Building a better social media messaging site

Published on Friday, January 8, 2021 By Brad Wardell In Personal Computing

Given recent events, I was thinking that maybe there's an opportunity to rethink social media platforms entirely.

The current leaders like Twitter and Facebook are actually fairly primitive from a technical point of view.  Not their fault. They were built when they were built.

If you were to create a new social media site today you wouldn't be able to get away with being like those sites because the distributors of the social media apps (Apple, Google) won't tolerate the growing pains that are inevitable with any clone of Facebook or Twitter (especially Twitter).

Until recently, Twitter was a dumpster fire of hate and extremism. It's still pretty terrible but nothing like it used to be.  But now Twitter has the capital and processes in place to actively moderate their dumpster fires.  It's a brute force way to do things and very expensive but it works adequately.  

So if you can't just clone Twitter and Facebook how would you make a social network today? 

Here are my thoughts on what it would take and I don't expect my ideas to be acceptable by most people but that's OK.  It would only need to be acceptable to enough people to be commercially viable.

First, The Problem:

Social media lends itself to tribalism.  Once a given group has sufficient numbers to dominate those with a minority opinion they can effectively kick them off the platform.  

Reddit has done a pretty good job by recognizing this and allowing people to silo themselves in subreddits.

But messaging networkings like Twitter have few features to prevent the site from becoming tribal. Indeed, the site is almost seemingly designed to magnify the power of tribes (sub-tweeting to your followers for instance and mass-reporting for example).

The answer to tribalism isn't to create a site for your tribe.  Instead, the solution is recognize that tribalism is a thing and build a platform that allows people to silo themselves in whatever ways they want.  

Moreover, the tech today is so much better and so much has been learned about social media that it could be done better. Much much better.

Here's how:

API Centric

Rather than being a monolithic service like Twitter or Facebook that retrofitted some APIs after the fact, you'd make the service start with APIs.  The apps would be built on top of the API.  That means that third-parties could create their own apps mixing and matching the APIs they think their users want.  This would distribute out the development.


Just like Teams and Slack make it easy for third parties to create extensions, so would this.  If you've ever used something like Jira you can imagine an entire App Store for people to extend their experience.

So rather than trying to have its own photo album service or its own file system or whatever, it would have extensions that users could use add that to their system.


You don't want your service at the mercy of a single ISP.  Instead, you'd make it kind of a Mesh.  Not pure peer-to-peer but rather allow people to host bits and pieces of the platform.  All the data would be encrypted partial blocks. 

Audience Control

Facebook has a pretty good (but rarely used) feature for controlling who can see a given post.  Stardock's own had audiences from the beginning but in 2001, it was just too CPU intensive on the servers to not slow down the entire site.

In the case of this platform, you'd have posts that are either Public, for Followers and a new tier called For Subscribers.  Subscribers are followers that have been approved by the Poster.    If Twitter had this feature, a lot of the viral hate dog piling would disappear.  Quote "tweeting" would only work for public posts for instance.


Instead of only having blocking, you could deflect. This would prevent someone from following you but would not prevent them from seeing your public posts.

Two-Factor Required

Accounts would be required to have two types of authentication. This is to make sure every user is an actual person and not a bot or alt.  The data wouldn't necessarily need to be stored. The goal is to prevent armies of phantom accounts.


Not the feline variety but rather Category tags. When making a post, the site would use AI to suggest a Category tag (#games, #political, #actualcats).  There would be a finite number of these cats with the purpose being that the public or even your followers can filter out cats that they don't want to see.  

So for instance, you're following someone who, 90% of the time, posts great info about tech.  But 10% of the time he goes on a rant about some political issue.  You're not interested in seeing the political crap.  You just want to see the tech.


255 characters. This wouldn't be a platform for writing articles. Quick and simple.  I am not looking to replace Reddit or Facebook. I just want to get and/or distribute information. 

Not for chatting

There'd be a 30 second delay between posting and it being sent out to give people time to change their minds and discourage people from trying to use the service as a real time chat program.  Plus it gives time to edit it before it goes out.

Moderation Event Logs

Moderation is an unfortunate requirement these days.  The problem with moderation is that they aren't transparent and come across as arbitrary.  That's mainly because they aren't transparent and are often arbitrary.

Any post that is reported would go to a moderation channel that volunteers and non volunteers can discuss.  Just like with Wikipedia's discussion pages, moderator notes and comments would be logged and which rule it violated would be listed.  

Monetization for Creators built in

You're not going to get anyone heavily invested in Twitter to use a new service unless there's a really strong incentive.  That's where the Subscriber stuff I mentioned earlier comes in.

News sites, content creators, etc. would still post most of their content for the general public to get users. But they could also provide additional content that's only available to subscribers.  

Because this platform would be handling (in theory) hundreds of thousands of creators eventually you could have subscriptions that might costs trivial ($0.10 for instance) per month because the actual overhead for charging and such would be far upstream (i.e. the subscriber would be paying a total minimum per month and then distributing it out).  Sort of like combining Patreon + RedditGold.  

More rules

This might sound bad but it's not.  Instead of a handful of vague, inconsistency enforced rules you would have what amounts to a very, very long list of rules that are very very specific.  So for example there would be no rule about "racism" but rather several specific rules concerning writing that would be prohibited so that everyone is on board (or not) with it BEFORE people invest thousands of hours into a service only to find out that something you wrote was deemed "racist".

So in this case, think of "more rules" as being more akin to a programming language where an error (rule violation) is specific and obvious.

I'd rather have 548 specific rules than 6 vague, open to interpret rules.

Understandable business model

Rather than the user being the product, revenue would be made not through ads but instead through the extensions I described earlier.  In a world where F2P Fortnite lives off of skins, it doesn't take much of an imagination to come up with many ways to monetize social media in a way people would happily volunteer to pay a little bit for something they want.


So that's my quick thoughts on the subject.  Back in 2001, we developed two platforms: and  We didn't have any venture capital behind us so naturally could not compete with Facebook or Twitter.   But we do know how to build these things.  

We sold our digital distribution platform, Impulse to GameStop in 2011 and we're currently pretty busy building an unannounced cloud-based (not web but a game for PCs and consoles) game takes advantage of many of the same techs such a service would require.

I'd never make the mistake of trying to build such a platform on our own again.  But I wouldn't rule out building one provided that it was properly capitalized.