Second Life: Towards Consolidation and Cloud Computing
A recent discussion on an article on PandoDaily explaining that Second Life “did not fail” provided many insightful comments on the current level of expectations by some of SL’s residents.
One of them has been on the SLogosphere for quite a while — a recurring meme, being repeated by all doomsayers, since it became quite clear that nobody can deny any longer that the Second Life Grid landmass is shrinking. And that means less income for Linden Lab. For a while, they will remain profitable, but there will be a tipping point. Perhaps well before that there might be a generalised panic; think of the following question: “When will Second Life become so small that you will make the decision to pack and go?”
It’s a worthwhile question to ask. Since I was used to a Second Life with less than a thousand regions and perhaps 5,000 residents, the “size” doesn’t matter so much for me. But different people have completely different expectations. So it’s reasonable to ask this question.
Now, some commenters on that are article might be mislead by the idea that “more people” means “more transactions”, and that “less land” means “less transactions”. As Catherine Fitzpatrick (handle: catfitz on Livefyre and elsewhere), much better known as Prokofy Neva, states, it was quite clever of Second Life’s creators to implement IP protection models and the ability to sell digital content safely and securely. But this should give people a hint why the “number of users” and the “landmass” are not directly related to “sales”. In fact, it’s all the other things that catfitz/Prok has mentioned — establishing relationships, the large number of participants in certain groups, and so forth, as well as a sense of confidence that the digital content economy actually works, that provides more sales (and no, it’s not the Marketplace).
What happens is that we’re watching consolidation happening: successful businesses are absorbing failing ones. The latter drop their land, so landmass shrinks. But the successful ones lose competition — or, by absorbing smaller (failing) ones, actually grow their content sales — so the result is more transactions. If we jump to a big brand shop in SL these days, we’ll see vendors from many different content creators, all under the same roof, brand, and even style, and sometimes we can see signs announcing that the brand owner is interested in “joining forces” with more designers. So we’re slowly moving away from the “single designer”-“single brand”-“single island” model (content creators have huge egos, and this is what everybody wanted: fame in SL!) and going towards the “single brand”-“multiple designers”-“possibly multiple islands” (but less than before!) model, where franchising a brand is now much more frequent than before.
This has a lot of implications. For the designers, it means deflating their ego a bit, in return of being part of a larger business which is more successful; but it also means less competition (and thus more sales). It means reduced costs: instead of being “forced” to buy their own islands and do their own promotion, now they rely on an overarching brand which promotes the shop and the brand first, and the designers next, but at a larger scale that the individual designers managed to attract. This, in turn, means more customers, more brand awareness, more sales.
But for Linden Lab it means less islands, and, as such, a smaller revenue — because, except for a small commission on the LindeX when money flows out of the economy, Linden Lab doesn’t directly benefit from a larger GNP. On one of my recent articles, at the bottom, I joined the suggestion made by some residents (which I first heard through Scarp Godenot) that LL could apply a sales tax to start benefitting directly from more overall sales, even if the landmass shrinks. My own suggestion is a bit more complex to allow for the psychological impact of taxes and I propose something akin to a “tax refund” to compensate for the effects… but I’m not going to repeat myself.
So, consolidation is the new trend (not really new, but perhaps more visible these days) which allows content creators to cut costs and increase sales. Successful businesses in Second Life are not stupid: they wish to cut down on unnecessary costs if they can. If three designers, each with their own island, barely manage to turn out a profit, instead create a joint brand and move to a single island, they can keep up the sales but drop the costs of two islands. With part of that money they can spend in more promotion mechanisms and even increase sales further. They can even lower the prices, become more competitive that way, attract even more customers, and expand their market share at the cost of kicking smaller content creators out of business (or absorb them under their new brand). So, overall, everybody wins — except a few content creators who utterly refuse to “join forces” with anyone else, and, of course, Linden Lab, who see a lower income from less islands. The side effect, of course, is a reducing landmass, which has a large psychological effect…
This, in fact, replicates exactly what happens in real life, where we moved from Mom & Dad neighbourhood shops to huge retail shops and megamalls. The principle is exactly the same. And this ties neatly into catfitz’s last paragraph: why don’t people study this effect further? Well, it’s not true it’s not being studied — it is, but it could be more! — but let me quote (by heart) what Castronova once said, when he was asked why he didn’t study Second Life as deeply as he studied other virtual economies: it’s just too similar to real life to be “interesting”.
So… what will the future hold? The psychological effect on having a landmass shrinking too fast might prompt content creators with solid businesses to fear that, all of a sudden, everybody panics and runs away, thus cutting their sales to zero, thus they might “jump ship” as well before it sinks. But first they will continue to consolidate more and more. Now Linden Lab has a problem. Their current architecture implies high running costs, and they cannot lower tier prices very easily (which would allow not-so-successful businesses to remain in SL, and, through that, pay for more tier, instead of giving up or being absorbed). They can obviously redesign the infrastructure, and, as I’ve pointed out on another comment on PandoDaily, I believe that they’re feeling the psychological effects of a reduced landmass — and reduced income! — to make them accelerate the change, but it won’t be easy. To cut the costs by one third means becoming 30% more efficient with the current hardware and networking infrastructure — and the announcement that they would make a huge investment in higher-density servers might be a strong hint that they are considering that route. But will people be happy with a 30% reduction in tier? I don’t think it’s enough. For someone who is used to spend US$300/month, a “reduction” to US$200/month is not really significant. It’s still a lot of money. Unless they can offer full sims with 15,000 prims (or the equivalent in Land Impact) for as little as US$30/month, they will not be able to revert the trend. But that means cutting infrastructure costs to a tenth — an order of magnitude! Throwing hardware at the problem will never allow them to cut the costs that much. No, it means far more dramatic measures. They’re not impossible to achieve, but they imply a huge redesign — it might mean implementing what I call “blink regions” (regions that are empty are simply offloaded from the active servers, and only when someone teleports to them, they will be instantiated — pretty much what Kitely is doing on their own grid, using Amazon’s cloud paradigm as an economic model on how to do this). Since 90% of SL is empty most of the time, this could work out, cost-wise. Technologically, they have all the tools for that — there are plenty of open source alternatives to deploy “cloud computing”: distributed virtual machines across many servers (but far less than they have nowadays) which are only loaded on demand. Or, as an alternative, they can completely change their business model: move away from tier and go for sales taxes. As said, I’ve alluded to a few suggestions on how this could be accomplished. In fact, both things could be done simultaneously.
So the interesting point of discussing all this is that I’m watching with some expectation what Linden Lab comes up with. For the past year or so, they have tried to ignore the problem and not take it too seriously — who can blame them, since the European Union does the same, and handle much higher figures and many more people? Now has come the time for applying drastic measures. LL has already mentioned that they are, in fact, going for an infrastructure overhaul. But the question to me is if they thing that “reducing slightly the tier prices” (and for me, a 30-50% reduction is simply not enough) will be enough, or if they realised that this cannot go on further, and that they need a completely different business model. Plunging into that is highly risky, but might become very rewarding. But since the risk is so high, what LL has to evaluate is what is more risky: applying small measures in the hope that things will change, and continue to see an eroding land mass and less profits (even though the in-world economy might grow more and more, and reach a billion annually in 2013 — but without LL getting a share of that, what’s in it for them?) until the psychological effect is too big to prevent a panic and a mass exodus — or, well, start experimenting with a completely different model and see what happens?
The two are not mutually exclusive. I would start deploying “blink regions” with a sales tax on them as an option (see my last article for details) and compare the income and costs of operation from that with the rest of the grid…