The raw averages are less interesting than the distributions here. Facebook is the prototypical example of the importance of network effects. The mean bids here

The raw averages are less interesting than the distributions here. Facebook is the prototypical example of the importance of network effects. The mean bids here are in the thousands, but the medians are $100-$200. That means you could pay its users $100-200 to leave for a year, and it would lose about half its users immediately. But after the first half of users leave, it's probably worth a lot less to the remaining users than before the first bunch left. Even if 90% of their users are willing to pay some price to use it, charging that price might lose them most or all of their users. I think that's super interesting.