Is your cash flow forecast any good? How to score it weekly
Score your cash flow forecast weekly: compare actual closing cash and receipt dates to forecast, then build your own error band from the gaps.

You built a cash flow forecast. Then it went in a drawer, and the drawer became a work of fiction. That is the quiet failure mode with cash flow forecasting: a number nobody scores every week is not a forecast, it is a guess with formatting. The fix does not need new software or a finance hire. It needs twenty minutes on a Friday and two numbers tracked in a log.
This post shows you how to score your forecast weekly, starting with the two numbers that matter: actual closing cash against forecast, and forecast receipt timing against actual. You will build a simple variance log with two small tables, see a four week worked example for a stock-holding brand with real numbers, and learn why week three exposes what the closing balance hid. You will also find out why the timing gap deserves investigating before the balance gap, which misses mean your assumptions are wrong rather than your sales, and how to set an error band from your own log instead of borrowing someone else's. There is a Friday routine at the end, plus answers to the questions people ask most.
Two numbers tell you whether your cash flow forecast is working
Your cash flow forecast is only as good as the gap you score each week. Run two comparisons: actual closing cash against what the forecast predicted for that day, and the date each large receipt actually landed against the date you forecast. Do the receipt date first if you hold stock, because lateness usually breaks the model before volume does.
A closing balance within a few hundred pounds can still hide a broken week. Two opposite misses can cancel out and leave the next four weeks wrong with nobody noticing. Unscored forecasts decay: someone edits a number to look sensible, nobody logs it, and by month three you cannot tell drift from real change. Ten minutes on a fixed day beats a quarterly rebuild. Watch the eight-week trend, not one Friday. If you are still building the forecast itself, our guide to cash flow forecasting covers the basics.
Build a weekly variance log with two small tables
Put both logs on one sheet. Two four-column tables beat one wide one on a phone at 11pm.
Table one:
| Week | Forecast closing cash | Actual closing cash | Gap |
|---|---|---|---|
| 1 | £42,000 | £41,600 | -£400 |
Cash gap = Actual closing cash minus Forecast closing cash
Table two:
| Week | Forecast receipt date | Actual receipt date | Days late |
|---|---|---|---|
| 1 | 3 Feb | 3 Feb | 0 |
Days late = Actual receipt date minus Forecast receipt date
Gap percentage = Cash gap divided by Forecast closing cash
Use it only where months differ sharply in size. It stops a quiet January looking tidier than a heavy November.
Take actuals from one point: cleared bank balance at Friday close. Score against the forecast version you actually used that week, never a later edit.
Keeping forecast, supplier bills and stock plan together makes this cheap. Cushion does that with read-only connections to Shopify, Xero and QuickBooks. Use the 13-week rolling cash flow forecast to build it first.
A four week variance log for a stock-holding brand
Here is a four-week run to show both tables working together. The numbers are illustrative and the brand is made up: it imports garden planters, holds about eight weeks of stock and pays suppliers on 30-day terms. This is not a client result. Check your own cover with the days inventory outstanding calculator. It sells through its own Shopify store and a marketplace, so money arrives on a payout schedule rather than at the moment of sale. One wholesale receipt of £18,000 lands somewhere in the four weeks, the kind of lump most stock-holding brands carry. Read both tables together. Either one on its own gives the wrong answer.
Table one: closing cash against forecast
Here is the table itself, four columns, filled in with the illustrative brand's four weeks. Gap is actual closing cash minus forecast closing cash.
| Week | Forecast closing cash | Actual closing cash | Gap |
|---|---|---|---|
| 1 | £42,000 | £41,600 | -£400 |
| 2 | £38,500 | £38,900 | +£400 |
| 3 | £31,000 | £27,400 | -£3,600 |
| 4 | £24,800 | £22,100 | -£2,700 |
Weeks one and two look excellent. A 1% miss either side of zero, then a small positive. If you stopped reading here you would conclude the cash flow forecast was working beautifully. That is the trap this piece exists to close. The -£400 and +£400 cancel to nothing, which is exactly why a signed gap flatters a forecast. Look at the actual miss on each week and the pattern changes: 400, 400, then 3,600 and 2,700 as the drift sets in. For an example of what a healthy eight week pattern looks like when dates hold, see this cash flow forecast example.
Table two: receipt dates against forecast
Now the second table, and the one to read first on a Friday.
| Week | Forecast receipt date | Actual receipt date | Days late |
|---|---|---|---|
| 1 | 3 Feb | 3 Feb | 0 |
| 2 | 10 Feb | 17 Feb | 7 |
| 3 | 17 Feb | 24 Feb | 7 |
| 4 | 24 Feb | 28 Feb | 4 |
Days late = Actual receipt date minus Forecast receipt date.
Same weeks, same brand. The dates were slipping the whole time the balance looked fine. Weeks one and two looked tidy in the cash table, but the main receipt was already a week behind by week two. That gap is why the benefits of a cash flow forecast for a stock-holding brand depend on dates as much as balances.
What week three shows that the balance hid
If table one is all you read, week two looks fine: an 18,000 receipt slipped seven days, a supplier payment slipped the same days, and the gap was 400. Week three is where the slip becomes cash. The money for the stock payment run was still in transit, so closing cash finished 3,600 under forecast. By week four closing cash is still 2,700 under forecast, with 18 late days logged across the month. Dates are the reason; balance is the symptom. A few hundred pounds of balance miss on a week of date miss is not small. It is a timing miss you have not paid for yet. Run the Can I afford this reorder? check before your next payment run.
Investigate the timing gap before the balance gap
Before you touch the volume line, ask one question of every miss: would this still be a miss if the money had arrived on the forecast date? If no, it is timing, and chasing volume will not fix it. In a stock-holding brand the usual culprits are processor or marketplace payouts shifting around a weekend or bank holiday, customers paying from statement date rather than invoice date, and supplier terms that start at delivery rather than at order. Lateness repeats quietly. One customer shifting by four days becomes three weeks of cash flow forecasting that is quietly optimistic, and every stock decision after that is made on a soft number. When lateness is consistent, move the assumption instead of logging it forever: if receipts always land four days late, shift the schedule four days and watch whether the gap closes. Supplier terms and payment runs are the usual source, so check those first.
Which misses mean your assumptions are wrong, not your sales
Lateness is a timing symptom. To know whether the cause is your sales or your model, look at the shape of the miss.
A sales miss arrives late but arrives, tracks order volume, and moves when trading moves. An assumption miss shows the same gap, with the same sign, for three weeks or more, in a line you set once and never revisited. In ecommerce forecasts, the usual culprits are payout lag, supplier payment terms, shipping cost per order, and the date a VAT payment actually leaves the account.
To test it, change one assumption in a copy of the forecast and re-run those four weeks. If the gap closes, it was an assumption. If it stays, look at sales.
Give a gap three weeks before rewriting anything. Weekly edits are just a different way of never measuring, and a template you keep scoring against only works if the baseline holds.
Set an error band from your own log, not a borrowed one
Once the log has a few months in it, stop asking whether a gap is normal in the abstract. No industry benchmark is quoted here on purpose. A band borrowed from another business tells you nothing about your payout lag, supplier terms or stock cycle.
After eight to twelve weeks, take the absolute cash gaps and find the middle one:
Typical weekly error = median of the absolute cash gaps
Find your worst normal week too, the gap you exceed roughly one week in four. Two numbers make a band: a typical miss and a bad week. Treat it as a trigger, not a grade. Inside it, no action. Outside it, investigate before you publish the next forecast version.
Run the same calculation on days late. If the median is zero and one customer is the only late payer, that is a collections problem, not a forecasting problem. A band that widens over two months means the forecast is drifting, warning you weeks before the bank balance does.
What to do with this on Friday afternoon
Pick a fixed slot, say Friday at 4pm, and put both four column tables in the same sheet. Score against the forecast version you were actually reading that week, not one edited afterwards.
Read the timing table first, then the balance table. Explain every date gap before you explain a cash gap.
After eight weeks, set your own error band. Leave assumptions alone until a gap repeats three times, then change one line, not five.
If the forecast, bills and stock plan sit in three places, Cushion keeps them together with read-only connections to Shopify, Xero and QuickBooks. Join the waitlist to see it.
FAQs
How often should I score my cash flow forecast?
Weekly, if you hold stock and pay suppliers within 13 weeks. Monthly is too slow: three weeks of lateness can pass before your first score.
What counts as a bad cash flow forecast?
No universal percentage. Bad means misses that are one-sided, or outside your band for several weeks. One wild week inside a stable trend is noise.
Should I forecast cash flow weekly or monthly?
Both. A weekly 13 week view for stock and supplier decisions, a monthly one for the year and your accountant. Score the weekly one: it drives money leaving the account.
What if a receipt arrives early?
It shows as negative days late and flatters the week. Check whether it pulled forward money you planned for later. Log it anyway, because early usually means the invoice date was wrong.
Do I need software to keep a variance log?
No. A spreadsheet with two four column tables beats the nothing most brands do now. Software earns its place when the forecast, bills and stock plan live in different systems.
Conclusion
Scoring your cash flow forecast comes down to two checks: does the closing balance land close to forecast, and do receipts arrive when you expected? When the balance looks fine but the timing drifts, you have a warning you would otherwise miss. A weekly variance log, even a simple spreadsheet, turns that warning into something you can act on. Build the habit first, then let the pattern from your own error band tell you how much slack to plan for.
Start this Friday. Open the forecast, compare last week's closing cash, and log the first two lines. You do not need new software to begin. When your forecast, supplier bills and stock plan do start feeling disconnected, Cushion is being built as one place where the cash forecast sees the stock and the stock plan sees the bank. Join the waitlist.



