Friday 22 July 2011

The state of social media


You post to the social media system of your choice. Your friends - the friends you choose and no-one else, unless you choose to make your post public - see it on the social media system of their choice. If they choose to respond, they respond to a common thread which all your friends (and possibly theirs, if you've allowed that) can see, limited by the specific capabilities of their chosen system. For example users of one system might see the discussion as a branching threaded discussion like Usenet, while others see a single unbranching thread. Common conventions for tagging, reposting and referencing other users are used across heterogenous social media systems.

Impossible? Technically, no.

Parts of this vision exist. Firstly, let's hear it for Diaspora, which I shall review in more detail below. Diaspora already - by default - allows a post to be reflected out to both Twitter and Facebook; as Diaspora development moves fast, I expect Google+ to be added very soon. You write once, but it creates three separate threads, so that those who respond on Facebook do not see responses posted on Diaspora, and vice versa.

A plugin for Google's Chrome browser called 'Start G+' adds features to Google+ which provide a considerable degree of integration between Google+ and both Facebook and Twitter; not only does it allow forwarding of posts to Facebook and Twitter but it also integrates other people's posts from Facebook and Twitter into your Google+ Stream. At least, I believe it does. The problem is that to do so it depends in part on the author's own server, and either because that server is extremely overloaded or because my Internet connection is poor I can't get this to work.

Reality: Diaspora

Diaspora is really very good. The user interface is slick, clean, elegant, easy to use. Although there isn't a dedicated Android app, the web interface is adaptive to the device and works just as slickly on a hand-held as it does on a desktop. The developers are thoughtful, engaged and responsive. In terms of features, Diaspora is most similar to Google+ - your contacts are organised into groups called 'aspects' which are essentially the same as Google+'s circles.

Best of all, Diaspora is designed for a distributed world: the project is open source, and rather than one organisation owning all Diaspora servers the whole idea is that different Diaspora servers owned by different organisation should interoperate, and that if your home is on one Diaspora node your posts can by seen by friends on any Diaspora node. Furthermore, since the interchange protocol is open, there's no reason in theory why Google+ or Facebook or LinkedIn shouldn't write their own sharing interface which allowed them to look to Diaspora like any other Diaspora node, this achieving the vision I set out at the top of this essay...

No technical reason, that is. No reason except the commercial reason. We'll come back to that later.

In summary from a software point of view - from a usability, design, and technical point of view - Diaspora is the best social network system out there today. I really, really recommend it. Sadly, I think it's doomed to fail, which again I'll discuss when I get round to talking about the commercial background. But at present it's the best there is.

Full disclosure, I put seed money into getting Diaspora up and running. I'm glad I did. I think some good is bound to come out of the project.

Reality: Facebook

Facebook is the incumbent. It's very simple to use, works well and looks smart. Most importantly, it has critical mass - it has the users. All your non-geek friends - if they have a presence in social media at all - are on Facebook. And because it (deliberately) doesn't interoperate well  with other social media, to talk to those friends you have to be on Facebook.

Facebook isn't technically so shoddy, either. With the amount of money they have to spend they certainly should not be! They do have an answer to Diaspora's aspects and Google+ circles. It's called 'groups', and almost nobody uses it. Why not? The functionality is hidden and the user interface isn't very good. And that sums it up for Facebook, really. It's grown like topsy, without a clear vision of the architecture; it's got the users; and now, provided it offers those mainly non-technical users a consistent and familiar user experience, it really doesn't have to try very hard.

Reality: Twitter

Twitter isn't really trying to out-Facebook Facebook. It's trying to do something distinctive and unique, and has succeeded, probably, past its creator's wildest dreams. Twitter is not a platform for private discussions between friends - there's no way you can prevent a tweet being retweeted by anyone who can see it. Threaded conversations are handled poorly (although they are supported to a degree). And, of course there's that notoriously short post length.

What Twitter does, though, it does well. It does live events better than anything else at present (although Google+ clearly intends to challenge that).

Reality: LinkedIn

LinkedIn is the one of the current generation of social media systems that I've been on longest. So what to say about it? It's social media for business people. It makes a virtue of dullness. Everything on it is dull. The people are dull. Even the people whom one knows are emphatically not in the least dull - like my good friend Andreea - appear dull in their LinkedIn profile. No-one I'm connected to posts anything interesting there except me, and I only do because I have it automatically pick up my Twitter posts. It probably has some excellent features, but frankly who cares? It's dullsville. Dull.

Reality: Niche systems

LinkedIn might be seen as a niche system for dull people. At the other end of the scale, Fetlife is a niche social media system for interesting people. I'm sure there are a lot of other interesting and good niche systems in other areas of life, but I'm not aware of them.

Fetlife is social media for sadists, for masochists, for bondage sluts, for transvestites and for transexuals, for people, generally, of transgressive sexuality. It's unsurprising that many people don't choose to express that side of their personality on mainstream social systems; for many people their sexuality is at least to some extent hidden from their colleagues and friends.

Surprisingly, Fetlife don't use a generic SM engine (sorry!). They roll their own. And actually it's very good. It is in some sense a hybrid between a social media system and a forum/newsgroup system - the main feed shows your friends' activity like Facebook, but most activity takes place in discussion groups which are mainly public (although many are moderated). Look and feel is excellent; performance is excellent; user experience is very good.

Underneath it there's some very solid software engineering, with a lot of thoughtful data-caching and graceful failure code. Specific parts of the system, particularly in these geeky data-management areas, are released as open source, but the core engine has not been, yet. Like Diaspora, Fetlife is under very active development, with a sufficiently small user community (hundreds of thousands, not millions) to allow the developers to be responsive and engaged with their users.

Reality: Usenet

Usenet is a dinosaur - a beloved dinosaur, but a dinosaur - and it is dying. It was - once - very vibrant. It predates not merely the Web but the Internet itself, and its faults have informed the features of the social media systems of today. Essentially Usenet is an anarchic peer-to-peer amalgam of servers owned and managed by different institutions, which share with one another a vast collection of hierarchically organised discussion groups. Some of these are technical in nature, but many are in effect social clubs.

So, what are these faults and how do they influence the design of modern systems?

Firstly, the institutions which own the systems gain little financial return from running them. While some are now run on a subscription model, usually the service is provided for free or bundled in with other services. For this reason the actual owners of the core servers have little resource to dedicate to manage the system and have offloaded that responsibility onto 'democratic institutions' of users. In theory, that seems a good thing, but it results in control being put into the hands of people who have literally no interest at all in the success and vibrancy of the system and are often more interested in playing politics than carrying out the responsibilities to which they've been elected.

The remainder of the problems all come back to that first problem, because in fact all the problems with Usenet are technically fixable - if those with the power to do so had the will.

So the second problem is the lack of verified user identities. It's fairly easy to make a post to Usenet which appears to come from some other user. It's extremely easy to set up multiple posting accounts and to make attacks on other people from behind the barrier of anonymity. There are very few sanctions which anyone can use to dissuade antisocial behaviour, because the antisocial person can easily abandon one identity and reappear with another.

As a matter of design Usenet is a broadcast medium. You can choose not to read certain identities messages (by using a 'kill file), but you cannot choose who can read what you post. You send your message out to a specific group, but you cannot exclude anyone from reading that group. That's not a fault, it's a decision. But I think that modern social media's mechanisms for optionally limiting ones posts to circles of known, chosen people is a consequence of learning from the social consequences of this aspect of Usenet.

Finally, Usenet's user interface is horrible. To be precise, there are many client applications which provide a frame to Usenet, and those have better or worse user interfaces. But under it everything is plain unstyled text geared to a mainframe era 80 column teletype. For the upcoming generation, it's simply bizarre that anyone should choose to read such ugly and poorly formatted text.

Reality: Google+

Google+ is the new contender. It has learned a lot from a series of relatively-failing social media systems which Google has tried. It's feature-rich and slick. It looks very good. Features are mostly intuitive and easy to find and to use. It has all of Facebook's functionality, all of Twitter's, most of Diaspora's; and, in addition, a built in realtime chat/video conferencing system which is significantly ahead of the Facebook chat system (although I don't yet know how useful this is). Furthermore, it integrates seamlessly with Google's existing mail, talk, photo storage (picasa) and calendar systems and I suspect will soon integrate with Google's groups system.

Google is putting a lot of investment into this, a lot of good design and a lot of infrastructure. It already all works - and it all works fast.

Of motivation and potential

I've said already that I think Diaspora is the best of what's out there. I've said also that I don't think it will win. Why not?

Microsoft have, over the past two decades, captured and held the world market in word processing. A major part of the reason is the appropriately named 'network effect' - in order to exchange documents with people who use Microsoft's word processor, you have to use Microsoft's word processor yourself. Markets with network effects are intrinsically monopoly-generating. It's in the interest of commercial companies to seek to establish monopolies, of course; they're profitable. Governments can seek to oppose this - and, in the case of social media, easily could by requiring a common, open, discussion interchange API - but Governments move extremely slowly and this market moves fast.

To provide a useful service you have to capture the majority of users within a population; otherwise, your users' friends will not be on your system and your users won't be able to communicate with them, so will move to other systems where their friends are. This is the 'network effect'.

To capture the majority of users, not only do you need to have the buzz and brand stature to capture a lot of interest quickly, you also have to have the infrastructure to provide responsive performance for all the users who choose to adopt your system. You have to scale very fast. And it helps if you have a lot of users to start with.

To pay for that infrastructure, you have to have a means of monetising the system. The infrastructure does not come cheap. Diaspora's model does allow some nodes to carry advertising. It does allow some nodes to be paid for by user subscription. But it also allows some nodes to be free and without advertising, and there's a lot of users who would prefer to use such nodes. Many such servers will be run by amateurs, will have low power and low upstream bandwidth. There's no way obvious to me that Diaspora can load balance. Slow servers in the network will, I suspect, slow the performance of the whole system.

Furthermore, there's no powerful central body with sufficient media clout to promote the system. Ten million people have joined Google+ in a fortnight; I doubt that number have even heard of Diaspora, in a year.

The Diaspora team's motivation is to build cool software. They are building cool software; they're succeeding. But their other motivation is to provide people with the means to keep their personal data out of the clutches of the big corporations. Both Google and Facebook have the same fundamental: to aggregate data about every person on the planet in order to be able to sell us as market-fodder to other corporations. This is lucrative; it makes money. It makes money which helps Google and Facebook hire good engineers and build good infrastructure which can support a fluid, responsive user experience.

Knowledge is power; knowledge about us is power over us, and there are very real civil liberties issues about corporations having so much power. As we've seen with News International, for one corporation with its own political agenda to control the majority of media through which flow the key day to day information within a polity is inherently corrupting.

Which brings us, finally, back to Diaspora. Diaspora offers a mechanism to support a distributed and potentially heterogenous cloud of 'pods' each providing a home for different groups of users. No one organisation controls, and, provided Diaspora have got their security right, no one organisation can see the data of all the users. But all the users on all their different home 'pods' can nevertheless collaborate with one another, share the information they wish to share with the people with whom they wish to share it, and engage in conversations as they choose. They can also discover their friends across the Diaspora cloud.

As I've argued above, there's no technical reason why Facebook or Google+ could not configure their service to appear to the Diaspora cloud like just another Diaspora 'pod'. Then, users on Facebook could share information with, and join in discussions with, their friends on any other social network which also conformed to the Diaspora APIs (including, obviously, Diaspora). Such an arrangement would also make it easy for users to choose to migrate their 'home' social media system from one provider to another - you could move all your contacts, all your circles, all your pictures from one home to another, and continue to share and converse with the same friends.

There's no techical reason. But there's a strong commercial reason why Facebook, at least, won't do this (and why Google probably won't either), unless they're forced to. Everyone uses Facebook, because of the network effect: your friends are there, so you have to be there. But overwhelmingly we don't feel warmly towards Facebook itself. The company is widely distrusted, and the software doesn't seem much loved. If it was easy to move, if it was easy to share with friends on Facebook from outside Facebook, Facebook would lose users; probably, quite a lot of users.

Google, on the whole, is still a more trusted brand than Facebook. But the days of 'don't be evil' seem to be fading into a more naive past, at least in the public perception. I've started to be concerned about how much Google know about me; not because they now know more - they've known a very great deal for very many years - but because I now trust them less. And I think many other people trust Google less than I do. They, too, stand to be losers from a more heterogeneous social media landscape.

Government - either the US or the EU - could, of course, require major social media systems to open up, through a mandated open sharing API possibly derived from Diaspora. That's what would, I believe, happen in an ideal world. If it does happen in the real world, however, it's likely to be botched and too late. Government processes do not operate at the speed of modern technical innovation.


Which means, in the end, I come down reluctantly on the side of Google+. It offers a slightly more fluid user experience and a slightly wider range of slightly better functionality than Facebook, from a slightly more trusted monopolist. Most of my friends are not geeks. My chances of dragging my friends to Diaspora are slim, because they won't find their friends there. And, if there were to be a mass migration from the corporate systems to Diaspora, I think it would be extremely hard with no central organisation and no significant revenue stream to make the system scale. So I think Diaspora will probably lose, and that Google+ will probably over time supplant Facebook as the place where tout le monde et sa femme hang out.

No comments:

Creative Commons Licence
The fool on the hill by Simon Brooke is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License