tw-rl

Putting up the Bat Signal A progress update on @PlaybyteInc as we near the 2-year mark - Our Big Theory - What we’ve learned building a new social media platform, “TikTok for Games” Quick stats - 25,000+ games - Users spending 21 min a day in the app https://link.medium.com/HfzjO9lSNib

Any new social media platform needs to bring together 3 elements 1. A Constrained Format — the format is limited in length + degrees of customization 2. Dedicated Tools — set creators up to succeed within the format 3. Owned Distribution — help them reach a responsive audience

Selecting the constraints that ultimately define the Playbyte format has been an interesting journey Some choices we’ve made: games are 2D, each environment is one phone-screen in size, and you can make a game with up to 5 environments (we call them scenes)

Game objects can be emoji, uploaded images, or text Each game scene can have its own custom background image From these very simple primitives you can construct very different game formats

It’s been a similarly fun journey to build dedicated mobile tools for creating games within these constraints Really proud of the work the team has done to build something that feels of a kind with TikTok and Snapchat

In recent months we’ve introduced a really cool tool: the Rule Editor It’s essentially a no-code scripting system that seamlessly builds from the simple default creation UI while keeping everything readable

Power users can make new assets with the Rule Editor and other users can then drop them in without having to touch their custom logic

Ok, so a unique, constrained format and tools that will make cool stuff in that format ✅ and ✅ Now, owned distribution: to succeed we have to match games with the right players via a personalized algorithmic feed

In earlier implementations, when we had fewer games, we could naively propagate whatever got engagement, assuming that if it got played at all it was better than other content

Over time we’ve had to build out a proper “interest graph” — working out our users’ tastes and interests by looking at how players engage with content that’s been manually hashtagged at publish time or manually tagged in our CMS backend