I’m currently under a lot of pressure to do my work, and every time I post anything on my blog, I get a lot of complaints that I should be working instead of blogging :) The complainers are right, of course… but every once in a while, I also need to rest my mind, and do that by focusing on completely different and unrelated issues. Just to prove that I’m alive and well, and will resume my analysis of Second Life and virtual worlds in general, as soon as my current work is completed (alas, it should take another year…), here are a few thoughts about a recent discussion that captured my attention on the SL Universe forums. My comments make more sense if you take your time reading that discussion first. I tried to post an answer there, but unfortunately I’m too long-winded for common forums; I had to resort to a blog post to overcome the character limitations :)

Thanks to the very intelligent community that regularly discusses at SL Universe. You guys are truly an inspiration for me!

The more time passes, the more I tend to believe a few basic assumptions about SL, some of which, as you will see, are contradictory:

1. Everybody just has a parochial view of SL, not a global one

Put it simply, one’s opinion of SL is basically revolving around one’s navel — we experience SL and form an opinion about it based on where we are, where we go, the people we see and interact with, and what we see. This is unavoidable (and of course the same applies to RL as well!), but, because SL is so extremely complex and has such diverse groups of people, our own ‘opinion’ is always biased and limited to our own world-view. Meanwhile, just across the grid, there might be a completely different group with a radically different view of SL. Both groups, of course, will claim ‘their’ view is the correct one, even if they defend the opposite view…

Here’s a typical example. I’m part of a 9-year-and-something old small community which shares the cost of leasing 5 sims in return for giving everybody the right to vote on all decisions of the community. Outside our community, we hear about horror stories of communities dwindling and disappearing, people reducing rent prices trying desperately to engage new customers and tying the remaining ones to continue to pay rent, and engaging in more and more events to try to ‘stand out’ and attract new customers. In our case, we have few events (there are certainly a few that are regular), our tier fees are above average, the buildings are nothing special (most have been done in the period of 2004-2007… with some remodelations here and there), and we have a huge waiting list — as soon as a new sim goes online (which takes a year of discussions!), it tends to get sold out in a matter of days. The community makes a profit, and has done so in the past 9 years (with just a few months in the red, when whole sims were closed down for rebuilding).

So, someone from that community, if they never travelled outside it, would be very surprised to learn that ‘SL is dying out’ or that land barons are getting more and more desperate to attract customers. By contrast, anyone having been part of a succeeding sequence of failing rentals, being bumped from one to another as their owners go bankrupt and close shop, would never believe that anyone in SL could actually make a running profit from high rent tiers, few events, and outdated content, in a community where everybody is always complaining about the lack of parcels for rent…

These are edge cases — you cannot form an overall opinion of SL just by visiting edge cases. The problem is that most of us will form our opinions exactly by experiencing the edge case with which we are more familiar. And that means that this experience biases our opinion, and we have a hard time believing it can be different elsewhere on the grid. This is especially true if we happen to travel always to the same locations when we wish to spend quality time in SL, and are so familiar with those locations that we start to believe they’re representative of the whole of SL.

Now, from our perspective as residents and customers of LL, this is not really a problem. The problem, I think, is mostly at the LL management level. As we all know, the top brass at LL doesn’t spend enough time in SL, and, when they do, they tend to a) wander around (which, like a RL tourist, will give you completely different and probably wrong idea of your surroundings); or b) stick to certain communities and areas (which obviously will give them a good idea of how that community/area works, but gives them no means to extrapolate to the rest of the grid).

This is really a serious issue when considering that LL does not merely run the show, they’re supposed to do some governance as well. And to do that, you need to know how your customers think and what they need. The problem is that LL has few channels allowing them to do that. The only ‘referendum’ they have is this occasional questionnaire that you get (or perhaps not any more) asking: ‘Is your SL experience better this month than last month?’ This is so subjective that the reply is mostly meaningless (for making business decisions). But any more complex questionnaire will only be answered by a tiny group of people. A post in any forum (LL’s, SLU, wherever) will only gather replies from a fez dozens, perhaps a few hundreds, but that’s just a tiny part of the resident population — and generally the one that is most active and vocal. As I tend to say, the residents who love SL and enjoy it so much have no time to participate in surveys, questionnaires, and long-winded forum debates. They’re just having fun in SL. They care little about information-gathering for management purposes. In fact, they wish to be left alone and in peace. And perhaps they are exactly the vast majority of residents — the ones logging in every day, paying rent and tier, being immensely happy with SL, but completely silent about their opinions. They don’t blog, don’t go to forums, don’t go to meetings.

My point is that anyone who complains too much is not going to stay around in SL for long — but they might be very vocal about it. The happy people don’t have time or patience to answer questions about their ‘SL experience’. So any information that LL retrieves from the resident population is so biased and so out of touch with reality (or should I say ‘virtuality’?) to be completely meaningless and hopeless to make good management decisions. That’s why LL is (in)famous for making bad management decisions — they have so little partial data, most of which wrong, that it’s impossible to make any good decisions. They have to rely on the most fickle tool of business management — intuition. And that’s very hard to have about SL unless you’re completely immersed in the environment (which they are not), and, as said earlier, that will still just give a parochial view of SL, not a global one.

The only area where LL does, indeed, have some data, is on the technical side of things — because they track down viewers’ statistics. The only thing they can be sure of is that over the years people have bought better computers and upgraded graphic cards, and have a good technical experience with SL. They can profile data about people’s movements and know, on average, what places are more visited (traffic data). But they have no clue about what people think about SL and how they interact with each other and with the environment.

They don’t even know how many of their residents are actually ‘residents’. That’s another point worth mentioning separately.

2. We’re clueless about sign-ups and retention data or how hard the UI is

Three managements have gone through the exec board in LL, and the fourth one is in place, and nobody, in 10 years, has been able to ‘fix’ the issue about user retention. We all see the statistics — 10.000 residents or so register per day, and just a handful of new residents actually remain around. We have all blamed the impossible-to-use UI and the difficulties of understanding the SL culture, as so many of you have reported on the SL Universe forums (and aye, even after 9 years, I’m still quite clueless about the intrinsic difficulties in buying land as well — that’s why I just rent, too, even though I also have a Linden Home somewhere…). All past CEOs have promised to do something about it, all of them have failed, some of them have invested massive amounts of money and resources into fixing that issue, but nothing has worked.

Why? Last year, Pussycat Catnap has proposed a theory that new signups are not residents at all. Instead, they are merely forum spambots. Anyone who has run a popular forum (like SLU!) knows the issue: you get thousands of fake registrations every day, many for spammers which look legit and will remain ‘dormant’ for weeks, months, or even years, until they suddenly get triggered to spew out spamming messages on all threads. Forum spammers are incredibly sophisticated and target popular platforms (like LL’s Lithium — anyone who can afford Lithium will have hundreds of thousands of users to justify the purchase cost of that community platform). The more popular a forum/community is, the more likely it will attract forum spammers (it pays off). No wonder, though, that SL gets half a million of fake registrations every year from forum spambots (ask any forum admin about how many they get every day, you’ll see what I mean!). These are not alts. They’re not even ‘SL bots’. They most certainly are not clueless people trying to figure out SL. They’re not people who log in thinking that SL is a web-based virtual world and give up after being offered the choice to download a viewer. They are, in fact, not people, but merely automated attempts of getting a valid username to log in to the community forums to spam them.

So LL apparently has been wasting years of money and resources to fix a problem that doesn’t really exist. Instead, all they need to do is to create a better ‘human validation’ system, and not count any forum spambot attempt to create a username as being a ‘registration’. Poof, problem disappears. However, as a side-effect, LL would now have to explain why they only get a handful of new registrations per day from real humans; maybe they prefer to give the idea that there is a user retention issue which they are attempting to solve than to explain that their only ‘problem’ is lack of interest from humans in SL…

Of course there will still be humans who will find SL difficult to use and give up quickly enough, but it’s totally different to address a problem that affects, say, 10 or 20 new registrations per day, than to imagine there is a huge problem affecting 9.990 new users per day.

3. It’s worth repeating: SL is a niche market. But what niche is it?

LL recently seems to grudgingly admit that they are in a niche market and that expanding it to the mainstream, as they hoped to do in 2006/7, is probably impossible (I would certainly agree with that). Companies catering to niche markets are not necessarily a ‘bad thing’. In fact, I have often put forward the hypothesis that a successfully explored niche market is very lucrative, and I tend to point out the classic examples of Harley-Davidson or even Apple, but also many examples from the software industry, like Oracle or Autodesk. These companies cater for markets with hundreds of thousands or even millions of users, but not to billions. They are not mainstream and deliberately wish not to be so.

LL is definitely lucrative. It doesn’t make billions of dollars per year, but half a hundred millions per year is not bad. It’s just that there is this crazy idea that any ‘Internet company’ these days has to be mainstream to succeed, when the opposite is actually true: few ‘mainstream’ Internet companies actually become long-lived (although the few that do are certainly multi-billion-dollar companies), while the ones addressing niche markets are long-lived and stay around for decades. Sometimes you can move from niche markets to the mainstream, as we are currently seeing with WordPress (which already drives over 20% of all websites in the world, from being little more than ‘just another blogging platform’ a decade ago). But these are exceptions: they make nice headlines in business magazines, but they’re most certainly not the whole story.

It’s fine to operate in a niche market, drive away all competition in that niche, and have effectively 99% market share. That’s pretty much what LL has. As I have often predicted in the past (and been always laughed at), no other virtual world environment will be able to make a dent on LL’s market, no matter how big the brand and how large the investment is. If Google’s Lively failed, why should Cloud Party succeed? The key issue here is ‘business model’. Nobody wishes to emulate LL’s business model — because everybody says is inherently flawed (as we can see on the SLU thread linked above!) — but all the approaches that appeared since 2007 have failed to produce a valid business model that is long-lived. All relied on venture capital to remain afloat. Venture capital usually lasts for 2-3 years without a business model, and then the company shuts down, gets sold, or the programmers simply move elsewhere. We have now a track record of a decade of commercial virtual worlds and enough examples to present a solid case on how virtual world companies ultimately will all fail unless they have a strong business model. None so far have been able to do create an alternative to LL’s Second Life.

There is perhaps one exception to this rule, and that’s why I think that LL just has 99% of the market share: OpenSimulator. The reason for that is simply because OpenSimulator has exactly the same business model — and practically the same features — as Second Life. So it’s like ‘competing with yourself’, for the same market, same features, same business model — just slightly different tech and lower prices (for a while). The reason why most independent grids don’t grow much is simply that it’s easy to keep prices low while you’re a small company, but, as one’s grid grows, company owners suddenly realize that their costs grow exponentially, and come closer to LL’s — which forces them to go up with their prices while essentially providing the same service as LL. There comes a point where the costs of providing region space on an OpenSimulator-based grid are as high as LL’s, and then there is no point for anyone to stay there (except intense hate of LL!). This comes rather soon: because LL has so many servers, they can negotiate hardware and bandwidth costs, while start-up OpenSimulator grids cannot. There is an exception called Kitely, which has a completely different business model, one that allows the company to continue to be profitable even with just a single customer, and which can grow exponentially without breaking down… or contract dramatically without being forced to close shop… so I’m watching their progress with some interest. Still, they’re at the very fringe of this niche market — they address the people who love SL but hate LL (or LL’s prices), and there are not so many out there as everybody expected.

Because, well, the virtual world market is saturated, and it has been so for years and years. What this means is that there are simply no more people interested in virtual worlds with user-generated content and visual contiguity (as opposed to 3D chatrooms like IMVU). We — that is, SL and OpenSim — have already attracted everybody that is interested in those. There are no more!

(Obviously this is not entirely true: there are always new people who have suddenly found out about SL/OpenSim, being completely in denial for a whole decade, and who join, being astonished why they haven’t found it before. But as time goes by, this number will diminish. That’s what ‘market saturation’ means: no more new users).

Why are there so few people interested in virtual worlds?

Answering this question truly requires a 64-million-dollar answer. Or, from my pedantic point of view, I’d say it’s worth some academic studies in sociology, anthropology, and psychology, and some related areas. I subscribed to Pamela Galli‘s suggestion that, while every resident is completely different, and has different goals and expectations about SL, they all share a common characteristic: the love to learn (we call it the spark — I guess we’re reading too many webcomics). This is something that teachers experience every day: a very tiny majority of their students actually has this ‘spark’, this love to learn, and the teacher’s job is to push them to learn what is useful to their studies. Teachers also expect to trigger that ‘spark’ in students. In fact, our whole education system, specially at the high school level (but unfortunately also on higher education), is based on the notion that you can trigger the ‘love to learn’ in anybody, if you just have the right conditions.

Unfortunately this is a completely distorted (and over-optimistic) view of the human species. The ‘love to learn’ is something only a relatively small minority of people actually possess. It might be genetically encoded; it might be triggered by the appropriate environment (this is what educators believe in); it might be the way people get educated by their parents; it might be a combination of all; it might be something else. The truth is, nobody seems to be actively researching what triggers that ‘spark’, and what conditions are necessary to turn someone who hates to learn new things into someone who suddenly realizes that learning things can be fun. The best that science can say at this juncture is: ‘we really don’t know. We hope it’s X or Y, but we cannot be sure’.

I claim that it’s a combination of all factors, some of which are known (education, environment), some that we cannot change (genetics), and some that we have no clue about. Based on that hypothesis, I also claim that the market for virtual worlds with user-generated content and visual contiguity is limited to people who have the ‘spark’, i.e. all those people who love to learn things. Everybody who hates learning new things will never become users of virtual worlds with user-generated content: because, well, there is no other way for people to enjoy such worlds, unless they love to learn how they work (I’m not speaking about ‘learning the tech’ — I’m speaking about learning how to interact with the environment; obviously, out of all those people, some — a minority among minorities — will also love to learn the tech and become content creators).

If my hypothesis is correct, then any attempts to turn SL into a mainstream product will ultimately fail, because there is no way to enjoy SL for an extended period of time without loving to learn new things. My assumptions are based on the trend that technology is taking: the more dumbed-down the tech, the more people use it, because they don’t need to learn much. TV is universal because anyone knows how to press a button and change channels. Tablets and smartphones are popular because you just press a switch to turn it on and press a button to connect to Facebook, which is ridiculously simple to interact with others. There is so little to learn (except for security and privacy issues…) that it can attract billions of people. Interestingly enough, too dumbed-down products or services tend to push people with the ‘spark’ out of those platforms (there is nothing interesting to learn there, so they tend to get easily bored). But this doesn’t affect multi-billion-dollar companies that operate a business where people do not need to learn anything to enjoy themselves — losing a few millions to niche markets is hardly important for them (and, yes, most SL residents also watch TV and have Facebook accounts).

The reverse, however, is not true. If you have a product or service that only appeals to a tiny majority of the world’s population that has the ‘love to learn’ spark, then any tiny little thing that changes expectations will make them leave, and there will be no others to replace them.

What we have seen in the population surge of 2006/7 was that SL was presented as a mainstream product and thus enjoyed popularity for a while — until people figured out that to enjoy SL you really need to spend some time learning things. While, by contrast, watching TV or being on Facebook doesn’t require any time spent in ‘learning’ anything. So all those people went away. SL didn’t ‘die’ in 2007 because the amount of people ‘loving to learn’ is still considerable to make SL stay afloat. The huge success of mesh-based content, after a slow start (which didn’t affect the success, it just made people pause instead of being rushed), and the explosion in content quality, creating a visually different world inside the same platform, shows that it pays off to ‘invest’ in things that require ‘learning': content creators, after a dwindling revenue, suddenly were able to sell oodles of new content (assuming they were able to learn how to do it — but that’s part of the fun we have in SL!). We all had to learn how to dress our avatars again. But because we love to learn new things, the change didn’t affect our decision to stay or to go. Rather, it gave us a new thing to learn and, because of that, made SL so much more interesting.

Of course, mesh becomes more and more commonplace — although obviously mesh content is by far not ‘exhausted’, there are gazillions of new ways to use it (like there are gazillions of ways to find uses for prim-based content) — but the ‘novelty’ value of mesh is not gone. Nevertheless, one hint to LL is that by introducing new features that make residents learn new things they can maintain a population of residents with the ‘spark’ always interested in the platform, no matter if all else fails (e.g. tech issues, high prices, lag, etc.)

Personally — and, again, I know I’m biased — I haven’t seen a better explanation about why SL attracts just a special kind of people, who are very reluctant to go elsewhere (meaning: go back to TV or Facebook), but who are not many and don’t seem to grow in number, no matter what LL does.

What does that mean long-term? The LL ads that some SLU members have posted are targeted to the mainstream, but SL cannot be a mainstream product. Even if LL creates a dumbed-down version of the viewer — one that is so easy to use and understand that even someone in a Central American jungle, not speaking English and never having seen a computer before, can use and operate after 5 minutes — that will not remove the need to learn about the environment. SL’s environment is not ‘tech’. Tech is just one aspect of it. There is a whole load of things to know about the culture, about the interaction, and ‘how things work': learning how you wear a meshed dress, how you need to buy and activate an AO, why it is so hard to align prim lashes onto your eyelids (almost as hard as wearing falsies iRL!), how to use a dance/pose ball, how to tip someone at a venue, how to join a group (and how groups are used for), even how you figure out where people with similar interests to yours hang around — those kinds of things. A simpler viewer cannot address those. Obviously you can still enjoy SL somehow if you’re unwilling to learn anything. But these will be edge cases again. The majority of the mainstream users will not have any ‘fun’ with SL because, at some point, they will realize (very likely unconsciously) that they have to learn something to have fun in SL, and ‘learning’ new things is not what the mainstream users see as ‘fun’.

4. The pricing model is unlikely to change dramatically

Facebook is not making as much money from ads as before, and social media/marketeer magazines and e-zines everywhere are giving out early warning signs that ad revenues on Facebook might be falling, and recommend companies to move elsewhere to get more ‘bang for the buck’ with their ads. Imagine that Facebook’s reply to their eventual decrease of revenue is to change its pricing model, and forcing every Facebook user to pay a monthly fee (even if it’s something as low as $0.50 per month or even per year). What do you think that would happen?

The thread on SL Universe claimed that for SL to survive, the ‘free culture’ had to be abolished — meaning that residents would have to be re-educated to understand that the availability of quality content and events requires someone to shell out money to pay for the costs — ‘There Ain’t No Such Thing As A Free Lunch‘. They proposed that, over time, people would have to learn that the best way to enjoy quality content/events is to learn to share the costs, instead of the current model, where everybody expects everything to be free. While we know that isn’t the case.

The discussion went further to put the blame on LL’s pricing structure: it’s because the price of land ownership is so high that the costs are too high; quality of content adds value, but that’s  on top of everything we already have to pay! The huge financial crisis, which is just slowly being reverted here and there (and not everywhere at the same rate), also meant that not only residents had less available money to spend on content and events, but also that venue owners were affected and had to cut down how much they spent in SL. It didn’t help that LL made absolutely no effort to adjust their prices to a downfalling economy. The result? More people wish for better content and better events at even lower prices (or, basically, free) than before, but much less people are able to afford the financial costs of providing those… especially not for free.

The solution, advanced on those SLU threads, and echoing so many discussions and blogs over the subject, seems to be that LL ought to change the pricing model. It’s not just merely ‘dropping prices’ for whole sims. It’s changing how sims are charged. Suggestions ranged from having more ‘kinds’ of private islands, something between Homesteads and full sims, being able to have some fine-tuning control over its features, e.g.: ‘I wish to buy a 100×150 place without neighbours, 7650 land impact, and enough CPU and bandwidth to accommodate 65 avatars’. The reasoning behind a more flexible pricing structure is that LL, with their ‘one-size-fits-all’ approach, raises the stakes too much, and, in certain cases — probably the majority! — event hosts and content creators might not need the full features of a private island but require much more than a Homestead — or even a plot on the mainland. A comparison has been drawn (once more!) with what is being offered in terms of Web hosting, where customers have often many possible choices, to fit exactly their requirements to their budget, and spend not more than what they need.

And then there were discussed the inherent technical limitations of SL, one of which is of major concern of event hosts: the so limited number of simultaneous avatars in a single region. The good news is that at least this technical challenge can be solved: we know that OpenSimulator can dynamically adjust region size to handle an increasing amount of avatars in the same visual space. Effectively what happens is that you allocate more CPUs to a visual 256×256 region, by splitting it down (internally) into, say, 4 128×128 regions, each having 100 avatars. If more avatars come in, then you split it even further into 16 64×64 regions — i.e. effectively running a visual 256×256 region on 16 CPUs — and that way you can accomodate a thousand avatars. Practical implementations of this have been demonstrated by Intel long ago, and the code is open source and part of every distribution of OpenSimulator. Most grids, however, do not implement it. Second Life’s own simulator software is of course completely different (it’s not even written in the same programming language as OpenSimulator!), but that doesn’t mean that the same approach could be employed by Linden Lab for handling huge masses of avatars — it would be just a question of porting the code and implementing it. It takes time, certainly, but it’s not impossible. Better still, nothing needs to be done on the viewer side — it can handle those thousand-avatars-in-a-region easily, assuming, of course, that the viewer is running on sufficiently powerful new hardware.

The problem is really one of pricing. To provide a 256×256 region for, say, 1000+ avatars, LL would have no choice but to charge people something like $4800/month for such a super-region. There would be no takers. So would the investment in such a technology have any return? I hardly believe that. No, the only way to handle that situation would be by having a dynamic model of pricing, not unlike what the cloud operators do: pay for the resources you consume. If your region is empty 99% of the time, then you just pay 1% of the monthly tier — i.e. just $3/month. Your region becomes popular? You pay more and more. Are there some hours with 1000 avatars? Then you pay for those hours — but only for those hours! — when there is really a massive amount of resources being consumed.

Resources exist to be consumed. And consumed they will be, if not by this generation then by some future. By what right does this forgotten future seek to deny us our birthright? None I say! Let us take what is ours, chew and eat our fill.

— CEO Nwabudike Morgan, ‘The Ethics of Greed’, fictional book quoted in the Sid Meier’s Alpha Centauri game (1999)

However, I side with those that argue that Linden Lab will never change their pricing structure (except for small details here and there), at least not substantially so. The reason is that Linden Lab doesn’t exist in a vacuum. Any price changing means a loss of revenue, because it cannot be proven that there is an elastic effect in price changes. In laymen terms, if LL slashes the prices in 50%, that doesn’t mean that it will automatically get twice as many regions instantly (and that the running costs of having twice as many regions remain the same). Rather, what it means is that a few people will indeed buy more regions (‘Two for the price of one! Get them while they last!’), but they would have to face their stakeholders and explain why suddenly they have reduced their revenue in half — and if they keep the labour costs (they would have to, since they would have more regions to support!), it would mean becoming dramatically unprofitable the year they slash prices. One might argue that a 10% decrease in prices would just affect LL’s profits for a year — but the same argument goes for residents, too. If LL announced a 10% rebate on full regions, how many new full regions would you buy? It’s extremely unlikely that the landmass would grow by 10% if the prices went down by 10%. To see some serious growth in the landmass, the prices would have to go down dramatically. And, in essence, that was what happened with the ‘first generation’ Homesteads. LL had to put a stop to them because, frankly, so many people were dropping their full regions and buying the low-priced ones instead, that LL’s revenues were dropping dramatically.

The lesson LL learned is that if they offer regions that have the visual equivalence of a full region, but are far cheaper (just 25% of the price of a full region!), they will sell like hot pancakes. But the problem is that all those full regions will quickly be ‘converted’ in the lower-priced regions — few, in fact, would remain. That completely ruins LL’s business model. In fact, they’re completely stuck with their pricing structure: reduce the prices, reduce the revenue; raise the prices, and more people will go away, thus reducing their revenue too. They have no margin for any changes!

Forcing everybody to be Premium (like it was in 2003/4) is also not a viable solution — it’s like Facebook starting to charge for having an account. SL’s population of ‘a few millions’ (whatever that means) would be reduced to something like just 100,000 or so (the actual Premium residents). Here we stumble upon another saturated market: inside this niche, which already has so few people, there is an even smaller niche, of those who are willing to pay a monthly fee just to be allowed to log in to SL and have a few extra things (one reason I still keep my Premium account is because I’m still one of the lucky ‘grandfathered’ accounts that get L$500/week, no matter if I log in or not).

Linden Lab could get some extra revenue from Premium accounts if they actually gave them far more features for the money. Right now, Premium accounts get just a handful of ‘privileges': they get a weekly amount of free L$ (I believe that it’s just L$300 these days); they get a Linden Home (useless, as most people will prefer to pick their own place); and technically they get ‘priority’ access, i.e., when the grid is ‘almost down’, all non-Premium residents are supposed to have to wait on a queue to log in, while Premium residents are allowed immediate access. To my recollection, this only happened once. Premium residents also get free tech support, but we all know how useless tech support is. Oh, and sometimes LL sends us a cool gift (some content). But, in truth, most Premium users are just Premium because, well, like Apple fans, we feel that we’re contributing in a small way for LL to be a profitable company for many years, but there is really no incentive for us to be ‘Premium’ users.

There could be some artificial constraints placed on Basic users: for instance, a limit on their inventory, or in the amount of attached items they could wear for an extended period of time. Discounts on the Marketplace, or reduced fees on the LindeX. All these have been proposed and discussed in the past, and, ultimately, discarded by Linden Lab, because most (Basic) residents will strongly complain about LL’s ‘greed’ and claim they’re already paying for a lot of things and contributing to SL in several different ways, and it would be quite unfair (even corporate suicide!) for LL to antagonize all those people who are actively supporting SL. Which, in a sense, it’s true: you don’t need to be Premium to be a top content creator or event host. I even believe that you can be a huge land baron without being Premium.

My preferred suggestion so far has come from Scarp Godenot: taxing all transactions. Every time L$ exchange hands, LL gets, say, 10% of the transaction. Tier could thus be kept as a bare minimum. Scarp’s reasoning comes from the fact that there is half a billion US$ in in-world transactions every year, of which LL effectively sees almost nothing (just a tiny bit from what goes through the LindeX, and another bit from the SL Marketplace). A 10% tax on transactions would give LL US$50 million/year in revenues. That’s not enough, but it would mean that they could set the cost of leasing land to realistic values, e.g. something like US$10-30/month on a single-CPU-region (which would mean a full region) and much less for Homesteads. The landmass would explode, growing to unthinkable proportions, but LL would not operate as a loss: they would get increased revenue from all those new regions, and still keep the 10% tax as a ‘cushion’, ensuring they would still be profitable even with highly reduced prices.

Probably the main two reasons why LL doesn’t use such a model are cultural, and the risk in exchanging those L$ for real US$. While SL is not ‘mostly American’ these days, Americans still make up a huge proportion of its overall population, and the most talented content creators and event hosts are still American. And Americans hate taxes. Europeans, by contrast, would consider 10% taxes ‘low’, since they’re used to much higher taxes on transactions in their own countries. But LL would most definitely not antagonize almost half of its population (and its strongest supporters) by forcing such a measure upon them. The second issue is that Linden Lab would effectively earn, say, US$ 50 million, but they would be Linden dollars, which they would have to exchange at the LindeX to get some revenue — and there is a risk on that. It’s not unfeasible — more than that gets exchanged on the LindeX annually! — but it means a risk. The current pricing structure has no equivalent risk: the only ‘risk’ is people getting tired of paying so much and leaving, but, at the current rate of decrease, LL’s good for another decade.

No, instead, what LL has done under Rod Humble’s direction, was to go the ‘corporate route': expand and diversify. Launch new products. Acquire new companies. Increase their revenue outside of Second Life, without needing to ask for more funding — use part of the profits to invest both in SL, but also outside SL, with the hope that the decreased revenue in SL is compensated by the new products, services, and corporate acquisitions. Unfortunately LL does not provide us with many numbers, but since we know they have retired Versu from their portfolio, I might suspect that the overall results are not (yet) good enough to compensate for the loss of revenue strictly from SL. We have to see yet what Ebbe Altberg will bring to LL as new ideas for increasing their revenue; it’s a bit early to say, but Second Life definitely looks like the proverbial white elephant: too costly too maintain, no easy way to increase revenues, too sensitive to changes, but with a reasonably large user base committed to pay enough to keep Linden Lab profitable for at least another decade (even if the profits are not expected to increase). It’s tough to be a CEO in that position.

When one’s unable to increase profitability by growing the user base or raising prices, there is just one alternative: cut costs. LL did that under Rod Humble, moving part of their infrastructure to Amazon’s cloud service. But they’re still stuck with thousands of aging hardware, as they had a corporate policy to buy everything instead of leasing (thus saving monthly recurring payments). However, I think that these days this strategy is not really profitable in the long-term. A decade ago, a top server with LL’s specs would easily cost them US$5k (yes, I know you could build servers at much lower prices, even back then, but there is a huge difference between a ‘home PC used as a server with DIY off-the-shelf components’ and a corporate-class rack-mountable server) and co-locating that would easily cost $300-500/month. They started to break even when they managed to buy $5k servers with 4 CPUs and rent bandwidth for $300-500/month — but sell 4 regions and get $800 for that (2006 prices!). As servers continued to maintain that price, but add more and more CPUs, LL’s profitability increased. Also, the more servers they have, the more leverage they have with their co-location facilities, and get reduced prices both for hardware and monthly bandwidth, so the more profit they make.

These days, however, thanks to aggressive pricing from cloud computing, the costs of hardware leased on data centres have reduced dramatically. To give you an example, I can get an unmanaged server from, say, 2012, with 4 CPUs, 4 GB of RAM and 500 GB of RAID space — enough to run 4 full regions with, say, 65 avatars on each — for as little as 30 Euros/month. This is ‘bare metal’ — not a cloud service, not a virtual private server, not anything like that. It’s really ‘my’ server, rack-mounted, with genuine Intel CPUs, with specs as good as LL’s own servers. There are no setup fees and I can stop paying whenever I wish. Certainly, using virtual private servers or cloud computing, the prices could be even lower, but at least some OpenSimulator specialists advise that those ‘virtual’ servers are not good for ‘always-on’ solutions like running virtual worlds, but merely ‘good enough’ for Web-based services. Kitely is the exception proving them wrong, and LL’s move to HTTP communications, to be able to use Amazon’s services instead of placing a burden upon their servers, tend to show that, at least in certain circumstances, you can effectively have virtual servers providing a virtual world, if you do things right. In either case, a move away from buying expensive hardware and leasing exactly the same hardware (which can be easily upgraded in terms of service, instead of trying to squeeze the last atom of performance from end-of-shelf-life hardware from 2007 or 2008, like LL does) for a tiny fraction of the cost would mean that LL could remain profitable even with reduced revenue. The question is mostly psychological. An infrastructure costing a tenth of the current overall costs could indeed mean that LL would be able to slash their prices to merely a tenth of what they are today — and still remain as profitable as before, and start to see a huge landmass growth again. The problem here is that while the profit would be the same, or even grow a little bit, overall revenues would not. And LL is proud of being a company showing revenues and a US$ 35-65 million profit. They would abruptly become a company having a tiny revenue but a huge profit. That might not be the right image they wish to pass along to their stakeholders and the media. These days, when Internet companies are measured in numbers, not in profitability, LL might be very reluctant to change the numbers too much.

Still, they could do a few experiments: move Homesteads to leased servers at first, offer new solutions using those leased servers, give a wider range of options, while still keeping the ‘core’ grid on their current infrastructure. They would save some money here and there, enough perhaps to compensate for the loss of revenue and still achieve the same profit levels. To be honest, I don’t think that LL can afford to do more than that.

5. Conclusions

There are plenty of reasons to explain why Second Life — as a platform, as an environment, as a technology, as a business model — is so unique, and why certain things are as they are, and cannot change easily. The whole uniqueness of its users and the business model that allows LL to be a profitable company shows that it’s pretty much impossible to ‘replace’ it, and that’s why all other competitors are doomed to fail. The only exception is perhaps competing with LL with the same technology and the same business model, which is what the OpenSimulator grids are doing — and facing exactly the same problems as LL is facing, albeit at a smaller scale, and having an additional handicap: in 2004, it was fine to have a hundred or two regions with 5,000 users and the content these people can produce, but in 2014, looking at the huge amount of content and people in Second Life, you simply cannot ‘start from scratch’ and expect people to be happy with it. Most of the major commercial OpenSimulator grid operators today have far more users, regions, and content than LL had in 2004/5. The problem it that it’s not enough, when compared with the ‘giant’ that LL is. Consider a small-scale email operator, boasting technology better than Google, Microsoft, Yahoo, or any other, and saying that they have ‘30.000 users’. I remember Gmail when it just had tens of thousands of users; it seemed strange to move to such a ‘small email operator’, compared with the giants Hotmail and Yahoo with their hundreds of millions. Nowadays, if someone wished to enter this highly competitive market, without a huge brand name, they would have no way to compete with all the giants. It’s unthinkable that it might happen. The same applies to Second Life, at a much smaller scale.

It seems to be clear that Second Life ‘cannot attract new users’, but the reasons for that are unclear, and successive directors at Linden Lab have failed to understand why. Most reasons given are related to the complexity of the user interface, the high curve of learning, and the high costs involved in leasing land in SL. I advance the hypothesis (well, with Pamela Galli — she deserves credit too!) that the major reason is that few people, out of the world’s population, has an innate ‘love of learning’, and that residents Second Life (and virtual worlds with user-generated content and visual contiguity in general) require that unique skill to make SL appealing to them. If I’m correct, there is next-to-nothing that LL can do to ‘improve’ their chances of attracting new users; even if we’re speaking to targeting their marketing efforts towards that group of people, they are spread across all layers of society. There are no specific media for people who ‘love to learn’. They have no identity — at least, not a community or corporate identity. They are often found hovering around open source projects, the Wikipedia, artists, and, of course, on the research and academic communities (no wonder that one of the worst decisions by LL was to kick academics out of SL…). ‘Dumbing down’ the UI for them is a waste of time and effort (they will learn pretty much anything); but the high costs of tier most definitely are a hurdle to overcome!

Changing the price structure of Second Life might be ultimately impossible. Not because of any ‘physical’ impossibility: there are certainly ways to reduce the costs of operating the Second Life Grid® dramatically. There are technological solutions that allow a much better usage of existing hardware and bandwidth, and I’m certain the LL’s developers are able to implement them. The biggest hurdle here is Linden Lab’s corporate structure, and its image as a profitable company, which is sensitive to dramatic revenue drops (even if profitability remains high). They might know what solutions to employ — both technically and commercially — but frown upon them, because any price cuts (or cheaper alternatives) will mean a revenue decrease, no matter what, and — most importantly! — even if that revenue decrease might actually make SL grow, both in landmass as well as number of residents and in-world transactions, and ultimately achieve even higher levels of profitability, at the psychological cost of having less revenue, and having to explain that to stakeholders and the media at large.