Startup College Startups 101

Module 6 of 9 · 18 minutes

Get your first hundred customers

A growth experiment board with one metric to move, three time-boxed experiments, a channel bet and a date to review the results.

Start with work that doesn't scale

This module covers the stretch from your first few users to roughly a hundred customers who use your product regularly and, ideally, pay for it. A hundred is a useful milestone, not a promise. Many products stall well before it, and finding out why is part of the work. You'll learn where to focus, which numbers to watch, how to learn from people who leave, how to test ideas quickly and how to set your first prices.

At this stage your job is to win customers one at a time, by hand, in ways you could never keep up with a thousand users. That isn't poor planning. Hand work shows you what customers actually need, and you can automate the parts that matter later.

In practice that means:

  • Message each new user personally and offer a short call to help them get started.
  • Set up their account or import their data for them.
  • Go where they are: visit the shop, join their team meeting, sit beside them while they use it.
  • Fix a problem for one customer the same day, even if the fix is a manual workaround.
  • Recruit users one by one from your interview list, your network and the communities you've joined.

A founder building a shift-swap app for hospital nurses spent her first month visiting wards during night-shift breaks, installing the app on nurses' phones and adding their colleagues for them. Each ward took hours. It also showed her that nurses didn't trust a swap until a manager had approved it, which became her next feature.

Keep a simple list of every user: how they found you, what they use the product for and when you last spoke. With fewer than a hundred customers, you can know each one.

Bet on one or two channels

A channel is a repeatable way of reaching customers: a particular online community, cold email, partnerships, articles people find through search, events, paid ads or referrals from existing users. Trying ten at once spreads you too thin to learn whether any of them works.

Choose one or two, based on:

  • Where your first customers came from. If most early users arrived through one community or one kind of introduction, start there.
  • How your customers choose tools. Small-business owners often ask peers, developers search and read documentation, and larger companies tend to buy through people they meet.
  • What you can afford. Paid ads cost money from the first day; outreach and community work cost time.
  • What you're good at, or willing to get good at.

Give each channel a fair test, a few weeks of steady effort with a number attached, before you judge it. Write down why you're betting on it, so you can check that reasoning later. If it produces nothing after a fair test, drop it and try the next one on your list.

A simple funnel, and the number that matters now

A funnel is the path from someone first hearing about you to paying and telling others. A widely used five-stage version, which you'll see under several names, works well at this stage. Give each stage one number:

  1. Reach: how many people in your target group came across you. Track new visitors or people contacted each week, by channel.
  2. Activation: how many had a real first success. Define the moment, such as the first lesson booked, and track the share of new sign-ups who reach it.
  3. Retention: how many keep using it. Track the share of each week's new users who are still active four or eight weeks later.
  4. Referral: how many came because another user told them. Track the share of new users who arrived through a recommendation or invitation.
  5. Revenue: how many pay, and how much. Track paying customers and the share of active users who pay.

Which number matters most depends on where you are:

  • Before people stick around, focus on activation and retention. Sending more people into a product that doesn't keep them mostly produces more people who leave.
  • Once a group keeps using it, focus on revenue. Will they pay, and how much?
  • Once people pay and stay, focus on reach and referral: finding more people like them, and making it easy for happy customers to bring others.

Pick the single number that matters most for you this month. Your experiments will try to move it.

Retention is the honest signal

Sign-ups measure curiosity, and compliments often measure politeness. Retention shows whether the product is worth coming back to.

To measure it, group users by the week they started (each group is a cohort; see cohort analysis) and track what share of each group is still doing the core action one, two, four and eight weeks later. Define "active" by the job your product does, not by logging in, and match the product's natural rhythm: a booking tool might be used weekly, a tax tool once a year.

You're hoping for a curve that drops at first and then flattens, showing a group of people who keep using it. There's no universal good number; it varies by product and customer. A curve that keeps sliding toward zero means people try it and leave, and more marketing won't fix that. With small numbers, skip the charts and look at a list of names: who came back, and who didn't?

Talk to the people who leave

Users who stopped using your product are one of the most honest sources of feedback you have. Contact them within a few days of going quiet, keep the message short, and make it clear you're not trying to win them back.

Hi Sam, I noticed you haven't used the booking page since early March. I'm not writing to sell you anything; I'm trying to understand what didn't work. Would you be willing to tell me, in a reply or a ten-minute call, what you hoped it would do and what you use instead now?

In the conversation, ask what happened the last time they used it rather than what they think of features. Don't argue or explain. After five or more conversations, sort what you heard: were they the wrong customer for this product, or the right customer let down by something you can fix? The second group tells you what to work on next.

Run small experiments, including on price

A growth experiment is a small, time-limited test of one idea for moving your metric. Before you start, write it down in four parts: hypothesis, action, metric and stop date.

Hypothesis: new teachers who get a personal welcome call on their first day will set up their booking page more often. Action: call every new sign-up within 24 hours for the next three weeks. Metric: the share of new sign-ups who publish a booking page within seven days, now 35%, with a target of 50%. Stop date: 28 November.

A few rules keep experiments useful:

  • Run no more than three at once, so you can tell which change did what.
  • Decide in advance what result counts as success, and what you'll do either way.
  • With small numbers, look for large, clear changes. Going from seven out of twenty to eight out of twenty is probably noise.
  • Stop on the stop date. Keep what worked, drop what didn't, and write down what you learned.

Price your first paid customers

Asking for money is an experiment too, and payment is the clearest sign that the product matters to someone. Many founders wait too long to ask.

  • Charge earlier than feels comfortable. Once the product reliably does its one job, ask your most engaged users to pay.
  • Price from value, not cost. What does the problem cost customers today in time, money or mistakes, and what do they pay for current alternatives? Your costs set the floor, not the price.
  • Keep it simple. One or two plans are enough at first.
  • Make early discounts explicit. If early customers get a lower price in exchange for feedback, say what it is and how long it lasts.
  • Test the price. Quote a higher price to your next few prospects and see whether the answers change. If nobody ever hesitates, you may be charging too little.

Find out whether you need to charge sales tax or VAT where your customers are; the rules vary by country, so check with a local accountant. Revenue from customers is also how many good companies grow: most never raise outside money.

Do the work

Use the growth experiment builder below to make your board, a one-page plan you can pin up, share with your team and revisit on your review date.

  • Choose one metric to move, using the funnel section to find the stage that matters now.
  • Write its current value honestly, even if it's small or zero.
  • Write three experiments, each on one line as hypothesis → action → metric → stop date, and keep each short enough to finish within two to four weeks.
  • Name the channel you're betting on and why, in a sentence or two.
  • Decide now what you'll stop doing if nothing moves, so you don't carry on out of habit.
  • Set a review date shortly after your last stop date.

You're done when someone else could read the board and run the experiments without asking you a question. On the review date, update the current value, keep or drop each experiment, and write the next three.

Your document for this module

Growth experiment board. A one-page board for your next few weeks of growth work: the one metric you're trying to move, three small experiments with stop dates, the channel you're betting on and when you'll review the results. Pin it up or share it with your team.

Open the growth experiment board builder (+100 XP)

Check yourself

1. You have eight clinics using your scheduling tool. Which is the best use of your time this week?
2. You get 200 sign-ups a week, but almost nobody is still active after a month. What should you work on first?
3. What's the most useful way to learn from users who stopped using your product?
4. Which of these is a well-formed growth experiment?
5. Your first users like your free beta. What's a sensible way to start charging?