Making billions of dollars from the federated social web

August 28, 2012 | 4 comments

Diaspora was pretty exciting. A Kickstarter campaign that promised it all: a platform created “for everyone to have full control over their data and to empower people in to become responsible, secure, and social Internet dwellers”. They raised over $200,000, which at the time was the largest Kickstarter funding round ever. Neat!

Yesterday, Diaspora announced that they were becoming a community project. That means that the wider open source community will be responsible for its direction; the original team will be involved to some extent, but are now involved in projects like, which just made its way through Y Combinator. ( is released under the same open source license as Diaspora: here’s the GitHub repository.)

The federated social web is a holy grail for many of us. It’s a simple idea: rather than ceding control of our data and activities to a monolithic, monopolistic third party, why don’t we create an open market for services we can use, which can all interoperate? Those of us willing to set up our own servers would own our own data; others would use the servers set up by their schools or employers; others still would join the hosted service of their choice. The result would be a more innovative, web-like application web, where niche interests could be better served and people could choose the applications and interfaces that best suited them.

It goes far beyond Facebook-style social sites: the potential is for deep interoperability between every application on the web and beyond – and, potentially, identity spaces where we could keep our personal information safe.

Back in 2005, I tried to make the case in a letter to the social web hating journalist Andrew Orlowski:

If a thousand sites depend on Flickr, what happens if Flickr goes down? Wouldn’t it make a great deal more sense to think about standards for data transfer and availability to allow for Flickr-like data all over the Internet, rather than in one place – effectively a peer-to-peer network (or networks) of tagged resources? That way one can still grab information and manipulate it, but with zero dependence and a high tolerance for network failure. Rather than there being one nebulous Microsoft Office 12 For Web, there could be ten thousand office applications that all use the same standards and allow for data transfer perhaps with storage services elsewhere. The answer to “how bad do you think it will get” is “exactly the same as now” if we allow one service provider to create a monopoly for a particular type of service. As ever, the solution is open standards with a framework that anyone can use. This is obviously not going to lead to billions of dollars for any one provider, so it’s probably up to the open source movement to create.

Returning to that letter seven years later, I think I was wrong about one thing: I believe it is possible to make billions of dollars from interoperable social web applications – and there’s no need for mass public adoption of apps using federated social web standards to do it.

I’m convinced that business software should be using federation protocols, as long as those protocols have built-in access permissions. Imagine if your organization – a government department, say, or a corporation – could selectively create shared spaces for joint projects with other organizations, for as long as the project existed. Imagine if you could collaborate in more fluid ways, where applications were more like documents that you could introduce into a project when you needed them. Imagine if email was smarter. Whereas decentralization in public consumer apps is an ideological feature, in enterprise apps it has a solid business case and adds measurable value. Businesses will become more profitable by using federated apps.

On hearing the news about Diaspora, StatusNet’s Evan Prodromou made a very sensible plea:

I only have one favour left to ask: please, for the rest of us who are still working full-steam on federated social networks, don’t fall into the comforting fiction that the problem is insurmountable. It’s not; it deserves our attention and support. You didn’t waste the last two years on something pointless and unattainable. Your work matters.

It might sound like hyperbole, but I still believe it could change the way we do business, and ultimately how we all communicate.

Update: In the comments, Johannes Ernst points to his post about why decentralized software is 10x harder. I think he’s right about the problems that need to be overcome. But check out his note at the bottom: watch this space.


August 22, 2012 | 6 comments

Tent appeared out of the blue today: a protocol and reference server implementation for individual-to-individual distributed social networking. Or to put it another way, Tent is a way to host your own social data – posting and reading from as many apps as you want. Here’s their announcement, and here’s the GitHub repository.

The Tent manifesto is right-on:

Every user has the right to freedom of expression.
Free speech is a necessary feature of all open societies. Speech can not be free if communication is centralized or intermediated. Users must be able to say anything to anyone they want on their own terms.

Every user has the right to control their own data.
This includes who can access the data they create and how that data is later used.

Every user has the right to choose and change their social services providers.
This includes the right to negotiate reasonable terms of service collectively or individually.

Of course, this is hardly the first open source social networking product – and many people are already asking why Tent doesn’t use the OStatus protocol. (StatusNet also includes an individual installation mode.) These are valid questions, but while there’s a slight air of Not Invented Here Syndrome, it’s an elegant idea and the API is very clean and simple, which means there’s every chance an app ecosystem will emerge. If any one of those apps is simple and elegant, we may see a very different kind of social networking community begin to develop.

Even more interestingly, I also think there are real commercial implications for this protocol. More on those in another post. For now, my takeaway is: Tent has the potential to disrupt the entire social web.

Building a distributed social network? You’re doing it wrong.

June 4, 2010 | 7 comments

Here are some distributed social networking platforms and technologies designed to facilitate distributed social networking:

Wow, that’s a lot! And following Diaspora’s flurry of both coverage and cash, you can bet there’ll be plenty more to come. But of all the projects listed above, I’d argue that only is orientated around consumer need. As a result, it’s the one most likely to survive, become self-sufficient and prosper. Several more – including DiSo and DSNP – are seeking to build out technologies that can support such products, rather than the products themselves. DiSo is certainly working with other vendors and projects, is full of super-smart people, and should do very well.

However, the others are basing their product on ideology and technology rather than a human use case. I worry that a lot of these projects will disappear – which is a shame, because they’re all doing great work.

Here’s a use case distinction I’ve been thinking about:

  • A social networking platform allows you to communicate and share with a specific group or community.
  • Distributed social networking software allows you to store and organize your own content and – optionally – share it with whoever you like.

Or to put it another way, in social networking platforms, sharing is the feature. In distributed social software, sharing is a feature. The two use cases are genuinely different: rather than being a competition between “monolithic” social communities and distributed social software, they’re used for different things. There is a place for both in the ecosystem – and there’s no real reason why they can’t work together.

As I pointed out in The Internet is People, in order to be successful, any social software you build either has to plug into an existing community, or be useful for the first user who joins. In distributed social software, you only ever have one user: distributed sharing, then, should be a piece of infrastructure that can be plugged into any kind of software.