There is no shortage of lists of steps for starting a business. What most of them get wrong is the order, and the order is the part that matters, because the early steps are cheap and reversible and the later ones are neither.
Step one: establish that someone will pay
Everything downstream depends on this and nothing substitutes for it. Not interest, not encouragement, not a survey in which people say they would buy. A sale, at a price, to someone with no reason to be kind.
For a service, that is a first paying client. For a product, a pre-order or a pilot. For retail or food, a market stall or a pop-up. The point is not the revenue; it is that the answer changes what you build next, and it changes it before money is committed.
Successful startups differ from unsuccessful ones far less in the quality of the original idea than in how early they found out what was wrong with it.
Step two: work out the numbers that constrain everything
Two figures. Break-even: the volume at which revenue covers costs, and runway, the months you can fund before reaching it.
These are not forecasts. They are constraints, and they should be calculated with pessimistic inputs, because an optimistic break-even is simply a wrong number. If the volume required looks implausible at the price you can charge, the business model has a problem that effort will not fix.
Step three: confirm it is permitted
Where the business will operate from, whether that use is allowed there, and what licensing the trade requires. This step is free and fast and it is the only one that can end the plan outright, which is why it comes before anything is signed.
The licenses and permits page covers the general shape; the business guides cover six trades in detail.
Step four: choose the structure
Sole proprietorship, partnership, LLC or corporation: the decision that is genuinely expensive to reverse. It turns on two questions: is there more than one owner, and does the business create liability that insurance will not cover. The forms of business organization page works through the alternatives.
Do this before taking on debt, signing a lease or bringing anyone else in. Converting afterwards means renegotiating everything already signed.
Step five: fund the gap
Most new businesses are funded by the founder, by a community lender, or by both. Equity is a narrow instrument suited to a small number of businesses, and the venture capital page is blunt about which.
The important discipline at this stage is to borrow for something that generates a return: equipment, inventory that turns, a fit-out that opens a location, not to cover a gap created by a weak model. Debt extends the runway; it does not fix the economics.
Step six: build the minimum that lets you trade
A bank account in the business name, a way to take payment, a record-keeping system set up before the first transaction, insurance for the work, and the licenses in hand.
Everything beyond that list is deferrable and most of it should be deferred. The branding, the office, the software stack and the website that takes three months are all ways of being busy without selling.
What goes wrong, and when
The failures cluster at predictable points.
Between step one and step two a founder who has made a few sales concludes the model works, without ever calculating the volume required at scale. The early sales came from people who already knew them; the hundredth will not.
At step three a lease gets signed before the zoning is confirmed, which converts a flexible plan into a fixed monthly obligation on premises that may not be usable.
At step five money is raised to postpone a decision, not to fund a plan, which buys time at the cost of ownership and ends in the same place later.
The sequence above is not a guarantee. It is an ordering that keeps the cheap decisions cheap and delays the expensive ones until there is evidence to base them on.
What to do in the first ninety days of trading
The steps above get a business to its first sale. The next period has its own short list.
Set up the records before there are many. A separate bank account, a bookkeeping method, and a habit of reconciling weekly. A business that reaches month six with a shoebox spends a weekend fixing it and has learned nothing in the meantime about its own numbers.
Measure where customers come from. Ask every one, write it down. After a month the pattern is visible and it is nearly always concentrated in one or two sources.
Find the real cost of delivery. Not the materials: the time. Most new businesses underprice because they never counted the hours a job actually takes, including the parts that are not the work: the quote, the travel, the chasing.
Diary the obligations. Sales tax filing dates, license renewals, the annual report. These are the things that produce penalties for no reason other than having been forgotten.
The failure modes worth naming
Building for a year before selling. Common in product and software businesses, and the year usually ends with something nobody wanted. The first step on this page exists to prevent it.
Confusing activity with progress. A website redesign, a new logo, a better accounting package: all of these feel like running a business and none of them produce a customer.
Not charging enough. Underpricing is recoverable and every month it continues makes it harder, because the existing customers are the ones who will notice.
Hiring too early. A hire is a fixed cost against revenue that has not yet proved durable, and it is the commitment most likely to end a business that was otherwise working.
Refusing to stop. The hardest one. A business that has not found demand after a fair test is giving information, and continuing to fund it from savings or from debt is a decision that should be made deliberately.