What great platforms do, and why it works.

Steal the mechanism, not the look.

You're building a subscription. A subscription only works if people come back. Nobody pays for month two if they never returned after month one. Building it is one job. Getting people to come back is the actual business.

These four are called retention mechanics. Each one has a name, and I'll give you those too. Underneath, all four are the same thing: reading and writing your users' progress to a database.

Platform 1

Netflix

Persisted progress state · "resume"

Pick up where you left off.

Why it works: it removes the decision. You never open Netflix and face a blank screen wondering where to start. It already knows, and it hands it back to you.

On your platform: they log in and see "Welcome back, here's lesson three." No deciding. No friction. Coming back is easy.

Copy this · swap in your expertise

I'm building a [platform type] for [your expertise/audience]. I want to add a persistent progress state, the Netflix "continue watching" pattern, so people pick up right where they left off. Based on what I'm building, suggest 2 or 3 ways this could work for my users, tell me what I'd need to store to make it happen, and recommend which one you'd pick and why. Don't build it yet.

Platform 2

Spotify

Aggregation query · a "rollup" of their activity

Show them their progress.

Why it works: Wrapped reflects your own progress back to you. It makes you feel seen and tells you a story about who you are. People post it everywhere.

On your platform: "You've finished four lessons this month. You're 60% through." People stay when they can see they're getting somewhere. When progress is invisible, people quit.

Copy this · swap in your expertise

I'm building a [platform type] for [your expertise/audience]. I want to add progress visualization, the Spotify recap pattern, so people can see how far they've come. Based on what I'm building, suggest 2 or 3 ways this could work for my users, tell me what I'd need to store and total up to make it happen, and recommend which one you'd pick and why. Don't build it yet.

Platform 3

Duolingo

Counter with a reset rule · a "derived field"

Streaks.

Why it works: it's not that people want a streak. It's that people won't break one they already have. That's loss aversion. Losing something hurts more than gaining it felt good.

On your platform: "You've been here three days in a row." Now they have something to protect.

Copy this · swap in your expertise

I'm building a [platform type] for [your expertise/audience]. I want to add a streak counter, the Duolingo pattern, so people keep showing up. Based on what I'm building, suggest 2 or 3 ways this could work for my users, tell me what should count as "showing up" and what should break the streak, and recommend which one you'd pick and why. Don't build it yet.

Platform 4

Brilliant

Interactive component · input + scoring logic

Make them do something.

Why it works: doing feels like progress. Watching feels like consuming. People quit things they consume, but they stay with things they're getting better at.

On your platform: this one matters most for you. Experts default to teaching by talking, so the temptation is a video library. But a quiz or a practice rep will keep people longer than your best lecture.

Copy this · swap in your expertise

I'm building a [platform type] for [your expertise/audience]. I want to add an interactive component, the Brilliant pattern, where people do something instead of just watching. Based on what I'm building, suggest 2 or 3 ways this could work for my users, tell me what I'd need to capture and score to make it happen, and recommend which one you'd pick and why. Don't build it yet.

Read this twice

This is a menu, not a checklist.

That was four things. You are not building four things. Pick two, maybe three, for your first version.

Build all of it and you'll spend three months building and launch nothing. Better to launch with one great retention piece than spend a season perfecting four.

You can always add more later. You can't get back the months you spent not launching.

Your job before the next video

Pick two or three. Know why.

Ask one question: what makes my people come back? Then match the mechanism to your actual problem, not to whatever sounds coolest.

Your thing is longPick "continue where you left off." Your problem is people losing their place.
Your thing is a habitPick streaks. Your problem is consistency.
Your thing is a skillPick active learning. Your problem is people watching instead of practicing.
Your people need momentumPick the progress recap. Your problem is progress feeling invisible.

By the end of this video

Here's where you should be.

You know what you're building (from Video 1).
You know the two or three things that will make people come back and keep paying.
You know why each one works, so you can apply it to your expertise.

Next video: we stop planning and start building. You'll feed all of this into Claude and watch it build your platform, live.