How do you build a 13-week cash flow forecast?

A 13-week cash forecast is a spreadsheet with five rows: opening balance, money in, money out, net movement, closing balance. Every customer payment and every supplier bill sits on the date it falls due, not the date you hope it will. The one rule that matters is that no closing balance line is allowed to go negative, because a negative line is when you find out, not when you can fix it.

Build it in an afternoon. Maintain it in twenty minutes a week. It is the cheapest operating tool in the business, because it turns a feeling about money into a date.

What is the difference between profit and cash?

Profit is an accounting opinion about work you have completed. Cash is the day the money moved. For a small business those two dates can be months apart, and the gap is where most surprises come from.

The items that create the gap are ordinary ones:

• Work delivered and invoiced, but not yet paid for. This is the biggest one.
• Money received as a deposit, counted twice if you are not careful.
• Stock bought and paid for now, against a job invoiced later.
• Wages and PAYE, VAT and tax landing before the revenue does.
• A loan repayment that comes out of the same account as the rent.
• A customer who has not paid, so a supplier bill you expected to cover has to wait.

None of these are problems on their own. A business that understands when they will arrive is not short of money, it is simply early.

What columns does the spreadsheet need?

Thirteen columns, one per week, with a Friday or month-end date at the top. Every week column holds the same set of rows in the same order, so the eye travels straight across.

One rule about the sheet itself: use a formula for every total, never a typed number. A typed total that disagrees with the rows above it is how a forecast starts lying.

Why must the closing balance never go negative?

Because a negative number on that line is not a warning. It is a post-mortem.

When the closing balance goes negative in week 9, the money left the account in week 7. By the time you can see it, the options have already closed: you cannot ask for the money you already spent, and you cannot get a customer to pay faster by discovering the problem later. What you can do about it at week 9 is decide which payment to miss and who to call.

That is the whole argument for the forecast. It moves the discovery point back to the week where a decision still exists.

Then treat the lowest line as a real instruction, with a deadline:

weeks of warning = (weeks before the lowest closing balance) x (your ability to move a date)

Three things move a date, in order of how quickly they work:

1. Ask a customer for an earlier payment, with a reason. Cheapest, and usually works once.
2. Ask a supplier for a later payment, with a plan. Cheap, and it costs you a relationship if you do it more than once.
3. Delay or split a purchase that is not yet committed.

What does not work is assuming next month's invoice lands this month, which is the single most common way a forecast gets broken by its own author.

How much money in should you count, and how certain should it be?

Split the money in into three columns and keep the certainty visible. Most forecast mistakes come from mixing a signed purchase order with a hopeful guess in the same cell.

• Committed. A customer has confirmed in writing, or there is a purchase order number against the invoice. Put the date they gave you.
• Expected. An invoice is out and nobody has objected. Place it using your own measured days-to-pay for that customer, not an optimistic one.
• Chased. The invoice is already overdue. Place it on the date you have asked for payment, and treat it as uncertain until it lands.

Your expected column is only as good as your measurement. If you do not know how long a particular customer takes, use the average across your customers and then move that one invoice into the committed column the moment they confirm.

Where does the gap actually come from?

When the closing balance dips, the cause is almost always in the money-in row rather than the money-out row. Three places to look, in order.

Overdue invoices. Take the total of everything past its due date and put it in the chased row on the dates you have promised. If that total is large enough to fill the gap, then the gap is a collection problem, and no amount of cutting expenses fixes it. Work out how much of it is genuinely recoverable:

recoverable overdue = (total past due) - (invoices with an unresolved dispute)
recoverable overdue = recoverable overdue x (share of past-due invoices you have collected before)

The last figure is one you can compute from your own records: take last quarter's past-due invoices, see what actually got paid, and use that share.

A cluster in one week. Several large bills landing together is a timing problem, not a shortage. Look at the date pattern before you panic.

Money out that is not a cost of the business. Director's drawings, a loan to a director, a personal expense on the company card, a loan to a family member. These belong in the forecast even though they are not a trading cost, and they are usually the line nobody wants to look at.

How do you run the weekly ritual?

Friday morning, twenty to thirty minutes, before anything else happens that day. The ritual is the forecast — a perfect sheet nobody maintains is worse than no sheet, because it creates false confidence.

1. Replace last week's projected figures with what actually happened, and note the difference. One column of differences is worth more than a full quarter of projections.
2. Add the next week's column and roll the opening balance forward.
3. Update the chased row with anything paid and anything newly late.
4. Look at the lowest closing balance in the next thirteen weeks and write the week it falls in at the top of the sheet.
5. Ask one question: what is the earliest week I would have to act?

Then leave it alone until Friday. Do not open it daily. The value is in the trend, not the fluctuation.

What do you do in week one?

Do five things, in this order, and stop.

1. Pick one operating bank account and use only that account's balance as the opening figure. Combining several accounts hides the problem.
2. List the money in that is already committed for thirteen weeks: anything with a purchase order, a signed quote, or a written confirmation of a date.
3. List the fixed money out: wages, PAYE, rent, loan repayments, insurance, and the confirmed tax dates.
4. Add the supplier bills you already know about, on their due dates. Include the ones for stock you have already ordered.
5. Enter a zero in every cell you have actually checked. A blank cell reads as zero to the eye and as optimistic to the mind; a typed zero reads as checked.

Then find the lowest closing balance and write its week number at the top of the sheet. That is a finished forecast, built in an afternoon, out of records you already keep.

What should you never put in a cash forecast?

Four things will break it, and all four are tempting.

• Profit, and anything derived from it. Profit is an accounting view of completed work. Cash forecast is a list of dates.
• A supplier's verbal promise. If the date is not in writing, treat it as unknown and plan without it.
• A tax date you have not confirmed. Confirm it, then enter it. Do not estimate it and then act on it.
• Stock value, goodwill or a sale you have not agreed. Non-cash items do not pay suppliers.

A forecast is not a prediction. It is a list of dates, so that a bad week is a decision instead of a surprise.

The companion checklist is the build sequence, in order, from setting up the sheet to the Friday routine. The templates give you the exact column structure to type in, a worked week, the gap worksheet to fill in when a closing balance dips, and the Friday update block you paste into the top of the sheet.

Not sure where to start? Book a free 30-minute strategy session and we will name the one gap worth fixing first: https://agentcy.co.za/#book-audit