A cash forecast you never check against reality slowly becomes a guess.
Each week, compare last week's forecast receipts and payments with what actually happened, line by line. Calculate the variance in amount and percentage, and explain the large ones. Track accuracy over time for each line and for the closing balance. Lines that are wrong week after week need a better method, not just a better explanation.
The value of a cash forecast is that people can act on it: delay a payment, draw on a facility, chase a customer. That only works if the forecast is reliable. Measuring accuracy tells you how much buffer to hold for forecast error, and which lines to improve.
Save each week's forecast before you roll it forward. A week later, put the actuals beside the forecast for that week:
| Column | Content |
|---|---|
| A | Line (customer receipts, payroll, suppliers, tax, loans, other) |
| B | Forecast |
| C | Actual |
| D | Variance: =C2-B2 |
| E | Variance %: =IFERROR(D2/ABS(B2),0) |
| F | Explanation |
| G | Timing or permanent |
For each line, keep a running list of weekly variances. A simple accuracy measure is the average absolute percentage error over the last 13 weeks:
=AVERAGE(ABS(E2:E14))
entered as an array formula in older Excel versions. Also check bias: if the average of the signed variances is consistently positive or negative, the forecast is systematically optimistic or pessimistic.
=AVERAGE(E2:E14)
The 13-week cash flow forecast guide covers the forecast itself, and the minimum cash buffer guide explains how forecast error should shape the buffer you hold.
Do the comparison at the same time every week, before rolling the forecast forward, and keep it short: a table of variances, a sentence on each large one and a note of any method change. The person who prepares the forecast should do the comparison; the finance director should read it. Over a few months, the accuracy history becomes the best evidence of how far the forecast can be trusted, which matters when you are deciding how much cash to keep in reserve or explaining your position to a lender.
Some variances come from your own actions, such as a payment run moved by a day. Others come from customers or events you can't control. Tag each one, so you fix your process where it is the cause and widen your buffer where it isn't.
The method is the same in both countries and any currency. For multi-currency forecasts, measure accuracy in each currency before converting.
Cash flow from your bank statement. Your balance day by day, money in and out by week, where the money goes, every recurring payment, and a 13-week forecast that warns before cash runs short. Runs in your browser on your own computer: nothing is uploaded.
Work out DSO and the cash tied up in late payers. Free to use online.
The Cash Flow Forecasting & Liquidity Control System tests every week of a rolling cash forecast against your minimum cash buffer and lending covenants, and flags a breach before it lands. Excel, no macros, one-time price.
The finance director or controller, so the person relying on the forecast understands how much to trust it.
It depends on the business, but you should know your own accuracy and hold a cash buffer that covers your typical forecast error.
A consistent tendency to over- or under-forecast, shown by the average of signed variances. Bias needs a method change, not just explanations.
Yes, if you can. Comparing the forecast made four or eight weeks earlier shows how reliable the longer horizon is.
Last updated 21 October 2026. General guidance for finance teams; check it against your own policies, systems and local rules.