With the new layout, I have more horizontal space, but most of that goes into the column for the name of the show, which will hopefully cut down on names wrapping into a second line.
FULL TABLE:
| Info | Show | Timeslot | True | |||||||
| A18-49 | Skew | Last | LeLa | Rank | y2y | TLa | Ty2y | |||
| Shark Tank (R) | 1.0 | 30% | -9% | -9% | 1.0 | |||||
| Wipeout | 0.9 | 38% | -18% | -0.2 | n/a | 14/15 | -25% | -14% | +38% | 1.0 |
| Rookie Blue | 1.0 | 27% | +0% | +0.0 | -0.2 | 7/13 | n/a | +0% | +67% | 1.2 |
| ABC: | -8% | +23% | ||||||||
A18-49 - From Nielsen. Live+Same Day Adults 18-49 rating. Percentage of US TV-owning adults 18-49 watching the program. This is the publicly available TV ratings currency that correlates best with renewals and ad rates, so most of the work here is conducted in 18-49 terms. For more on why I use this and not total viewers, I recommend the Demos & DVRs portion of the Intro to Nielsen Ratings and the Peetooplus posts.
New for 2013-14! Skew - Calculation. Percentage of adults 18-49 within the show's total viewership. This year I'm dropping total viewers out of the table. Bottom line is that the "shelf space" is relatively limited and I want everything in these tables to be something I use. And I don't use the total viewers, at least on their own. (Total viewers, along with A18-49 share, can still be found on each show's SpotVault page.) But comparing skews can actually accomplish something: it gives us a picture of how lead-ins and lead-outs are interacting with each other. Skew comparisons are also now included in the formula for the timeslot metric True. If there's a ten-point or more difference between two adjacent programs, those shows are probably not compatible.
Last - Calculation. A18-49 difference (percent and numerical) from the show's previous episode. This has become sort of the standard-issue number in the great race to make daily ratings interesting. I'm not someone who obsesses over these for every single show every single week, because most individual fluctuations are meaningless, and quite a few are misleading, but it's still a pretty reliable go-to way of explaining what happened last night.
New for 2013-14! I'm putting the "Last" number in two different forms. I'm kinda worried it's too redundant, but we'll see how it feels as the season goes on. Both forms are here for a specific reason. I've always used percent difference in the past. The percent difference certainly has its uses, especially for gauging situations where we have a fairly well-formed idea of an expected percent difference. (For example, week two of a new show typically drops in the vicinity of 15%, so comparing against that is nice. A viewing depression due to a holiday is another situation where comparing percent changes is valuable.)
But I thought adding the numerical change would also help. First, it gives you a good idea of which fluctuations might just be sampling error. If a CW show is up 33% week-to-week, that looks amazing percent-wise, but this second number lets you know it's just a 0.1-point difference. Generally those one-tenth differences don't mean much (unless there are a bunch of them in the same direction). Second, the numerical change also coordinates better with the new "LeLa," which has also made the move to numerical for reasons that I am about to explain!
LeLa - Calculation. A18-49 percent difference between the show's lead-in and its lead-in for the previous episode. Typically, if the listings are the same as last week, you can just look to the "Last" of the program above it to get this number. This number is at its most useful when a show has a drastic change in lead-in. It's meant to be used in tandem with "Last" as a possible explanation for week-to-week fluctuations. Since I don't have half-hour breakdowns till next week, LeLa compares full program ratings; so if last week's lead-in had a 2.0 average but a 2.5 in the last half-hour, the LeLa compares to the 2.0. (The last half-hour numbers can be seen later in the SpotVault.)
New for 2013-14! I've changed "LeLa" to a numerical comparison rather than a percent comparison. This is to line it up with my timeslot metric True, which judges the lead-in impact based on numerical rather than percent changes. (Specifically, a one-hour show is expected to drop/gain 0.1 points per 0.6 points dropped/gained by its lead-in, while a half-hour show is expected to drop/gain 0.1 points per 0.5 points dropped/gained by its lead-in.) So it's better for analysis to say that a show got a +2.0 lead-in boost rather than a +200% lead-in boost. By saying +2.0, you know that True thinks the 18-49 rating should've gained 0.3 to 0.4 points.
Rank - Calculation. The A18-49 rating's rank among the show's episodes so far this season. I kind of avoid getting too heavily into the "season/series low" stuff in the spring because I just don't find it all that newsworthy when it's happening to everyone. Still, I think I need some number that refers back to the rest of the season beyond last week, and especially early in the season I prefer this to the "Avg" number used last year.
y2y - Calculation. Percent difference between A18-49 and the show's rating a year ago. This is roughly the "SHOW" portion of my daily "Demos Year-to-Year" posts of past seasons, although the formula is more complicated. I have always been hesitant to have the year-to-year numbers in the mix all season because I don't think they're all that helpful during non-sweeps periods. But it's also been kind of awkward setting them aside in separate posts. So maybe I just won't refer to them during non-sweeps periods!
Anyway, the algorithm for "y2y" is kind of complicated, and this is one of the things I'm still working on. Here is the process the y2y number goes through for now:
1) Premieres and finales are always compared against premieres and finales, regardless of dates.
2) During non-sweeps periods, ratings are compared against the same Monday thru Sunday week last year.
3) During sweeps periods, ratings are compared against the same Thursday thru Wednesday sweeps week last year. This is the number used in the "Demos Year-to-Year" posts last year.
4) Failing to find a comparison in any of those three steps results in "n/a."
I might tweak this a little more. I would like to try to weed out premiere-vs.-non-premiere and finale-vs.-non-finale comparisons. But I'm not sure if I will get that portion done.
TLa - Calculation. Percent difference between A18-49 and the network's rating in the timeslot one week ago. This number is not useful all that often in terms of evaluating the show itself. I will probably make the most use of it to note changes in competition for other shows. I thought about just trying to resurrect the "CoLa" (last week's competition) number I used when I first started these posts, but the problem is I can't make that number accurate without having this week's half-hour breakdowns. So as with LeLa, the best bet is to consult the SpotVault for more accurate competition numbers.
Ty2y - Calculation. Percent difference between A18-49 and the network's rating in the timeslot one year ago. This is the "SLOT" portion of my daily "Demos Year-to-Year" posts of past seasons. Again, this is not that useful during non-sweeps periods when the comparisons are largely original-vs.-repeat, so I'll try to point to it when I think it's at its most relevant.
A note on the two timeslot numbers: Even in scenarios where it's the same show in the timeslot both times, these comparisons will sometimes look a little different from the "Last" and "y2y" since the timeslot numbers compare vs. my half-hour breakdowns. That means it'll compare vs. an average of half-hour numbers that round to one decimal point rather than a full program rating that rounds to one decimal point. For example, a one-hour show could have a 2.5 average and 2.6 -> 2.5 half-hours, meaning it's treated as a 2.5 in the "Last" comparison and a 2.55 in the "TLa" comparison. I came up with a fix for this in the Demos Year-to-Year posts to try to minimize confusion, though I think it had a couple bugs, so I'll work on a fixed version of that if I have time.
New for 2013-14! In addition to the TLa and Ty2y numbers for each individual show, I'm now including them for each network as a whole and for all the broadcasters combined, to provide a larger-picture look at where broadcast TV is compared to a year ago (or a week ago). I've had these on my spreadsheet for awhile and it's something I like looking at each day, so we'll see if it's something I actually want to talk about.
True - Calculation. A metric that adjusts the A18-49 rating for overall viewing levels, competition and lead-in. Trying to explain it any more than that is probably hopeless without writing pages, so I'll just point you to the True Index.
New for 2013-14! The preliminary True numbers included in the tables will be taken to just one decimal point, which is the "approximate" notion I've always used in War of 18-49 tables. Besides, since all the ratings used in the prelims are program ratings to one point, I probably shouldn't claim to be any more statistically significant than that. (At least in the finals I have half hour breakdowns, which hone in on the actual ratings a little better than full program ratings.)
As last year, all Monday to Friday True numbers in daily posts (and sometimes Sunday) are PRELIMINARY. These are recalculated next week when I get half-hour info, and the FINAL versions go up exclusively in the SpotVault.
3 comments:
Not too happy with the new tables–their width has always been annoying on mobile, and now it's worse. Oh well.
You know, a few months ago I was looking at some scripts that adjusted tables based on screen size, and I kinda forgot about it. Once I get through the massive rush of the next couple weeks I'll try to research that a little more. Thx for the feedback.
I know that you don't like it when people fixate on the week-to-week numbers going up or down a tenth, so why don't you take the last four or five episodes' ratings, add them together and see if it says anything. Call it Trend or something.
Post a Comment