Is SL's Growth Rate Being Choked By Its Own Growth Rate?

Why is Second Life still experiencing a plateau of active users?  Morris Vig of Second Arts has a theory, elegantly summed up in a single sentence:

The technical quality of the user experience in Second Life ™ is inversely proportional to the social quality of the user experience.

In other words, the best SL experiences are social, but the more social they become-- say, over 40 avatars in a single contiguous space-- the more laggy and generally frustrating they become.  "Second Life will not meaningfully grow," Mr. Vig concludes, "until they reverse the technology-based fundamentals that underlie my theory."

Is proximate concurrency as important as he thinks?  I'm not entirely sure, especially when you look at virtual worlds with far larger growth rates, which offer contradictory lessons.  In Habbo Hotel, it's possible to have hundreds of avatars in the same space (see screenshot above right), and that seems to be a significant factor to its massive popularity.  By contrast, Gaia Online also has millions of active users, but the CEO once told me that typically, just 20 or so avatars share a single space. 

So, where do you stand?  Is SL more likely to grow, when it becomes possible for hundreds or even thousands of avatars to inhabit the same region?  (And be sure to check the full post at Second Arts.)

Comments

[Life]

Sure, it will turn SL in a real virtual WORLD!

[Ann Otoole]

Agreed. Linden Lab needs to determine why more than 10 avatars in a region renders it unusable.

Used to be we could have 80 avatars in a region.

Linden Lab needs to remove whatever it is, even if it is windlight, that is causing this issue. Sunsets are nice but they are rather unimportant to social systems.

And a sim with 10 avatars with no attachments (ala hair fair) is as laggy as when the avatars have prim attachments so it is server side and has exactly zero to do with the so-called "ARC".

Linden lab needs to bring in external professional consultants to evaluate the code and personnel involved and determine a path to a correction that does not contain any personal bias.

Business and customers come first. If that is an issue then whoever thinks it is needs to reevaluate their life priorities in a abysmal job market.

[Alberik Rotaru]

Linden Lab also needs to think very seriously about system requirements. It was one thing to demand rather high system requirements before the economy started lagging like a sim with 50 avatars. It is quite another when new computers can no longer be everyone's priority, especially when a machine will run everything except SL well. At NPIRL, Bettina suggests the crash will lead to a boom in SL use because its much cheaper to socialise and entertain yourself in SL than RL. That idea fails if you need high end graphics to get a reasonable experience.

[Osprey Therian]

I'm more interested in seeing the things individuals and small groups do, and seeing the success of niche marketing. Right now anyone can compete; there's a lot of variation in things because no big players are swooping in and seizing most of the possible audience.

I don't consider an event better because it's bigger. Sometimes that would be useful, but we'd have to give up our village structure to get it, and I'm not sure the consequences of that would be so wonderful that we should wish for it to happen sooner.

[Sophrosyne Stenvaag]

I don't see any reason to believe that the lack of mega-events is holding SL's growth back.

I suspect that the vast amount of time that Residents *want* to spend in SL is in small groups - couples to a half-dozen friends, with another statistical bump at about 25 or so.

At 25, a club is full enough to seem fun, a discussion event small enough that everybody who wants to talk can. Past 40, you're not really interacting with everybody anyway, lag or no lag.

We certainly haven't noticed the lag issues Ann O'Toole mentions: we can get 50-85 high-ARC people at events in Extropia Core with negligible lag - but people have the most fun at our 25-40 person events.

Just for us in Extropia, if we had the capacity to run 500-1000 person events, we wouldn't: SL works best at the human scale, the village structure, as Osprey calls it, or Dunbar's Number.

Events at the 25-75 person scale are democratic, social, networked. Past that, and you're back into a broadcast model with a few producers and many consumers.

The whole point of SL is to enable us all to create and to share - you want broadcast, turn on your television.

[Ian Lamont]

Back in May of 2007, the former LL/SL CTO said "69% of the development staff at LL are currently on scaling and stability." Despite these efforts, the situation has deteriorated, at least in terms of scaling. Can it be reversed, while growth slowly continues?

Ian Lamont
The Industry Standard

[Morris Vig]

Thanks for the mention, Hamlet. There's a lot more behind this theory - from the way we market the Second Life experience to each other, all the way to how Linden Lab would likely be marketing Second Life to corporate clients.

When I get a little more time, I'll be devoting some effort to suggesting how my theory plays in a host of ways...none of which, I can see, bode well for growth patterns for Second Life until lag is brought under control.

[Samantha Poindexter]

"Past 40, you're not really interacting with everybody anyway, lag or no lag."

...you know, I'd really like to find out whether this holds water. It's hard to say when all we've ever known in that situation is lag.

I tend to agree with Morris's thesis.

[Samantha Poindexter]

To elaborate a bit, there are also different kinds of interaction beyond conversation. I have a wardrobe full of excellent outfits. I would like to be seen in them, and I would like to see others in their own finery. When every event with a decent number of people involves spending much of one's time watching particle clouds coalesce into grey people who gradually gain color, much of the fun gets taken out of that...

[Alberik Rotaru]

Density drives the random encounter. Small numbers, whether Ann's 10 or Sophrosyne's 25, are not 'democratic, networked' they are elite and closed.

The Dunbar number is 150, not 25-75. It is contested and other researchers have proposed 231. The Dunbar number is not prescriptive. The Dunbar number describes describes relationships, not an optimum size for social events.

The very words in 'democratic, networked' contradict each other. The 'networked' event is, by its definition, one where you gain access by who you know and that is not a democratic process. It is perhaps, also not a happy process for new entrants to SL as they wander though a string of unpopulated sims.

[Troy McConaghy]

The hypothesis that "The technical quality of the user experience in Second Life is inversely proportional to the social quality of the user experience." sounds nice because it's *testable*.

How do you measure "technical quality"? How do you measure "social quality"?

Did Morris Vig gather any data? Did he check if his data bolsters or refutes his hypothesis?

Or did he just do an Aristotle, and say something is true because it sounds reasonable? Then people start throwing in anecdotes and other logical fallacies and before you know it, it's "common knowledge."

[Morris Vig]

@Troy McConaghy: OK, you nailed me. I'm a liberal arts grad, so I went the Aristotle route.

Considering the scientific/academic nature of Second Life's userbase, I'm surprised that this didn't come up earlier.

You're welcome to disprove my notion. It sure would make for a fascinating set of experiments!

[Morris Vig]

One further clarification, Hamlet...you ask, "So, where do you stand? Is SL more likely to grow, when it becomes possible for hundreds or even thousands of avatars to inhabit the same region?" That's not my point.

My point is that even under current circumstances (40 users/sim max on mainland, throttling up to 100-ish on islands), the lag and associated technical shortcomings make the SL social experience less than ideal - definitely less satisfying than other virtual spaces.

So I'm not even making your leap of thinking that you need greater sim concurrency. Just make SL's biggest social functions tolerable for the users!

[Hamlet Au]

Noted, Morris, that was my own inference, I didn't mean to attribute it to you. So you're arguing for better experience quality at existing sim concurrency levels, i.e. 40-100?

[Morris Vig]

I'm simply arguing that the social experience is related to the quality of the technical experience.

Let's see - in other words, the most amazing social/cultural experience in Second Life will be directly affected by the infrastructure. Attend a Komuso show, or reflect on Burning Life. Tell the that lag in these instances didn't make you enjoy the experience less than you would have?

Now compare this to other virtual world engines where technical glitches aren't so pervasive.

If you're a new resident, used to a quality of social AND technical experience like those worlds, I'm guessing you will find SL frustrating and less likely to hang around. Thus, LL's effort to grow its userbase is self-limiting because it can't deliver the equivalent experience of today's modern games.

[Two Worlds]

YES. THIS. This is what I keep saying, and many other residents and former residents are saying!

If there is ONE THING you and your Linden-esque syncophants take away from the naysayers, Hamlet, if Second Life is truly to be this great massive virtual world revolution empowering thing that you like to claim it to be, the learning and technology curve need to appeal to the MINIMUM, not the maximum.

Windlight and all that is great for PR and visual appeal, but when I fear for my 2-year old laptop bursting on fire from overheating everytime I'm in-world for anymore than 20 minutes...I'm not getting anything out of Second Life. It's not some great immersive experience. It's just a waste of hard drive space.

You may not understand this, Hamlet, you and your bandwagon-jumping Web Bubble 2.0 technorati pals, but there are people out there who DON'T have the latest six-month old or less bleeding edge technology. That's why stuff like World of Warcraft and Habbo Hotel are so wildly popular and Second Life isn't. With WoW or Habbo, it can be either a casual thing, or a more involved thing. You can pop in a CD and go, whether you're wanting to be a casual gamer, or poopsock it looking for boots of +2 virginity or whatever. It doesn't have to be this huge process where you have to take out a freakin' loan just to get the latest in graphics technology, just so you can see some freakin' clouds on a beach or something. Second Life needs to be democratically empowering, not something for the technologically elite.

I'm sorry to Prok out on you like this, but you NEED to get this through your thick little ivory tower skull. I know you like to pretend you have the ear of the Lindens, and if I had a dime for every fluff piece you wrote for them I could freakin' buy and sell the Goddang thing...so get this out. If you want Second Life to be successful, they (the Lindens) need to accomodate the user's needs...not the other way around.

[Melissa Yeuxdoux]

I must admit, I did build a new computer in March, mostly because of Second Life... but it is far from bleeding-edge technology. I regularly run with the graphics knobs cranked to the maximum, using an nVidia 8600 card: PCI Express 1.0, not 2.0, and sneaking up on a year and a half old technology. Looking on newegg.com shows 8600 cards going for between $50 and $100, most hovering around $80. I can see the clouds just fine.

SL supports objects made by users of wildly varying levels of expertise, but it's compared with games where far more can be precalculated and prerendered and with content made by professionals. Other virtual worlds are breathing down LL's neck--take a look at the Nurien demo videos, and compare them with an SL fashion show.

LL desperately needs to improve performance, yes, but the bottleneck now lies with the asset server--just watch objects and people dribble into existence when you TP or first sign on, no matter how good your graphics are--and with the fixed assignment of CPU cycles to areas of virtual land, so that crowded regions grind to a halt while empty sims twiddle their thumbs.