Globalising the mainstream

Meeting at Locus AmoenusSir Tim Berners-Lee’s article on the online version of the Scientific American might just have been shared by less than 7,000 people on Facebook, but that’s possibly the whole point of the article: it makes a plea to keep the Web neutral. A tough call which only appeals to what Prokofy Neva calls “technocommunists”, privacy-conscious netizens, and a few oldtimers who might have debated Tim’s views on the first USENET forums about the emerging Web protocol, in the early 1990s — discussions around the tag and why it wasn’t such a good idea 🙂

The rest of the world is pretty much joining Facebook like crazy, sharing all their personal data, and the trend doesn’t seem to be stopping so soon — even though Facebook grows way faster than the Internet, there is an upper limit of users: the world’s population. Still, it’s fair to say that pretty much everybody who is on the Internet regularly has a Facebook account (or several), even though they might not use it much. And certainly aren’t spending any money on Facebook.

On the other side of the coin we have niche markets. In an interesting article about the role of the mainstream appeal, Robin Harper, formerly VP of Marketing for Linden Lab, explains how the change from a niche market to the mainstream might force companies to develop different strategies. In her new role at Playdom she will most certainly have to deal with a mainstream product — new Playdom games on Facebook attract millions of players just days after being launched. Second Life, by contrast, just gets 300,000 new registrations per month, most of them never returning again. The business models behind both approaches have to be completely different.

There is no “best” choice (e.g. the mainstream is not inherently “better” or even “more valuable” than a niche market), just different strategies to pursue the same goal: profitability. BMW makes far less cars than Ford or Toyota, but that doesn’t mean that BMW is losing money, or that they don’t have an interesting market. Ralph Lauren, Hermès, Dior, Chanel all sell to niche markets, and they’re not bankrupt; they don’t need to follow American Apparel’s or Zara’s lead. Microsoft Windows has 90% of the operating system market, and Nokia still dominates the mobile phone market, but the most valuable company in the world is Apple (well, almost), not Microsoft, nor Nokia. So clearly there is a case to be made pro and against “being mainstream”; it’s definitely not an easy recipe for success, just one of the possible strategies.

How to become global: get journalists!

Facebook is pretty much a brainchild of the dot-com era: get people to match profiles, invite their friends, let the network grow, sell ads on it. So long as the media are happy with the numbers, they will write articles about it. Articles with numbers — in this case, registered users — attract attention. This brings even more users to Facebook, but, more important, it makes economists go ho-hum and venture capitalists, business angels, and investing companies drool. Having completely forgotten why the dot-com bubble burst, they’re eager to jump on anything that attracts the attention of the media, and journalists, being hardly trained economists, just know about one thing: numbers. Since the early 1990s that’s all they care about: how many people use a service.

I cannot blame them entirely. Understanding business models is hard. In the entertainment business — as well as on most businesses — “lots of users” usually means “lots of money”. Just because the Internet is actually very different (most of the Internet use costs little besides bandwidth) is hard to understand. In the so-called real world, we don’t have many examples where “lots of non-paying users” are a relevant metric. A shopping mall announcing hundreds of millions of visitors, but nobody actually buying anything, would be very strange. In some cases, a few examples exist: for instance, on a rally against Government, having millions of people protesting on the streets make a difference. A charity event which is free to attend can feel the importance of showing a huge attendance, even if most people didn’t contribute to the funds; but at least they were present. And on marketing events, attendance can certainly be high but few might actually become buyers; that’s fine, because you can still measure “brand awareness”. Finally, in the advertising business, you’ll be pushing ads to TV or billboards and have over 99.99% of all people seeing the ads pretty much ignoring them; nevertheless, brand awareness is important to measure, and the few that actually will buy the product will compensate for the costs of actually buying the ad space.

This is the closest that we have on the “mainstream” use of the Internet. Spam works because after sending out hundreds of millions of messages for as cheap as US$50 (say, buying the massive-spamming services from Facebook’s part-owners, Russian-based, you might actually get a few people buying your product. Hundreds of millions have to suffer from spamming just for that handful of people willing to spend more than US$50 enlarging some anatomical part of their bodies to have a more, uh, “fulfilling sex life”. Web ads fare much better, and are as cheap as spam, but they rely on the same principle: get millions to watch, and a handful to buy, and it’ll work out.

Mainstream services on the Web still leverage on ads — and profiling, which is far more interesting in terms of financial returns — to provide some income. To “go mainstream” means mostly to deal with ratios of millions of non-paying users to a handful who pay something. Since the Internet has an estimated 2 billion users, there are plenty that pay something; get enough to join your service and pay a little bit, and you’re a financial success.

Sadly, it’s not so easy to create a business model that actually works. In fact, I pretty much believe it’s incredibly hard, because unlike other businesses, non-paying users really cost a lot. Just think about the massive amount of bandwidth and servers that Facebook has to keep to be able to provide ultrafast service to 500 million non-paying users. If all of them just paid one US$ per month for being allowed to log in to Facebook, would probably make 50 times their 2009 revenues. But if people had to pay just a single dollar for Facebook, they wouldn’t join. That’s what ruined most companies during the dot-com bubble, and is ruining a lot of small-scale companies which get millions in funding one year to disappear without a trace the next one. Facebook was just one of the lucky ones.

Contrast that to the usual “good examples”: and eBay. Both took a decade until they became lucrative — by making sure that most of their registered users are actually buyers. Both just take a tiny slice from the transactions on their online marketplaces, so they also need millions of users to be profitable. expanded their services to move away from the one-product-company (always a danger) and might be making some money out of their cloud services — or from selling Kindles. eBay cleverly bought PayPal (which has a quarter of a billion users — and PayPal gets a slice of every transaction) and Skype and diversifies by acquisitions. Obviously many registered users will never buy a single book; and similarly, most eBay users (and definitely most Skype users!) are not spending any money there. Nevertheless, the ratio is much smaller, and there is a direct correlation between their traffic and their revenues: more users mean more transactions — more people shopping, more merchants selling — and that means higher infrastructure costs, sure, but also a larger slice of the transactions. A contraction of the economy doesn’t mean much for the overall impression we have of or eBay — they haven’t published their number of users recently — would just mean they’d save on infrastructure costs and keep being profitable.

While an announcement that suddenly people are flocking out of Facebook by the millions would spell doom for Facebook — like it happened with MySpace (which had, at some point, a quarter of a million users) and its predecessors. That’s the tough problem when being in a market that doesn’t really exist except in the minds of journalists; few read about how much money those hyped-up companies actually make, but just care about how many non-paying users they have.

In 1999, the secret recipe to create a successful dot-com company was very simple: give accounts for free (a departure of earlier models, where “cheap” was the keyword, not “free”) and make sure it’s very easy to tell your friends — specially your friends on the media! — to join the service, and start announcing exponential growth curves. Don’t care about the infrastructure costs; make sure you reach “millions of users” and then put the company for sale. Someone is going to read those articles and make an offer to buy, paying off the venture capitalists or business angels. This actually worked until the services had to be shut down when the new parent company had no way to squeeze money out of their “millions of users”. We didn’t learn anything in the past decade; we’re still doing pretty much the same mistakes when looking at a so-called “successful” company.

But here is the recipe for success in the 21st century: get a service that journalists use, and use all the time. My journalist friends will hate me (even more) when I tell this little secret: journalists spend all their time in front of a computer, and whatever they’re running, will become the “Next Big Thing”. In 1995, it was email. Via email they could contact sources, fellow journalists, and friends — and get contacts that way which might give them a lead. Once they managed to consistently use those leads to write articles which their editors loved, they got encouraged to use email even more, and reach out further. Soon they would be telling fellow journalists about how powerful email can be, and how quickly they could validate sources, send pictures, and get information first — which they would turn into news, far faster than using traditional approaches. Editors loved it, and encouraged them to use email more. So they started writing about email as the fantastic new communication tool of the era; other journalists picked up the news and started using email on their own. And that meant that soon every journalist was using email, and, to justify spending the whole day sitting in front of the computer exchanging emails with friends and colleagues (and now and then some sources, too), they wrote even more articles about the power of email, which in turn got even more journalists to get an address, write more articles about it, and so forth. Mobile phones and the Web came next: the first allowed them to get bits of news even faster than email, no matter where they were; the Web allowed them something even more important: background material. Large news agencies and TV networks routinely employ experts to give background material; so, when, say, NASA launches a new shuttle, journalists would consult with the expert in astrophysics to get a clue of how the Space Shuttle works and why the launch was important. Expert consultants are very expensive, though. A more clever journalist would just launch NASA’s website, get the relevant information from them, and avoid paying a single cent to an expert — and make the editor happy. Soon, the combination of journalist + web became a source of information on all subjects at very low cost; and naturally, journalists used the pretext to browse the Web all the time, hoping to find leads that way — and writing more articles on how great the Web was.

The problem was how to get that information more quickly. When search engines were launched, journalists had finally a way to search a massive wealth of information — for free — and get down to business immediately. With the advent of the dot-com bubble, things became even more interesting: now journalists would have access to the Wikipedia for way more reliable information, get images from Photobucket or Flickr to illustrate their articles, look at videos before writing a story… all that information was literally swamping the Internet, all easy to grab, easy to search and locate, and re-use as part of an article. Overnight, all major media offices became digital — I would be very hard pressed to believe that there is a single major newspaper or TV station in the world today (yes, even in Kenya or India) which doesn’t give their journalists a computer with Internet access (and very likely has their own website, too). The traditional way of doing things is simply too slow for this fast-paced world we live in.

When social media started hitting the Internet, journalists were among the first early adopters. The problem with the Internet is that it has far too much garbage, and little information: the role of the journalist is to pick the gold nuggets out of the swamps of junk. But even Google with its awesome searching abilities is not perfect: single-word searches come up with way too much noise for the few nuggets of precious information. No, the best source for reliable news still are people. Reaching people, however, was hard. With Twitter first, and Facebook next, it became child’s play to be in touch with thousands and thousands of people, and sift through their bits of news. No wonder that the media was swamped with tales of success of all social media networks, promoted by the early adopters who immediately benefited from them: journalists in search of the next breaking news. The “mainstreaming” of social media thus had two major pillars: word-of-mouth growing exponentially as each new user sent invites to a handful of friends, and strong promotion on the media, because the more Facebook and similar tools are used, the more information journalists can get, and the more they can justify writing articles about it… and that means more willing editors interested in publishing those articles. The effect feeds on itself and grows and grows.

And, of course, while journalists wait for some interesting bit of news to pop in on their walls or timelines, they can play FarmVille (also owned by It’s a win-win situation!

A high number of users is nice, but what are they actually doing?

I very humbly admit to be very conservative when evaluating a company. If I cannot figure out the business model, I don’t see a bright future for the company. You can certainly run a company without any business model if you keep slicing parts of it and feeding it to prospective buyers, so long as you can have the media on your side.

As long as people are happy with a mainstream product, and that it has a successful business model, it’s not important if the product/service is “useful” at all. TV is perhaps the best example. Being mainstream means also lowering expectations. MySpace’s success was based on the theory that most people might like to share pictures and videos with friends on a website, but had absolutely no skill to create a web page or even a blog. MySpace allowed them to create it very easily. Facebook just added three important changes to the concept:

  1. Don’t allow visual creativity. Stick to a simple, uniform interface. Stifling creativity means less ugly pages (like MySpace!) and this strategy has proven to be far more wise.
  2. Embed applications, and allow anyone to create applications to embed. This allows “pages”or “profiles” to do things previously unheard of on a social networking website, and it still remains very simple for the mainstream user to use. Nothing could be simpler then playing a game on Facebook!
  3. At least 20% of all users are just interested in one thing: sex. (The remaining 80% are probably lying if they say “no”). True to its origins, Zuckerberg correctly identified that the biggest problem with the dating sites out there is, well, that people lie so much about themselves. By forcing users to create accounts with their real identities, and displaying as much personal information as possible (including GPS-based physical location), it becomes a much more attractive environment for searching for potential sexual partners.

Nowadays, specially on point 3, we like to shrug it off and explain that there is another category that is also interested in “real life” information — besides profiling companies and marketing agencies, of course. It’s the professional user — specially the one that relies on credentials (e.g. diplomas) and not on reputation. Professionals with low self-esteem need to show off their credentials, and a social networking site which encourages presenting as much personal information as possible are welcomed by them. It’s ironical that people like Bill Gates or Steven Jobs never finished their degrees, and their previous work experience was between negligible (definitely not enough to be hired as CEOs for any of the top ten largest companies in the world!) and non-existent. This is not to say that credentials are not important — they are, and they are mandatory for a lot of job applications — but reputation and actual experience are far more important. Nevertheless it cannot hurt to show off your credentials publicly even if you have a good reputation and a lot of experience!

I have no idea how many “professionals” actually use Facebook that way, but I can imagine there are far more than LinkedIn or Plaxo users (where the focus on publishing real credentials is even stronger). Perhaps they’re also 20% of all Facebook users. The rest, well, they are part of a different mindset, which is more deeply disturbing: by publishing their information, they feel they’ve got an audience, and the reality shows on TV show us how eager everybody is to become famous. The Internet gives everybody an ersatz fame, by showing off to how many cool people you’re connected, and by how many interesting videos about pets you share. Everybody likes to show off.

This false sense of “fame” is very perverse. It falls neatly into an overall political strategy where people think they’re empowered, because they are able to post their ideas and opinions online “where the world can see them”. In fact, they’re simply ignored. While everybody produces a continuous flow of data about their opinions all the time, and all of it gets indexed by Google and Bing, the truth is that nobody will even remotely care about what you do or say — and the more date is fed into the endless void of the Internet, the less relevant it becomes. This is an excellent example of perverting the whole “freedom of speech” paradigm — you are free to say whatever you please because nobody will listen to you anyway. And anyone following thousands of “friends” on Facebook will not manage to keep up with the constant stream of meaningless rambling, unless, as many do, they spend 24h/day on it. As said, my journalist friends do exactly that, and label it as “work” — keeping in touch with what others say.

Where all of this becomes dangerous is when naive governments all of a sudden think that this is the future of online presence — a vast walled garden, ruled by Zuckerberg, where everybody is connected to their friends and colleagues, and is constantly pouring out gigabytes of data. Surely that means a “paradigm shift” — citizens were unable to communicate that much and so frequently as ever before. Thus, some countries think that it might be a good idea that relationships between citizens and their government ought to be done inside proprietary websites like Facebook. Australia is one example where court orders have been served by Facebook, while the European Parliament is toying with the suggestion of making laws if a million EU citizens “like” a page, and this the trend will probably continue, as the younger generations admit that they don’t read email but are keen on sending Facebook messages or tweets to each other.

This has been labelled as merely governments being “modern” and thus using technological tools that are popular. The hidden catch is that those organisations are not under the scrutiny of the citizens. Just suppose that, on the Australian example above, for some reason Zuckerberg decided to cancel the user’s account after the court order has been served, but before it has been read. The Facebook Terms of Service, like pretty much every other online service, allows immediate termination of any account without prior notice. Was this a successful notification? The Australian police might think so (since they would have seen the message being delivered), but the offending party might claim never to have seen it, and complain that they have no control over what the Facebook owners are doing. And on minor things this might be true as well. How can the Australian government, as users of the Facebook service, make sure that it complies to Australian laws and protects Australian citizens under their constitution? The trouble is, they can’t. They can try to “forbid” Facebook to offer services to Australian citizens — similar cases exist for online gambling, for example — but it’s pretty much impossible to refuse people to access whatever sites they wish on the Internet (as China is slowly figuring out). If countries like China or Iran or even Saudi Arabia are unable to control what their citizens do or don’t do with the Internet, how much can democratic governments actually do? The answer is, of course, “very little”.

Why privacy is important and how does the “real world” deal with it

To make sure that you don’t think I’m on a personal anti-Facebook crusade — which would be silly; I use Facebook every day, and love some of the games offered through Facebook — I’ll give a different examples. In Portugal, in the early 1980s, when ATMs started popping up, the first networks were tied to a single bank, pretty much like everywhere else. Then at some point a few banks thought it was silly to put so much effort in their own network, and instead proposed to join forces and launch a common network. They founded a new company which had their own data centres safe in some bunkers (only one has a publicly known location) and interconnected themselves with this company, which provided the ATMs for all member banks. As a result, debit cards from different banks would work on ATMs installed by the competitors. Soon every bank was buying shares of this company and become part of it — a network explored under the brand name “Multibanco” — and even Visa and Mastercard in Portugal joined the network. As you can imagine, ATMs and portable payment systems became insanely popular; today, every tiny coffee shop with two or three tables and two employees has a portable, wireless terminal which can be used to accept payment of every debit card and every credit card in the world. Everything, from hospitals to groceries, including parking lots and tolls on motorways and bridges, is connected to the same, single network. Systems like invoice payments, direct debit (for regular monthly payments to the utilities), money transfers between banks, check clearances, and so forth, all travel the same network as well. And it’s also connected to the Internet: you can log in to the network, no matter what your bank is, and get a temporary credit card in nanoseconds based on your bank account, to use safely on a single online payment; it’s one of the safest methods for shopping online. And yes, you can even buy Linden dollars through the network — there is a ‘bot in Second Life which will accept your money, interconnect via the network to your bank, update the account, and send you L$. It’s an ubiquitous system which pretty much controls all financial transactions in Portugal — a single, uniform system.

Some banks in the late 1990s loved this idea so much that they set up virtual banks. They are just P.O. boxes and have not a single human being in an office. When joining the bank network of Multibanco, they will immediately give their clients access to dozens of thousands of ATMs nation-wide, as well as hundreds of thousands of points of sales in the whole country. ATMs can both deliver money and accept deposits, so there is nothing else that needs to be provided by a bank. The online “homebanking” systems are little more than interfaces to the Multibanco network with some bells and whistles. You could call Multibanco “the Facebook of Portuguese banking” in the sense that everybody uses it daily for pretty much everything, and new applications are constantly being announced and integrated into the vast network.

Naturally enough, a bit over a decade ago, government started to use Multibanco as well, because it’s so convenient: a sequence of digits can be used by anyone with a bank account, no matter what bank, to make payments to the State, so no citizen would be excluded; for the ones without bank accounts, Multibanco was able to provide a “virtual account” (not tied to any bank) to allow safe payments and universal service for anyone. VAT, revenue taxes, property taxes, vehicle taxes, and all sorts of fines and court expenses are paid via the network. At some point, politicians started to toy with the idea of getting rid of “regular” treasuries where people had to be physically present to make a (cash) payment, since the number of citizens actually still doing that is very small, and let the Multibanco network collect taxes automatically on behalf of government — after all, they have access to all bank accounts of all citizens and can extract automatic payments from everybody in the country; switching bank accounts would be of little practical use, as they would still be connected to the same network, and, of course, your VAT number is the universal identifier to all bank accounts you’ll ever have.

But this poses a problem. Multibanco is run by a private company which is not publicly traded. Delegating tax collection to a private entity which is not under the control of the government (except marginally, through laws regulating financial transactions) is pretty much the same as “tax farming”, which any modern democracy is supposed to have abolished. It’s putting too much power in the hand of a few individuals — bank owners — without public scrutiny. So the idea was abandoned, although it’s still true that a lot of encouragement is being made for citizens (individuals and corporations) to use the network to pay their taxes and fees to the State that way; for instance, if you wish to be incorporated, it’s now mandatory to have a bank account. Even for individual citizens, tax refunds are processed much faster — several months earlier, in fact — than manual refunds using a regular Treasury check. The result of all this is that citizens have no way to hide their money from their Government; and with the current laws, the Government can give Multibanco an immediate order to seize all your bank accounts without even a court order or a due process. Anyone without access to their bank account has automatically fewer (financial) means at their disposal to, say, hire lawyers to sue Government, banks, or the network; once the money has been seized, there is no way to get it back, except after literally decades of court fights which will usually be archived.

Nevertheless, the system is not that bad because at least it is regulated. For instance, the Multibanco network has far more information about all citizens than any other company in the world. They know exactly where you live (your bank account will say that), where you work (from the payments you receive from your employee), how much you earn, how many taxes you pay. But they also know what medicines you buy and what doctors you consult. They know what kind of things you eat (from bills from restaurants and coffee shops) and what groceries you buy more from. And of course all these expenses are geo-tagged — so they can know very accurately where you usually are located. But if you’re a car driver, it gets even worse: every gas station in the country is geo-tagged and shows exactly where you’re buying gas for your car; every toll on all motorways is connected to the system as well. And so are almost all parking lots; the same network deals both with large amounts of money as well as micropayments. On top of that, of course, every ATM and point of sale is connected to the network as well, so there is really little you cannot do without leaving a “paper trail”. It’s almost impossible just to withdraw a large sum from the bank, put it under your leesa mattress, and try to use cash for everything. It’s just barely possible to live without being known by the system — if you have no bank account, no job, no home, and no car.

As you can imagine, this tremendous amount of data is incredibly valuable, and it reveals an insane amount of information about individual citizens and corporations alike. So it’s also very strictly regulated: not even courts can get access to the whole set of data. Financial secrecy has been abolished, so banks might be forced to reveal all the data they have on their customers via a court order, but what the banks actually know is just a part of the whole data. If you have two different bank accounts, only Multibanco will know the relationship between both — courts might just try to get access to all bank accounts individually, but might not be able to make much sense of the complexity of financial transactions between all. Only Multibanco has all the data, and thus they are very thoroughly checked in the amount of information they can actually reveal — which is surprisingly little, and perhaps it might even be more surprising that there were no leaks ever reported in the past 25 years since the system went live.

Facebook as the next (totalitarian) world government

Now imagine that in a near future, Facebook (or something like Facebook) becomes a major player in financial transactions. They already offer “Facebook credits” for some games. Although I’m sure that the “Facebook credit” is much less transacted than the Linden dollar, imagine, for a moment, that it becomes a common currency among the 500 million Facebook users. It is thus conceivable that banks start connecting to Facebook’s “financial servers” and allow their clients to freely exchange “Facebook dollars” among bank accounts; and that online shops start accepting “Facebook dollars” as currency; and soon governments might start accepting tax payments and processing refunds via Facebook at well. It might not even take much time until that happens. But at this point, governments completely lost control of the whole situation. They cannot legislate what Facebook does or doesn’t; except for US citizens, who might be able to sue Facebook, the rest of the world has no protection against Zuckerberg’s whims — and we know he’s hardly the most honest and trustable CEO of the planet. Just imagine what happens if your taxes are due to be paid next week, but Facebook cancels your account because it doesn’t like the picture on your profile (because it shows a cute kitty instead of your real face). Now you’re unable to pay taxes, and also unable to complain — because Facebook will just shrug and point at their ToS, which allows them to cancel accounts at their whim without needing a “cause” to do so; they are, after all, a private company, and allowed to engage in any kind of agreement with their clients. You’re free to go if you dislike their policies; but if tax payments become semi-mandatory via Facebook, you might get kicked out of the system — and since Facebook tracks your IP address and location, they might prevent you from creating a new account — and not be able to deal with your country’s revenue system any longer, and be fined (or, in some cases, even go to jail). And there is nothing you can do about it. You won’t even be able to complain about it on Facebook because your account has been cancelled!

This is, for me, the problem of private quasi-monopolies without direct competition, without public scrutiny, and too much power: at some point, it becomes impossible to prevent abuse, and we just have to go with it. In my country, for instance, it’s impossible for someone to set up a second financial network, just because all the banks are on Multibanco, and the costs of starting a new competing service are staggering — and not worth the expense anyway, since the banks won’t be joining it. Why should they? They are already co-owners of the existing service, and protective of it. The only advantage for the citizens is that the system is very strongly legislated and controlled, because Government has power over it. On multinational, global organisations, this is simply not the case. Even if the US started to worry about Facebook’s perceived power, all that Zuckerberg had to do is to relocate his HQ to some other country, or, like PayPal, establish lots of separate companies all over the world (which globally use the same brand) to avoid any restrictions. PayPal is able to terminate accounts and seize money at will because they are not under a single jurisdiction; some jurisdictions might actually forbid that without a court order, but PayPal neatly evades the issue by simply forwarding the money to whatever jurisdiction is more appropriate and seize it there. You cannot complain to a single entity — either in your country or elsewhere — because the money is not in any “real” (physical) place any longer. It flows across servers in the world, jumping around jurisdictions, and eludes any attempt of being tracked.

Now I’m not paranoid, neither do I like too-strong, too-controlling governments either. I’m not even paranoid enough not to trust governments. So long as I know where I can complain to, or where I can exercise my freedom of speech, I’m happy — I cannot demand much more. With Facebook, if my account gets terminated, or my comments are cancelled, I have no recourse, and no place to complain to: in effect, I have no rights, no privacy, and no way to protest about the lack of them.

This is what worries me about the tendency of pushing everything into the mainstream and letting a handful of global corporations set the rules, over and above any governments, and without any interest in the rights of individual citizens. It sounds like a scenario from a cyberpunk novel, and perhaps what surprises me is that how quickly we have come to this point…

Walled gardens vs. the open Web

So, Tim Berners-Lee might have a lot of facts wrong, but at least he’s right on one. Pushing everything into “walled gardens” where nobody has any kind of control, except the business owners, is a step back from privacy and individual rights — it’s trading off the commodity of belonging to “the system”, which requires accepting pretty much what is imposed on the terms of service, or making sure some of your rights as an individual are still respected. Prokofy talks about “walled gardens with governance” as being the way to go, but I think it’s an utopia — the whole point of a corporation setting up a “walled garden” is exactly to prevent governance, but instead to keep control. There are a few exceptions, of course. eBay has its faults too, but most merchants are honest because if they’re not, clients will rate them badly and alert potential customers on the forums; and eBay provides at least automatic arbitration system to try to solve the trickiest issues arising between customers and merchants. eBay’s community is not perfect, but it’s an example of a semi-walled garden where there are at least some emergent community-based governance tools in effect, and they influence the way that eBay, the corporation, defines its policies and rolls out new services. But there are not many similar examples, not even bad ones.

By contrast, the Internet model is the total opposite: by standardising protocols (and the standardisation bodies are free to join; anyone can define a new Internet protocol, all they need is enough community support), services that interoperate can compete for customers, instead of forcing them to adopt a single provider with a limited set of rules. Let’s take email as an example. Half a billion Internet users might prefer to have a Gmail account, but that doesn’t mean that Google “controls” email. If by any reason Google cancels your account, you can still send and receive email from another address; and you can even set up your own mail server and address, and, so long as you refrain from spamming, your emails will be accepted by everybody in the world. Before using Gmail I used other providers (and even self-hosted solutions), but my email address has remained the same since 2004. In fact, I use in real life an old email address since 1994, which I have never changed again — email providers come and go, but my email address is always the same. This means that I am in control of my email addresses, not Google, not my ISP, not anyone else — not even the Government!

We take this for granted because we’re so used to it. But every application we use on the Internet works pretty much the same way: freedom of choice, by giving you freedom to use whatever service provider you wish, and switch among them at will. Switching providers will not give you “limited” access. Even if Google gets angry at me and starts boycotting anything that comes from my IP address, I won’t be shut out from the Internet — or stop paying taxes to my government. I would just use another email provider. Google Chrome might not work any longer on my computer, blocked by Google, but I have plenty of choices among browsers; and if I wished, I could create my Google Chrome “clone” by simply downloading the WebKit sources and compile my own version. I might be unable to use Google search, but I’d still have Microsoft Bing to do some searches. I might have to get rid of all my YouTube videos (there are not many anyway), but I could upload them to Vimeo or instead. I seldom use Orkut, but if that got cancelled by Google too, I’d be happy to switch to something else (like… Facebook!). Since I can own my own domain name, with some clever DNS tricks, I could even have addresses like,, or, and point them all to the appropriate page (in fact, I did just that 🙂 ). Of course it’s not just about Web pages and DNS addresses, but you get my point: the freedom to pick any provider according to my tastes, the price of the service, or the rules they impose on me, is mine and mine only. Having choices is great.

I have to admit that I very reluctantly joined Facebook, Twitter, Plurk, or Digg. Among all those, the only one that I use regularly is Facebook — because I enjoy some of its games. The rest I use mostly as ‘announcement’ venues: I use to distribute information on all social networking tools, instead of updating each and every one separately. Why? Because for most people, and that certainly includes myself, time is at a premium. It’s next to impossible to keep in touch with everything, assuming you have a reasonable amount of “friends”. So I need a one-size-fits-all solution. The irony is that I already had one which was perfectly acceptable: email. But these days, we’re all supposed to communicate asynchronously using these social networking sites, of which every week you get a handful of new ones… and with each, one shares more and more information online, in return for… what? Having to upload all your profile data again, and finding all your “friends” once more, and connect with each in turn. Of course there are now mechanisms to make this process easier… but still there will always be some ‘friends’ missing, since nobody ever registers for everything.

Invent your own wheel: the fragmentation of the social networks

The result is fragmentation: instead of having this wonderful feeling of being connected to one huge community of 2 billion Internet users, one is actually connected to separate, individual websites, each with its set of things you can do in them, and each with a different set of ‘friends’. Some might argue that this model allows social networking sites to compete for people’s attention, and that the ‘best’ will win at some point, thus benefiting everybody that way — we just get the ‘best’. The truth, of course, is quite different: the more options you’ve got, the more fragmented it gets; Facebook’s success can be partially explained for filling a lot of roles at the same time, and thus allowing different kinds of people to use it in different ways. For some, it was just a “Twitter replacement”; for most, it’s just a way to share links; for a lot, it’s the largest free game distribution mechanism.

Contrast that with the way some things in the world work, where standardisation and protocols gave us an uniform way to share an environment without, however, constraining us in picking a specific company. Let’s move out of the Internet. In the real world, all cars have similar specifications — there is a reduced amount of gas they can use — and all have to conform to a set of possible sizes, so that you can create roads where all can drive. But each car is different; consumers gravitate towards the brand they know best, fully aware that they will be able to enjoy the same roads and pump gas on the same stations. Imagine a world where the automobile industry would create specific brands which would require special roads and a different kind of gas.

The same applies to the telecommunication network. One reason why the US has lagged behind for a while ws because most mobile operators decided to go with their own protocols, requiring a specific model of mobile phone to connect to each network. Change operator, and you need to change the phone, too. Nowadays there is a much more limited amount of protocols in use; nevertheless, the US still has mostly non-GSM operators, while the rest of the world pretty much uses the same protocol, and you can buy a phone everywhere and be part of the global, 3-billion-large community of mobile users. But you still have a choice of what phone to use; more or less sophisticated or not, it’s up to you — in any case, you’ll be still connected to the world-wide mobile network.

The success of the widespread use of personal computing was due to Microsoft’s operating systems, which could be used across pretty much any kind of hardware. Except for consoles and very specific hardware, you can run Windows on any personal computer — even on Apple’s own hardware. This, again, allowed PCs to become mainstream, while still providing consumers the option to use whatever brand and features they wish and not be stuck to a single hardware manufacturer.

The Internet is not “a network”: the name itself implies what it is — “interconnected networks”. Each individual network might be run under completely different hardware and protocols, but to communicate, they all use the same set of protocols. That’s why you can email anyone from anywhere, no matter what device, operating system, or provider you’re connected to. And every web site (assuming they conform to strict HTML specifications, of course!) is viewable by any browser. As an Internet user you still get the option to use whatever application you prefer. Some email applications don’t even look like email applications; Gmail dramatically changed the way we think about “filing” email messages. But we’re not “forced” to use what Google thinks best; I’m a traditionalist and get far too much email to find Google’s method of sorting useful. If you get few emails per day, keeping conversations together might be better, but if you get hundreds of messages per day, it’s impossible to keep track of them, and the old method of piling up new messages on top of old ones work best for me. But I can still enjoy Gmail and not use Google’s method of viewing messages: I have a choice. I can even choose providers. If I dislike Google’s Terms of Service, I can still get access to email (elsewhere), fully confident that my messages will still get to my friends, if I send them through a different provider (and I can become my own provider, too).

By contrast, “walled” social networking sites remove all choices, deliberately so. One either agrees to those pre-set choices, and the Terms of Services, or one is simply excluded. There are no “Facebook-compatible” sites, e.g. where you log in to an alternative provider but get full access to Facebook’s innovative features — at best, you can pull and push data to Facebook, but you cannot recreate the whole experience of being in Facebook (unless you’re willing to get sued 🙂 ). You have either the choice of being part of it or being out of it. But as time progresses, due to peer pressure, it gets harder and harder to avoid it. For instance, a lot of companies make LinkedIn profiles a requirement for all employees. Other companies might “force” their employees to use Plaxo or a Facebook page linked to the company’s page.

And at some point, government might “force” individuals and corporations to use service X over service Y because it happened to be the choice that most people made. Now I have no problem in using technology to be in touch with my government; but I resent the notion that I have to be forced to use a single solution (or a set of solutions) just because it happens to be “popular”. For instance, I might be “forced” to use a bank account, but I have the choice to pick the bank I prefer, fully aware that all banks will transfer money correctly in the same way.It’s very different from saying that I have to use Facebook and connect to a fancy application that will interchange data with government (for instance, for paying taxes), and if I don’t want to do that, I have to fill my tax forms manually.

“Forcing” to use a quasi-monopoly sanctioned by the State is pretty much what I do not wish to happen. And it’s little relevant if there is a small number of choices: the point in this case is that “someone” (at the government level) will pick a set of choices according to their subjective perception of the market, which might not be aligned with mine. Another example: until recently, most government-based web sites in my country “assumed” that everybody was using Internet Explorer (because there are a lot of protocols between the Portuguese government and Microsoft) so official sites would discriminate against anyone “not wishing” to use IE. But sometimes all you had was a mobile phone. When IE finally stopped having the dominant position in the web browser market share, government suddenly realised that it was not fair just to allow a group access to those websites and excluding the rest. No matter how many protocols were signed with Microsoft, governments simply cannot impose a monopoly on top of their citizens, specially one that is not under government control, much less under the citizens’ scrutiny. These days, all official websites require to be fully compliant with the current set of HTML specifications, and web designers (some did complain a lot) just had to adapt. The result is full access to all government-based official sites no matter what browser or device is used.

This is what I expect to remain in the future. No matter how “mainstream” a product becomes, imposing it as the “only” choice is simply not fair. The Internet was not the “resulting network” after “competing” with proprietary networks (like Compuserve or America Online). Rather, the Internet protocols allow those proprietary networks to communicate across boundaries, and specify a set of available applications that will universally be supported — Web, email, remote access, remote file transfer, and so forth. “Walled gardens” are the opposite; they’re going back in time and trying to establish an independent view of what it means to be “digitally online”, and compete to exclude all the rest.

Are (social) virtual worlds ready for the mainstream?

Now we come to Second Life, virtual worlds, and Linden Lab’s stance. Here we started to see the push more towards an open “Metaverse Protocol”, exactly because right now every virtual world is a walled garden. If you join SL, you’re stuck with it; even if SL became much more mainstream (which it won’t), it would mean that you’d have to forfeit all your content and profile and ‘friends’ on all other virtual worlds if you just wished to connect to something “officially” done in SL. OpenSimulator, while being technically “compatible” (in the sense that the same protocol is used for communicating between the SL client and an OpenSim server), doesn’t mean “content-interchangeable”; or at least, not yet. In late 2008, we were all expecting that at some point Linden Lab would make a serious effort to interconnect their own grid(s) and allow third-party grids a safe way to access (even if only partially) the content available on the Second Life Grid; not only 3D objects and scripts and textures, but also identity and profile information associated with an avatar, as well as their inventories. The old Open Grid Protocol (superseded by the IETF’s standard VWRAP) was expected to be adopted by mid-2010, and LL showed off an early implementation which did indeed interconnect with an OpenSim-based grid run by IBM. While a first draft of the protocol was presented in July 2010, the VWRAP group will only finish its work on a complete set of protocols by February 2012. With Linden Lab’s lack of enthusiasm after the break-up with IBM (?) in May or June 2010, one can only wonder if LL will continue its efforts at standardisation.

In the mean time, OpenSim developers have released an alternative, “quick-and-dirty” interconnection protocol called HyperGrid (about which I have written extensively 🙂 ). This allows OpenSim grids to fully interconnect in the sense of allowing avatars with their profiles and inventory to jump across grids, buy/create content outside their “home grid” (assuming the permissions allow it) and hyper-teleport elsewhere to use that content. It works. But it’s highly unlikely that Linden Lab will ever remotely consider using HyperGrid as an alternative interconnection protocol.

And they might have good reasons for that. At some stage, Linden Lab was seen as the “Evil Corp” which wanted to insist on a walled garden with little entry points; there are very few APIs that allow communication with Linden Lab’s servers. On the OpenSim side, we started to see more and more tiny grid operators popping up and happily interconnecting each other — until they grew and became conscious that they were, indeed, attracting enough customers to sustain their operation, mostly because they were giving them good service. At that point they dropped the ability to teleport in from other grids (their competitors). So the irony in this case is that the small grid operators create their own walled gardens… when they would benefit much more from interconnection!

But such is the strange consequence of this niche market. Instead of focusing on expanding the number of people that can be interconnected, they apparently prefer isolation, thinking somehow that it’s a “better” way of attracting customers. I’m sure that the reasoning behind it is that they work from a (misguided) belief that over time enough competition will force a handful of “winning” grids to emerge, and that all residents will just connect to those, forfeiting the rest — and each wants to be one of the “winning” grids, of course. However, the fallacy here is that the number of grids can grow much faster than their ability to attract customers. This is the problem in niche markets: since the number of users is very small, if there is an excess of offers, all will suffer from excessive market fragmentation.

By contrast, the market for social websites is insanely huge. It’s relatively easy to launch a new site and get a few million users — and that’s why there are millions of social websites, most of which might never hit the media even with millions of faithful users. That’s just because the market for social websites has, well, 2 billion potential users. 3D virtual worlds have perhaps just a tenth of that; and Second Life, with over 20 million registrations, is one of the largest social virtual worlds still around (Habbo Hotel very likely still is the largest one). OpenSim-based grids don’t even come close to that — not yet, and if they persist in their “isolation” policy (which is very likely), they will very unlikely grow much. The irony is that it’s conceivable that at some time most people might be connected to an OpenSim-based grid at some point — that is, adding up all OpenSim-based grids might one day have far more than the 20 million registered users of Second Life — but none will dominate. For that to happen, a well-known brand would have to give it a hush push: just imagine if IBM/Intel announced the launch of a massive infrastructure to support a global OpenSim-based grid. That would certain catch people’s attention! However, it’s unlikely that this will happen soon; they prefer to sponsor grids like ReactionGrid and slowly catch all Linden Lab-rejected universities and educators, one by one, on that grid. It’s also conceivable that at some point large-scale OpenSim-grids are funded by governments. Imagine a NSF-funded academic OpenSim grid, for example. A few attempts have been made in Europe to get funds for a Trans-European Academic Grid, but, so far, and to the best of my knowledge, none have managed to raise enough funds. But the trend might continue, and when that happens, it’s very likely that all research and academic work will be done on one of those grids.

So, ultimately, this is the challenge for Linden Lab’s new CEO, Rod Humble: define the strategy for Second Life. It’s ironic that Second Life started as a very specialised niche market application — a development platform for 3D modellers (and programmers) to develop their own 3D games. From that it morphed into a 3D social virtual world where users happened to be able to create and share (and later sell) their own content. It’s peak of media popularity seemed to be at the time that Second Life struggled to leave its niche market status and become mainstream; it failed to achieve mainstream status because, well, it is not a mainstream product that appeals to everybody, but just to a very special kind of people — I’m still waiting for some anthropologists and sociologists to develop a “Second Life resident profile” and publish a major paper about what makes people not only log in to SL, but remain there as residents for many years.

Attempts to push SL into the mainstream have failed so far, after two years of looking at several options: more academics? More business? More integration with social tools? None of these seemed to have worked, until Philip came back from the void and just told everybody to focus on making the good bits of SL better (meaning: more stable). SL is nowadays more stable and faster than ever before, but it remains a slowly growing niche market, which the mainstream doesn’t “get”. Journalists are still baffled by it: while every journalist “gets” Facebook and can easily come up with a dozen stories of what Facebook is good for, few can say the same about Second Life. The best they can do is to provide examples of successful uses of SL, but they’re so different and mostly unrelated to what (most) people are interested in, that it’s very hard to grasp the concept of “social virtual worlds as a mainstream product”. Even as a dating/cybersex platform it seems to lack some appeal — SL’s stance towards privacy is the total opposite of what Facebook (and others) are doing. SL is about absolute anonymity, where you have to work hard to tell others who you are. Facebook & friends are all about revealing too much of yourself without worrying about how secure is your data. The endless list of reasons why Second Life is so different from a “mainstream” product seems to grow, as social networking sites copy each other and become more and more similar. One huge, major difference is that SL is mostly about real-time (synchronous) communication, while almost all social networking sites are asynchronous. This doesn’t mean that mainstream products cannot be real-life — just take Skype or MSN Messenger as typical examples (as well as all the webcam group chat sites) on how real-time communication is important for so many people. However, Robin Harper argues:

The fact that the connection is asynchronous doesn’t matter. Asynchronicity facilitates the connection — a meaningful way to stay in touch or reconnect.

Well, perhaps. Nevertheless, Skype has quite more users than Facebook, a fact that gets little mention in the news; and, of course, there are 3 billion mobile phone users world-wide.

So, it I were Rod, I’d start listing what major features Second Life has and contrast it to other products, mainstream or not; based on that, he might get a good idea on how to develop a strategy that makes sense. It might mean that SL will just remain a niche market, but one where all LL’s efforts are focused on making SL the best product ever in its niche market — and make sure LL’s profitable that way.

Just to give a clue on what kind of features I have in mind, here are a few examples:


  • Runs on the Web and doesn’t need installing anything; goes through firewalls
  • Runs on low-powered computers (or mobile devices) and doesn’t need any specialised hardware
  • Very easy to use, even for an untalented person
  • Incredibly easy to search for friends, people, events, etc. even if you don’t have their exact names/addresses
  • Very easy to upload images
  • Fun & entertaining games are easy to install (self-entertainment)
  • Inflexible layout (no personalisation)
  • Presence is limited (you usually don’t know what others are doing, which pages are more seen, etc.)
  • Asynchronous communication, which however reaches out very quickly to many users
  • Little privacy concerns
  • Constant stream of communication can be overwhelming, but can be ignored (most often it is)
  • Very easy to share content, even content outside Facebook
  • Moderately good API to interface with other platforms; good developer support
  • Free, ad-sponsored (as well as sponsored by selling profiling data)

Most social networking sites would share almost all features above, although content sharing is usually harder, and searching might be way harder than on Facebook (mostly due to privacy concerns); not all have the extensive APIs that Facebook has for integration with other sites.


  • Requires installing an application (and setting up a firewall)
  • Requires specialised hardware (camera, microphone); however, this hardware is available on most devices; not available on most mobile devices
  • Simple to use
  • Moderately hard to find friends (unless you have their exact data, nicknames, etc.)
  • Very limited profile, no user-generated content (but it also means higher privacy)
  • Synchronous communication, mostly one-on-one
  • Sharing requires text chat, no special tools available
  • No plugins, etc. (I don’t know how their APIs work)
  • Fun & entertainment provided by communicating with others (no self-entertainment possible)
  • The 600 or so million users can communicate with 3 billion users on mobile phones for a fee (fees from paying users pay for the service)

Second Life

  • Requires installing an application (and setting up a firewall); however, things like the Canvas viewer might change that
  • Requires highly specialised hardware (high-end graphics card) for good results; low-end hardware might give such unconvincing results that new users will be disappointed
  • Very hard to use (compared to Facebook or Skype!)
  • Limited profile due to security/privacy concerns (anonymity is the default setting)
  • Everything is user-generated; no templates, no rules; freedom of choice on how you and your environment looks
  • Space is visually shared: you know where people are and can see what they’re doing, all the time; on the other hand, you need first to travel to where people are! (which might be difficult to find)
  • Building tools inside the viewer (well, most; this might change once mesh becomes popular)
  • Synchronous communication, mostly on small groups. One-on-one is definitely possible. Audiences of more than 40-100 avatars (to be realistic) are only possible through group text chat
  • Rich immersive multimedia experience
  • Way too many ways of entertainment (in the sense that there are more ways to get entertainment, either alone or with others, than classifications on the types of possible entertainment…), but it’s hard to find it, or to understand how it works, or who else is providing entertainment or getting entertained
  • Few APIs to interface with the external world, but inside the virtual world, everything can be programmed by the users
  • Service is free except for storing persistent content (land tier fees)
  • Content in Second Life is (mostly) not free, but it has created a huge economy worth over 600 million US$ per year

I think this is a good start 🙂 It already shows a trend: niche markets are concerned about user-generated content, giving users a vast array of options, and allowing users (all users, not just a few selected ones) to make money by providing services. To allow for that to happen, the interface needs to be complex, and the hardware requirements are way high. Mainstream products limit choices and keep requirements low; Facebook trades this off by allowing people to do a lot of things at the same time and making application integration inside Facebook so easy that millions of developers are constantly spewing out new applications.

A “mainstream” 3D social virtual world might require very simple graphics, an easy-to-use interface, a good way to search for friends, lots of entertainment opportunities, and more facilities for developers to create content. Can Second Life become that kind of platform? Well, I guess that Habbo Hotel already has taken over that market (188 million registered users, of which 15 million are active each month, although they don’t use Habbo Hotel that much — about 45 million hours/month compared to SL’s 35).

It shall be interesting to see what approach Rod Humble will take.

Print Friendly, PDF & Email
%d bloggers like this: