From Business Idea to First Customer

Most business ideas do not fail because they were bad. They fail because eight months went by and nobody was ever asked to pay. The founder was busy — the logo, the entity, the website, the pitch deck, the second version of the logo — and the whole time the only question that mattered went unasked.

This is the ninety-day version of asking it. It assumes you have an idea, a job or other obligations, somewhere between a few hundred and a few thousand dollars you can risk, and roughly eight to twelve hours a week. It is written for service businesses, software, and small physical products, and it flags where those diverge.

The goal is not a launch. The goal is one person, not related to you, choosing to send you money.

The 90-day plan at a glance

The costs above are the ones that buy you information. Money spent outside this table before day 90 — brand systems, ads, custom software, an office — is usually money spent to feel like a founder rather than to become one.

Days 1–7: the week before you spend anything

Write the idea in one paragraph, in this shape:

I help [specific group of people] do [specific thing] so they can [outcome they care about]. They pay me [amount] [per what].

If you cannot fill in the amount and the per-what, you do not have a business yet, you have a topic. That is fine on day one. It is not fine on day thirty.

Then do the three things that cost nothing:

Name 30 real people. Not "small business owners in Tampa". Thirty rows in a spreadsheet with a name, how you would reach them, and how you know of them. If you cannot get to thirty, that is your first real finding — either the customer group is too vague or you have no route into it, and both need solving before anything gets built.

Spend ninety minutes finding out who already does this. Search the terms a customer would actually type. Look at marketplaces, local listings, industry directories, the trade groups on Facebook where people complain. You are looking for evidence that money already changes hands for a worse version of what you want to sell. Finding five competitors is good news. Finding none is usually bad news, and the reason is almost never that nobody thought of it.

Write down what "no" looks like, before you are emotionally invested. A kill number and a date. Something like: if by day 90 I have not collected 500 dollars from three people I am not related to, I stop or change the offer. Write it in a file with the date. This single sentence is the difference between a ninety-day experiment and a three-year hobby that quietly drains a savings account.

What not to do this week: form the entity, buy the logo, register five domains, build the deck, design the packaging, or pay anyone. The average first-two-weeks spend on that list runs 800 to 3,000 dollars and produces exactly zero information about whether anyone wants this. You will do some of it later. Later is cheaper.

Days 8–21: talking to people who might pay you

Twenty conversations in two weeks. Fifteen to twenty-five minutes each. That is about two a day, and it is the highest-value work in the entire ninety days.

Getting the conversations

Start with people you can already reach: your phone contacts, your sent email folder, past colleagues and clients, the group chats, the trade association, the parents at the same school, the alumni list. Warm routes get replies. Cold ones mostly do not, at this stage, and you do not need them yet.

The ask should be short, personal, specific and easy to refuse:

Hi Dana — you have been running the shop for years and I am looking at something in that space. Could I ask you fifteen minutes about how you handle [specific thing]? Not selling you anything. Happy to work around your schedule.

Roughly a third to a half of warm contacts will say yes to that. Expect to send fifty messages to book twenty conversations.

What to ask

Ask about the past, not the future. People are unreliable about what they would do and quite reliable about what they did.

1. Walk me through the last time you dealt with [the problem].
2. What did you do about it?
3. What did that cost you — money, time, or a customer?
4. Who did you pay, or what did you use?
5. What happened next? Did it come back?
6. Who else deals with this worse than you do?

Question six is quietly the most valuable one — it is how twenty conversations become forty without cold outreach.

Never ask "would you use this?" or "would you pay 200 dollars for it?". People are polite, and enthusiasm in a conversation costs them nothing. A friendly yes from someone with no budget line has bankrupted more founders than any competitor ever has. This is the same discipline covered in more depth in how to validate a business idea.

What you are listening for

The strong signals are behavioural, not verbal:

• They already pay someone for a worse version of this.
• They built a workaround — a spreadsheet, a part-time hire, a scary manual process.
• They ask when they can have it, unprompted.
• They try to hand you money in the conversation. Take it.
• They refer you to someone before you ask.

The weak signals: "that is a great idea", "I would definitely use that", "let me know when it launches". Those are politeness. Log them, weight them near zero.

Keep a simple sheet: name, how you know them, the problem in their exact words, what they currently pay, urgency from one to five, and whether they would take another call. That last column becomes your first sales list in phase five.

If you get through twenty conversations and nobody has a budget, a workaround, or a story about it costing them something, you have found a real problem that nobody will pay to fix. That is a genuine result. It cost you two weeks and no money, which is the entire point of doing it in this order.

Days 22–30: deciding what to build first

Split what you imagined into three parts:

• The promise. What the customer gets. The clean roof, the finished books, the hired candidate, the working integration.
• The delivery. How it actually gets done this month.
• The machinery. The software, automation and systems that would let it happen a hundred times without you.

Build the promise. Deliver by hand. Skip the machinery entirely until customer ten.

This is the part founders resist hardest, and it is where the money goes. A dashboard nobody has asked for costs 8,000 to 30,000 dollars and four months. The same job done with a shared spreadsheet, a scheduled email and a Zoom call costs nothing and teaches you what the dashboard should have been.

Two hard limits for the first version. If it will take more than four weeks of building or more than about 5,000 dollars before anyone can pay for it, you have picked the wrong first version — go smaller. The exceptions are real: physical goods need samples (commonly 300 to 3,000 dollars and three to eight weeks), and regulated work — food, childcare, health, anything touching other people's money or homes — has licensing you cannot manually work around. If that is you, sell first and deliver on a stated future date where the law allows it, and get the licence in parallel. The sequencing question is worked through properly in from idea to MVP.

Days 31–45: the minimum you actually need to take money

There is a list of things you truly need before a stranger can pay you, and it is much shorter than the internet suggests.

A realistic total for a service business is 150 to 800 dollars. A small physical product with a first inventory run is more like 1,000 to 5,000. Compare that to the 6,000 to 15,000 dollars people commonly spend on a "proper launch" before a single customer exists. The full breakdown lives in how much it costs to start a business.

None of this is legal or tax advice, and licensing rules vary enough by state and trade that you should check yours directly. But note what is not on the list: a logo, a trademark, business cards, an office, a CRM, or a finished product.

The one item people underweight is the written promise. One page stating what is included, what is not, when it happens and what a refund looks like prevents the specific disaster where you and your first customer had different pictures in your heads, and you work six extra weeks free to protect a testimonial.

Pricing the first offer

Do not launch at your lowest possible price. Cheap does not attract customers; it attracts the worst ones, and it caps everything you can afford to do afterwards.

Three ways to set the first number:

Replacement cost. What does solving this cost them today, including their own hours? An owner spending six hours a month on the problem at an effective 60 dollars an hour is losing 360 dollars a month to it. That is your reference point, not your floor.

Share of the outcome. If the work plausibly creates a measurable gain, ten to twenty percent of the first year of that gain is defensible and easy to say out loud.

Competitive anchor. Find three published prices for something similar. Sit in the middle unless you have a specific reason to sit at the top. Being new is not a reason to sit at the bottom.

Then pick the shape:

• One-off project. Simplest to sell, hardest to live on. Good for a first ten.
• Retainer. Better business, harder first sale. Ask for it once you have delivered a project well.
• Per unit or per job. Natural for trades and products, and the easiest to quote badly — count your hours on the first three jobs before you set the standing rate.
• Subscription. Only if the value genuinely recurs. Monthly billing for a one-time outcome creates cancellations, not revenue.

For the first customers, use a stated first-client price rather than a permanent discount: the price is 1,200 dollars; the first five clients pay 750 in exchange for a testimonial and permission to reference the work; that ends on the 30th. This gives you a real number to raise to, a reason for the discount that is not "I am new and unsure", and a deadline that creates a decision.

Two mechanics matter more than the number itself. Take a deposit — half up front for project work, all of it under 500 dollars. And watch the yes rate: if everyone agrees immediately and nobody questions the price, you are too cheap. Raise the next quote thirty percent, and keep raising until roughly one in three says no.

Days 46–75: getting the first ten customers from people you already know

Your first ten customers will come from your existing network and one narrow channel. Not from ads, not from an audience, not from a launch post. Accepting this early is worth thousands of dollars.

Build the 100-name list

Go through your phone contacts, your sent email, your professional connections, past employers and clients, group chats, neighbours, the gym, the school, the alumni association, the trade you used to work in. Get to a hundred names. It takes about two hours and feels awkward, which is why most people skip it and then buy ads instead.

Sort each name into one of three buckets:

• A — could buy this themselves.
• B — knows several people who could.
• C — credibility. Their forwarding it means something.

Ask directly

Personal messages, sent one at a time. Never a mass blast, never a broadcast post as your primary move. The shape is: a genuine personal opener, one sentence about what you are doing, one specific ask, and an easy way out.

Hi Marcus — congratulations on the move to the new unit. I have started doing [specific thing] for [specific type of business], and you are the person I most wanted to ask. Two things: would this be useful to you right now, and if not, is there someone in your group I should be talking to? Completely fine if the answer to both is no.

For a B or C contact, the ask is a referral, not a sale. Make it easy: tell them exactly who you are looking for, in one line they could forward.

As a planning assumption — not a promise, and it varies enormously by offer and network — plan for something like a hundred personal messages producing thirty to forty replies, ten to fifteen real conversations, and a small handful of paying customers for an offer priced between 500 and 5,000 dollars. If your numbers come in far below that, the problem is usually the offer, not the list.

Pick exactly one channel beyond your network

One. Run it for a full thirty days before judging it. Reasonable candidates: local search and a Google Business Profile for anything geographic; one community where your customers already gather and where you participate honestly; a partner who already sells to the same people; narrow cold outreach to a list you built by hand; showing up in person where the customers are.

What not to do in the first ninety days: paid ads before you know which message converts (you will spend 1,000 dollars to learn what twenty conversations tell you free); building an audience as an acquisition plan (six to twelve months before it pays); a podcast; a rebrand.

Ask for the referral at delivery

The moment a customer says the work was good is the moment to ask who else needs it. Not next quarter. Write it into your own delivery checklist so it happens every time, and it becomes the second ten customers.

Days 76–90: reading the result honestly

Go back to the kill number you wrote on day five and compare. There are three outcomes and they are easy to tell apart if you are honest.

It works. Three or more paying customers, at least one of whom is not a friend. Somebody pre-paid, repeated, or referred. Delivery hurt but not impossibly. Now build the second version, raise the price, and start on the machinery you deliberately skipped. The launch checklist covers what to formalise at this point.

It half works. People pay, but only friends. Every sale needs a discount. Delivery takes four times what you priced. The customers are happy but never come back. This is a positioning and offer problem, not a demand problem, and it is fixable — but not by working harder at the same thing.

Nobody paid. Twenty-plus conversations and a hundred asks produced no money. This is real information, and it cost you a quarter and a few hundred dollars instead of two years and your savings.

What to change when the answer is no

Change one variable, not all of them:

1. Same customers, different problem. Usually the best move. You now know a specific group of people and they will take your call. Ask them what they would have paid for instead.
2. Same problem, different customer. The problem may be real but concentrated in a group with more budget or more urgency.
3. Same everything, different delivery. Done-for-you instead of a tool. A tool instead of done-for-you.
4. Same everything, different price or package. More common than founders expect — a monthly fee that should have been a project, or 200 dollars that should have been 2,000.
5. Stop. A legitimate outcome, and the cheapest one available if the answer is genuinely no.

The asset you built in ninety days is not the idea. It is the list, the twenty conversations, the language your customers use, and the fact that you now know how to run this loop. All of it carries into the next attempt. The idea itself was always the least valuable part.

Sunk cost has no vote here. The money and the weeks are gone regardless of what you do next. The only live question is what the next ninety days should contain.

What most people get wrong, and what it costs

The pattern in every row is the same: spending money to avoid the discomfort of asking someone to buy.

The 90-day checklist

Do these in order. Anything you have already done, tick and move on.

1. Write the idea as one paragraph in the "I help X do Y so they can Z, they pay me N" format.
2. List 30 named potential customers with a contact route for each.
3. Spend ninety minutes finding out who already sells something similar, and at what price.
4. Write down your kill number and kill date. Save the file.
5. Send fifty conversation requests to warm contacts.
6. Complete twenty conversations of fifteen to twenty-five minutes.
7. Log every one: their words, their current spend, urgency, willingness to talk again.
8. Identify the strongest recurring problem across those twenty.
9. Decide the offer: what they get, by when, for how much.
10. Split it into promise, delivery and machinery. Cut the machinery entirely.
11. Set the price using replacement cost or share of outcome, then add thirty percent.
12. Write the one-page promise: inclusions, exclusions, timeline, price, refund terms.
13. Set up a payment link and an invoice template.
14. Put up one page of copy with a way to contact you.
15. Get email on a domain and a phone number you answer.
16. Sort out the entity, bank account, licence and insurance your trade actually requires.
17. Build the 100-name list and sort into A, B and C.
18. Send a hundred personal messages, one at a time, over three weeks.
19. Choose exactly one channel beyond your network and work it for thirty days.
20. Deliver the first jobs by hand. Track how long they really take.
21. Ask for a referral the moment a customer says the work was good.
22. On day 90, compare against your kill number and make a written decision.

The short version

• Two weeks of conversations before a dollar is spent will tell you more than four months of building, and the people who skip it commonly lose 4,000 to 15,000 dollars finding out.
• Ask about what people did last time, never about what they would do. Politeness is not demand.
• Build only the promise. Deliver by hand until customer ten — the machinery costs the most and teaches the least.
• The real minimum to take money is 150 to 800 dollars for a service business: a page, a payment link, an email, a written promise, and whatever licence your trade requires.
• Your first ten customers come from a hundred people you already know plus one narrow channel. Ads do not fix an offer that has not sold yet.
• Write the kill number on day five. It converts an open-ended dream into a ninety-day experiment you can afford to lose.