# How Do I Start a Business? A Build-In-Public Playbook

> Source: [https://botensten.com/articles/how-to-start-a-business-startup-bootstrapping-playbook](https://botensten.com/articles/how-to-start-a-business-startup-bootstrapping-playbook) (canonical)
> Author: Botensten — Botensten, https://botensten.com
> Published: 2026-08-20

## TL;DR

To start a business, pick a specific problem people already pay to solve, then confirm demand by getting a real payment or signed commitment before you build. Register a simple legal entity (usually an LLC), open a business bank account, and keep fixed costs near zero. Price for profit from day one and reinvest revenue instead of raising money you don't need. Ship a small paid version fast, talk to customers weekly, and let real usage — not a business plan — decide what to build next.

Roughly half of new U.S. businesses survive to their fifth year, according to Bureau of Labor Statistics data — so starting one is less about a dramatic launch and more about surviving long enough to learn what customers actually want. Pick a problem people already pay to solve, sell the solution before you build it, register the simplest legal entity you can, and keep fixed costs near zero until revenue is real.

## What Does It Actually Take to Start a Business?

Starting a business takes three things: a problem worth solving, a way to charge for the solution, and a legal wrapper to collect the money. Everything else — the logo, the perfect website, the 40-page business plan — is optional at the start.

You don't need permission, a co-founder, or a novel idea. Most durable businesses solve an old problem slightly better or for a narrower group. "Bookkeeping, but only for dentists" beats "a revolutionary new platform" almost every time, because a specific buyer is easier to find and easier to charge.

The odds reward patience over hype. [Bureau of Labor Statistics data on business survival rates](https://www.bls.gov/bdm/entrepreneurship/entrepreneurship.htm) shows about 20% of new businesses close within the first year and roughly half are gone by year five. Most of those deaths come from spending money before earning it, not from a bad idea. The renter-to-owner move is simple: own the thing that makes you money instead of renting a stack of tools that raise prices every year.

## How Do You Validate an Idea Before You Build?

Validate an idea by collecting money or a firm commitment before you write a line of code or sign a lease. If nobody will pay a deposit, pre-order, or sign a letter of intent, the demand isn't real yet — it's just a compliment.

Cheap ways to test demand before building anything:

1. Sell a "concierge" version by hand — deliver the outcome manually for your first three to five paying customers.
2. Put up a one-page site with a real Buy or Pre-order button and send a little traffic to it.
3. Offer a paid pilot to one company at a discount in exchange for honest feedback.
4. Post the offer where your buyers already gather and count replies, not likes.

Talk to ten real potential customers before you build for a hundred imaginary ones. Sahil Lavingia's *The Minimalist Entrepreneur* makes the same case: reach profitability with a small group of paying customers before you scale anything.

## What Should You Spend Money On First?

Spend on the few things that let you collect payments and talk to customers; skip everything that only looks like a business. The [U.S. Small Business Administration's business guide](https://www.sba.gov/business-guide) covers the boring-but-required steps — structure, registration, taxes — and most of them cost little or nothing.

| Spend on this first | Skip until you have revenue |
|---|---|
| Registration & business bank account ($0–$300) | Custom logo and branding agency |
| A domain and a simple landing page (~$12/yr) | Office or coworking space |
| Payment processing (Stripe, ~2.9% + 30¢/charge) | Paid ads before product-market fit |
| Time with 10 real potential customers | SaaS tools you "might need later" |

## How We Ship a Software Product With AI in a Week

At Botensten we ship production software with AI every day, so here's what a real first week looks like instead of theory. On day one we write the offer as one sentence and put up a landing page with a working payment link — no product yet, just a promise and a price. If a few people pay or pre-order, we build; if not, we change the offer and try again.

By midweek we have a thin, working version. Our stack is deliberately cheap to own: an AI coding assistant for the heavy lifting, Bun and SQLite for the backend, Stripe for money, and a small server that costs a few dollars a month. Total fixed cost is often under $50 for the first month, which means we can be wrong cheaply.

What breaks is always the boring part. The AI writes a feature that looks right but silently fails on the second user, or a payment webhook fires twice and double-charges someone. So we test the real path — actually clicking "buy" as a customer — before calling anything done. The lesson we relearn constantly: shipping the happy path is easy, and the money lives in the edge cases you'd rather ignore.

## Should You Bootstrap or Raise Money?

Bootstrap unless you have a proven, capital-hungry model that can't grow any other way. For most software and service businesses, funding growth from your own customers keeps you in control and forces the discipline that outside money removes.

| Bootstrapping | Raising money |
|---|---|
| You keep full ownership and decisions | You trade equity and control |
| Customers fund growth | Investors fund growth |
| Slower, steadier, profit-first | Faster, riskier, growth-first |
| Fits solo and small teams | Fits winner-take-all markets |

Books like *Profit First* by Mike Michalowicz, *Company of One* by Paul Jarvis, and *Rework* by Jason Fried and David Heinemeier Hansson argue the same thing from different angles: profit and freedom usually beat vanity growth. Take the money route only when the math genuinely demands it.

## What Are the First Steps to Take This Week?

Do the smallest set of steps that lets a stranger pay you. You can finish most of this in a weekend for under $300.

1. Write your offer in one sentence: who it's for, the problem, and the price.
2. Register a simple entity (an LLC is fine for most) and open a business bank account. Tools like [Stripe Atlas](https://stripe.com/atlas) bundle incorporation and banking if you want it fast.
3. Put up a one-page site with a real payment or pre-order button.
4. Message 10 people who have the problem and ask them to try it.
5. Get one real payment, then ask that customer what to build next.

Start small, charge from day one, and let paying customers — not a plan — tell you where to go. That's how you become the owner of your business instead of a renter of someone else's tools.

## Related reading

- [How Do You Bootstrap a Startup? The Honest Playbook](/articles/how-to-bootstrap-a-startup-2)
- [What Is Bootstrapping in Startups? An Honest Answer](/articles/what-is-bootstrapping-in-startups)
- [Bootstrapping vs Venture Capital: Pros and Cons](/articles/bootstrapping-vs-venture-capital-pros-cons)
- [Agency vs SaaS: Which Business Should You Build?](/articles/agency-vs-saas)

## Frequently asked questions

**How do I start a business or startup?**

Pick a specific problem people already pay to solve, confirm demand with a real payment or commitment, register a simple entity, and keep costs near zero while you charge from day one.

**How much money do I need to start a business?**

For most online or service businesses, under $300 covers registration, a domain, and payment processing. Keep fixed monthly costs below $50 until revenue is real.

**Do I need an LLC to start?**

An LLC is a common, low-cost choice that separates personal and business finances, but you can validate an idea and take payments before formalizing. Check the SBA guide and your local rules.

**How do I know if my business idea is good?**

You don't know until someone pays. Collect a deposit, pre-order, or signed pilot before building; if no one will commit money, the demand isn't proven yet.

**Should I bootstrap or raise venture capital?**

Bootstrap unless your model genuinely can't grow without heavy upfront capital. Customer revenue keeps ownership and control with you and forces profit discipline early.

**What percentage of new businesses survive?**

Bureau of Labor Statistics data shows about 20% close within the first year and roughly half survive to year five. Surviving long enough to learn matters more than a big launch.

**How fast can I actually launch?**

You can register an entity, publish a one-page site with a payment button, and take a first order in a weekend. Building the full product comes after someone pays.