Just saying "focus group" doesn't really provide much insight on the level and utility of feedback networks receive.
How big of an opportunity do you feel there is for more/better data?
But more to the point, the real magic isn't in the data networks may or may not be able to use ahead of time. It's in the data they should capture from the shows: how they're trending, how they're being talked about, what channels people use to discover them, conversion rates from various channels, marketing efficiencies in various channels, cohorts among viewers based on spending patterns, sentiment, word-of-mouth likelihood, etc.
My overall thesis is that less time and money should be spent on (80-90% useless) development research, and more money should be spent on analytical marketing post-launch. I think TV networks can learn a great deal from video game publishers, for example: similar business model, similar hit dependency, similarly huge budgets -- but a complete mastery of distribution and marketing channels.
Nielsen ratings, of course, get looked at and scrutinized daily. But those are lagging indicators, they're approximations, and they're not actionable. And changes made to shows can't be directly attributable in the following ratings.
A strong counterargument to this logic -- and I feel I should give it its due say -- is that overanalysis will lead to design by numbers. And that's a very real danger. But I think the greater present danger is on not being analytical enough. We've got a long way to go before we're designing shows by numbers. That'll be a bridge too far, but the bridge isn't even in eyeshot yet.