Speculative Splitting

Well, drama aside, I just wanted to refresh a suggestion for the Road Ahead for Linden Lab®. It is deeply buried inside the long list of comments on my recent post. I cannot take credit from the overall concept, because I saw it suggested somewhere else — and sadly forgot to keep a link for it, because I didn’t think it was worth considering at that time.

A couple of years ago, Linden Lab launched their website for the Second Life Grid®, with a different logo and all. When it was launched, some of us were intrigued. It pointed to something quite different from the stance LL had been adopting up to the day that site went live (that Second Life “was a virtual world created by its residents”), namely, pointing out that Second Life (Grid) was a platform for creating virtual worlds, and showing how this technology was being employed by organisations, corporations, and academic institutions for their myriad projects. About that time, the Grid Status blog also went up, where we could see information strictly concerning the technological environment — from the current status of the grid to announcements of technical changes to be launched soon.

This seemed to be tied to a new strategy for “business” in Second Life — real business, that is. A lot of things were launched, from Developer Programmes, to the Gold Solution Provider programme, the closer ties to education, and so forth. The list goes on and on, new Lindens were hired, whole new relationships were forged, and, well, we started to look at Linden Lab as a body with two heads: one was looking at the “virtual work created by its residents”, the other was looking at Second Life as a platform for doing business. The ultimate destination of this strategy seemed to be the licensing of the simulator software (the Second Life Enterprise boxes) and the Second Life Work Marketplace for real-world businesses. Strangely enough, though, the old “Second Life Grid” site seemed to have been dropped in this process, and a new site, Second Life Work, was launched to replace it.

In terms of business strategy, this was accomplished at the same time that LL created a third co-location facility in the East Coast and opened further offices in Amsterdam on March 7, although the latter are, for now, just to handle marketing.

So far, this is just the expected corporate growth. With Mark Kingdon (a.k.a. M Linden) at the helm, Linden Lab now views Second Life as something even vaster than “merely” a “virtual work created by its residents”. It’s still that, of course — it’s LL’s main source of revenue, after all — but it’s so much more. It’s a technology. It’s a platform (in the sense that you can develop your own technology on top of Second Life). It’s a product (thus the brands “Second Life Grid” and “Second Life Enterprise”). It’s a service (like in the SL Work Marketplace or the LindeX). It might even be just a communication protocol for the future intergrid (or metaverse). This meant recruiting quite different people for the ‘Lab, and not limit themselves at merely “running the virtual world”.

Now nobody knows what M Linden thinks in terms of the long-term future (but the ‘Lab, since mid-2009, started — again — to publish roadmaps, something they had stopped to do for a while as it has been hard for them to meet timelines and goals…). But I think that this might all be related to an unified whole, and possibly have a single explanation for all the latest decisions.

There.com, if you remember its history, opened its doors to the public at about the same time as Second Life, and for a year or two, had about the same amount of users, with a difference: there were more paying users in There.com. When SL started to allow free registrations, it grew exponentially from a few dozens of thousands to a couple of million residents in just a single year. This made a difference — it made SL enter the “Hype Era”, where growth was exponential, with all its advantages (in terms of the economy) and disadvantages (in terms of technology unable to handle exponential growth). There.com, by contrast, remained with a strong user base (estimated at around 160,000) which were all paying users, and grew slowly during that time.

There Inc., the company behind the virtual world, went through a subtle change in 2004. They split the virtual world from the technology, and created two new companies to explore the technology. Makena Technologies would handle the virtual world There.com and Forterra Systems Inc. would deal with government and military contracts. This allowed them to create further virtual worlds based on There.com’s technology — the most notable one being MTV’s Virtual Laguna Beach — as well as license the virtual world technology for the US military, among others. I guess this wasn’t very peaceful for the “old” There.com employees. The founder of There.com, Will Harvey, left the company and created IMVU in 2004, which already has 100 million registered users (perhaps five times as much as Second Life) and 6 million monthly active users (about four times as much as Second Life). The co-founder and main designer, Jeffrey Ventrella, went to work for Linden Lab for a while as Senior Designer, and developed “avatar puppeteering“, a technology not yet implemented on the current SL viewers.

Ultimately, however, the decision was sound. There.com, the virtual world, closed its doors recently. However, the technology behind the virtual world continues to be developed (and you can see from the in-world pictures shown on that site how much it has been enhanced since There.com!), and Forterra was actually recently bought by SAIC, a company worth US$10 bio. and quite likely to continue to develop and enhance the virtual world technology (named OLIVE) for government and military simulations.

The lesson that we learn from There.com is that in business sometimes you have to adopt a very detached view on what you wish things to be, and what real assets your company actually has. Against my own predictions, There.com, after all, didn’t have such a strong business model (based on paid subscriptions). But the point is that the virtual world might have disappeared, but the technology didn’t.

Now, Second Life, the virtual world, has a way different business model. It all relies on the in-world economy. If residents are happy, they buy content. If they are willing buyers, there will be content creators to provide that service for a fee. Content has to be deployed on land to be made permanent, so both the buyers need land, as well as the content creators (for their shops). And virtual land is ultimately LL’s core business. Forgetting Premium users for a while, LL draws income from “merely” 32,000 or so islands they lease, which get them a revenue four or five times higher than There.com ever managed at its peak, and possibly four times more than IMVU did in 2009 (in spite of the huge difference in numbers of registered users).

There.com and IMVU are relatively good comparisons — not in the technology itself, but in a certain similarity on what makes people return to those worlds. There.com allowed user-generated content. IMVU even went further with a website showing 4 million items for sale. The more recent batch of social virtual worlds (like Frenzoo) all have user-generated content somewhere. The difference is that There.com’s revenue was mostly from paid subscriptions, while IMVU and almost all other new virtual worlds, like the Facebook games, make their revenue from selling virtual currency (and from ads). This means that they all rely on content creators (and yes, event hosting and community management in SL is also “content creation”; it’s not all about clothes…) to keep the rest of the residents happy, so that they continue to invest in the internal virtual goods economy, which in turn will send some revenue to the company running the virtual world.

Now Second Life is just slightly different because Linden Lab has a different model. They definitely generate revenue from LindeX fees, XStreetSL fees, classifieds, and Message of the Day ads, and even from subscriptions from the Premium residents — but it pales in comparison to what they earn from leasing virtual land. That’s the big difference really. However, like the other companies, they’re too dependent on what the residents do in SL. If they don’t find anything interesting at all to do in-world — they leave. If they leave, they don’t buy content. If they don’t buy content, creators (and community managers) shut their operations down, and stop paying LL for the precious tier fees. Keeping residents happy, however, is very hard — and depends a lot on listening to what they want 😉

A cleverer strategy seemed to be to “go Forterra” — that is, to license the simulation software to military, government, academic institutions, and Fortune 500 companies. These are not dependent on users’ tastes and whims, and just pay a license regularly to keep using LL’s product for their internal use. If LL is generating a revenue from this area or not, we have no way to know — they don’t publish the numbers. Nevertheless, it’s a very interesting area of business: SL Enterprise boxes don’t require LL to provide any infrastructure. They also don’t provide in-world support. Thus the costs are infinitely low — there is just an investment in developing the technology and marketing & sales costs. But these are far, far lower than running the whole grid infrastructure and its support team — both technical and administrative.

So it seems to be a reasonable assumption that, at some point in time — perhaps not yet in 2010; LL still has a few cards to play (introducing meshes; having a plug-in based SL client that would allow client-side scripting; and possible interoperability with OpenSim-based grids) to keep interest in the virtual world high enough to keep the major source of income tied to leasing virtual land — it would make sense to split the whole operation in three big areas. Let’s call them, for the sake of simplicity, Linden Software, Linden Grid, and Linden World. Linden Software would keep the software development team (possibly both the SL client and the simulator software), and they would be a standard software development company, just like Unity 3D. The client would continue to be free and open source; the simulation software would be licensed, under a model not unlike the one used for SL Enterprise. This would be at the beginning the most profitable branch of the ‘Lab: they would sell licenses for 32,000 or so sims to their sister company, Linden World. It’s a steady income 🙂 And, of course, they would sell licenses to everyone, e.g. here is where the current team dealing with SL Enterprise would end up working.

Linden Grid would be the co-location facility manager. They would lease servers and maintain them to run the current Second Life Grid — or any other. They would handle grid maintenance, server updates, and so forth, but give no end-user technical support. They would buy or lease servers, spread them out in the world, and offer co-location facilities for third parties to co-locate their own grids (e.g. SL Enterprise boxes), offering the maintenance services for them as well. Thus anyone could buy a simulator license and host it with Linden Grid to create their own “SL-compatible” grid, but using LL’s software — and announce themselves as a “virtual grid operator”. This is, to an extent, what could be seen as a similar market than what happens with “virtual mobile phone operators”: they just buy mobile capacity from mobile carriers, but under their own brand, without having their own infrastructure.

The third company, Linden World, would be such a “virtual grid operator”, and they would run the virtual world known today as Second Life. They would own no servers, no infrastructure, and no software — they’d lease servers from Linden Grid, and buy licenses from Linden Software. They would offer technical and billing support to the end-users; they would run additional services like the LindeX, XStreetSL, and so forth, which would generate extra income. And, of course, they would establish policies to connect to “their” grid which would be much stricter than the ones we have today: for example, they might run sub-grids only for adult content, others for “family content”, and allow just one single client to log in to their grids (thus allowing things like, say, ads on the SL client for non-paying users). They might be far more aggressive in protecting content on some of the grids. They might have different policies on grids co-located in the US or in Europe, thus allowing them to better manage the kind of use residents wish to give to SL. Through interop all these grids being run by Linden World would possibly allow easy jumping from one grid to the other, but with strict policies (disallowing adult content on “family” grids, for instance). They would share a single currency — in fact, currency-as-a-service might be something developed independently, and licensed for use by third-party operators (meaning you might be able to use L$ on Facebook games… or on OpenSim grids).

Why would this model be “better” than the current one? If, right now, SL suddenly loses a critical mass of residents because of bad policies or business decisions, the entire operation might fail. On the other hand, with the three-company model, this would only affect Linden World. The technology would be continued to be developed by Linden Software. Linden Grid would just try to attract more companies to their hosting service; but even if they failed and disappeared, the applications developed by Linden Software could simply be deployed on other non-LL co-location facilities (and the current teams working for the ‘Lab would just be hired by anyone needing that know-how).

The model is also expandable, of course — i.e. why limit it only to three companies? Specialised Linden sister companies might just run grids for, say, the education sector; or just for business. They might run grids that are not connected to anything at all (in effect, managing a large array of interconnected SL Enterprise boxes, that however are not connected to what we know as Second Life today). They might create new revenue models, like a grid where access and land are for free, but you can only connect with a special client that constantly shows ads. Other grids might have no user-generated content at all, and everything would have to be bought from the grid operator, who would also employ designer teams to constantly put out content for sale. Or they might have a mixed model on a “family” grid, where user-generated content requires pre-approval, like on many competing virtual worlds. Even better, they might play around with all those models and see which ones generate revenue enough to be profitable; if they have to be abandoned, they wouldn’t affect the overall operation, or the overall virtual world…

I see this scenario as the one providing the most flexibility in terms of business operation. To be honest, I don’t know what the current trend is. Historically, companies like IBM or GM wanted to run everything under their umbrella corporation and brand. In the late 1990s, the trend was to split off companies every time a company felt they should explore alternate business models without fearing the impact of one of them failing (that daughter company would just be dropped). The There.com scenario probably shows that this is a sound business decision — the technology survives, even if the virtual world does not. But companies like Microsoft or Google have shown that the “megacorp” scenario is still sound: they constantly invent new business models, and when they fail, the “mother” company is rich and powerful enough to absorb the negative impact of having one of their departments utterly failing. Perhaps Linden Lab thinks that they prefer to go that route and risk everything in the current business model.

My personal problem with this is that conflicting decisions can push to stagnation and discontentment, and when that is directly tied to what paying customers are willing to pay, the risk is too high. Balancing “features” vs. “grid stability”, for instance, is quite tough. We suffered for years without features as grid stability was the focus. By splitting in two companies — one providing features, the other grid stabilty — and having a third company to supply several grids with a mix of those, we could (literally) get the best of both worlds. Users would just pick the grid they liked best.

Print Friendly, PDF & Email