These days, I have to humbly admit that I cannot keep up with the whole of Second Life. So I just read what others, who still can, write about SL, while remaining at my little corner of the virtual world and do my own things which nobody cares about :)

But there is something I have noticed in the past few months. Like a crescendo finishing at SLCC, Linden Lab has been rolling out a lot of features. Some are very innovative, like meshes. Others are curiosities which we still don’t know what they will be used for, like the new profiles integrated with a “social stream”. Others still, like the Havok upgrade, or the new communication layer that deals with Group Chat, are “improvements” that we have waited for half a decade and that were finally implemented. And then there is what I call “pure research” — like the new user experience — which are “experiments” to see what works and what doesn’t.

I feel I’m back in 2006 again.

Now don’t get me wrong. During the Philip Linden days, LL had been very criticised when they started to roll out too many features and didn’t fix bugs. Then Philip announced a whole year of bug fixing, and no innovation: the end result was that not all bugs were fixed, and the virtual world sort of lagged behind in innovation. My own definition of innovation, in Second Life, means mostly new features that can be used by residents to create completely new things that were not possible before. Gestures were perhaps the most innovative thing back in early 2004 — it created a whole new industry for dances and Animation Overriders. Then we got XML-RPC, which allowed objects to interact with web servers — that gave us GOM (the first US$-to-L$ exchange) and SL Exchange (allegedly the first web-based shopping experience), among trillions of other services. Later we got flexies, creating whole new opportunities for fashion design, and finally sculpties, which gave us the best-looking shoes ever — as well as a lot of new avatar design and much more realistic-looking content. I’m sure I’m skipping a lot of innovative features which changed SL’s landscape, but I wished just to give an example of a few that completely shifted the way we perceive SL.

That was… 2007. After that, we really had little innovation. Sure, the rendering engine got better, and with that, we got much better-looking machinimas. We eventually even got shadows and a new illumination model — more realism, which in turn also encouraged content creators to increase realism (in buildings and avatar accessories). But these are “improvements” — like Havok getting upgraded — and not truly “innovation”.

2007 was perhaps the highest peak of the so-called Golden Age. Ironically, we have grown 5, 6, 7 times after the “peak”. The economy has grown like crazy; the landmass has grown even more; but because concurrency calculation has changed (no more bots!), and some users have actually left SL since 2007, we tend to see SL in 2011 as being much worse off than in 2007. I think that everything that happened in 2006/7 was incredibly premature. And LL struggled hard to deal with infrastructure designed for an exponential growth.

The major reason that exponential growth was so enthusiastic for everyone involved in SL was that demand vastly outnumbered supply, and to such an extreme level, that anyone with a newly registered avatar could call themselves “expert content creators” (or “expert virtual world consultants”) and get away with it — and even make a lot of money. I won’t get back to this point; I’ve written about it thoroughly before. Let me just summarise it that back in 2006/7 pretty much everybody put a lot of expectations in a constantly, exponentially growing virtual world, which, except for infrastructure problems, makes pretty much everybody happy.

Exponential growth gives one a false sense of security. Because there is so much demand and so little offer, low-quality content and services are in huge demand, and while the exponential growth lasts, everybody can get away with it. This, in turn, encouraged more and more people to do exactly the same, and thus foster even more exponential growth. The problem — as we have seen during the dot-com bubble — is what happens when the growth stops being exponential but becomes a much more sensible linear growth.

In that case, what happens is that the exponential revenues tend to hide the fact that the business model is simply broken. Let me give you a simple example. Imagine a land leasing operation during exponential growth. Everybody wants some land, but the supply is limited, so you can charge high, expensive “setup fees” — because there is demand for it. Then the leasee starts paying their monthly fees, but you’re not interested in those, but focus on attracting new clients for the next parcel — charging them high for selling the land again. During exponential growth, since there is so little supply, you can base your whole business in just the high “setup” fees and forget about the monthly rent.

So in effect your business model is based on how many new customers come in. Existing customers are little worth. Under this model, which we had in 2005-2007, buying land was expensive, but tier fees were low — sometimes even forfeited. Linden Lab also did something similar, and some calculations done back then showed that LL was making a whole lot of money out of the silly “setup fees” — specially after they got fully automated and didn’t require manual intervention (which came at a cost) — but couldn’t care much about tier fees. That’s why they so nicely gave discounts.

When you’re basing your whole business on the number of new users that come in every day/month, and that number is growing (which is the case during exponential growth), this means that the right strategy is to get as many people to join every month. But to make it even more attractive, the point is being able to get, say, twice the customers in a month than on the month before. During those years, some tended to see Second Life as something very similar to a pyramid scheme. Well, we know it isn’t, but during highly exponential growth, it certainly looks a lot like one!

What happened in late 2007 is that the exponential growth started to slow down, then was replaced by linear growth. Suddenly everything was upside down! New users were starting to become rare. In business models where you place all the focus on “getting new users” and charge them a lot for content/services in high demand, this means a catastrophe is ahead. Suddenly, the only working business model is the one where customers are retained and pay a monthly fee because they understand the service is worth it.

Something new then happens: competition. In a model with exponential growth, competition is not really important, when demand outstrips supply. But during linear growth, only the best, fastest, or cheapest (or better: all three) will survive. The next stage is one of quality: the ones successfully selling a product or service that is consistently good, quick to be deployed, or comparatively cheap, will have an edge on the others who will have to work more to achieve similar results. We have seen this happening first on land sales: suddenly, all the small-scale land barons simply couldn’t compete. At some point, they preferred to sell their land below cost, because keeping it for longer than a month worked at a loss. As less and less new users came regularly in, the demand for land stagnated, and this means that only well-established land barons could still make a profit. There simply was no market for “newcomers”.

In content creation, the same happened, although with a twist. Demand was so high that anyone with Photoshop could create something quick and still expect to make a profit. Then, with linear growth, and low content prices, consumers just wanted to buy “the best of the best”. Newcomers with little talent were simply shut out of the thriving market; and among the high-end content developers, quality increased to staggering heights — with prices remaining about the same — as designers fought aggressively to get more and more sales. The first years of linear growth brought to us things like special groups or blogs to keep in touch with existing customers.

The “new economy” thus changed the whole way we thought about SL — now it’s a highly competitive market with little margin for errors, dominated by professionals and individuals with high reputation and some sort of business sense. And they all understood where the money is: upselling to existing customers. What this means is that instead of betting on hordes and hordes of newbies flooding into SL, they all focused on their existing customer base and tried to improve their offerings in order to get repeated sales. This is common for pretty much everyone who is still serious about business in SL.

The scales can be tipped, though. When sculpties were first introduced, there was exponential demand for them, since few were able to create them. For a while, content creators with the required knowledge to design 3D models and export them as sculpties throve, as they could have an edge over the competition, and this meant more sales than possible with linear growth. I believe that sculpties were the last trend in SL where content creators enjoyed the benefits of exponential growth again. But it was in a completely different market, one where the amateurs had already been cast out.

We might see that happening again with meshes, but perhaps not with the same impact as when sculpties were first introduced. I have some reports from professional 3D designers that are so incredibly busy with constant demand for meshed content — and meshes are being rolled out officially on the grid as I write this — that they certainly are blessed with unexpected income. This will last for a while, specially in the early stages, when content creators are yet to discover the “killer application” for meshes (I’d say that shoes — and then hair — were the “killer application” for sculpties, until they became widespread everywhere). Then, when all avenues of application for meshes have been exhausted, Linden Lab needs to be innovative again.

Spending almost four years without any true innovation from LL, during a period where exponential growth was replaced by linear growth, hurt the economy very badly. Still, we have not only survived, but content is looking incredibly good these days, and more and more ideas have been put into practice. The result? In spite of the lack of LL-fostered innovation, the economy has grown, even though more and more users leave, and the ones that don’t, tend to use less hours. But nevertheless the introduction of technological innovations will make a lot of difference to the economy, in a positive way. Sooner or later, everybody will be running a SL 2.X or 3.X viewer or derivative, which are able to see the full range of technological innovation that LL has put into it. That’s not limited to meshes: it also means access to the destination guide or the “social stream”. How these will be put to good use remains to be seen, but I’m sure that it will be exploited very successfully by a few. And even some things that have been neglected will very likely attract new businesses. Let’s take Windlight as an example. We had Windlight for eons (does anyone remember the name of the company which developed Windlight and that LL bought to get the technology?), but since the settings were not shareable, there never emerged a serious business around it. But now Windlight can (finally!) be set on a sim by the sim owner. This means that besides content, sim owners will soon wish to want personalised settings to give their sim the proper atmosphere, and this will create a small business opportunity for a few.

As LL is rolling out more and more innovative technology here and there, all this will have a positive impact. Here and there, demand will outstrip supply once more, for a specific product or service making good use of the technology, and, for a limited time, the few embarking on that business will enjoy exponential growth again. The good news is that all of a sudden, Linden Lab under Rod Humble has been aggressively pushing out lots of new features. And I mean lots. I have five LL-compiled viewers installed, each with a different aspect of some new technology they’ve been experimenting with. The rate of innovation and experimentation is now pretty close at what it was in 2007. The good news is that the reports of “lack of innovation” have sort of been dying lately, and even though people will always complain about lags and bad teleports, the number of those reports have also decreased — at least, on the whole SLogosphere. I don’t mean that everybody is “excited” (some certainly are) but it seems rather unusual to be going through a period in time where there isn’t a massive riot against LL and its policies.

Even on the policy side, LL has been careful. To the ‘nymwars,  Rod Humble sort of explained that LL was on the side of privacy and pseudonymity — a solitary stand against giants like Google and Facebook. Google and Facebook will ultimately win, of course — they’re rich; LL is not — but this will only mean that LL will be seen as the Last Bastion of Pseudonymity and attract a few more users. Not many, but a few.