The short answer is enough to cover every fixed outgoing for the number of weeks it would take you to fix a serious problem, with the tax money kept separate and excluded from the total. For most owner-managed businesses that lands somewhere between six weeks and four months of fixed costs. The range is wide because the right figure depends on how concentrated your income is, how quickly you get paid and how much of your cost base you could actually stop if you had to.
I am an accountant by trade, and the version of this conversation I have most often starts with an owner saying the bank balance looks fine. It usually is fine, on that day. The question that matters is what the balance would look like eight weeks after the largest customer went quiet, and almost nobody has done that arithmetic.
Start with what you actually have to pay
Take last month's bank statement and split the outgoings into two piles. In the first pile, everything that leaves the account whether you trade or not: rent, rates, salaries, the finance agreements, insurance, software, utilities, the accountant, anything on a contract with notice. In the second pile, everything that moves with the work: materials, subcontractors, stock, delivery, commission, the extra hours.
The first pile is the number the buffer is built on. Add it up for a normal month and write it down. If the figure surprises you, that is information in itself — the fixed cost base is the thing that decides how long you can survive a bad quarter, and it is the pile that grows quietly, one small subscription and one salaried hire at a time.
Then decide how many weeks you need
Three questions set the number, and you can answer all three in ten minutes.
- How concentrated is the income? Work out what share of last year's revenue came from your largest customer. If that is a third or more, your risk is a single phone call and you need a deeper buffer than a business with a long tail of small accounts.
- How long do you wait to get paid? Add up what you are owed and divide by your average daily sales. That gives you debtor days. Sixty days means two months of work is financed by you before the money arrives, and the buffer has to carry that.
- How much of the cost base could you stop? Go down the fixed pile and mark what could genuinely be paused inside a month. Usually far less than owners expect, which is the point of doing it on paper rather than in your head.
A business with spread income, thirty-day payment and a cost base it could cut quickly can run on six to eight weeks of fixed costs. A business with one dominant customer, sixty-day terms and a salaried team should be closer to four months. Most sit in the middle, which is where the familiar three-month rule of thumb comes from.
An illustration of the arithmetic, not a client. Fixed outgoings are £28,000 a month. The largest customer is 30 per cent of revenue and pays on sixty days, and about £6,000 of the fixed costs could be paused inside a month if it came to it. That business needs roughly three months of cover, so £84,000 — and it holds £61,000, of which £19,000 is VAT collected and not yet paid over. The real buffer is £42,000, which is six weeks. The gap between what the balance looks like and what it is comes to £42,000, and nothing about that is visible from the bank app.
Take out the money that is not yours
The balance in the current account usually includes three amounts that belong to somebody else: the VAT you have charged customers, the PAYE and National Insurance deducted from your team's pay, and the tax due on profits you have already made. None of it is available, all of it feels available, and this is the most common reason a business that looks comfortable in March is scrambling in April.
The fix is mechanical. Open a second account, move those amounts out on a fixed day every month before anything discretionary is paid, and treat the remaining balance as the only money the business has. Ask your accountant for the percentage to move if you are unsure; running slightly high and refunding yourself later is far easier than catching up. Buzz Accounting sets this up with clients as part of cashflow and budgeting work, and it changes the tone of every month afterwards.
If you are short, fix it in this order
Almost every owner reading this is under their number. The order of work matters more than the effort, because building a buffer out of a business that leaks is slow and demoralising.
- Stop the leak first. Margin and pricing come before saving. A business running three points of margin below where it should be cannot save its way to a buffer, and the same work that fixes the margin usually produces the cash to build one.
- Get paid faster. Terms, deposits, stage payments, and a real chasing routine rather than an intention. Ten days off your debtor days on £600,000 of revenue is roughly £16,000 back in the account, once.
- Then build it automatically. A fixed percentage of every receipt, moved the day it lands. Percentages survive busy months; plans to save what is left over do not.
- Set the floor in writing. Decide the balance below which something changes, and what changes. A number with no decision attached is a worry rather than a control.
What good looks like after six months
You know your fixed monthly outgoings without looking them up. The tax money sits somewhere else. The buffer has a target, the target has a date, and the transfer that builds it happens without anyone deciding to do it. And the question “can we afford this hire” gets answered from a forecast rather than from the balance on the day.
If you want to see where your own numbers sit before changing anything, the Profit Improvement Calculator shows what small moves in margin and volume would be worth, and the profit and cash guide covers why a profitable month can still drain the account. The Financial Coaching page explains how we work on this with owners month by month.
Peter Allen is Co-Founder and Finance Director of Buzz Accounting, and runs Financial Coaching at Buzz Coaching. He has spent twenty years in finance and is licensed by the AAT. If your figures are not currently good enough to answer the questions above, that is a management accounts problem before it is a coaching one, and it is fixable in a month.
Get the next one in your inbox
One practical, plain-English guide for business owners each week. No spam, unsubscribe any time.
Frequently asked questions
Is three months of costs the right amount of cash to hold?
Three months is a starting point rather than an answer, and the right number depends on three things about your business. How concentrated your income is: if one customer is a third of your revenue, you need a deeper buffer than a business with two hundred small accounts. How quickly you get paid: sixty-day terms mean you are financing your customers and need more in reserve. And how much of your cost base you could actually stop: rent, a lease and salaried people carry on regardless, while subcontractors and stock can be slowed. Work out the number from your own figures and treat three months as the midpoint you are testing against.
Should the tax money sit in a separate account?
Yes, and it is the single change that makes the most difference to how a month feels. VAT you have charged, PAYE and National Insurance you have deducted and the tax owed on profit are all sitting in the current account looking like available money. Move them out on a fixed day each month, before anything discretionary is paid, and the balance you see becomes a balance you can actually use. Owners who do this stop having the conversation where a quarter looks strong until the VAT falls due. Your accountant can tell you the percentage to move; it is usually easier to run slightly high and refund yourself than to catch up.
What should a seasonal business do differently?
Hold the buffer against the trough rather than the average. A business that takes half its income in three months has to fund the other nine, so the useful number is the total cash cost of the quietest stretch plus a margin for the season starting late. Build it during the peak with a fixed transfer from every good week, because a percentage moved automatically survives a busy period and a plan to save what is left over does not. Then plan the quiet months as a budget rather than a hope: what has to be paid, what can be paused, and the date you would act if bookings are behind.
Can an overdraft or a loan do the job instead?
A facility is useful and it is not the same as cash, for two reasons. It can be reviewed or withdrawn at the point you most need it, and borrowing costs money whether you draw it or not. It also does nothing about the underlying problem if the business is short because of margin or payment terms rather than timing. A sensible order is to fix the leak first, build a real buffer second, and keep a facility behind both for genuine timing gaps and opportunities. Borrowing to cover a structural shortfall buys a quarter and makes the following year harder.