Fixing our social sharing tools

Have you ever decided not to share something even though you thought it was awesome? Maybe you found a great article breaking down a complicated conflict in a foreign country, or a Particularly Funny Internet Cat Video, or a thoughtful piece about the economy. You wanted to say “I enjoyed this,” but you didn’t, because you were worried you might offend some friends, or lose a few followers.

Every day I see Twitter users lamenting or belittling the followers that left them because of too many political tweets. Or bickering among old friends on Facebook over a blog post about gun control. The problem is that our social tools are fundamentally broken, and it’s shaping the way we represent ourselves online.


Many of the social tools we use today are designed as binary waterfalls: Either you’re in or you’re out. If I want your tweets about web development, I also get your tweets about politics. If I want your status updates from our hometown about our high school friends, I’m also opting in to your opinions about the latest episode of Real Housewives of Wherever. The problem with our social tools is that if you want to share something publicly, you must share with everyone.

Google attempted to solve the binary waterfall problem with the Google Plus “Circles” feature, but they ended up with a targeted sharing platform instead of a great social publishing platform (and maybe that’s what they wanted, as limited as that may be). On Facebook, they make the flood of information more palatable with a system called EdgeRank, which tries to guess which content you’ll “like” more. Twitter doesn’t try to filter anything. Instead, they leave it up to you to curate your feed by following or un-following.

Social tools also fail to recognize that individuals change. Our interests change over time, but our tools don’t automatically adjust without manual input from us (This was recently highlighted by Matt Haughey). The people you friended on Facebook 5 years ago are still there in your timeline, even if you rarely talk to them. This issue is possibly one of the reasons why social networks can be so volatile: You add the people you are friends with at the time, but after a few months, the people around you have changed enough that visiting the site is no longer interesting to you. Twitter leaves it up to you to curate your stream by making it easy to follow people who are interesting right now or un-follow people who tweet too much about things you don’t care about. As mentioned earlier, Facebook is trying to solve this problem with their EdgeRank algorithm. Rather than have you un-friend people, they try to show you things they think are relevant to you right now based on various signals like what you’ve recently “liked.”

The last problem I’d like to highlight is more related to social sharing sites like Reddit or Hacker News or the original version of Digg. These are sites where submissions fight for space among the homepage by gaining votes from the site’s visitors. In theory, it’s a fun idea: Social filtering brings most popular content for that day to the top of a list. Unfortunately, as these kinds of sites grow in popularity, you end up with two problems. First, the content on the homepage grows increasingly more homogenous in order to appeal to the broadest possible number of users. In order for a post to reach the homepage, it has to get more votes than all the other content, so niche subjects start to disappear from the top of the list. Second, these sites often have commenting systems where everyone can comment in the same space. With a small audience this is fine. You might recognize some of the names of the users leaving comments and even establish a rapport with them, but as the number of users grows, your monkey brain can’t keep up with the thousands of users. At that point, the users might as well be anonymous and some interesting group dynamics start to kick in (Clay Shirky wrote a fantastic article on the subject). This behavior can be partially summarized as: Normal Person + Anonymity + An audience = Total Fuckwad.

So how do we fix these issues?

First, our social tools need to recognize that people are complicated. We have many friends of varying closeness and many interests of varying intensity, and trying to communicate all of that through a single output isn’t natural. Paul Adams has compiled a bunch of fantastic research on how people interact with groups and has even written a book about it. Google Plus interpreted this research and came up with the idea of “Cirlces,” a tool that lets you categorize all your friends into groups in order to share things with them. Circles turned out to be an interesting way to share things with people you already know, but it takes a lot of work to maintain, and it doesn’t easily let you share with strangers whose interests are similar to yours.

What Google should have realized is that the important part about sharing content online is not who you share it with, but who you share it as. We all have various personalities. Mine might be my work personality, my photographer personality, my hometown-highschool personality, my video gamer personality. These interests are bigger than my small group of friends who also share these interests, but it’s really, really hard to express my various interests online without managing a bunch of distinct social networks. Our social tools need to allow us to share whatever we want, whenever we want, and not worry about pissing off our friends and followers.

Second, social feeds need to be more dynamic. The people I interact with are changing all the time, so why should my social networks be comprised of a rigid list of people? Alexis Madrigal recently said “These tools are only as good as the network you create on them,” but requiring me to constantly curate the people in my networks doesn’t seem like a fun way to spend my free time. Our social tools should be smart enough to know who and what we like and be able to adapt automatically over time. Facebook is making great progress in this area, and I’m really excited about what Google will do when they start to integrate Google Now-like features with Google Plus.

Last, we need to stop building tools that lump everyone together in one big group. Sites like Reddit and Hacker News and the original version of Digg are guilty of this design flaw. The result is that the Reddit homepage ends up being full of meme images and other one-off joke content or other widely accessible things. Of course you can customize the page by choosing categories, but that requires work—Reddit already knows what I like, why not make it automatic? Reddit’s attempt to solve this problem are Sub-Reddits, which are just sub-categories, and categories fail to address the root of the cause and instead treats the symptom. As a Sub-Reddit grows, the same large-group dynamic occurs and the content again shifts towards the more homogenous submissions. There’s a recurring discussion on Hacker News about how to “fix” the perceived lack of quality and politeness in the ensuing discussions, but I’m convinced that the answer isn’t buried in some fancy algorithm. I think the solution we need is more fundamental than that. We need to design our social tools with human behavior in mind.

It’s just not natural for humans to interact in huge groups. Our brains have evolved to handle only a few hundred social connections, so when we are thrown into an environment where we are interacting with hundreds or even thousands, we might as well be hanging out with a giant anonymous mob. That makes us not care about the people we interact with, and we might even change our behavior without realizing it.

The good news is, not all the social tools are broken. Two sites stand out among the crowd to me at the moment are Pinterest and Medium. Pinterest solves the “binary waterfall” problem described above: If someone is sharing pins that I don’t find interesting, I can easily un-follow the board they are posting to without un-following them entirely. Pinterest has succeeded where other social sites have failed because they let us share all of our interests, not just what our followers might like. The result is that we share more content.

Medium is another site that is doing interesting things with content categorization. When you post content, you are forced to post it to a “collection” (the same way that Pinterest forces you to pin to a “board”). This results in loosely structured categories that contain ranked lists of content. I think it’s too early to tell if this model will be successful, I think it has a lot of promise and I’ll be watching the product closely as it evolves.

Both of these sites have made fundamental product decisions that allow them to scale their user numbers very effectively. Rather than scaling vertically, like Reddit or Hacker News, Pinterest scales horizontally like Twitter does. The total user base can grow and grow while your own network of friends and followers can exist on their own without much adverse effects (the exception is their global categories, but those aren’t the main focus of the site).

All of this thinking about social sharing led me to try and build a tool that puts some of these ideas into practice. I’ve teamed up with a friend and ex-coworker from my YouTube days to build a new sharing site that we’re calling Personafy. Personafy lets you create a set of personas and share links “as” that persona. You can then build a distinct audience for each of your personas. When sharing on Personafy, you no longer have to worry about whether or not your social network would enjoy a link you want to share—if it doesn’t fit in to any of your existing audiences, you can easily create a new persona and share it, and perhaps find some other users with similar interests in the process.