The first answer is: primary polling is historically difficult. For a variety of reasons, especially in an open-primary state like New Hampshire, it's really hard to get reliable results. Being off by double digits isn't exactly common, but it's not all that rare either.
That said, last night's results really were at the high end of unusual, and none of the obvious possibilities seem to explain it. For what it's worth, Time's Jay Carney, via "a social scientist friend of a colleague" who did some comparisons of polls vs. actual turnout, seems to have the most plausible explanation:
What he found...is that a certain percentage of Democratic voters in the last days of polling presumed Biden (especially) and (to a lesser degree) Dodd hadn't dropped out. By and large, come election day, those Biden and Dodd supporters ended up casting ballots for Hillary. Also, of the 5 percent or so who were still undecideds in the last polls, almost all broke for Hillary.This makes sense to me. None of the "big" explanations seems to pan out, so it's most likely a collection of little explanations: a few points from Biden supporters, a few points from Dodd supporters, a few points from undecideds, a little bit better turnout from women, and perhaps a bit of polling error in the post-Iowa polls. Add it all up and you get a 10-12 point swing. It's not a sexy explanation, but it seems like it's probably the right one.