Do we see the road ahead?
Now, how do I start this story? Once upon a time, sometime in mid July to be more precise, I was approached by a group of guys from a young start-up company who wanted me to assess their in-work project both in terms of conventional SEO as well as from a more generic angle and all because I once wrote an article on Pownce. My first reaction was to avoid this since I didn’t at the time consider myself to be an expert in Social Networks and Online Communities at all. Before continuing with the story perhaps I should give you some more details about the project itself. Initially the project was conceived as Yet Another Social Network Profile URL Aggregator capable only of storing URLs of user profiles on various Social Networks. Somehow logically it progressed to the idea of gathering friends from those social networks thus creating a sort of representation of a Social Graph or better User Circle that can be viewed, accessed, managed and controlled by user from one centralised location. The system is supposed to issue alerts and inform the user of new or orphaned friends and other discrepancies in the user circle. At first I took these guys for a bunch of deluded and overexcited amatures, arguing that to amass and manage this kind of data would require computing resources available only to the likes of Google itself. However, I had to radically change my opinion after a rather impressive demonstration of their working prototype. The guys had already a spider capable of reading friends circle with minimal delay simply by parsing XHTML code and a database server running already several hundreds thousand user circles gathered from several major social networks. Most importantly they had a clear idea about scaling this prototype and some sketches of user interface allowing a very detailed management of this user circle. They also produced a draft version of API and data exchange protocol. Suddenly it all had begun looking doable, realistic and interesting. So I decided to accept the challenge and started working with them. Needless to say that I am continuing working on the subject to this day providing some sort of theoretical and even ideological background for those managing the project and with their kind permission I am going to publish some of my conclusions in a series of articles in my personal blog here.
Now, let us continue with the story. Eventually we’d set up a timeline and all went working on our respective tasks, me doing research and thinking, them doing their stuff which I suppose mainly consisted of some kind of hardcore programming of which I honestly know nothing. Back then, nothing spelled a disaster that was about to happen when all over sudden there comes Brad Fitzpatrick with his now infamous Thoughts on the Social Graph article. Although with the benefit of the hindsight being quite predictable at the time it came as a bolt from the blue and looked like a terminal deathblow for the whole project. The team spirit went so low that I nearly thought it will never recover. Pity, you couldn’t see the expression on their faces after they’d heard the news. The closest analogy that comes to my mind is what perhaps every one of us had experienced or at least observed while still at school. You are about to sit down on your chair and just at the very moment you cunning classmate quietly pulls the chair from under your bum. It’s that split second when you realise that the seat is not where it is supposed to be and now there is nothing that can stop you from falling and hitting the floor in front of the whole class as well as your much hated teacher and neither yourself is now able to stop the force of gravity, that can produce this truly unique and inimitable facial expression. In our circumstances I’ve been given a pleasure to observe this very expression on their faces for several days in a row and this is something that happens not very often in your life. Their reasons for panicking are not so difficult to understand as when the almighty Google decides to move into the domain you rather foolishly considered to be your own, you really need the nerves of steel not to run away and duck and cover. Should I mention that the fact that Brad now works for Google was even back then just two clicks away from the article itself?
It took me some days to convince the shell-shocked guys that not everything is lost and in fact this sudden intervention may even work to their advantage. First of all I had to unveil my treasured theory of second-mover advantage trying to convince them that time trial isn’t at all the worst kind of racing game. Moving in after the leader had already ploughed the virgin land makes following the furrow a much easier task. Besides Brad’s article itself didn’t look very convincing. Not only had he managed to spoil the whole impression by inelegantly trying to promote the Dopplr project to which he personally is seemingly related, but also it had a number of weak points we are about to discuss here. But before we begin with the critique it is customary to acknowledge all the positive sides and achievements of the article in question. The strongest point of the article is obviously the magnitude and the power of the vision: the vision of the project, the vision of the future of the internet and most impressively the vision detailed enough to include such practical things like API, servers, databases, browser apps and so on. The vision that is so close to my own that I must admit I nearly went frustrated seeing it written by somebody else.
Undoubtedly, the most attractive part of the picture drawn by Brad is the openness of user data, but being at the core of the idea this is at the same time is the weakest point in this whole construction and precisely at this point of course our visions differ. However wonderful the idea of a non-profit body taking care of the promotion, maintenance and proliferation of user data might look, this part is the most questionable in the whole article and no wonder it remains the least elaborated. Please, understand me correctly, I am not defending the greedy corporates trying to get hold of everything that’s alive, neither I am an advocate of paid services, and quite to the contrary, I would love to see such a project as being completely free and open for all, but at this stage it looks merely utopian, especially considering the amount of work to be done and the size of obstacles ahead. The starting point for the project, to which I am now related, was certainly different. From the very beginning it was thought to become an independent business venture with a rather banal but clear business model based on selling ad space. And by all means the project is intended to be self-supporting starting from phase two. After digesting the open data idea, obviously there was an attempt to adjust the current business model to the one offered by Brad. To my infinite disappointment it turned out that there is none. The problematic issue here is sustainability of the whole project. To achieve the desired sustainability and growth one simply must have a clear business model; even non-profits do have one. Keeping the project data free at the user level one must have at least an idea where those money needed to set up and maintain servers, to program APIs and browser add-ons etc, etc will come from. Avoiding even remotely any possible outline for this side of the project is either an evidence of a complete carelessness concerning the business side of the idea (something I find myself hard to believe into) or, more likely that the funding had been already promised at least for the initial stage of the project. All subsequent events only confirmed my suspicion. It is not incidental that soon after the article Google itself comes with their Open Social initiative. By all means Google is probably the worst possible starting point for a project of this kind. Let me explain why. Structurally as well as commercially Online Communities, no matter who owns or operates them, in their very nature are antithetic to the concept of Search Engine (more about it in my next posting). Google being a Search Engine is eager to gain a foothold in the competing business area and get at least a partial access to Online Communities data. Hence, an attempt to gather free community effort only to provide Google with free qualified labour and ideas is certainly bound to fail. We’ve been there, we’ve done that. In addition, one needs to ensure a rapid uptake of the service in circumstances where the initial data collection depends largely on a good will of big players in the Online Communities field, and those players are not likely to display any good will at all, clearly seeing this as an attempt to lessen their competitive advantage. Merely, there is no pretext which would convince large Online Communities hosts to share their data with anyone, but to share those data with Google would be plain stupid if not suicidal on their side. To sum up - I can see three highly questionable points in what can be seen as a practical side of Brad’s plan:
- absence of a clear business model independent from Google funding and support
- consequently his call for community effort in a project that would ultimately benefit Google looks to me a bit hypocritical, to say the least, and by all means not convincing enough to gather independent volunteer developers under those banners
- over reliance on willingness of big Online Communities hosts to share their data – just get real, it simply ain’t going to happen, and the only way out is to focus instead on building a robust spider engine capable of gathering public domain User Circle data from those hosts regardless of their willingness to cooperate
However, much more fundamental follies of Brad’s article are not in the field of practical implementation but rather in the field of theory and general methodology. Like many others before him Brad falls under the magic spell of the word social that is ever present in thinking about online communities. Trying to narrow the problem to a mere representing of pre-existing social relationships in digital form we are running a risk of missing entirely the meaning of the ongoing online interaction. A somewhat longer discussion of this issue can be found in my previous posting while here I’d rather point out some technical issues that will inevitably arise as soon as one has to deal with the actual reality and not with imaginary Social Networks.
Social Networks and People Search are perhaps the most evil terms currently circulating in our lexicon. Both are equally deceiving making us to believe that we are dealing with real-life people and real-life social interaction. Undoubtedly most of the data we are dealing with is a result of actions performed by humans, as even robot-generated content ultimately comes from human activity, but it is a serious mistake to represent our dataset as if it contained data on real social relationships between real people. All we can analyse is a dataset containing user accounts and relationships between them (mind you, only positive relationships are on record). Firstly, there are no grounds to make one man - one account equation, as for various reasons most individuals have multiple accounts on the same service and have their own reasons for not declaring this fact. To illustrate the point ask yourself how many email accounts you currently have and have had since the introduction of this wonderful way of communication. Isn’t it correct that you’d prefer not to mix your work email with your friends and family one and will use a third separate email account for registering at free services that are likely to send you tonnes of spam and virus attachments? Consider also that the same individual may abandon one account simply because of a lost password and open a new account instead. Now think about institutions, organisations, companies and simply groups of friends having their joined accounts at various Online Community Services. How about registering an account for your pet? Facebook had to introduce a special policy dealing with accounts of deceased people, placing it in a so-called "memorialization state" but what about other Community Services? And mind you we are not even talking about things like spammer activity constantly generating thousands of accounts from one source and account hijacking that totally confuses the picture. Now, the relationships we are able to see in our dataset: apart from the obvious lack of negative relations that are always present in real world, we have to deal with a completely different set of ties. Nobody in their right mind would go and list all of their rather boring relatives, friends and acquaintances from real life on their internet profile while one has a wonderful opportunity to find new and exciting online friends. Perhaps one of the strongest intents for having an internet identity at all, is the ability to find new friends, not to get stuck with your old ones. Real world friends sometimes can even be a burden online. Generated and supported by offline social environment, real world relationships are rarely needed to be actualised online, as the only reason for displaying your offline relationships publicly is when they may add to your online status. This brings up yet another distinct feature of online friendship – it is all public, and not all of those with whom you communicate online are supposed to know about each other. Sometimes multiple accounts are created precisely for the purpose of keeping one’s relationships of one kind separate from another. More importantly, online friendship is so attractive precisely because it does not function in the same way as the real-world one does. It is capable of traversing geographical and class barriers, state borders and time zones. It requires neither spatial co-presence nor even temporal synchrony as online communication allows for a reasonable delay. Sometimes it can even break through the language barrier thanks to online translation services. On some occasions online friendship can be transferred to the offline world and I assume this happens more often than the other way around. Far from being just a mirror image of real-world ties, this is a distinct type of relationships, the meaning of which should be analysed in its own right. On aggregate all of the above not only introduces a loud noise that distorts otherwise "pure" picture, but creates a completely different one – it is an altogether different phenomena. Once again, the idea that Online Communities are mirroring offline social relationships with each account representing single individual and friendship relations reflecting real-world relationships is no more than a phantasm, a product of wishful thinking and a sick fantasy of marketing executives and election campaign managers. The stronger the tendency to bind online profiles and accounts to personally identifiable data the stronger the backlash to be expected from internet users. The stronger the threat and the pressure the more often Internet users will evade positive identification of their accounts with their real-world identities and the more effective this evasion will be. The fact of the matter is that nobody wants yet another Big Brother, this time not in a form of a potentially accountable state agency, but worse, in a shape of unaccountable private corporations watching over every aspect of their online activity. As recent Facebook privacy scandal reminded us, users tend to become quite aggressive, active and inventive when their privacy is at stake, and you don’t mess with an angry user. The internet is a place that provides the means for effectively anonymous communication and as soon as it looses this ability it becomes no different from any other way of communicating and therefore redundant and hence it is not likely that this will ever happen!
To give a roundup on this:
- in our Social Graph or better User Circle dataset we are NOT dealing with people and their relationships, we are dealing with user accounts and a sort of very primitive connections between them
- one person – one account equation if fundamentally flawed, instead we are dealing with all sorts of accounts, from accounts representing just a segment of an individual (segmentary accounts) to accounts owned by a group of individuals (collective accounts)
- relations between accounts are only of positive value and most often have no direct correspondence in offline world; it is only the fact of a relation that can be registered, as there is no reliable and uniform mechanism of classification (microformats is a step in right direction, but it is too limited to provide a reliable solution)
Ultimately, to answer the challenge we need not to reduce the complexity of online interaction to a mere representation of offline social relationships but finally understand the futility of such attempts and start dealing with the subject the way it deserves. We have to accept online interaction not as a secondary phenomenon but as an activity in its own right and analyse and deal with it accordingly. To illustrate this I can give you an example of a problem that belongs to the same set but had not yet been mentioned precisely because it seems to have a workable solution already. I am talking about a problem of establishing correspondence between accounts held by the same user at different online services. If left on its own this functional feature may easily get out of hands as nothing prevents me or any other user listing account of Bill Gates (or Brad Fitzpatrick for that matter) as my own at some other online service. Fortunately there is a way to verify these connections either with microID or with OpenID. These two solutions are precisely of the kind we are looking for. Instead of attempting to relate Internet accounts to an external real-world identification marker (be it SSN, credit card number or whatever else) it elegantly connects one internet account to another via internal mechanism of the internet itself. It seems the solution to problems described above can be found along the same lines.
And now, back to where we started. The project continues, as gradually all fear and panic happened to settle, all boiling down to the fact that it would be altogether completely stupid to abandon such a wonderful idea only because we are not alone in this field. Considering the enormous complexity of issues one has to work with, dealing with this completely untapped field there will be some room for all - from giants to some petty traders like ourselves.
PS Merry Christmas to you all! To those who celebrate it, and to those who don’t, regardless!
tags: big brother, pownce, social graph, social networks, user, user circle, api, brad fitzpatrick, facebook, google, online communities, open social, openid, opensocial, people search, social graphs, url aggregator, users circle, microid, microfirmat, online friends, online relationships, profile aggregators, public domain, second mover advantage, social relationships, user data
Posted by LZZR under Social Networks |
































[…] 23.12.2007 When I was in the middle of writing my next post I realised that I desperately need another term to describe adequately my new idea. Once again, I […]
Pingback by LZZR » Online Communities - Notes on Terminology — March 9, 2008 @ 6:06 am