Find ten customers before you build anything

Before you register a company, build a website, design a logo or write a line of code, go and have thirty conversations with people who could plausibly pay you. Then try to take money from ten of them.

That's the whole method. It costs you a few weeks and some awkwardness, and it will tell you more than six months of building will.

Most small businesses that fail don't fail because the product was bad. They fail because not enough people wanted it at a price that covered the cost of delivering it. The founder found that out on month nine, with savings gone and a garage full of inventory, when they could have found it out in week three for the price of a lot of coffee.

Compliments are not demand

The first thing to get straight is what actually counts as evidence.

Counts: money in your account. A signed agreement with a start date. A deposit. A purchase order. Someone rearranging their calendar to make the thing happen.

Doesn't count: "That's a great idea." "Definitely send me something when it's ready." "I'd absolutely use that." Email signups on a landing page, on their own. Your brother-in-law's enthusiasm. Anyone saying "you should build this" who isn't offering to buy it.

People are kind. When you describe your idea with your face lit up, most of them will not tell you it's weak. They'll say something encouraging and get on with their day. You'll walk out feeling validated and you'll have learned nothing.

The way around this isn't to be pushy. It's to stop asking about the future and start asking about the past.

Ask about last time, not next time

Hypothetical questions get hypothetical answers. Behavioral questions get facts.

Instead of "would you pay for a service that handles your equipment maintenance," ask:

• "Walk me through the last time a machine went down. What happened?"
• "Who did you call?"
• "What did that end up costing you, including the downtime?"
• "What did you try before that?"
• "Who signs off on spending like that?"

You're looking for three things. Is this a real, recurring irritation? Have they already spent money trying to solve it? And do they, personally, control a budget?

That last one catches a lot of people. The person who feels the pain is frequently not the person who can approve the invoice. If you spend four weeks delighting an operations manager who then discovers she needs a director's signature, you haven't validated anything. Ask early and ask plainly: "If you wanted to go ahead with something like this, whose approval would you need?"

Where to find thirty people

Start with the industry you just left. If you spent eight years in commercial flooring, you already know who runs the crews, who complains about what, and whose number you can call without an introduction. That's not a small advantage, it's the single biggest one available to a bootstrapper.

Beyond that:

• Trade associations and their regional meetings. Unglamorous, poorly attended, full of exactly the people you need.
• Suppliers who already sell to your customer. The rep who sells parts to forty workshops knows which forty. Buy him lunch.
• Trade groups on Facebook and LinkedIn. Don't post a pitch. Ask a genuine question about how people currently handle the problem, and follow up privately with the ones who answer thoughtfully.
• Referral chaining. End every conversation with "who else should I be talking to about this?" One good contact turns into five.

Say clearly what you're doing. "I'm looking at starting something in this space and I'm trying to understand the problem properly before I build anything. Can I have twenty minutes?" That framing gets a yes far more often than a sales call does, and it removes the pressure that makes people flatter you.

Then ask for money

Thirty conversations will teach you the shape of the problem. They won't prove anyone will pay. The pre-sale does.

Pick the sharpest version of the problem you heard. Write one page: what you'll do, what it costs, when it starts. Take it back to the ten people who leaned forward hardest and ask them to commit.

You've got options on how hard you push, and they're not equal:

• A paid pilot at a discount. Real money, real signal. Best available.
• A signed letter of intent with a price and a date. Decent. Weaker than cash, stronger than talk.
• A refundable deposit. Honest, and it gets you past the people who are just being polite.
• A free pilot. Worst of the lot. Free pilots convert badly, and a company will accept anything that's free. You've tested nothing.

If you can't get anyone to commit, you've learned something that cost you almost nothing. That's a successful experiment, not a failure. Most ideas die here, and they should.

Service businesses get a shortcut

If you're selling a service rather than a product, you can skip straight past validation and into revenue, because you're allowed to do the work manually and badly at first.

You don't need software to run a bookkeeping practice for three clients. You need a spreadsheet and your evenings. You don't need a fleet to start a delivery service. You need a van and a phone. Do it by hand, for money, for real customers, and let the systems come later when the manual version breaks under the weight.

This is the part people skip because it doesn't feel like founding a company. It feels like a job with worse hours. It is. It's also how you find out whether the work is something you can stand doing five hundred times, which matters more than most people admit before they commit five years to it.

What goes wrong

Selling only to friends. They buy to be supportive. It tells you nothing about whether a stranger will.

Averaging thirty opinions into one bland product. You're not taking a vote. You're looking for the small group with the sharpest, most expensive version of the problem. Serve them specifically.

Falling for the one enthusiast. There's always someone who loves it. One person with an unusual problem is not a market. Look for the pattern, not the outlier.

Building the demo first. A demo shifts the conversation from their problem to your interface. You'll get feedback on button placement instead of on whether anyone cares.

Taking the first no as the answer. Sometimes the no is about timing, or budget cycles, or the wrong contact. Ask what would need to be true for it to be a yes.

Give it a deadline

Open-ended research becomes procrastination wearing a suit. Put six weeks on it. Thirty conversations, then ten asks. Write down before you start what result would make you proceed and what result would make you walk away, because you will not be able to judge that fairly once you're emotionally invested.

Three paying customers from thirty conversations is a business worth pursuing. Zero from thirty is information worth having, and you got it for the price of six weeks and some petrol.

The founders I've watched do this well weren't smarter than the ones who didn't. They were just more willing to hear no early, while no was still cheap.