Second Life's absentee population Second Life's absentee population CLAY SHIRKY — Second Life released a bunch of figures last Friday, including the cumulative number of users, as part of their "effort to drive toward complete transparency and openness", as they put it. I've been critical of Linden Lab's population figures in the past. And it turns out I was right, about all of it.

Now that Linden is publishing actual user numbers, we can see that the Residents figure, as expected, is a big overcount over actual people (about 50% inflation, in fact, accounting for over a million ersatz users). Second Life doesn't have two million users. They have had two million users over the life of the service, and they've lost most of them. Of those users, the majority — something like 5 out of 6 — bailed in the first month. What we don't know is what the other sixth are up to, but after Friday's post, we can guess the answer is "Not much." As John Zdanowski, the Linden employee who posted the figures, notes, "Approximately 10% of unique users have logged in for 40 hours or more."

He doesn't caveat this — it isn't current users, or 40 hours per month. The plain meaning of that sentence is that fewer than 200,000 people have given Second Life even a cumulative work week of their time, over the history of the platform. (After revealing this figure, Zdanowski immediately offers two separate rationales for having so few committed users, and two separate analogies for why poor adoption is no big deal, in a single paragraph.)

As any illusionist will tell you, the trick is mainly in getting the audience to look at the wrong thing. In Linden's case, they want you to think that cumulative users matters when it doesn't. A new user won't care one whit that, as of last year, 1,422,846 people had tried Second Life. What they want to know is how many of those people will still be around to interact with now?

This is the question the press should be asking — "How many of those users from 2006 have logged in recently?" Linden won't answer, of course, but it might be interesting to hear how they square the invisibility of the one population number that actually affects user experience with their stated goal of transparency and openness.