How to find your first B2B SaaS customers
The first ten customers are not a marketing problem. They come from people who already have a reason to take your call, and almost every founder has more of those than they think.
In short
- Nothing that works at a hundred customers works at ten. The first ten come from people who'll take the call for reasons unrelated to your product.
- Separate the **user, the buyer and the champion** — they're one person at 10 staff and three departments at 1,000.
- Company size changes the sale more than the job title does.
- Ask about the problem, not for the sale. Enthusiasm is politeness; details are evidence.
The first ten are not the next hundred
Nothing that works at a hundred customers works at ten. Paid acquisition needs a conversion rate you don't have yet. Content needs months. Outbound needs a message you haven't earned the right to write, because you don't yet know which words land.
What works at ten is people who will take the meeting for reasons that have nothing to do with your product: they know you, they owe you a favour, they're curious what you've been building, or they liked working with you six years ago. That's not a scalable channel. It doesn't need to be — it needs to get you to the point where you know what you're selling.
Your first customers aren't buying the product. They're buying you, and tolerating the product.
Separate the user from the buyer
This is where most network searches go wrong. Founders search for the person who'd *use* the thing, then get stuck because that person can't authorise spend.
In a B2B sale there are usually three distinct roles, and one person rarely holds all three:
- The user — feels the pain daily, can describe the problem precisely, often can't buy anything
- The buyer — owns the budget line, may never open your product, cares about risk and outcome
- The champion — wants it to happen and will spend internal credibility to make it happen; may be either of the above
At a ten-person company these collapse into the founder. At a thousand-person company they're three departments. Your ideal-customer description needs to name which one you're targeting, because “Heads of Operations at logistics firms” and “operations coordinators at logistics firms” produce completely different lists from the same network.
A job title means almost nothing without headcount attached. A “VP Operations” at a 30-person company is a hands-on generalist who can decide in a week; the same title at 5,000 staff runs a department and buys through procurement. Same words, two entirely different sales.
Mining your own network for them
Export your connections and work through them deliberately. The useful signals in the file are narrow but real:
Before opening the file. Job titles, company size band, industry. If you can't write it in three sentences, the list won't be any sharper than the description.
Include the seniority variants — Head of, Director of, VP, Chief. Exclude the ones that sound right but can't sign: coordinator, associate, assistant, intern.
This is the manual bit, and it's the highest-value step. Half the names that pass the title filter will work somewhere far too small or far too large.
A slightly-off-profile person who genuinely likes you will take the call. A perfect-profile stranger won't. Rank by relationship first for the first ten; invert that once you're looking for the next hundred.
What to ask for
Not a sale. For the first ten, ask for the problem — and mean it. “Does this happen at your company, and how bad is it?” gets replies that a pitch never will, and it gives you the language to sell to the eleventh person.
The conversion happens naturally when the answer is yes. Someone who has just spent fifteen minutes describing how painful a process is will ask what you're building, and at that point you're not pitching, you're answering.
One warning: friends will tell you your idea is good. That's not data. Push for specifics — what they currently do instead, how much time it takes, what they've already tried and abandoned. Enthusiasm is politeness; details are evidence.
Running the conversation
You've got the meeting. The most common way to waste it is to spend it presenting.
Ask about last time, not about generally
"How do you usually handle vendor onboarding?" gets you an idealised description of a process nobody follows. "Walk me through the last one you did" gets you the spreadsheet, the three chase emails and the two weeks lost waiting on a certificate. Specific past events are memory; general questions are theory.
Find out what they've already tried
This is the single highest-value question in early customer discovery, because it separates a real problem from an annoyance. Somebody who has built a spreadsheet, trialled two tools and hired a contractor has a budget-worthy problem. Somebody who has done nothing has an irritation they'll happily discuss forever and never pay to fix.
- What do you do about it today? — every problem has an incumbent, even if it's a spreadsheet and a person
- What have you tried that didn't work? — tells you the graveyard you'll be compared against
- Who else does this touch? — surfaces the buyer and the champion you hadn't identified
- What would have to be true for you to change it? — the closest thing to a straight answer on budget you'll get early
Closing without pitching
If the problem is real, the natural next step is theirs to take, and you should let them. "Would it be useful to see what we've built?" is enough. If the answer is a polite deflection, you've learned something more valuable than a demo would have taught you — and you've kept the relationship, which matters because these are people you actually know.
The output of this phase isn't customers, it's language. After ten honest conversations you'll be able to describe the problem in the words your buyers actually use — and that description is what makes every later channel work.
Common questions
How many people do I need in my network for this to work?
Below about 300 connections there usually isn't enough there and you're better off with direct outreach in communities where your buyers already gather. Above roughly 1,000 there is almost always something worth finding.
Should I offer the first customers a discount?
Prefer a design-partner arrangement over a discount: more access to you, input on the roadmap, full price or close to it. Discounts anchor your value low and attract exactly the customers least likely to stay.
What if my network is from a completely different industry?
Then use it for introductions rather than sales. A connection in the wrong industry may still know three people in the right one, and asking for an introduction is a much lighter request than asking someone to buy.
See who’s already in your network
Upload your LinkedIn connections export and every row gets read against what you sell. Free scan, no account, no card.
Scan my network — free