"the way we watch television has changed; Nielsen has not"
Actually, that's not quite true. These days, Nielsen takes into account the viewing of shows that have been time-shifted via VCR/DVR/PVR, up to 7 days after the show originally aired. AFAIK, however, they don't take into account viewing via VOD (where shows are available the day after airing on TV) or via the Internet. Although I suspect that the networks keep track of the viewings via their web sites and make that data available to advertisers as well.
"The Nielsen Twitter TV Rating will serve to complement Nielsen’s existing TV ratings"
Well, at least this will have one benefit as far as us Canadians go: since a lot of (most?) shows air on Canadian TV at the same time, or within hours, of when they air on US TV, Canadian viewers will *finally* be counted in the ratings for US shows. It's been a *long* time coming...counting Canadian viewers could probably have saved several shows from being cancelled.
"While it doesn't address many of the problems with the current Nielsen system -- notably, the significant underrepresentation of DVR users"
That's hardly the notable problem with the current system, given that DVR time-shifted viewing is counted in the ratings these days every bit as much as live-viewing is. The notable problem with the system is the same as it pretty much has been since the start of the program: the system only takes into account the viewing habits of "Nielsen families", and each "Nielsen family" is considered to represent the viewing habits of a large number (4,580 as of 2009, according to Wikipedia...25,000 "Nielsen families" out of 114,500,000 homes with a TV) of American families. The fact that anyone would attach any significance at all to such a small sample, let alone decide whether a show lives or dies, is laughable. Technology these days is such that it's possible to take into account what show *every* cable/satellite connected TV is tuned to. Nielsen is just far too slow to adapt to technology.
"Of course, if we're looking at it objectively, ratings are all about advertisers ... might a smaller group of more engaged viewers be more valuable than a larger group of more passive viewers? It's possible"
An interesting question. I think one potential problem with your theory is that advertisers are not interested in just overall numbers, but also in the age demographics those numbers fall into...which can't be determined from Twitter. If it was just overall numbers, you could well be right. Of course, since most Twitter users probably fall into the prime demographic group (18-49), it's possible the overall number of Twitter users talking about a show would be enough for advertisers...especially when combined with the data from Nielsen. Another potential problem with your theory is: when are these viewers that are more-engaged with the show doing their tweeting...especially if it's so-called "live tweeting"? Probably not while show content itself is on, because then they'd be missing parts of the show and thus wouldn't be as engaged. Most likely they'd be tweeting during the commercials...which would be bad news for the advertisers, since that means the viewers aren't watching the ads. Looked at from that point of view, the advertisers might consider the more-engaged viewers as *less* valuable than the more-passive viewers. Of course, an argument could be made that even though the more-engaged viewers aren't *watching* the TV, the sound is most likely still turned up, so they're still *hearing* the ad pitch...which could have the same result as watching the ad.
Overall, I agree with you. This idea is very interesting and has lots of possibilites...not the least of which is turning the rating system on its ears, which would most definitely be a good thing and one that is long overdue. I guess we'll see just how well it works next fall.