In the past three weeks, I've seen press releases, news stories, and at least one press conference touting Second Life's millions of users. The NBA touted Second Life's 5 million users at a press conference to announce their new area in Second Life. Kraft included the number in an e-mail about its virtual supermarket in Second Life. BusinessWeek uses that number. But the number is bogus, technically accurate but highly misleading. Linden Lab, which develops and operates Second Life, knows it, but they highlight that number anyway.

Here's what's really going on: The Second Life home page prominently lists usage stats. At the top: "Total Residents." That number is, as I write this, 6,074,095; it topped 6 million this week.

But what's a "resident?" For the purposes of this discussion, a "resident" is simply a Second Life account. When Linden Lab claims Second Life has 6 million "residents," what they're really saying is that 6 million accounts have been created on Second Life since it was founded four years ago.

Many of those accounts were abandoned -- 90% of them as a a matter of fact, according to Linden Lab. Moreover, many of Linden Lab's most dedicated users have more than one account, which they use for different purposes (the secondary accounts are called "alts" in Second Life jargon).

But the average person isn't going to think of those things. Most people, seeing that number, are going to say: "Wow! Second Life has 6 million people using it!" Which is exactly what the NBA, Kraft, BusinessWeek and hundreds of other businesses and journalists have done.

How many people are really using Second Life? It's hard to say. Of course, it depends on how you define a user, and that's a problem in any online medium.

For example, let's say I have a blog with Google's free Blogger service. If I haven't updated or visited the blog in a day, am I still a user? How about a week? A month? A year? How often does someone need to log on to a service to be an active user?

Second Life's front page contains the statistic that 1.7 million residents logged in in the last 60 days, and 34,352 residents are logged in now. Those are better stats than the number of total residents -- but, still, how many of the accounts logged in in the last 60 days are alts? How men of those are people who logged in once, decided Second Life wasn't for them, and never came back again?

Linden Lab needs to publicize more detailed and nuanced figures about usage -- aside from the total-residents number -- and let observers make up their own minds about how many users they have.

Now here's the weird part: Linden Lab already does that! Linden Lab produces monthly traffic reports. They sometimes lag -- the most recent detailed report was for March. But they're there -- if you know where to look for them. I couldn't find a link on the home page, so I went to the corporate blog, clicked the economy category, and followed the link from there to the economic stats page..

There, you can find all kinds of interesting statistics on the Second Life economy, many of them in downloadable spreadsheets and charts. The key metrics spreadsheet includes unique users -- the number of actual, unique, flesh-and-blood human beings who've created Second Life accounts. At the end of March, the last month for which public figures are available, that was about 3.2 million people. At that time, the total number of residents was 5.1 million.

Linden Lab makes no secret that it has a high churn rate. Only about 10% of users are still using the service after three months.

Therefore, we can conclude that Second Life had fewer than 320,000 active users at the end of March. How much fewer? We don't really know, but probably a lot fewer -- remember, the active-users number is for people who've spent a total of three months online in a service that's four years old; presumably, many of them are people who dropped the service after three months. Of course, the number of formerly active users who dropped the service are offset by the number of people who should be counted as active users, who are logging in several times a week or more but haven't racked up three months yet.

None of this is new. Researcher Clay Shirky called Linden Lab on its bogus usage numbers back in December.. Since then, Linden Lab has released more detailed statistics, but hasn't backed off of reliance on the number of total residents as its main measure of usage.

Why is Second Life still leading with the sexy, technically accurate, but misleading number?

I asked Linden Lab for an interview, but they declined my request. (By the way, that's a sentence I'm going to program as a keyboard macro so I don't have to keep retyping it; Linden Lab executives are tough to get a hold of.) A PR person said in e-mail: "Linden Lab has gone out of its way to shed light on the active usership, which remains at about 10% of the total registered number." Which is true -- but they're doing even more to get that 6-million-user number out there, and that's the number the overwhelming majority of people will see.

Second Life is a great service, and Linden Lab has a lot to be proud of. The number of unique users grew 20% month-over-month in February, 21% the month before that, and 39% the month before that. More than half of Second Life's unique users signed on since New Year's Day. 264,000 people spent money in-world in March. Users spent $1.4 million in the last 24 hours. The entire world has been created by the users, in a highly concentrated outpouring of creativity.

I love it and I recommend it highly. But Linden Lab needs to cut the crap and stop highlighing that misleading usage statistic.