The number you actually need is the one that carries you through that period. Work backward from monthly outgoings rather than forward from a launch budget, because the launch is a single event you can control and the months afterward are not.
The three things to add together
Start with what the business costs to exist each month, whether or not it sells anything. Software subscriptions, insurance, any licence or registration fees, phone, hosting, and anything else that arrives regardless of activity. This figure is usually smaller than owners expect and it is the easiest to establish precisely.
Then add what you need to live on, which most plans omit entirely and which is the largest number in the calculation. A business that cannot pay its owner is subsidised by savings, credit, or another income, and pretending otherwise produces a plan that looks viable and is not. Include the tax on that draw, because a portion of what you take is owed rather than yours.
Then add the one time costs of starting: registration, equipment, initial inventory if you sell goods, a website, and any deposit or upfront purchase. This is the figure people usually mean by startup costs and it is frequently the smallest of the three.
Multiply the first two by the number of months you expect to reach reliable revenue, add the third, and that is the honest requirement.
How many months to assume
Longer than the plan says. Most first year businesses take somewhere between six and twelve months to reach revenue they can depend on, and the ones selling to other businesses take longer than those selling to consumers because the buying cycle is slower and the first sale requires more trust.
The useful adjustment is to take your own estimate and add half again. Not out of pessimism but because the estimate was made before you knew how long the first sale would take, how many prospects would go quiet, or how much time would go to work that does not produce revenue. Every business discovers those, and none of them appear in the version of the plan written in month zero.
If that total is unreachable, the answer is not to assume a faster ramp. It is to reduce the monthly requirement, start alongside other income, or begin with a smaller version of the business that needs less runway.
What to spend on and what to defer
Spend early on the things that are painful to change later and cheap to get right now. Your domain registered in your own name, business email, a bank account, accounting software connected to it, and a password manager. Each takes an afternoon and each is expensive to unpick once a year of history sits on top of it.
Defer almost everything else until there is evidence. Advertising before anything converts, an office before the work requires one, inventory bought deeply for a better unit price, and software bought for a scale you have not reached are the four most common ways a first year budget disappears into things that felt like progress.
The test worth applying to any early purchase is whether it produces revenue, prevents a loss, or is required by law. If none of the three apply, it can usually wait until you know more than you know now.
Keeping the requirement low
The lowest risk version of starting is the one where the monthly requirement is small enough that a slow quarter is survivable.
That usually means starting alongside existing income if you can, which buys time rather than money and is worth more. It means variable costs over fixed ones wherever possible: contractors before employees, monthly subscriptions before annual commitments, and a coworking desk before a lease. And it means selling something before building the infrastructure to sell it at scale, because the infrastructure built first is frequently built for the wrong thing.
Then track the two numbers that tell you where you stand. Your monthly burn, calculated from your bank balance rather than from a budget, and your runway in months. Businesses that run out of cash are almost never surprised by the arithmetic. They simply were not doing it.