The ol virtual worlds numbers game.

January 10, 2008 at 9:23 pm (Uncategorized)

Numbers of avs in virtual worlds are hard to figure out, but the other day i was reading this article by Rick Van Der Wal and something didn’t ring true. Sure enough it was the numbers, which I’ll repost here.

1. Barbie Girls (4 million)
2. Coke Studios (4 million)
3. Gaia Online (3.5 million)
4. Whyville (1.7 million)
5. Second Life (1 million)
6. IMVU (1 million) *Techcrunch
7. Disney’s Toontown/Magic Kingdom (1.4 million)
8. There (1 million)
9. Active Worlds (900,000)
10. Entropia Universe (600,000)
11. Virtual World of Kaneva (600,000)
12. Red Light Center (250.000) *Wikpedia
36,550,000 active accounts (not users) in non-gaming virtual worlds.30,000,000 in social worlds for children.

These are estimates as provided by Fabric of Folly

Size Active Digital Media Universe: 350,000,000

Active users in ‘online 3d environments without a gaming focus’: 36,550,000

(very) Rough percentage active internet users in Virtual Worlds: 10.44%

You’ll notice I’ve set a few in bold – they’re what I consider serious multi-purpose free-form virtual worlds, as opposed to games or kids stuff. I’m creating a distinction between a VW’s, MMOG’s and avatar based chatrooms here. I’ll also concentrate mostly on Activeworlds and Secondlife, as they’re the worlds i’ve been most active in and am familiar with.

First onto what I’ve left out so you understand my reasoning.

Barbie, whyville, Disney etc are pretty much gaming worlds with a social focus. They’re avatar, chats and so not revolutionary. As a matter of fact they’ve been around since the early ninties, and even in primitive forms in the late 80s. In that way they’re not particularly virtual worlds.

Kaneva is facebook or myspace in 3d, so not a general purpose virtual world either. In that sense IMVU is similar but just scrapes it in on user generated content. From what I can tell WhyVille less so.

As for Entropia, often lauded as a general VW just like the big boys, well I’ll let their homepage speak for itself:

The Entropia Universe is more than a game. The Entropia Universe is for real. Real people, real activities and a Real Cash Economy in a massive online universe.

… looks promising so far…

Join people from around the globe who use the Entropia Universe currency, the PED, to develop their characters everyday on the untamed planet of Calypso.

… erm spend spacebucks on your character on another planet. Nope, it’s a game.

So what about the real virtual worlds?
I’ve never seen good statistics from so I can’t comment precisely. What I have heard though is they’re about on par or slightly better than activeworlds, which we do know about.

Activeworlds says on their site:

Since Active Worlds is primarily an Internet-based platform, the
potential market for all Active Worlds-based products is global. We
currently have over 2 million individual users worldwide who have
downloaded the browser
and visited the one UniServer we operate. We
generally receive more than 1,000,000 hits to our universe server per
day, and more than 500 new users download our browser each day. To
date more than 70,000 users have registered to be a “citizen”
of our Active Worlds
universe. These users receive enhanced
capabilities in the Active Worlds environment. Users who do not register
are called “tourists.”

OK let’s get the same statistics from Linden Labs and see what the difference is (keeping in mind that activeworlds has been around a lot longer so their total premium accounts figure is going to be higher.

Linden Labs economic statistics page for today says: Total Residents 11,849,438. That’s sign ups, so not counting alternate accounts (2 million by the best estimates i’ve heard) it means that secondlife has attracted just under ten million individual users to sign up for an account. That’s a guestimate, but a pretty good one. So unique logins alone that puts SL about five times the size of activeworlds (which claims 2 million). This doesn’t take into account the fact that activeworlds has been online since 1995 (see history) whilst SL has only been active this century, making it somewhat younger for such a startling figure.

The difference in growth rate can easily be seen by comparing signup rates. Activeworlds boasts 500 new users each day, which may well be mostly unique. From what I can see Secondlife boasts about 15,000 signups a day, which is down from the peak average around 20,000 a day last year, but even so. If we don’t get tricky with the numbers somehow (eg. throw a few away for alt accounts) it simply looks like SecondLife is growing at thirty times the rate of ActiveWorlds.

Total accounts also doesn’t factor in that people leave. All those eleven-million sl accounts have arrived since I’ve been in world (just over 18 months). Activeworlds has a long history, and has been trading since 1995 as a virtual world, and 2001 as a commercial one. The average churn rate for VW users who do become residents is about six months according to some. We must assume that many of those accounts have left activeworlds over that time and a significant proportion of their two million residents are no longer there. This has happened to secondlife as well, but over a much shorter period.

To summarize this point – SecondLife has about as many fake accounts as Activeworlds has total accounts. When it comes to real accounts the difference is staggeringly large. If you imagine that over the 12 or 7 years the activeworlds figures represent, some users have left, then the difference in size between SL and AW are obvious. You’ll notice in the figures from the top ten at the start of this post, all the virtual worlds I’ve pointed out are about the one million users mark.

So if it’s not the total accounts figure it must be something else. If it’s premium accounts or active accounts… let’s see.

Well it’s not as simple as that – to be a member of activeworlds you have to pay to do anything apart from stand around in orange overalls (from memory). So since 2001 there have been 70,000 users registered to be a “citizen” of activeworlds. Many may no longer be paying. We can compare that to two figures in SecondLife, neither of which is particularly illuminating except to create a stark contrast.

If we compare paying secondlife customers then we’re onto “premium accounts”. I only have the published economic statistics for last september, but at that time 91,015 residents were paying for premium accounts. That was for that month, not over the entire history of SL. Some come, some go. If we knock off the churn for activeworlds it looks quite different.

Pulling the “6 month average churn” figure out of the aether (well out of the mouth of Corey Bridges from Multiverse actually) we’d assume AW’s active premium accounts would be much closer to 10,000. As SL’s premiums have continued to grow since september (with a brief dip recently) we can assume that SL has about ten times as many paying customers as AW. However you don’t have to be a paying customer to live in SL, earn virtual cash and do stuff….

The other figure that probably compares better with activeworlds residency is SL “active users”. They can do everything premiums can do apart from own land. We can be generous with AW and say that – back of the napkin – there might be quite a few folk who might wish to run around in world without any powers or rights and just chat. Let’s say that’s 20,000 or so. We compare that to the number of folk who’ve run around SL in the last fortnight and we get Residents Logged-In During Last 14 Days – 508,561. So there’s a big difference again. LL defines active residents as those who have accumulated at least 60 minutes in the last 60 days so it’s probably closer to 1.25 million in the figures they publish in their PR.

So there you have it. Based on solid figures from Linden Lab and some guestimates from Activeworlds, the active user base of SecondLife is about sixty times the size of activeworlds. Q.E.D. … well not quite.

These figures however, are patently wrong. This is partly due to the difficulty of getting propper statistics out of these companies, and partly cause of something I spilled on the napkin while working them out. Then again my estimates here are in stark contrast to those of the original article which says they’re both about a million. I know my figures aren’t exactly right, but I suspect neither of us has the whole picture. My best guess is that his stats are also wrong. Given that activeworlds still services their population with “one UniServer” while LL bought online “over 9,000 server CPUs” this year alone, I’d say my figures are closer to the truth.

This is not to poke fun at Mr Van Der Wal, simply to alert anyone who cares to read it that virtual worlds figures aint what they put in the PR. There’s plenty of very poor press going around with very silly numbers in it, both from the news makers, and the virtual worlds folk themselves. Given the poor acuracy of the numbers in the press you’ll often find reputable sources getting the stats very wrong indeed.

I encourage anyone following virtual worlds for professional purposes, and SecondLife in particular, to keep an eye on Mis Tateru Nino’s posts in Massively. She can be relied on to cut through the lard when it comes to silly numbers.

Tags: , ,



  1. Digado | Mapping the Metaverse said,

    I agree on all points – But I’m using VW’s differently then you are as explained in my article. You give another intrepretation to my intrepetation of Gartners statement – which is fine, but we are trying to prove two different points.

    Activeworlds may have been off by a lot on active numbers – and this not an excuse – but i went with the best numbers that where available to work Gartners Thesis. I think my SL estimate is pretty accurate to be honest, close to 0.5/1 million of active users at this moment.

    The point was to make a rough estimate of the percentage of the internet population ‘actively involved’ in VW’s, this is taking benchmark numbers to see growth, and wether Gartners claim (80% by 2011) is achievable. The numbers are rough estimates, correct, but they help with indicating the growth of virtual worlds none the less when taken cosistantly (ie, meaning i will use the same formula next benchmark moment).

    I appriciate all the input on the numbers I use though, and thank you for the feedback – it will help get those benchmark percentages to show ‘real’ numbers but fact is even when we finetune like this, but there is no way we can get the active users of virtual worlds entirely accurate.

    – Rick

  2. pavig said,

    aha! you used the Gartner figures. Now it makes sense. I haven’t seen reasonable virtual world figures out of any of the major consultancies yet – which is somewhat disapointing. At the time the Gartner report was published for ITxpo 2007, it is true that LL statistics were a point of contention in the community.

    All the VW’s were publishing signups (which don’t represent active use) and the ones that were publishing regular figures were making a mess of it. Linden Lab was at the start of their major hype cycle and copped a lot of flak for messy numbers, so refined their reporting to it’s current state. Most of the other VW’s were not so forthcoming.

    Even so, at the time Gartner should have realized even from the publicly given figures by activeworlds inc that their estimates on that were way out. At the time there may have been about that many AW registrations (though I doubt it as they’ve always been a bit vague and weasely with their public stats), but LL had gone from 250,000 to over 2 million in a year with no sign of slowing in the growth. If they sat on the report for a week under those conditions it would be out of date, which they surely would have known had they not simplified down to a static top ten.

    From memory many consultancies were comparing VW’s at that time based on a misunderstanding of how the “number of users” should be calculated to provide a meaningful figure. For a VW in it’s initial growth spurt, the signups figure is the easiest to sell, but to a virtual world with any significant population, active users becomes the stat to watch.

    Some worlds churn very rapidly, especially content related ones such as the branded worlds – I can’t imagine becoming a permanent resident of coke world or a bank VW. This doesn’t mean high churn VW’s are a failure, just that the churn rate should match the application. The last half of 07 was characterized by think tanks missing out on some of the basic elements of the VW equation, and from the looks of it Gartners predictions inherited some of that shaky intel.

    This is not to say that their prediction is wrong either, VW uptake is exceedingly rapid. In South East Asia the uptake is much more rapid than the U.S. and Europe – they’re ahead of us by a year or so. Predictions that ignore a market which is ahead of your own can’t be complete. Based on Gartners figures (with a bit of a reality check on what those figures really mean) I’d say their figure is ambitious, and likely to plateau earlier. Looking at adoption in Asia however it might not be too far off the mark.

    Hopefuly the metrics for virtual worlds will start to become more meaningful now that there is a meaningful market to track. Investment into VW’s now the bubble has broken should start to be saner and more performance based. Investors need to know how many folk are in them and how long they’ll be around so they can choose the VR that suits their needs.

  3. darkhotte/gohtrocker said,


  4. Northwards said,

    Somehow i missed the point. Probably lost in translation 🙂 Anyway … nice blog to visit.

    cheers, Northwards.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: