Win the first customers

The first customers are won through proximity, specificity, and controlled proof. You do not yet need a funnel at scale. You need a few people who will expose the truth.

A polished website can launch into complete silence. That silence does not prove the product is poor. It proves that publishing is not distribution. At the beginning, nobody owes the new business attention, and the founder has little reputation to borrow. The work is manual: identify likely buyers, earn conversations, ask about real behaviour, propose a small paid result, and stay close enough to see what delivery actually requires. The goal of the first customer is not maximum revenue. It is maximum learning within a safe, fair commercial exchange.

Start with the trust map

Build a manageable list—perhaps thirty to fifty names as an operating heuristic, not a universal target— across these circles:

1. people who have seen your work directly; 2. introductions from people whose judgment the buyer trusts; 3. organizations that show the problem signal publicly; 4. professional communities where the problem is discussed;

5. adjacent providers who serve the same customer without competing for the same outcome.

Specific requests travel. “Do you know anyone who needs software?” makes the helper search the entire world. “Do you know one operations lead at an independent recruitment firm handling high-volume screening?” gives memory a usable index. Approach the first conversations as research, not concealed pitching. State why you chose them, the pattern you are trying to understand, and the small amount of time requested. Ask for the last occurrence, the workaround, the consequence, the decision owner, and what they have already tried. Do not count praise as demand.

Climb the commitment ladder

Different actions prove different things:

The first paid proof should be small enough to control and large enough to matter. Define the customer, workflow, baseline, scope, exclusions, success measure, time-box, price, responsibilities, data boundaries, and decision after the test. “Free pilot” often removes the very commitment you are trying to observe. If a free diagnostic is appropriate, bound it tightly and make clear what it can and cannot establish.

Payment is evidence, but it is not permission to overpromise. A customer may pay for a weak solution because the problem is urgent, the buyer misunderstood, or the salesperson concealed risk. The ethical test remains delivery and informed choice.

Give each early customer a different learning job

Use a sequence, not as a law but as a way to resist premature scale:

The first customer tests access and truth. Choose a buyer who will show the workflow, discuss failure, and tolerate a narrow scope. Optimize for a clean learning relationship.

The second tests repeatability. Stay close to the same segment and problem. Identify what was universal, what was configuration, and what was custom work hiding inside the first result.

The third tests buying and onboarding. Improve qualification, scope, payment, access, kickoff, and time to first value. A product that works after heroic onboarding is not yet a repeatable offer. The fourth tests delivery without founder rescue. Use checklists, acceptance criteria, exception handling, and explicit owners. Observe where judgment remains trapped in the founder’s head.

The fifth tests continuation. Compare the result with the baseline and ask whether it earns renewal, expansion, a reference, or a referral. Accept “no” as information.

The number five is a teaching device, not statistical proof. Five customers can expose a pattern. They cannot establish the size of a market, the long-run churn rate, or channel economics.

Reduce risk without giving away the company

An unknown provider asks the buyer to take vendor, delivery, adoption, and reputation risk. Reduce what you can:

• • show relevant work and its limitations;
• • narrow the first scope and environment;
• • agree on a baseline and acceptance test;
• • separate test from production where consequences are material;
• • use milestone payments and clear stop conditions;

• • identify customer responsibilities and access needs;
• • document confidentiality, data handling, ownership, and exit as appropriate;
• • keep human review where an error could cause material harm;
• • never promise an unlimited remedy the business cannot survive.

Risk reversal should assign risk to the party best able to control it. It should not transfer every uncertainty to a small supplier merely to force the sale.

Ask for proof you have earned

At the end, write down what changed, how it was measured, what remained uncertain, and what the customer would tell a peer. If the customer agrees, ask for one specific introduction. Make the context easy to forward: problem, result, limitations, and the kind of person for whom the conversation would be relevant.

Do not request a public testimonial before the customer has lived with the result. Do not write the praise and ask the customer to endorse a claim they cannot verify. One accurate reference from a suitable customer is more valuable than a wall of vague approval.

· TABLE pp. 13–16 and 34–36 · SALES pp. 4–13 and 35–37 · ASC pp. 86–94 · GE ch. 14. The five-customer sequence and list size are operating heuristics, not market-validation thresholds; corporate anecdotes are not used as proof.