How far ahead should a 13-week cash flow forecast be accurate?

Weeks one to four should be close, usually within a few percent. Weeks five to thirteen are directional, and that is expected. If week two is regularly wrong by a wide margin, the problem is usually the receipts assumption rather than the model. Track variance by line each week and the near weeks tighten quickly.

Should the forecast be weekly or daily?

Weekly is standard and enough for most decisions. Move to daily when the balance is going to come close to zero or close to a facility limit inside the next fortnight, because within a single week the order of a large receipt and a payroll run decides whether you have a problem.

Can we build a 13-week forecast in Excel?

Yes, and most good ones are in Excel. What matters is the structure: inputs separated from calculations, visible assumptions, no hardcoded numbers inside formulas, and a clean way to drop in a new week of actuals. A dedicated tool is not necessary and often makes it harder to change an assumption quickly.

What a 13-week cash flow forecast should include

A 13-week cash flow forecast should include the opening bank balance, customer receipts modelled week by week from the debtors ledger and actual payment behaviour, supplier payments from the creditors ledger and your payment runs, payroll and superannuation on their real payment dates, tax payments including BAS and PAYG, loan repayments and interest, capital spend, any one-off items such as settlements or deposits, and the closing bank balance for each of the 13 weeks. It should also show headroom against your facility limit, a stated assumption for every receipt line, and a variance column comparing last week’s forecast with what actually happened. What it should not include is anything taken straight from the profit and loss, because a 13-week forecast is built on cash movements rather than on accounting revenue and expenses.

Why 13 weeks and not 4 or 26

A quarter is long enough to see a shortfall with time to act, and short enough to forecast at invoice level. Four weeks shows you a problem when the only choices left are expensive ones, such as paying a supplier late or an urgent call to the bank. Twenty-six weeks pushes most of the forecast past the invoices you can see, so the later weeks become estimates built on estimates. Thirteen weeks also lines up with a quarter, which is the period most lenders and boards already think in.

The direct method, and why the indirect method fails here

The direct method builds the forecast from the money that comes in and goes out: each receipt and each payment, in the week it lands in the bank. The indirect method starts from forecast profit and adjusts for movements in debtors, creditors and stock. That works for a monthly or annual view, but over 13 weeks it hides timing. A month of profit says nothing about whether payroll goes out before your largest customer pays. For a short-term forecast, use the direct method and reconcile it to the profit and loss separately if you need to.

Receipts: the line everyone gets wrong

Customer terms might say 30 days. Your debtors ledger tells you what customers actually do. Look at the last few months of receipts, work out how many days each large customer really takes to pay, and forecast from that. Put the largest accounts on their own lines, because one late payment from a top customer can move a week more than every small account combined. The smaller accounts can share one line with a stated collection pattern. Write the assumption next to each line so anyone reading the forecast can challenge it.

Payments: fixed, committed and discretionary

Split payments into three groups. Fixed payments, such as rent, loan repayments and leases, go out on set dates whatever happens. Committed payments are approved purchase orders and supplier invoices you have already received. Discretionary payments, such as capital spend, bonuses and some supplier runs, can move if they have to. When the forecast shows a tight week, this split tells you straight away which amounts can move and by how much. Without it, the forecast shows the problem and leaves you to work out the answer from scratch.

Payroll, superannuation and tax

Payroll, superannuation and tax are usually the largest single outflows, and their dates do not move. Put each on its own line, in the week the money leaves the bank rather than the period it relates to. Fortnightly payroll gives six or seven pay runs in 13 weeks, and a month with three of them needs planning. Superannuation, PAYG withholding and the BAS each have fixed due dates. Missing one brings penalties, and for super and PAYG it can make directors personally liable, so these lines are never the place to find room.

Facility headroom and covenant dates

The bank balance is not the constraint. The limit is. If you run an overdraft or a revolving facility, add a line for the limit and a line for headroom, which is the limit less what you have drawn. Watch the week with the lowest headroom, which is often a different week from the one with the lowest balance. If the facility has covenants, mark the test dates in the forecast and show the measure the bank will test, so a likely breach shows up weeks before the test date.

The variance column

Each week, put last week’s forecast next to what actually happened, line by line, and note the reason for any large gap. This is the part that makes a forecast believable to a lender, a board or an adviser. It shows the forecast is tested every week, and it shows you where your assumptions are off. Most of the variance in the early weeks comes from receipts. When a line misses, change the assumption behind it as well as the number, and the next few weeks get closer quickly.

A worked example of the line structure

The table shows the first four weeks of a simple layout, in thousands of dollars. The figures are illustrative. Receipts and payments sit above the net cash flow line, and the facility lines sit at the bottom so headroom is the last thing you read. Brackets show money going out or an overdrawn balance.

$k	Week 1	Week 2	Week 3	Week 4
Opening balance	(40)	(76)	9	(11)
Receipts, large accounts	85	60	90	70
Receipts, other customers	40	45	38	42
Total receipts	125	105	128	112
Supplier payment runs	(75)	(20)	(80)	(25)
Payroll	(60)	0	(60)	0
Superannuation	(8)	0	(8)	0
Rent	(18)	0	0	0
BAS and PAYG withholding	0	0	0	(35)
Loan repayment and interest	0	0	0	(12)
Total payments	(161)	(20)	(148)	(72)
Net cash flow	(36)	85	(20)	40
Closing balance	(76)	9	(11)	29
Facility limit	150	150	150	150
Headroom	74	150	139	150

Weeks five to thirteen follow the same rows. In a working forecast, each receipt line also carries its assumption in a note column, left out here to fit the page. Week 1 is the one to watch: the balance is lowest there and so is the headroom.

What to leave out

Leave out anything that never moves money. Accruals, depreciation, amortisation, provisions and unrealised foreign exchange gains or losses all change profit without touching the bank. Revenue recognised in the period does not belong either; the receipt from that customer does, in the week it arrives. Keep GST in, because you collect it and pay it out, and the BAS payment is a real outflow. If a line is in the forecast only because it is in the profit and loss, check whether cash actually moves that week.