Why did JANET Talk fail?

A couple of years ago JANET(UK) announced a pilot VOIP service for UK HE, JANET Talk. The pilot has now finished, and will not be extended to a production service, as there is little apparent demand for it. It was obvious to me from the start that JANET Talk wouldn't be a success. I remember pointedly suggesting as much at the Networkshop session where it was launched. The reason? Network effects - or the lack of them.

Imagine you owned the first telephone in existence. It is completely useless - there is nobody to call. It only becomes useful when someone else bought the second telephone. More precisely, as a telephone network grows, its usefulness increases in proportion to the number of potential connections between people in the network - the square of the number of people. 

JANET Talk was only for people in UK HE to talk to others in UK HE. In todays world - a networked, global world - there is no way that a communications service for such a small group of people will be successful. It will inevitably be overtaken by other services with a larger network of users. People will choose to join the larger, more successful service, as that is what their contacts already use.

There are ways you can let a small network benefit from network effects - by connecting it to other networks. The Internet itself came into existence as exactly that - a network of networks. For a VOIP service, the best way to do this is to allow calls to and from the PSTN. Unfortunately JANET Talk was restricted so that you couldn't receive calls from the PSTN. JANET did this deliberately. They don't want to become a public telecommunications provider, as that would potentially open them up to more regulation. It's a very understandable decision from their point of view but sadly it made JANET Talk irrelevant.

Why is this important? Network effects don't just apply to telephones. Everything these days is networked. Web 2.0 is all about social connections & communications. The same factors explain why only a minority of colleagues at Bristol use our calendar service. It's internal only. This is no help to you if if you collaborate with people outside the University. Instead people choose to use Meetup.com - a free, web-based service which individuals adopt of their own accord, with no involvement from the institution.

I've instinctively understood network effects for years. I learnt my lesson launching ResNet Chat. I watched the small, lonely procession of users log in to the chat room one by one, saw the tumbleweed howling through it, and log off, never to return again, while MSN beckoned brightly. 

Don't let it happen to you. When planning a new service, see if it has built-in positive network effects. It is doesn't have these naturally, find a way to connect it to larger networks so it can benefit from theirs. If you can't find a way to do this then you are dooming your project from the start. You're better off doing nothing, unless you want to see your service become irrelevant, pushed to one side by a larger, more popular one.

(image of network effects courtesy of Wikipedia)