2) This shouldn't be surprising as it tells us most people don't care that much, and a few people 'really do' - which is normal for most goods.
3) Since we're dealing with network externalizations, wherein the value goes up when more people use it (the value of FB is the network, less so the features) - then a lower price makes more sense. If all those people who bid $100 or less (say 2/3 of the population) quit FB tommorow, than those in the $1K range who 'really like FB' would probably leave as well, because their friends would not be there.
This gives us a hint that in product wherein the consumer surplus is not very evenly distributed - and there are strong network effects ... that the optimal price should be low.
4) Getting paid to leave is simply a different metric than having to get out your wallet to pay for something. Surely the authors know this but I didn't see it mentioned. 'WTP != WTL' i.e. willingness to pay is not the same as willingness to leave.