How do you retain users after they sign up?
Retention is won or lost in the first few interactions, and at the early stage the answer is almost always more human contact, not more automation. The founders who retain users best treat each new signup as a relationship to build, not a conversion to log. Do that intensively for your first hundred users and you will learn more about churn than any analytics dashboard will ever tell you.
Why most startups lose users in the first 30 days
The gap between signing up and becoming a habitual user is where nearly all early retention is won or lost. Most founders fill this gap with automated onboarding emails written at the level of abstraction they wish the product operated at — not the level their users actually experience it. The result is a drip sequence that answers questions nobody asked while the real friction goes unaddressed.
Paul Graham's account of Airbnb's earliest days makes the mechanism clear: roughly 30 days of founders going out and physically meeting users was the difference between the company surviving and collapsing. That wasn't a marketing tactic. It was a diagnostic tool that also happened to make users feel cared about. The meetings revealed what the product needed to do to become sticky before the founders had built those things.
The practical implication is that your first retention strategy should be a calendar, not a funnel. Block time every week to talk to users who signed up in the last two weeks and haven't yet done the thing your product is designed for. Find out exactly where they stopped. That single conversation, repeated across twenty users, will surface the two or three blockers responsible for most of your early churn — and those are the only things worth fixing first.
Make individual users unreasonably successful, then study why it worked
The instinct most founders suppress is the one that would actually save them: doing something absurd for a single user to make them succeed. Absurd meaning: jumping on a call to walk them through setup, manually completing a step they got stuck on, or sending a follow-up the same day with a specific suggestion based on what they told you they were trying to do. These feel like one-off heroics that don't scale, and they are — that's exactly why they work at this stage.
Graham's argument in 'Do Things That Don't Scale' is that founders who obsess over making individual users happy don't get lured into dead ends by doing so. The concern that this effort can't scale is backwards: if you make twenty users genuinely successful, you now know what success looks like, which means you can eventually build the product path that gets users there without your intervention. The manual version is the prototype for the scalable version.
Concretely: identify the moment in your product when a user has crossed from 'curious' to 'dependent' — the action that correlates with retention. For a project management tool it might be inviting a second team member. For a writing tool it might be completing a second document. Then instrument that moment, and when a user hasn't hit it within 48 hours of signup, reach out personally. Not with a template — with a message that references something specific about what they told you they were trying to accomplish.
Use a deliberately narrow focus to build felt ownership
One of the most durable retention mechanisms isn't a feature — it's the user feeling that the product was built for people exactly like them. Graham describes how Facebook's early decision to restrict access to Harvard students meant a critical mass of those students signed up precisely because it felt like their product, not a general-purpose tool they happened to be allowed to use. The narrowness created belonging, and belonging creates the kind of habitual return that no push notification can manufacture.
For a pre-scale startup this translates into a specific decision: resist the pressure to broaden your positioning to capture more signups. More signups with weaker felt ownership produces worse retention than fewer signups with strong felt ownership. The users who feel the product is exactly for them will use it more often, recommend it more credibly, and tell you more useful things when you ask what's missing.
In practice this means your onboarding copy, your empty states, and your first-run experience should speak to a specific kind of person solving a specific kind of problem — not to anyone who might conceivably benefit. When a new user reads your product's interface and thinks 'this was made for someone like me,' that recognition is doing retention work before any feature does.
Direct feedback from real users is your highest-leverage retention input
Retention analytics tell you what happened. User conversations tell you why, and more importantly, what would have had to be different for the user to stay. The earlier you are, the more the conversation is worth relative to the metric. When you have fifty active users, a thirty-minute call with one who churned is more useful than a cohort analysis of all fifty, because the cohort analysis raises questions you can't answer without the call anyway.
Graham notes that the feedback you get from direct engagement with early users is the best you will ever receive — the clarity and specificity of what someone tells you when you are sitting across from them, watching them use your product, is irreplaceable at scale. Big companies run focus groups as an inferior substitute for exactly this. You don't have to settle for the substitute yet.
Set up a dead-simple system: when a user hasn't logged in for seven days after signing up, send them one personal email asking a single question — 'What were you hoping to do with [product] that you haven't been able to do yet?' Not a survey. One question. The answers will cluster around two or three themes within two weeks, and those themes are your retention roadmap. Fix those things before adding anything new.
“I have never once seen a startup lured down a blind alley by trying too hard to make their initial users happy.”
— Paul Graham, source
The one thing to do
This week, personally reach out to every user who signed up in the last 14 days and hasn't completed your product's core action — ask one specific question about what stopped them, then fix the most common answer before anything else.
Frequently asked questions
When should I stop doing manual retention work and automate it?
When you understand the exact sequence of actions that predicts retention well enough to encode it — not before. Automating before you understand the mechanism just scales the wrong behavior. Most founders automate too early.
What is the single metric I should track for early retention?
Find the one action inside your product that correlates most strongly with users still being active 30 days later — often called the 'activation event' — and measure the percentage of new signups who complete it within their first week. Everything else flows from moving that number.
How do I retain users when my product still has obvious gaps?
Be honest about what the product does well now and compensate for its gaps with personal attention. A user who knows you are actively fixing a problem they flagged will stay through friction that would make an anonymous user churn.
Should I offer incentives like discounts to stop users from churning?
Only if the reason they're leaving is price, which is rarely the real reason at the early stage. Incentives mask the signal. Talk to the user first; if they tell you the product didn't solve their problem, a discount won't fix that.
Sources
- gstack: skillify/SKILL.md — Garry Tan
- Do Things that Don't Scale — Paul Graham