How to get your first 100 SaaS customers

You can build a product in a weekend. AI writes the boring parts. Pieter Levels said it in June 2026: everyone can build apps with AI. Almost nobody has an audience, the cash for ads, or a free way to get attention. Building is not the problem. Getting people to pay still is.

If you do not have ten people who paid yet, stop here and do that job first. How to get your first 10 paying customers without ads is names, a price, and a direct ask. This guide starts after that proof. One hundred customers is not ten times more features. It is the same offer, repeated through intros and one channel you can run every week.

Indiecity is for one person or a team under twenty. You do not need a growth team. You need people who already paid to bring you the next people, and one channel you can work every week.

What changes after the first ten

The first ten prove someone will give you money for a specific outcome. The next ninety prove you can find more of the same person without starting from a blank list every Monday.

Shipping more product does not do that job. You already know someone will pay. What you do not have yet is a repeatable way to reach the next person like them. Stop hunting for a new kind of buyer. Write down who the first ten are, how each one found you, and the words they used. Then ask those people for intros, and do more of the one thing that already produced a payment.

Ask for intros every week

Paul Graham wrote in 2013 that you cannot wait for users to come to you. You have to go get them. That line still holds in 2026. After ten paying customers, the cheapest place to go is the people who already bought. They sit next to more people with the same problem.

Do not wait for a referral program. Ask after they got a result: a saved hour, a mess they closed, a report that used to take a day. Make the ask small. One name. One intro email you write for them to forward. That is the same ask as getting customers from people you know, now run through people who already paid. Track it the same way you tracked the first ten: asked, intro made, booked, paid.

Glad [result] landed. Who else on your team, or in your circle, still deals with [pain in their words]? A name and a short intro is the most useful thing you can do for me right now. I can send you a note you can forward in one click.

Hey [name]. I use [product] for [outcome]. It is [price]. [Founder] is taking a few more [job titles] who still [pain]. Worth 15 minutes if that is still on your plate. Happy to intro.

If they stall, do not punish them. Ask once more after the next win. If every happy customer refuses, you may be selling something they will not put their name on. That is a product problem, not a marketing calendar problem.

Pick one channel and work it

Intros help, but they rarely fill the whole path to 100 on their own. You still need one repeatable way to start conversations with strangers who look like your buyers.

In Traction, Gabriel Weinberg and Justin Mares argue that startups fail more from weak reach than from weak product, and that at any stage one channel usually dominates. Their method is simple: test a short list, then put almost all of your time into the one that produces customers. Finding one channel that works beats dabbling in five.

Do not open ads, LinkedIn, SEO, partnerships, and a community at once. Look at the first ten. How did each one find you? Which habit already produced a paid customer? That is your candidate channel. Give it weeks of real volume, not a weekend of dabbling.

Good early channels for a solo or small team often look boring:

• Warm intros from customers and past coworkers (still a channel if you run it every week)
• One niche Slack, Discord, forum, or trade group where buyers already complain
• Outbound to a fixed list of titles in a fixed niche, with a human note each time
• A weekly answer in public rooms where the pain is already written down
• Partnerships with one adjacent tool or consultant who already serves your buyer

Write the offer in the words your first ten used. Send or post on a schedule you can keep. If you do outbound, write human notes every weekday for a month. If you sit in a forum, answer real questions every week for a month. A weekend test tells you nothing. Track sent or published, replied, booked, paid. If paid stays near zero after that, change the sentence or the room. Do not add a second channel to avoid reading the result.

SEO later, not instead

Search can become a strong channel. It is a slow one. Rand Fishkin has said for years that inbound and organic work are hard to start and useful later, once you have been publishing long enough for the next page to help the last one. You do not get a clean return in week one. That is fine when you already have cash and a clear message. It is a bad primary plan when you still need this month’s customers.

Use SEO after three things are true. You know who pays. You know the phrases they use for the problem. You can keep shipping useful pages without abandoning the intros and the one live channel. Then write for the questions your buyers already type: the error, the comparison, the “how do I stop X” job. One honest page beats a blog calendar full of fluff.

Until then, do not hide in content strategy. Pages that never lead to a conversation or a payment are another form of polishing the product. The next intro and the next note still get you customers. You do not need a public diary of someone else’s first hundred. You need the words your first ten already used, and a week you actually work.

Measure payments, not comments

Graham’s 2013 line still holds when you are small: you grow by getting a few more people by hand. You do not need a thousand strangers. You need people who pay, week after week.

Keep a simple sheet:

1. Paying customers (money moved).
2. Intros asked this week.
3. Intros completed this week.
4. Paid customers from intros.
5. Activity in your one channel (notes sent, posts, replies).
6. Paid customers from that channel.
7. The exact phrase buyers used for the problem (update the page with it).

Cool comments, waitlist emails, and launch spikes do not go on the sheet. A waitlist can feel like demand and still produce zero paid. Payment is still the proof.

A practical path from 10 to 100

1. Week 1. List every paying customer. Note how each found you. Ask the happiest ones for one intro using the script above.
2. Week 1. Pick one channel from what already worked. Write the weekly habit on a card: who you reach, where, how often, what you say.
3. Weeks 2 to 4. Ask for an intro after every real win. Run the channel on a fixed schedule. No new features unless a paid customer cannot get the outcome.
4. Weeks 2 to 4. Rewrite the page using only phrases paid customers used. Put the price where a stranger can see it.
5. Month 2 and after. Keep the same channel until the numbers tell you it is dead. Only then pick the next candidate. Start one SEO page only when the message is stable and you can afford slow returns.
6. Every Friday. Count paid. Count intros. Count channel activity. If both intros and the channel are empty, change the offer or the buyer. Do not open five new experiments.

This is still unscalable work on purpose. Graham’s point about doing things that do not scale was never “stay small forever.” It was: get the first customers by hand, then use slower methods once the offer is proven.

What to stop doing

• Building v3 because marketing feels hard.
• Running five half-channels so none of them get a fair month.
• Waiting for SEO to “kick in” while you skip intros this week.
• Buying ads before you can say, out loud, which sentence makes a stranger pay.
• Treating a launch spike as a customer plan.
• Adding a reward program before anyone has introduced you for free.
• Pitching founder rooms when your buyer is not a founder.

How you know it is working

Customers send you names without you begging. Strangers reply in the one channel you chose. Someone asks the price before you do. Someone pays, then pays again next month. Those are the signals. “Interesting product” is not.

If you have ten paying customers and still cannot get one intro or one channel reply, the product may not be worth recommending, or you are talking to people who do not sit near more buyers. That is useful. It is not a reason to hide in the editor.

Then you have a story worth writing

One hundred paying customers is proof that the offer works past your personal list. Building was never the hard part. In 2026 it is even easier. The hard part is still the next person: the intro, the note, the one channel you did not abandon after two quiet weeks.

When the product is real and the name is yours, write the story. Put it on the map. You can join Indiecity while you do the weeks. That gives you people to talk to. It does not replace the intros.

Until then, ask for the next name. Work the one channel. Leave SEO for when the words are already true.