This was gonna be the tail end of a Notes from the Vault post, but it got too long, so I'm just putting this up now and I'll try to finish up the Notes post sometime soon.
So when I started looking at half-hour breakdowns for sporting events, my goal was some grand unified theory about how much of a sporting event's rating is people showing up to watch this particular "matchup" and how much is people showing up because of interesting developments in the "game quality." My idea was that whoever's watching the first half-hour is there solely for the "matchup," and whoever shows up across the rest of the telecast* is there for "word of mouth" reasons, and that there would be some way to correlate the ratings build with the dramatics of the game.
*- My "Game Quality" number compares the first full half-hour of game action (usually from 8:30-9:00) with the game's full rating. This isn't the best possible measurement (that'd be first half-hour vs. last half-hour), but it's the best one I can get, since I don't have ways of weeding out the pre-game from the 8:00 half-hour or getting the after-primetime breakdowns.
Probably the most profound thing I've come up with so far is that it's much more about the sport than about the dramatics of an individual game. Yes, there's some correlation between game quality and ratings in Sunday Night Football, but it's always a relatively small deal. The Saints' 55-point win over the Colts had a -7% Game Quality (in other words, a 5.6 A18-49 at 8:30 and a 5.2 A18-49 for the whole game). The best Game Quality on Sunday was last week's close Steelers/Chiefs game at just +10% (7.2 at 8:30, 7.9 for the whole game). So a really massive blowout drops just 7%, and most "good" games increase somewhere from 5 to 10%. These games' ratings are much more about the appeal of the matchup than the game quality.
How's this compare with ABC's Saturday Night Football? Blowouts are still in that 0%ish range; the worst of the season was LSU's 26-point win over West Virginia at -4% (2.4 at 8:30, 2.3 for the whole game).
Good games? Massively different. There are seven weeks of Saturday Night Football that are up double-digits. Ohio-State/Nebraska was up 29% (1.4 -> 1.8). Stanford/USC was up 33% (2.1 -> 2.8). And then there's the epic night of November 19, in which ABC had USC's upset of Oregon and then immediately cut to the thrilling end of Baylor/Oklahoma. Up 42% (2.4 -> 3.4). That's even a bigger raw numbers gain (1.0) than the largest raw numbers gain of the biggest Sunday NFL game (0.9 for Dallas and the Jets in week one, a game that was over three times as big overall).
What's the difference? My two theories: first, college football is much more regional than national. This mostly explains why the beginning ratings are so much lower in college. The NFL has betting and the fantasy football phenomenon to create significant interest nationwide in almost every game. The people who are going to be interested are there from the outset. There are a lot fewer people who will sit down to watch any ol' college game. I know I'm much more inclined to watch some ACC game than some Big 12 game.
Second, and nobody probably wants to hear this: the BCS. Say what you want about the BCS, but there's no denying that it creates extreme regular season stakes, and some game between Oregon and USC can become a matter of urgent national interest if the game is close down the stretch. A game can go from regional to national over the duration. It's not really that way in the NFL. If you lose, you're one game farther from the playoffs, but 12 teams out of 32 still get in. Those aren't very definitive stakes, especially prior to the last couple weeks of the season. In college, two get in, and so a close game can draw in fanbases from lots of different teams on the outskirts of the BCS picture.
This is kind of rare in the regular season, but what if the stakes are hugely clear from the opening kick in college? See CBS' LSU/Alabama "Game of the Century," which got a massive 6.8 in the opening half-hour but "just" a 7.2 for the whole game (+6%), even though the game was really close. A more NFL-esque situation.
The conclusion: both college and NFL ratings are heavily about matchups. Only college ratings are heavily about what happens after the opening whistle. The raw number of people who will come in due to "game quality" is actually not that different; the difference is that it's a much larger percentage of the initial college audience.
I'll try to throw these Game Quality numbers onto the SpotVault pages later this month.