What it means
Prediction sits at the heart of any cash plan, and the hard part is rarely the amounts. It is the timing, because a sale recorded in March may only become cash in May, and a purchase agreed in March may only leave the account in April.
Most predictions that fail do so because they assumed money moved faster than it really does. The workhorse technique is the collection profile, which describes what proportion of a month's sales is collected in that month and in each following month.
Historic invoice data gives you the profile, and once you have it you can turn any sales forecast into a dated collections forecast in a single step. Costs are predicted differently.
Fixed commitments such as rent, salaries and loan repayments can be scheduled almost exactly, while variable costs are usually predicted as a percentage of sales and then shifted forward by the supplier payment terms. A prediction is only as good as its feedback loop.
Comparing predicted with actual cash each month and measuring the percentage error tells you whether the model is reliable, and it usually reveals a systematic bias, such as consistently optimistic collection dates, that can be corrected. Predictions also need a range rather than a single line.
Running a cautious case alongside the expected case, with slower collections and a lower sales figure, shows how much room the business has if things go wrong. The distance between the two cases is often more informative than either number on its own.
The useful nuance is that precision and usefulness are not the same. A prediction that is 5% out but produced within two days of month end will change more decisions than an exquisitely detailed model that arrives three weeks late, because the earlier figure still leaves time to act on what it shows.
In practice
Real-world examples.
Example
A digital agency predicts cash from a retainer book with 95% confidence because clients pay by standing order on the first of the month. Its project work is predicted far more cautiously, with an assumed 20 day delay after invoicing.
Example
A wholesale bakery predicts a cash dip every January and confirms it by modelling supermarket payment terms against the post-Christmas sales drop. It arranges a seasonal overdraft in November rather than in the middle of the dip.
Example
A construction subcontractor predicts collections using certified work values rather than invoice values, because main contractors routinely certify less than is claimed. The change cuts its average prediction error from 18% to 7%.
Think of it
“Cash flow prediction is forecasting what your future cash flows will be-educated projection.
Formula
Calculation
Predicted collections for a month = sum of each prior month's sales x that month's collection percentage.
A commercial furniture supplier finds that historically 30% of sales are collected in the month of sale, 50% in the following month and 20% in the month after that. Sales were $300,000 in January, $400,000 in February and are forecast at $500,000 for March. Predicted March collections are (30% x $500,000) + (50% x $400,000) + (20% x $300,000) = $150,000 + $200,000 + $60,000 = $410,000. When March closes, actual collections are $385,000, so the prediction error is $410,000 - $385,000 = $25,000, or about 6% of the predicted figure, which is well inside the 10% tolerance the finance team treats as acceptable.Case study
Seen in the real world.
This is an illustrative, fictional example. Marlowe Instruments, an invented maker of laboratory equipment, predicted cash by simply assuming every invoice was paid on its 30 day term.
The prediction was wrong every single month, always in the same direction, because the true average collection was 47 days. Once the finance manager rebuilt the model on the real collection profile of 10% in month one, 55% in month two and 35% in month three, the average monthly error fell from roughly $180,000 to under $40,000.
Nothing about the fictional company's trading changed. The improvement came entirely from predicting the timing that already existed rather than the timing the payment terms described.
Watch out
Common mistakes.
- Predicting collections from payment terms instead of from actual historic payment behaviour, which is almost always slower.
- Predicting only the total for a month, when a business can be comfortable on the 30th and unable to pay wages on the 12th.
- Never measuring prediction error, so a persistent optimistic bias goes uncorrected year after year.
Questions
People also ask.
How accurate should a cash flow prediction be?
Within about 5% for the coming month and 10% to 15% three months out is a reasonable expectation for most established businesses.
Do I need forecasting software?
No, a well-built spreadsheet driven by a real collection profile beats expensive software fed with optimistic assumptions.
How should one-off items be handled?
Predict them separately and clearly labelled, so a single asset sale or tax refund never gets mistaken for an improvement in the underlying trend.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%