Most guidance on starting a business concentrates on registration, logos and paperwork. The part that decides whether the business survives is messier: the first ninety days, when you are selling something nobody has bought yet and discovering how wrong your assumptions were. Treat that period as a test with an end date, not a launch with a ribbon cutting.

Week one: write down what you believe

Before you spend money, put your assumptions on a single page. Who is the customer, described specifically enough that you could point at them in a queue? What do they use instead of you right now? What do you think they will pay? How many sales per week cover your costs? These guesses will mostly be wrong, but written guesses can be corrected. Unwritten ones quietly harden into habits.

Keep that page visible. Every conversation over the next three months should confirm or contradict one line of it. When a line survives ten conversations unchanged, you can start trusting it.

Weeks two to four: sell before you build

The cheapest way to test demand is to ask for money early. A landing page with a clear offer, a few conversations with people who fit your customer description, and a simple way to pay will tell you more than a month of planning. If nobody will commit to a small deposit or a booked slot, the problem is the offer or the audience, and you have found out before buying stock or signing a lease.

Resist the urge to soften the test. Friends who say your idea sounds lovely are not evidence. A calendar booking is evidence. So is a refund request.

Weeks five to eight: build the smallest version that works

Deliver something real, even if it is narrow. One product, one service package, one type of customer. Narrow scope is not a lack of ambition; it is how you find out which part of the work you are actually good at and which part eats your day. Most first businesses fail on delivery, not on demand.

Track three numbers from the first week of trading: how many people asked, how many bought, and how long each job took you. Those three numbers will price your next month more accurately than any spreadsheet forecast written in advance.

Weeks nine to twelve: fix the machine, not the brand

By now you have a rough sense of where the friction sits. It is usually boring: quoting takes two hours, follow-up emails get forgotten, packaging runs out at the wrong moment. Fixing those frees more capacity than a new logo ever will. Write a one-page routine for enquiries, delivery and payment, and follow it for a fortnight before changing anything else.

The first ninety days are not about proving the idea is good. They are about finding out, cheaply, which parts of it are true.

Money rules that keep you solvent

Keep a cash buffer for at least three months of fixed costs, and know the number. Separate business money from personal money from day one, even if it is a second personal account at first. Invoice immediately, chase politely at seven days, and refuse to fund growth from a credit card you cannot clear monthly.

Set a spending threshold. Anything above it waits seventy-two hours. Most early purchases feel urgent and turn out to be optional, and the seventy-two hour rule costs nothing to apply. Write the threshold down, because the version of you that wants a new laptop at midnight is not the version that set the budget.

A ninety-day checklist

  • One page of written assumptions, reviewed fortnightly
  • Ten conversations with people who match your customer description
  • A priced offer that someone can actually buy this month
  • Business bank account and a simple record of every payment in and out
  • Three tracked numbers: enquiries, sales, delivery time
  • A named person to call when you are unsure about tax or contracts
  • A review date in the diary, with the honest question: continue, adjust or stop?

That last item matters most. Decide in advance what evidence would make you change course, so the decision is made by the plan rather than by exhaustion. A business that ends after ninety days with clear lessons is not a failure. It is a cheap experiment that saved you three expensive years.