Skip to content

Yoyo Code
MatyΓ‘Ε‘ Racek's blog


How to build better social media?

If you look at social media sites these days, they all boil down to basically the same thing: People post text, images, videos or links, and there are likes and comments (and sometimes other reaction-like things). That's basically it.

The question is - why are some sites more harmful than others when they all work basically the same way?

It seems like some slight differences in how they are set up play out differently at scale (like a butterfly effect, which is probably not a far-fetched analogy)

The differences can be huge. As I've written the last time, I had to delete Twitter, because it was unbearably bad for me. But Instagram is totally fine, I don't experience much harm using it.

You might object to this, and that's another perplexing problem - everybody has a different experience with each platform. Some people have much worse relationship with Instagram than me. Seems like the individual is part of the equation, too.

This is a topic I'm interested in, partially because I work on a social media site, too.

I can see some of these things play out on our platform. Seemingly minor choices in user interface design can shape people's behaviour on the platform, sometimes causing unexpected conflicts.

I'm trying to find some common patterns to see how can we set up a social media to improve the outcomes. Here's a few ideas that I think could help. Most of them are vague and imperfect - social media platforms are complex systems, so there's rarely something clearly good or bad. In most cases the outcomes are unpredictable, and we have to experiment to find the optimum.

1. Focus helps πŸ”—

Few observations from my experience:

  • Discussion in narrowly focused Facebook groups tend to be better than discussions on general posts
  • Communities formed around a specific hobbies seems to work well
    • Game related forums
    • My instagram, which is basically only yoyo content
    • The app I'm working on, which is focused on learning skill toys
  • High reach posts have crazy comment sections (also applies to news sites)
  • Twitter is very general - anybody can tweet to anybody and the whole thing breeds negative interactions
  • YT comments on small videos are generally better than comments on large ones
  • Reddit is organized in groups and I tend to find it mostly good
    • size of the group matters, though.

2. Require effort for contribution πŸ”—

If the platform incentivizes doing things that require more effort, it's often better. Here I'm thinking about YouTube. Making videos is expensive and YouTube rewards things that require more effort. I generally have good experience with the platform.

Contrast this with Twitter, which encourages you to post whatever is currently on your mind without thinking twice about it.

There are a lot of exceptions to this rule, though. People can spend a lot of effort on something negative - I definitely remember spending hours on writing a single negative comment, for example. People write very elaborate 'personal attack essays' or even make mocking videos.

3. Facts over opinions πŸ”—

If the media is set up to highlight some objective reality over personal opinions, it seems to be better. Here I'm thinking about Instagram, where main focus is photos and videos, and especially captured at the moment.

Notice how on Instagram, the vast majority of your screen is taken by the picture (or video), and only a tiny part is the description, usually not even fully visible. Comment section is often completely hidden, or shows only a single comment.

This greatly reduces the amount of judgement and reality distortion that people often do and puts their contentious opinions to the background.

Of course, some reality distortion is inevitable. There's always a selection bias with what people decide to share and a possibility of editing. Nevertheless, encouraging photos over text seems to make it less likely.

4. Real world action focus πŸ”—

If the media is directed towards some real world activity, it tends to work better. Think Facebook events or Meetup.com.

I think this has to be a big part of the answer to how do we handle social media as a society in the end. In my opinion, if the media is only a tool to aid some real world activity, then it works the best. Whenever it's too inward focused and captures too much of our attention, it seems to go bad.

5. Positivity over negativity πŸ”—

From my experience, social media has a bigger potential for negative social interactions than the real world at the moment. In real world, we have more bandwidth, more information about the individual and the situation, body language and a ton of social norms.

On social media, a lot of these default negativity blockers are stripped out. The bandwidth is narrow, context is lost or changed over time and the information shared is more permanent, so it's less easy to forget, forgive or tailor your reaction to the situation. There's also some negativity bias at play, where people often interpret comments and information more negatively than it was intended by the poster.

For social media, I believe that we should have much stronger bias towards positivity. I'd actually go even further and say that social media should heavily disincentive negative posts and comments, or even things that can be easily interpreted negatively even if not intended to. This might seem too heavy-handed, but we have to think about the large scale consequences of the environment we create, not only the individual interactions.

6. Norms and rules πŸ”—

In real life, we have a lot of implicit rules for common social interactions. I think we should have this for social media, too, but we should expect those rules to be different, because the medium is different. The same way, the real world rules are different based on context. Things that are ok in a swimming pool are unacceptable on a classical concert.

For example, I have a rule that before I post anything negative, I'll wait till the next day. I also include emojis in messages whenever there's even a slight chance of misinterpretation by the receiver. This helps to avoid unnecessary negativity.

Almost every Facebook group, forum or Reddit community has some specific rules to avoid problems. This seems to improve the situation.


Closing thoughts πŸ”—

None of the ideas are perfect or solutions by themselves. In the end, those sites are complex systems, and those are naturally chaotic and unpredictable. We should expect there to be no simple solutions to their problems.

You might notice that most of the ideas here are more about people's interactions and less about the platforms themselves. I think it's way too common to blame the platform, but in my experience, most of the problems on those platforms are caused the users. The platform only plays a role of a substrate for certain actions, which are ultimately taken by people.

I think that as a platform developers, we can't take responsibility for people's actions, but we have to take responsibility for giving them good options and propagating their consequences responsibly. It's like when you design a road - if you make the sidewalk too narrow, you're increasing the probability of accidents, so you have certain amount of responsibility for that, even though you didn't participate in the accident.