Tuesday, August 2, 2011

True Strength: 10:00 Shows as Competition

Before we come up with a "formula" for how much competition brings down TV ratings, we have to get a sense of what a "normal" level of competition is. And before we can do that, we have to get all competition on a level playing field. Now, I could approach this in a much more complicated way than I'm going to (more on that a little later), but in looking over the full season of competition or "Broadcast Persons Using TV" levels, there are two really obvious problems with using "bcPUT" as a level playing field stat:

1) The 10:00 hour. The basic problem here is that we don't have local programming numbers for Fox, at least not with any regularity. Fox is a big primetime presence in the other two hours but counts as a 0.0 in 10:00 "broadcast PUT" calculations. That makes the bcPUT look much smaller than it should be for 10:00 shows.

2) Sports. We've looked at this a bit in previous posts, but big sporting events are largely comprised of an audience that would not otherwise be watching something else in primetime. Many programs that air against big sporting events do not drop nearly as much as they do/would against entertainment programs of the same rating. In other words, when calculating competition or bcPUT, sports "count" too much.

The 10:00 Hour

When I started doing Competition calculations on the daily Spotted Ratings posts, it became clear that Competition as a stat was not very useful when comparing 10:00 shows with anything else because of the reason above: Fox (and to a much lesser extent the CW) were not included.

So how should we make the adjustment to put 10:00 on a level playing field?

My fix back in the daily posts was to multiply the Competition by 1.75. The thinking was that rather than facing two networks of competition, the 10:00 shows should be facing 3.5 networks of competition as they do in the other hours (calling CW a "half network" in adults 18-49 is actually pretty generous, but we'll go with it). So multiply the two-net competition by 1.75 to get what would basically be the equivalent of 3.5-net competition.

To do that with Broadcast PUT, which I've been using in these posts, that intuitive multiplier would be 1.5; broadcast PUT counts three networks at 10:00, but it should count 4.5 networks. 3 * 1.5 = 4.5.

But the problem with using a multiplier is that you're basing the other 1.5 nets on the first two nets; meaning, the higher they are, the higher the last 1.5 are. And that sort of flies in the whole face of coming up with a competition adjustment; the whole point is that big ratings drive DOWN the competition.

So let's take a look at how 8:00-10:00 and 10:00-11:00 compare in practice.

broadcast PUT

Week Wkd All
10 to 11 6.90 5.15 6.40
8 to 10 11.28 6.39 9.94
Change 4.37 1.24 3.54
%Change +63% +24% +55%

So it seems that the 1.5 multiplier (which would be +50%) is pretty close to how it actually shakes out in practice across the whole week (+55%). But again, it's not a very good way of doing it, because it almost certainly undercounts the number when the big three aren't programming much and almost certainly way overinflates it when there's something huge at 10/9c like The Voice or Sunday Night Football.

I think the best strategy is to come up with some constant number to add. And it looks like the 3.54 difference is as good a starting point as any. Maybe the local programming/cable influence is smaller on Friday/Saturday, but keep in mind that these numbers are combinations of "True Strength" which have already adjusted for viewing levels.

What does adding 3.54 to every 10/9c bcPUT mean? Well, in terms of trying to "equalize" competition across the night (which really shouldn't be the goal anyway, because we have no real reason to believe competition should be "equal" at 10:00), it won't really do that. bcPUT will continue to be a little smaller (about 0.83 on average) during the week at 10/9c, but it probably should be; most big hits air at 8:00 and 9:00. It'll be 2.26 points bigger at 10:00 than in the other hours on average, but that sort of lines up with how Fox's local programming is probably relatively stronger on that night than other nights. 10:00 is also consistently the most-viewed hour on the weekends (even using Old Methodology), meaning cable competition is relatively stronger too.

So, I'm not really sure about the merits of using 3.54*; if I were pulling a number out of my ass, it'd probably be something closer to 3.0 so the 10:00 hour wouldn't rise quite as much on weekend nights. But I'd rather go with a number that I calculated than pull something out of my ass whenever possible. We'll see how it works.

*- Incidentally, I have seen a very limited amount of total 10:00 local programming info. It usually has Fox scoring mid-to-high-1's in A18-49 and the CW getting something like an 0.5. I would imagine at the peak of the season (especially with an Idol lead-in for instance), it might get higher or even much higher than that, but I'm guessing on average it's probably a little less than a 3.54 PUT-TS. However, the added influence of cable (which usually tries hardest at 10:00) I think makes 3.54 reasonable enough.

Next time, we'll try to figure out how to calculate how much competition is "normal" in a given timeslot and use that to come up with a competition-adjusted True Strength. After we get the formula, we'll have to tackle the sports thing and then I think we'll be able to move on!

No comments:

Post a Comment

© SpottedRatings.com 2009-2018. All Rights Reserved.