Since this is such a new distribution model, they'll probably be patient and just see what the viewing pace is like for a show like this. The downside is that they'd have no real incentive to tell us what they're finding. Anyway, they're scheduled for two seasons of the show so that should give them a lot of data.
And then they can also use Arrested Development
to see if there's a comedy vs. drama difference, and use Hemlock Grove
to see if genre stuff plays out differently. Personally, I binge view comedy the most, genre stuff maybe somewhat less, but would be most likely to pace my viewing of a political drama.
They'll also look at the rates of new subscriptions and subscription cancellation around the time of a new series launch, vs usual patterns. These shows pay for themselves by attracting new subscribers and preventing cancellations, so any deviation from the expected can be credited to the new show (assuming everything else about their service is "normal" during that time.) But this is something else they won't tell us about, except maybe in generalities if the numbers are very good.
Beyond all this, there's the brand-building impact of a high-profile series starring Kevin Spacey. Netflix needs to keep building credibility with Hollywood (they are inundating deadline.com with ads for the show now) for future development efforts, so regardless of how well it does with customers, that's still important.