← All writingGrowth · August 2026 · 8 min read

Growth flat at
100 users? Don’t buy ads yet.

If your first 100 users don’t come back on their own, spending to acquire 1,000 more won’t fix the problem. It will make the signal harder to read.

By Monika MalikProduct Growth · San Francisco

You shipped it. The first week gave you 70, maybe 100 signups. Then they started leaving, and new signups don’t cover the churn.

You have maybe six months of runway and four engineers. You’re thinking: “Let’s run some ads” to get more users to sign up. Don’t. If your first 100 users don’t come back on their own, spending $15,000 to get 1,000 more won’t help. It just makes your numbers messier and your runway shorter.

You get one or two real fixes before you have to raise again. Use them on the hole, not the hose.

THE FRAMEWORKUIE

Understand → Identify → Execute

01

Understand the problem

Draw a funnel with just four steps:

  1. They know about the product → they enter your signup flow.
  2. They want it → they sign up.
  3. They get it → they hit that “oh, this is good” moment in the first session.
  4. They come back → day seven, without you begging them.

Most early products have a leak around steps three and four. Your job is to find which arrow is broken.

02

Identify the reason

Pull your last 100 users and look at the data:

  1. How many came back on day seven without you nudging them? If it’s eight out of 100, you have a leak. If it’s 30, you might have something.
  2. What did those returners do that everyone else didn’t? Look at the first hour. Did they invite a teammate, run a report, or create one real project? That’s your aha moment. Write it in one plain sentence.
  3. How fast did they get there? If returners hit it in five minutes and churned users never hit it, your onboarding is the problem—not your acquisition.
The data trap

Paid traffic can turn a readable 8% baseline into a confusing 4% average. You didn’t necessarily make the product worse; you mixed high-intent users with paid tourists and made the signal harder to interpret.

Once that happens, you can’t tell whether your onboarding fix worked, find the real aha moment, or understand which channel deserves more investment. You didn’t buy growth. You bought noise.

03

Execute: ship one fix toward the aha

Ship one small change that helps more people reach the aha moment this week. The exact ticket depends on the product:

ProductAha momentSmall ticket
B2B SaaSImported real work, created a task, shared it.One-click import, then guide users to create and share one real task.
AI toolReceived a useful output worth sharing.Replace the blank box with three concrete starting actions and a pre-filled example.
MarketplaceSaw relevant supply and posted easily.Start photo-first; generate the title, description, and price.
ConsumerMade something they want to show someone.Begin with a strong template, auto-save, and make Share the primary action.
AnalyticsSaw their own data produce an insight.Auto-connect one source or load realistic sample data, then pre-build one chart.
04

A story that stuck with me

Drew Houston didn’t even have a working Dropbox when he applied to Y Combinator—just a four-minute video showing how the magic folder would work. The video took off, and the waitlist grew far beyond what he expected.

Most of us would have opened the doors and announced, “We’re live.” He didn’t. He let people in gradually and watched the boring little actions: did they put one file in one folder? Did they install Dropbox on a second computer?

Only after that action began retaining people did Dropbox add its famous referral program. They fixed the bucket first. That’s why the fuel worked later.

Watch Drew Houston tell the story ↗
TL;DR

If people don’t come back, don’t pay to get more of them. Find the one thing returning users do on day one, make it happen faster, and ship that fix before spending a dollar on fuel.