How to get your first B2B customers: a founder’s guide to customers one to ten

You get your first B2B customers by hand: the founder chooses a small group of companies with a problem the product solves, reaches the people who own that problem and stays close enough to fix what breaks. Paul Graham’s essay Do Things That Don’t Scale (Jul 2013) makes the same point: nearly all startups recruit their first users manually.

Customers one to ten are a different job from customers 50 to 100. You are not running a funnel yet. You are proving that a specific kind of company has a specific problem, will make time for it and will pay to have it solved. Every early deal should teach you something about who buys and why.

1. Write a narrow profile: one industry or type of company, one buyer, one problem.
2. Build a short list of companies you can link to that problem today, with the evidence written next to each.
3. Map who owns the problem at each company, not only who signs.
4. Contact them yourself with a point of view on their situation, not a product tour.
5. Offer the best-fit few a design partnership with a date to decide on paying.
6. Write down why each company said yes or no, and narrow the profile again.

Your first customers should be companies that feel the problem now, that look like the market you plan to serve, and that you can reach and support closely. Revenue matters less than learning and a customer who will talk about the result.

Graham’s essay notes that Facebook was at first just for Harvard students, which helped it reach a critical mass in one place. For a B2B founder the narrow market is usually one industry, one company size and one role. Our ideal customer profile guide and ICP template help you write it down.

You should qualify first customers on a problem you can link to each company, because companies that only resemble your target rarely care enough to buy. Lookalike traits such as industry, size and tools tell you who could buy; a linked problem tells you who has a reason to talk now.

We saw this in SalesOne’s testing round: companies were under 70% relevant whenever they could not be linked to an existing problem. The fix was to qualify on a linkable problem rather than lookalike traits. For one consulting and strategy firm, more than 90% of the accounts found were a 90%+ fit to its own profile, and the research surfaced industries it had never considered.

Gartner’s B2B buying journey page (archived Sep 25, 2026; no sample stated) says that “99% of B2B purchases are driven by organizational changes”. A linkable problem is usually one of those changes: a new leader, a new site, a new owner, a new rule, a missed target. Our guide to trigger events in sales lists the common ones and where they show up.

Timing matters as much as fit. In SalesOne research runs, S1 agents checked 1,051 candidate companies against customer profiles and kept 249, about one in four. All 802 rejects carry a reason, and two in three failed on timing: no qualifying dated event in the window. Our guide to why accounts get rejected shows the reasons in more detail.

Founders find their first customers through their own network, warm introductions from investors and former colleagues, and direct outreach that names a problem the company has now. The founder should do it personally, because no one else can answer the hard questions or change the product on the call.

• Start with people who already trust you: former colleagues, customers from a past job, investors who know buyers in your market. Ask for one specific introduction, with a short note they can forward.
• Add direct outreach to companies you can link to the problem. A short email that names what changed and asks one question works better than a pitch.
• Reach the person who owns the problem. In an early deal that is often an operator, not the most senior title.
• Make it easy to start. Graham describes the “Collison installation”, where Stripe’s founders set new users up on the spot instead of waiting for them to try it.
• Hold every early call yourself and write down the words buyers use. Those words become your website and your next emails.

Our guides to founder-led outbound and founder-led sales cover the weekly rhythm, the tools and when to hand the work over.

The first message should name one dated event, the problem it usually creates, why you are working on that problem and one easy question. It should be short enough to read on a phone and specific enough that it could only have been sent to that person.

A sample first message from a founder

Fictional sample. Halvorsen Castings is a fictional foundry group. It announced a second plant on Sep 12, 2026 and named Ruth Adeyemi as COO on Aug 25, 2026. The sender is the founder of a fictional startup that schedules maintenance across plants.

Subject: Second plant and maintenance planning. Hi Ruth, congratulations on the COO role. With the second plant announced on Sep 12, maintenance planning usually moves from one team’s spreadsheet to two sites sharing parts and people, and that is where downtime tends to creep in. I am building a tool for exactly that handover and am working closely with a few manufacturers before launch. Would a 20-minute call to compare notes be useful, or is someone else leading planning for the new plant? Maya Lind, Founder.

Fictional sample: company, people and startup are fictional.

Notice what it leaves out: no feature list, no claims about her plant, no customer logos you do not have. Commercial email rules vary by country; check your program with your own counsel.

A design partner is an early customer who gets the product early, often at a discount, in return for regular feedback and a decision on paying by a set date. Bessemer Venture Partners (Atlas, May 6, 2026) writes that most practitioners recommend between five and 12.

The same Bessemer piece says that eventual conversion to paid is the whole point, and that indefinite free access is an advisory relationship, not a validation signal. Treat the partnership as a sale with a short trial built in. Our guide to design partners covers how to find, choose and convert them.

Graham also suggests a B2B founder pick one user and build for them almost like a consultant, until the product fits; other users with the same need tend to follow. Keep the custom work to what the next customers in your profile will also need.

You land a first enterprise customer by starting with one team that owns a problem, not the whole company: find the leader whose group feels it most, run a small paid pilot with a clear success measure, and map everyone else who must agree before it can grow.

• Find the change. A new leader, a new site or a new program gives a large company a reason to try something new.
• Start small and paid. A pilot inside one team avoids the full procurement process and still proves willingness to pay.
• Map the buying committee early: the user, the budget owner, procurement, security and the executive sponsor. Our buying committee map is a simple way to do it.
• Write the success measure and the expansion step into the pilot, so the next conversation is already planned.

Our guide on how to sell to large companies covers longer cycles and multiple stakeholders in more depth.

SalesOne does the research behind founder-led sales, so the founder’s hours go to conversations. S1 agents research your market, accept or reject each company against your profile with the reason, map the people who own the problem and draft a point of view for the first message.

A person approves every message before it is sent, and every approval is recorded with who gave it. Outreach rules decide who on your team may contact whom, and a reply pauses the account’s remaining steps. See how SalesOne works for startups and how account research works.