I have been an enthusiastic user of twitter for quite some time now, but I am rapidly coming to the conclusion that it is unlikely to survive, at least in its current form. This is partly because they will struggle to build a business model to start paying off their venture capital backers. But more importantly, it is because twitter is a closed system and that will ultimately constrain its potential.
For those of you not familiar with twitter, it is the most popular of the so-called “microblogs”, along with pownce, plurk, jaiku, kwippy and many others. While these sites vary in their specific features, they are all based around the idea of allowing users to quickly publish short messages (often only 140 characters to fit in a SMS message) viewable by all of their friends online. For users of Facebook, just imagine a site with status updates but no photos, vampires or scrabulous.
Ever since twitter launched, making its services available to users for free, people have wondered about their business model. There is a web school of thought that you should only worry about making money after you build a critical mass of users and Facebook currently appears to be an example of the wisdom of this approach. Twitter has certainly grown rapidly and now boasts some high-profile users such as Barack Obama and 10 Downing Street (the office of the UK Prime Minister). However, its current estimated user base of around 2 million is still well short of the size required to generate adequate returns on the capital invested in the company so far. The recent blog post Why Twitter will be Sold in a Fire Sale estimates that $20 million has been invested in twitter to date but, even by tapping into a wide range of revenue generating alternatives, the most they could hope to generate from their current user base is around $17 million. So they need a lot more users, and yet their growth rate appears to have slowed of late* and the frequent performance problems the site experienced in July this year cast doubt on their ability to cope with much more growth. The fact that twitter recently had to pull its international SMS service is an indication that financial considerations are starting to bite.
However, the real problem with twitter is that it is a walled garden. The current microblogging environment is like the early days of email. If you were a Compuserve user, you could send messages to other Compuserve users, but not to people on other networks. It was only once email was able to work across different networks that it was able to start realising its potential and become so entrenched in our online lives.
Instant messaging (IM), a successor of sorts to email, never managed to move beyond its initial fragmentation and remains split across AOL, Google Chat, ICQ, iChat and many others. Attempts have been made to bridge the gap with IM clients like Pidgin that work across multiple protocols, but these were really kludges that left the underlying problem unresolved.
IM history is repeating and today if you are a twitter user, your messages are not visible to pownce, twitter or jaiku users. Again the demand for cross-platform messaging is there and this demand has been partially met by sites like Ping, which allows you to post to multiple microblogs and FriendFeed which can receive updates from multiple platforms. However, the real answer lies in following the email example and developing an open protocol for microblogging. Luckily, that has started to emerge in the form of Evan Prodromou’s OpenMicroBlogging specification, which he has implemented in the open-source Laconica application. Just like an email server, Laconica can be installed anywhere to create a local microblog, but users on the local system can also connect to other microblogs that implement the OpenMicroBlogging protocol. This works in much the same way that you can send email not just to colleagues in your own firm, but to anyone anywhere in the world with an email address. Evan uses the phrase “federated servers” to describe this approach.
It is still very early days for Laconica, which was only released a few months ago, but it is already gaining momentum. So far the biggest installation is Evan’s own identi.ca site, which picked up quite a few twitter users in July when twitter was experiencing repeated outages. Like many others who stumbled upon identi.ca at the time, I intially thought it was just a twitter clone with fewer feaures (back in July it didn’t even support replies). But not only did Laconica (and hence identi.ca) start adding features at a furious rate, but I started to appreciate the advantages that the federated approach brings. Here are just a few:
- Scale: messaging is spread across a distributed network of servers (just like email) rather than having a single point of failure (like twitter).
- Privacy and Control: web 2.0 evangelists rightly point to the benefits of social tools like twitter, but many organisations would be reluctant to use a public site like twitter for internal discussions. Instead they can have an internal microblog which can also interact with the outside world.
- Enhancement: if twitter, pownce or plurk is missing a key feature, you are at the mercy of their developers for getting it implemented. Laconica is open source, so you can always add that killer feature yourself, although the development community is likely to do it for you.
An excellent way to get more of a sense of the enormous potential of the open model, is to have a listen to this FLOSS weekly podcast in which Leo Laporte interviews Evan Prodromou. Leo was so impressed by what he heard that he immediately launched his own microblog on Laconica, the TWiT army.
Whether it is the OpenMicroBlogging protocol, or something very much like it, an open approach to microblogging is clearly the way of the future. Of course, twitter could always respond by adopting the OpenMicroBlogging standard themselves, which would be great for their users. Unfortunately for their venture capital backers, that is likely to make it even harder for twitter to turn a profit.
UPDATE: laconica is now known as status.net.