The 13-Week Cash Flow Forecast, Updated Weekly, Done in 20 Minutes

Estimated reading time: 7 min

An FD I spoke to last month opened her cash forecast on a Wednesday morning. It was four months old. She'd been running the business on bank balance and instinct since February, and the business had just committed to €180K of stock for a Q3 push she hadn't modelled.

Nobody trains you for this. Most finance qualifications spend three hours on forecasting and six weeks on revaluation reserves. Cash is the thing that actually kills SMEs, and it's the thing most finance teams manage worst.

The fix is unglamorous. A 13-week rolling cash forecast, refreshed every week, in 20 minutes. That's it. Done properly, it's the single most useful thing a finance function produces in an SME; done badly, or not at all, it's the reason people find out they're insolvent on a Tuesday afternoon.

Why 13 weeks

Four weeks is payroll cash. Useful if you're already in trouble, useless for anything else. By the time a four-week window shows a problem, you're inside it.

Fifty-two weeks is fantasy. You're not forecasting, you're budgeting. Every line past week 20 is a placeholder, and the whole thing rolls forward by the calendar rather than by what actually happened.

Thirteen weeks is a quarter plus one. Long enough to see the big lumpy items coming: the VAT payment, the preliminary tax, the insurance renewal, the €40K supplier invoice that always seems to land at the end of July. Short enough that every line is still contestable. A payment due in week 11 is close enough that you know the customer, the invoice, the history, and whether they'll actually pay on time.

Thirteen weeks also maps to how CFOs and FDs actually think about cash: the next quarter, the current runway, not next month's payroll in isolation and not the annual plan. That middle distance is where the interesting decisions live (hire or hold, ship or delay, draw down the facility or ride it out), and it's where most SMEs have no model at all.

The four inputs

Most SME cash forecasts fail because they're built on aggregate numbers. "Expected sales receipts: €340K." Where from? Which customers? What terms? Nobody can stress-test that, and when a customer slips a fortnight, the whole forecast moves by judgement rather than by line.

Build the forecast on four categories, all at line level, and the structure holds.

**AR by customer.** Every open invoice on the sales ledger, mapped to the week it's actually expected to land, not the week the terms say it should. If Customer X has paid 42 days late for the last six invoices, forecast 42 days. Xero and QuickBooks expose the ageing in seconds; the judgement is in the override, not in the pull.

**AP by supplier.** Every open bill, mapped to the week you intend to pay. Intend is the operative word. A supplier on 30-day terms doesn't have to be paid at 30 days. The forecast reflects what you decide to pay, not what the terms say.

**Fixed costs.** Payroll, rent, software subscriptions, utilities, loan repayments, pension contributions, commercial rates. These are the boring lines that make the forecast usable, because they don't move much. Model them once properly and they stay put.

**Capex and financing.** The one-offs. Stock buys, asset purchases, dividend payments, loan drawdowns, tax payments, lump-sum bonuses. This is where most forecasts fall apart. The lumpy items are the ones that actually kill you, and they don't come out of the accounting system on their own. Someone has to remember them.

Four categories. Built weekly. That's the whole model.

The 20-minute weekly rhythm

The forecast only works if it's refreshed every week. If it's refreshed every month, it's a budget. If it's only refreshed when someone panics, it's a fire drill.

Twenty minutes, every Monday morning, before anyone else is in the office:

Five minutes: pull the latest AR ageing from Xero or QBO. Update expected receipt dates for anything that's slipped since last week. Flag anything over 60 days that you previously forecast confidently.

Five minutes: pull the AP list. Confirm what's going out this week. Decide what can be pushed a week without burning a supplier relationship. Mark the ones you won't move.

Five minutes: update capex and financing. New stock order been signed off? Add it. Invoice from the accountant or the solicitor arrived? Add it. A VAT or corporation tax payment falling inside the next 13 weeks that wasn't there before? Add it. This is the five minutes where the lumpy items get caught, which is also why it's the five minutes people skip when they're under pressure.

Five minutes: look at the closing balance each week for the next 13. Flag any week that dips below your minimum operating cash. That's the output. One column, 13 rows, one red cell if there's trouble coming.

If it takes longer than 20 minutes, the model is wrong, not the process. Usually because someone is re-entering data the accounting system already has.

Five mistakes that break the forecast

**Lumpy items buried in averages.** "Average monthly stock purchase: €60K." No. Stock doesn't work like that. You buy €180K in week 3 and nothing for nine weeks. Average lines hide the spike that actually matters.

**VAT forgotten.** The single most common omission in SME cash forecasts is the VAT payment. It lands every two months in Ireland, quarterly in the UK, and it's often the difference between a comfortable week and a scramble. Model it as its own line, triggered by the period end. Same for corporation tax and the October/November preliminary tax deadlines.

**Timing wrong by a week.** A €50K receipt pulled from week 6 into week 4 changes the picture entirely. If you're forecasting customer receipts by terms rather than by behaviour, you're off by one to four weeks on most lines. The bigger the customer, the more the timing error compounds across the 13 weeks.

**AR stale.** If the AR list you're working from is last month's, the forecast is worthless. Customers pay when they pay. A two-week-old ageing report is already wrong, and a month-old one is actively lying to you.

**Optimistic AP.** This is the quiet one. Finance teams under pressure push payment dates out on the model without pushing them out in reality. The forecast shows week 7 closing €20K positive; in practice the supplier gets paid in week 5 and the closing balance was negative all along. The forecast has to be the truth, not the story you'd like to tell the MD.

What the data side solves

Pulling AR and AP accurately every week is mechanical work. It's the part of the job that shouldn't take any time, and in most SMEs it takes hours because someone is copy-pasting ageing reports into a spreadsheet and fat-fingering dates.

We built the data side of this into Stratavor. AR and AP pull automatically from Xero and QuickBooks every morning, line by line. Payment behaviour by customer is tracked so the expected dates adjust to how people actually pay, not how the terms say they should. The 20 minutes a week go where they should: into the judgement calls about capex, timing, and which supplier you're going to ring about week 9.

The model is only as good as the person driving it. But the person driving it shouldn't be fighting the data.

What this means on Monday morning

Open the forecast. If it's more than a week old, it's wrong. Refresh it before the 9 AM meeting. Flag any week in the next 13 that dips below the floor, and walk into the exec meeting with a number, not a feeling.

If the forecast doesn't exist yet, build it this week. Four tabs: AR, AP, fixed, capex. Thirteen columns. One summary row. Crude is fine for version one. A bad 13-week forecast beats a perfect quarterly one that nobody updates.

The template below is a clean starting point you can drop into tomorrow's work. It won't tell you when to panic. But it will make sure you find out in week 9 rather than week 13.

Download the 13-week cash flow template →

Start a trial to see it pulled from your accounting system automatically.