I want to start a business but have no idea what kind. How do I choose?

Choose by constraint, not inspiration. Three questions eliminate most options: how much can you genuinely spend without fear, what skill or knowledge do you already have that people pay for, and how fast do you need income. A $2k budget and a need for income in 60 days points at services. $50k and patience points at e-commerce or buying a business. Our free calculator runs exactly this matching in four questions.

Do I need an LLC before I start?

No, and starting with paperwork is the most popular form of productive procrastination. In the US you can operate as a sole proprietor and form the LLC once revenue is real; most founders form one somewhere between their first client and their first $10,000. The exceptions are genuinely risky work (anything with liability exposure) and partners: form the entity before mixing money with a co-founder.

How much money do I need to start?

Less than you think, in most cases. Service businesses start under $2,000. A proper one-product online store runs $5,000 to $9,000. A staffed operation or acquisition wants $50,000. And with $0 you can still start by selling a skill, paying with hours instead of dollars. The budget mostly picks the business for you, which is a feature: fewer options means faster starts.

Do I need a website first?

Not first. Your first customers should come from outreach, platforms, marketplaces, or people who already know you: channels that need no website. The site earns its cost the moment you have real proof to put on it. At that point a fast one-pager with your results and a booking link does the whole job, and it costs hundreds, not thousands. Building the website first is decorating an empty shop.

How do I know if my business idea will work?

One real sale beats every method: a stranger paying money before you build much of anything. Pre-sell the service, take deposits, sell the first unit before ordering the thousand. If you cannot find one person to pay, more building will not fix that. Surveys, courses, and business plans are all ways of postponing this test, which is exactly why they feel safer than running it.

How long before I should expect income?

Services: first payment within 2 to 6 weeks of actually selling. E-commerce: 1 to 3 months to launch, longer to profit. Content businesses: 6 to 12 months. If someone promises faster, check what they are selling you. Then give whatever you pick six months of consistent effort before you judge it. Most 'failed' first businesses were abandoned, not disproven.

I Want to Start a Business. What Should I Do First? (2026)

When you want to start a business but do not know where to start: first pick by constraint, not by idea. Your budget, your skills, and how fast you need income eliminate 90% of options for you. Then validate with one real sale before spending on anything: no LLC, no logo, no website yet. The whole first month should cost under $100. Legal structure, branding, and tools come after a stranger has paid you, because that one payment is the only proof that matters.

“I want to start a business” is usually followed by six months of researching LLC formation, buying a domain, and watching other people’s income-report videos. Here is the shorter path, in order, with what each step costs. The first month costs less than $100, which is the point: the expensive-feeling steps come after proof, not before.

1. Pick by constraint, not by idea

The idea-first approach stalls because there are infinite ideas. Constraints are merciful: they eliminate. Your real budget, the skills you already have, and how fast you need income cut the field from everything to a handful. That matching is exactly what our what business should I start calculator does in four questions, free. If you would rather read the long form: $0 options, $10k options, $50k options. Notice the pattern across all three: fast money is services, big ceilings are products, and patience is the price of the second one.

2. Budget honestly, including the part after launch

One constraint that has genuinely moved since last year: software. A founder with no coding background can now build their own internal tools, and often their first product, with an AI coding agent; we wrote the honest guide to doing that with Claude Code.

Whatever the business, the startup cost has two halves: what it takes to open, and what it takes to be found. The second half is the one first-timers skip. A store without a marketing budget is a brochure in a drawer, and we wrote down the full real cost of an online store because almost nobody publishes the second half. Rule of thumb: hold back a reserve of 20 to 30% of your total budget, untouched, for the surprise that is coming either way.

3. Get one real sale before you build anything else

The step people skip because every other step feels safer. Pre-sell the service to one client. Take one deposit. Sell ten units before ordering five hundred. One stranger paying real money outranks every survey, plan, and course in existence. The SBA’s market-research guidance is solid on sizing a market, but a market of one paying customer is the only sample that cannot argue back.

4. Do the legal minimum, then move on

In the US: you can operate as a sole proprietor today, and form the LLC when revenue is real. Separate business bank account from the first dollar, basic bookkeeping from the first expense, licenses if your trade needs them. That is the whole list at this stage. Paperwork is the most respectable way to avoid selling, which is why it is so popular.

5. Put up the minimum online presence, not a flagship

When outreach and platforms are producing conversations, strangers will check whether you are real. That check needs one fast page: what you do, proof, a way to book or pay. It does not need a five-page site with an About video. We build exactly this minimum credible presence for new businesses, usually in a day, and our honest advice sits one guide over: do not buy it before step 3.

6. Spend in the order that compounds

After the first sales, spend where hours leave: the tools or automation that remove your admin, better photography if you sell products, a real site once proof exists, help once you are turning work away. The founders who fail with decent budgets almost always spent in the reverse order: office, brand, tools, then looked up and the runway was gone before the market ever answered.

7. Give it six months, on purpose

Decide the review date in advance and do not renegotiate with yourself weekly. Most first businesses that “failed” were abandoned in month three, right about when the novelty wore off and just before word of mouth would have started. Six months of one idea beats six years of six ideas. The market rewards the person still standing there.

Start with the calculator. Four questions, your real numbers, and step 1 is done in a minute.