Jun 182014

When my friend Warren posted his issues with commenting on posts shared across social networks, it got me wondering how commenting would work in my hypothetical “next-gen” social network. My initial write-up on the idea focused mainly on the idea of the main content itself, so it got me wondering how to integrate things like comments into a feed-based social network. As I thought about how comments might fit into this new concept, I also started wondering about just what purpose comments served in a social network. I don’t mean in a “don’t read the comments” sense, but rather what role are comments serving in a social network?

What are comments for?

To be honest, I don’t really have a good answer for this. Is it to support and encourage discussions and conversations? That doesn’t seem likely, just look at the tools for adding comments and sharing them. The editing support can be sparse, at best, not to mention the visible space with which to write them. Have you tried to follow a long thread on Facebook, Google+, or even Twitter? If so, it starts to become painfully apparent that back-and-forth is not what those apps were built for. If that was their intended purposes, these apps would look a lot more like a forum than what they do now. That tidbit opened up another question: what role should comments play in a social network?

Thinking about this got me wondering about whether the primary focus of social networks should be on back-and-forth conversations or the creation and sharing of content. Were social networks doing things wrong, and should they look more like a discussion forum, or is the current idea of focusing on the original post, with comments as add-ons, correct, in which case I just need to figure out how comments should work in my RSS-based social networking world. To try to come up with as correct an answer as I could, I wound up having to step further and further back until I ended up asking myself a very basic question that I had utterly neglected to ask in my original screed on the idea of social networks:

Just what is a social network for anyways?

Ultimately I realized that one is never going to design a “good” social network without first focusing on the root problem that social networks are trying to solve. As I thought about it, I realized that social networks are nothing more than a model of people and how they connect to 1 another. That’s it. Everything else is merely an app built on top that model. Those apps are what we’re constantly referring to as a “social network”. Facebook’s “social graph” is the actual social network, Facebook itself is nothing more than an app using that data. The same goes for Google+, Twitter, Pintrest, et. al. That’s also why people think these apps are valuable. It’s not because of the stuff being posted or discussed, it that’s map of how real people connect and how content moves amongst them that gets people to see dollar signs. You can tell Facebook understands this, with their acquisition of Instagram, building their Stories app, and promising to stabilize their API. It’s part of a trend towards letting users use whatever app they want so long as the model of who they are and how they connect with other people flows through them.

This raises even more interesting questions. Right now, social networks are nothing more than apps that let people post/share content, with the ability to add a few small notes. But getting back to 1 of the things that I was excited about with my idea on a new social network, what if we built a social network that was actually a network, not just an app sitting on top of yet another proprietary data model? What if we just built a good social network with API that other people built apps on, whether they’re like current social networks, apps that look more like forums, apps for photosharing, etc. No matter what app you use, you’d be on the same social network, and the updates from all of the apps could appear in whatever you’re using to read the RSS being spat out.

Yes, I know that’s already been tried

Before anyone points out the obvious, I’m well aware that what I’m proposing is basically what App.net tried, and I’m also aware of how that worked out. I’m maintaining that this was caused by their initial decision to charge everyone for access to the network, including users. Although they opened up a freemium tier, there were still limits on how many people users could follow before having to pay the subscription fee. As I said in my idea for a new social network (really just an app on a social network), charging regular users for access to your service isn’t something I believe would fly. App.net is one of the reasons I stand by that belief. Sure it’s still up and running and being maintained, but it doesn’t have the level of support needed for continued development, which makes evolving the platform incredibly challenging. So far, I’ve only come up with 3 models of development that a “pure” social network could use to stay afloat.

Anonymized analytics

The first is just biting the bullet and building in some type of anonymized data analysis for businesses to use to try and target ads. It flies in the face the idea of a completely ad-free social networking app that I originally discussed, but there a reality is storage space, hosting, electricity, etc. aren’t free, and people aren’t going to pay a monthly fee for a social network. Besides, given that this model is the valuable part of a social network, this is what businesses really intend to pay for. At least this way a couple of things can be guaranteed: companies aren’t getting your identity without your consent, and advertising still isn’t completely baked into the network itself. Instead, all the network can do is try to tell businesses what characteristics they should try to target, and try to predict how well targeting ads specific groups of people on the network would work. Since this is a pure social network, it could also help companies know what types of ads work best with what groups (i.e. who responds best to pictures, long posts about how they built something, etc.). Besides, even if ads weren’t part of this network, businesses would be clamoring for ways to do this type of analysis, and paying top dollar for it. It’s best they get an analysis tool that doesn’t expose individual user’s data. Individual app makers using this network would still be making the choice of whether or not to make ad-supported applications, but this keeps the barrier for building on this hypothetical social network as low as possible.

Developer subscription

Another way of sustaining a pure social network could be to charge for developer SDK access. I know my original proposal was to charge businesses a subscription and them just leave them to it, but that was for an app on top of a social network. Now that we’re talking about the actual social network itself, so we have to adjust who pays the subscriptions. This keeps the fees off the backs of the users, but now makes it harder to build new apps on this new social network, which is already hard enough to get off the ground because people who sign up for social networking apps have probably already signed up for at least 3 of them already. Still, this keeps ads off the network itself (although app-makers can still put them into their apps) and keeps the social network itself more “pure”.

FOSS-fully hope for the best

The last possible model I see for this sort of thing is a free and open source social network that the people building the apps on top of it maintain along with their apps. Basically this approach is built around the idea that enough people are building apps on this network that they have a vested enough interest in the network to keep it maintained themselves. This seems to be along the same lines as where App.net is heading, and I have no idea how that’s going to work out, although I hope it goes well. The downside to this approach is that it’s predicated on massive developer interest in a shiny, brand-new yet-another-social-networking app. That’s not really the kind of thing you’re going to have out of the gate.

In the end, I think a pure social network would need a combination of the FOSS model and either a developer subscription or anonymized user profiling to be viable. Most likely it’ll have to combine with an ad model. With all the social networking applications already out there and already at critical mass, any barrier to entry on a new is basically a kiss of death. That means adding costs to building apps on a pure and open social network is out. However, even if this network becomes ad-based, an open-source model allows people to audit just what of their data is being handed over to advertisers and how the rest of it is being protected. This lets the user data export be subject to audit, and helps add an extra layer of user privacy protection, despite the data being mined to better target advertisers.

So now what?

Thinking about building a pure and open social network, instead of just a social networking application, raises a lot more questions than it answers, even before getting into the technical issues how best to model this thing. 1 of the biggest questions, which I have no real idea how to crack, is how do you convince people to sign up for yet another 1 of these networks? I suppose using this pure social network as an ID service (similar to the way Facebook, Google, and Twitter let you sign into other applications using your account on those services), as well as a contact recommender could be a good start, but that’s still a hugely significant source of early friction. However, by building a pure social network, and making it open and accessible to anyone, opens the door to tons of apps, and app ideas, without having to try to do everything in 1 application. Want an app built around deep diving into discussions with your friends? No problem. Want another app built around posting pictures to family members? Sure thing. Have a new family member finally sign up for this network? No problem, they can already see the pictures simply by the nature of the connection between you and that person.

Want to believe what’s being said about protecting user privacy? Read the code, and/or what people are saying about the security of user data in that code. Want to understand just how sharing works? See the code, and even download a copy and step through it yourself. Curious about how ads are targeted? See the source and any analysis posted by anybody online. This network would be completely open source, so there aren’t going to be any mysteries about what data it collects, how it operates, or what data gets shared with advertisers. It’s a pure social network, with open, well-defined output, which means anybody can use it to build any social app they’re interested in.

It’s time social networks were actual networks, not just silos. It’s time we focused on better modeling our real-life social networks, not on making a good “app”. It’s time the app possibilities were wide-open, and the social data they use easily exportable between applications. It’s time we could publicly audit how social networks deal with user data. It’s time for a “pure” social network.

 Posted by at 5:30 PM