49 absolutely blows the "CW demo" correlation out of the water. Or leave in Supernatural with a Friday speculation figure of 1.00 in A18-49... still a great correlation at r = 0.96. Throw out Smallville along with Supernatural, meaning we've now gotten rid of both the shows that some people say don't target W18-34, and the W18-34 correlation is up to 0.88, but still much worse than the A18-49's 0.98.

So that's a lot of numbers, but the deal is that it looks like even on the CW, the best correlation is with adults 18-49. I don't have the "in-between" number of all adults 18-34 to test out, and that has long been the netlet's "official" target audience, but it couldn't possibly be much stronger.

It's a tough spot for me, having written so many posts about this network using the W18-34 numbers, but it looks like they just don't match up with the ad rates quite as well as plain old adults 18-49. So can women 18-34 numbers still be educational? I think so. There's still that Gossip Girl/Supernatural discrepancy that spawned this whole kerfuffle. Let's look at the price per A18-49 for the CW shows:

A18-49 Price ($) $/A18-49 W18-34
Vampire Diaries 1.7 74913 44066.47 2.67
Gossip Girl 1.09 47248 43346.79 2.45
America's Next Top Model 1.51 63285 41910.6 2.59
90210 0.95 38426 40448.42 1.86
Life Unexpected 0.94 33725 35877.66 1.63
One Tree Hill 1.1 39382 35801.82 2.19
Smallville 1.01 34163 33824.75 0.75
Supernatural 1.22 29100 23852.46 1.16

I noted in an earlier ad rates post that the 48 shows I looked at charged around $44,000 per demo point per spot this year. There are too many variables for me to proclaim with great authority that "A DEMO POINT IS WORTH $44K!" but it seems like it's in that very general area. And it seems like many of the shows well below that $44,000 per point mark have a pretty weak raw demo number as well. I figure the high-demo shows get more not just on the whole but more per demo because they often grab people who don't watch much other TV.  How does a netlet with generally pathetic A18-49's compete with that? It seems like they can sell that young female-heavy audience to get close to the typical "going rate," while a show like Smallville with very little of that doesn't manage to get nearly the rates you would think it "deserves." Just an educated guess, but it makes more sense than my old method of trying to pin a linear correlation to the women 18-34 number.

It is just one year, and there aren't a lot of data points, so I may revisit this one next year. I'm not sure when I'll be saying much about the CDub going forward, but all this info suggests maybe I ought to go back to the adults 18-49 well a little more often. Unlike the viewers/demos thing, where ad rates are mostly about "the demo" and total viewer/household counts really don't provide any supplemental information, this seems to be a case where there are actually two different numbers that are both helpful and readily accessible. So I'll try looking at both going forward.


kkkkkkk said...

you are right to note the linearity of correlation. The relation of ad prices to 18-34 does not look linear. If you want to do it properly, you might want to run a regression, where the explanatory variables would be all kind of ratings + ratings squared (to capture the possible non-linearity).

Spot said...

Thanks for the advice! You're right that I probably should've been more clear that there does appear to be a relationship, just not a linear one. I tried saying that in less technical terms toward the end, mostly because I am a little too rusty on regression analysis to be able to explain it technically. Since I only do 1-2 posts on correlation per year, I just try to get away with it. :(

Spot said...

The CW Network has it worse with 18 to 49 raitings on their network

Post a Comment

© 2009-2022. All Rights Reserved.