#4 Retention, from the ground up: what this module will teach you

Most people working on a product check their retention curve and ask one question: is this good or bad?

It's the wrong question.

A 20% D30 (30 days after sign-up) retention rate is a disaster for Duolingo. The same 20% is perfectly healthy for Uber. Same number, opposite conclusions — because retention only makes sense in context.

The following videos are about building that context. By the end of them, you'll be able to look at any retention curve and actually understand what it's telling you.

What is retention, really?

At its core, retention measures how many users remain active within a given period of time. Put another way: it tells you whether the people who tried your product found enough value to come back.

That sounds simple. But the moment you try to apply it across different products, it falls apart — because "coming back" means wildly different things depending on what the product is for. A daily language app, a ride-hailing service, and a CRM used by a 500-person sales team all have users who "come back," but the mechanics, the frequency, and the reasons are completely different.

So before we can measure retention, we have to understand the shape of use our product is built around.

Nature vs. Nurture

Every problem a product solves has a rhythm that already exists in the world.

We get hungry daily. We work out weekly. We go on vacation once or twice a year. We buy a home every twenty years. That rhythm is the Nature of the use case — it's there whether your product exists or not. You didn't invent it, and you can't change it.

Nurture is everything you build on top of that natural frequency: the notifications, the features, the re-engagement campaigns, the streaks, the emails. All of it is designed to amplify the underlying rhythm.

And here's where most products get it wrong. There's a Goldilocks rule at play:

• If your Nurture is more frequent than the Nature of the use case, you become spam.
• If your Nurture is less frequent, you get forgotten.

Staying in that sweet spot requires deeply understanding the use case itself — not what you wish your users would do, but what they actually need, and how often they actually need it.

This is the foundation everything else is built on. You can't design good retention mechanics, pick the right metrics, or even know whether your curve is healthy until you've honestly answered: what is the natural rhythm of the problem I'm solving?

What's coming in following videos

Once you have that foundation, the next step is recognizing that products don't all behave the same way. Over the rest of the videos, we'll work through the three core archetypes — habit products, utility products, and enterprise SaaS — each with its own retention logic, its own curve shape, its own risks, and its own metrics.

We'll look at why a flat, high-plateau curve is the goal for some products and a red flag for others. We'll get specific about which metrics matter in which context (and which ones are vanity). And we'll tackle the products that don't fit cleanly into one bucket — the hybrids, like Netflix, where the analysis actually gets interesting.

By the end of the videos, you'll have a mental model for diagnosing any retention curve: start with the use case, identify the archetype, then read the numbers through that lens.

The curve doesn't lie. But it doesn't explain itself either.