How to get your first customers in 2026: a 30-day plan

To get your first customers for a B2B product, go and get them yourself: choose one buyer, list 50 companies that show the problem right now, send each a short personal message, turn the good conversations into paid pilots and ask every early customer for an introduction. The 30-day plan below aims for ten paying customers. Lenny Rachitsky collected how more than twenty fast-growing B2B companies found their first customers, and his conclusion, in our paraphrase, is that only three routes drive that early growth (the founders' personal network, the places buyers already spend time, and press), that almost every company combined the first two, and that getting press "is rarely the way to get started" (Lenny's Newsletter, 7 July 2020).

Checked on 22 October 2026; prices are what a US buyer sees, before tax. Revexa builds Leo, which has its own section below. For founders with zero to ten customers; how to get customers covers the channels that scale.

AI made building cheap, so customers are the hard part

The slow part of a new B2B company is no longer the code; it is finding people who will pay. In a video reported by TechCrunch in March 2025, Y Combinator managing partner Jared Friedman said that for a quarter of YC's Winter 2025 batch, 95 percent of the code had been generated by AI, and that those founders were "highly technical, completely capable of building their own products from scratch" (TechCrunch, 6 March 2025).

The other half of a startup has not changed. In "Do Things that Don't Scale" (July 2013), Paul Graham told founders: "You can't wait for users to come to you. You have to go out and get them." (paulgraham.com). Our reading: AI rewrote the building half and left the recruiting half untouched.

The 30-day plan for your first ten paying customers

Give the month four jobs, one per week: list, write, talk, close. A second list starts in week 3, so the month covers about 100 companies plus your network. Ten paying customers is the goal, not a promise; if the first round brings fewer, run the four weeks again with a sharper list. The numbers are our targets, not a benchmark.

Week 1: one buyer. Write the role, the company type and the moment the problem appears. "Heads of support at SaaS companies of 20 to 200 people, in the month ticket volume outgrows the team" is a buyer; "businesses that want to save time" is not. A template is in our ideal customer profile guide.

Week 2: the network message. Tell former colleagues, past clients and founder friends exactly what you built and ask who has the problem today. As Lenny put it, the question is not which of the two routes to take, but "how far your own network will take you before you move on."

Weeks 3 and 4: talk, then deliver. Ask what the problem costs them before you show anything. After onboarding, ask: "Who else do you know who has this problem?"

How to find the first 50 companies that have the problem now

Start from the moment your problem appears, not from an industry list: ask what happens at a company in the weeks before it needs your product, then search for public traces of that moment. Five places a builder can search in an afternoon, free:

• Job ads that describe your product's job. Read the duties, not the title. "Own the weekly reporting spreadsheet" or "reconcile payouts manually" describes the work your tool replaces, at a company that just budgeted for it.
• Unhappy users of the tool you replace. Reviews and threads asking for "an alternative to" it name the pain and often the company type.
• People who reacted while you built. Those who commented on your build posts or joined the waitlist are your warmest list. Keep the ones at companies that fit.
• Companies like the one where you first saw the problem. Its look-alikes are your first rows.
• Changes that create the problem: a funding round, a new leader in the buying role, a team that doubled. More free sources are in how to generate leads.

Fictional example: you built a tool that drafts answers to security questionnaires. Your 50 are SaaS companies of 20 to 150 people that posted a job mentioning SOC 2 in the last 30 days, announced a first large enterprise customer, or whose founder asked in a community how others survive security reviews. At that size, write to the CTO.

What to write to them

Name the sign you saw, say what you built in one line and ask for a short conversation about their problem, not a demo. Say you are the founder and the product is new; an early buyer gets the builder on the call and a say in what comes next.

Three fictional examples:

To a company whose job ad describes the work. "Hi Maya, your ops analyst ad mentions reconciling payouts by hand every week. I built a tool that does that reconciliation automatically, and I am looking for five finance teams to shape it. Could I ask you three questions about how you handle it today? Twenty minutes, no slides."

To a founder who asked about the problem in a community. Answer in the thread first. If they want more and the community allows direct messages: "You asked how others handle security reviews. I built something for exactly that after filling in dozens by hand. Happy to show you what worked, even if you never use it."

To a former colleague. "You saw how many hours our team lost on vendor onboarding. I have spent three months building the fix. Who do you know who has that problem right now?"

Charge from your first customer

Ask for money from customer one, through a paid pilot with a fixed price, one measurable result and an end date. Our view: free users tell you what people will try, paying users tell you what they will buy.

A pilot shape for a small B2B product (fictional example): one team for 30 days, one result you can both measure, a founding-customer price below list and never zero, rolling into the monthly plan unless they cancel.

In the US, Stripe charges 2.9% plus 30¢ per successful transaction for domestic cards, with no setup or monthly fees, and its no-code Payment Links are included "at no additional charge" on standard pricing, so you pay the usual transaction fees and nothing extra for the link (Stripe). A buyer who likes the product but will pay nothing is telling you something: the problem is not urgent, or they hold no budget.

What does not work for the first ten

Anything that skips the one-to-one conversation:

• A launch as the whole plan. Lenny files a launch under press ("Get press: Orchestrate a launch") and concludes that press "is rarely the way to get started". In our view a Product Hunt or Hacker News launch buys a day of attention, mostly from other builders. Launch once you have customers to quote.
• Bought lists. A list tells you who exists, not who needs you, and it ages fast. ZeroBounce, which sells email verification, reports that at least 23 percent of an email list degrades every year and that only 62 percent of the addresses submitted to it in 2025 were valid (ZeroBounce Email List Decay Report 2026). Several European countries, Germany among them, require consent for a cold email even to a company address; see cold email laws by country. This is not legal advice.
• Mass sending. Gmail's sender guidelines say to keep the spam rate reported in Postmaster Tools below 0.10 percent and never to reach 0.30 percent (Google). Our reading: one blast to strangers can push a young domain over that line, and then your onboarding emails and invoices land in spam too.

From ten customers to a hundred

After ten, stop guessing: find the sign that came just before each customer bought, then watch for it at a scale you cannot cover by hand. Keep the calls; hand over the watching and the first drafts.

1. Rewrite your buyer sentence with what actually happened in the first ten deals.
2. Find the sign that predicted a sale. If seven of ten bought within weeks of hiring for one role, that hire is your trigger.
3. Make it weekly. Our founder-led outbound guide sets out a routine of 3 to 5 hours a week.
4. Add one channel that compounds, usually pages that answer your buyers' questions. Selling a service rather than software? How to get clients ranks the routes for agencies and consultants.

Doing this with Leo

Cursor writes your code. Leo finds your customers.

Leo by Revexa is an AI GTM agent that finds the B2B companies ready to buy and starts the conversations that turn them into customers, by email and LinkedIn. By default you approve every message before it is sent.

Leo builds your ideal customer profile from your website, then watches hiring, funding, headcount growth, job ads, new executives in the buying role, competitors' tools on a company's website and engagement with LinkedIn posts (your own posts and company page, and those of competitors and industry voices). It judges each company's business model, picks the right person and drafts a message in your voice, on LinkedIn from your own profile or by email from a connected mailbox. You approve or skip each draft in the portal, Slack or iMessage, in your own words, or switch on autonomous mode per channel. Leo handles replies and books meetings; with HubSpot, Pipedrive, Salesforce, Upsales or Close connected, your customers and open deals stay out of outreach. German leads are never sent cold messages, because of UWG §7.

Leo costs $99 a month in the US and the rest of the world, €89 in Europe and 999 SEK in Sweden, excluding VAT or sales tax, with the first 7 days free, no lock-in and one-click cancellation (pricing).

Is it 10x? Only against a hired SDR: our cost per held meeting calculation puts the agent route 12 to 23 times below a hired US SDR, assuming every route books the same number of meetings. A founder with no SDR to replace saves hours, not salary.

What Leo is not: a contact database to export lists from, or a dialer. The calls, the pilot and the price stay with you. Revexa is a Swedish company founded in 2026 with no published customer results yet. Leo is the wrong choice if you cannot yet say who your buyer is (talk to ten of them first), if most of your buyers are in Germany, or if you want case studies before a trial.

Related: Where to find clients.

Frequently asked questions

How do I get my first customers for a B2B startup?

Find them yourself, one conversation at a time. Pick one buyer, list 50 companies that show the problem this month, message the person who owns it at each, and offer a paid pilot when the need is urgent.

How do I get my first 10 customers?

Run a 30-day cycle: list 50 companies with a fresh sign of the problem, message them and your network, hold calls that end in paid pilot offers, then onboard and ask for introductions. Repeat if ten has not arrived.

I built my SaaS with AI: how do I get customers now?

By recruiting them by hand, as founders always have. AI shortened the building, not the selling, so spend the saved weeks talking to companies that show the problem now.

How do I get my first 100 customers?

Learn from your first ten. Note the sign that came before each purchase, let software watch for it every week, keep the calls yourself and add one channel that compounds.

Should I charge my first customers?

Yes. A paid pilot with a fixed price, a measurable result and an end date shows whether the problem is worth money. Offer a founding-customer discount, not a free plan.

How do I get customers for my app or startup without a big launch?

Start conversations with companies that have the problem now: job ads describing the work your product does, users unhappy with the tool you replace, people who followed your build. Launch later, with customers to quote.