3 tests worth running before trusting any cash flow forecasting tool

Three tests decide whether a cash flow forecasting tool is worth its subscription. Each takes minutes on a trial account, and each fails loudly on packages that draw a confident line from the wrong numbers.

The tools most often shortlisted for small business books are Float, Fathom, Cashflow Frog, Pulse, Jirav and the native forecast inside QuickBooks Online or Xero. Float and Fathom both forecast from the accounting ledger and add scenario modelling.

Cashflow Frog and Pulse sit in the same category at lower monthly tiers; Jirav bundles forecasting with budgeting and board reporting for larger firms. QuickBooks and Xero include a cash projection with the subscription, and it is the one most readers already own.

Check current pricing on each vendor's site, because tiers move.

Run the three tests below against whichever of them you are trialling.

What to take away

Test one: classify a loan receipt

Enter money arriving from a lender into the trial.

A tool that treats the deposit as revenue will do the same to a customer deposit and a tax refund. The receipt belongs in a liability account and appears in financing on the statement, and the loan balance on the balance sheet should move when it lands.

Watch the account suggestion the tool offers. A default that guesses income on a five-figure bank deposit is a default that guesses wrong elsewhere.

Test two: split a settlement

Enter a payment processor settlement that arrives net of fees and refunds.

The good answer is one deposit decomposed into gross receipts, the processing fee as a cost, and refunds against the original sales. A tool that only accepts the net figure buries your payment costs inside revenue permanently, and every margin you calculate afterwards inherits the error.

Then enter a loan repayment made of principal and interest. Interest belongs in operating, principal in financing, and the loan balance must fall. A package that cannot split the two produces a debt that never shrinks.

Test three: interrogate the forecast assumption

Produce a forecast, then find what it assumed about when customers pay.

If the assumption is invisible or uneditable, the forecast is a restatement of the aged receivable report wearing a chart. Reading it as a prediction is the mistake the accounts receivable guide warns about. The assumption should be stated, changeable, and ideally settable per customer.

Two checks follow from it. Ask the tool to hold a committed payment with a date, such as an annual insurance renewal, because a forecast built only from recurring history cannot see a payment made once a year. Then put last month's forecast beside what actually happened. A forecast nobody scores is decoration.

Why these three and not ten

The other candidate tests, closed-period stability, export quality and three-section statements, are worth running. They are also cheaper to fix.

A statement that will not agree to the bank balance is a reporting problem. A forecast built on the wrong payment assumption is a spending decision made on a number that was never cash.

The method behind the statement itself is set out in the cash flow statement guide, and it is worth knowing which method your tool follows before you compare two periods.

The classification decisions underneath all three tests are a chart of accounts matter. No package makes them for you.

Worked example: the annual insurance renewal

A business pays 12,000 a year for cover, billed once, in March.

A forecast built from the last twelve months of transactions sees no insurance payment in ten of those months and projects none for next March. The owner reads a healthy cash position in February and commits to a equipment purchase. The renewal lands, and the position was never real.

Holding the renewal as a dated commitment fixes it. So does a spreadsheet column listing known future payments. Either works; the automatic forecast alone does not.

Running the tests on a trial

An afternoon covers all six. The results tell you which work stays with you, and that is the number to weigh against the subscription.

Which amounts you hold rather than earn follows from your own obligations. Start with public guidance on paying business taxes, and keep the underlying records retrievable over the spans the federal guidance on record retention sets out.

Common questions

Do we need a separate forecasting tool?

Only if the ledger's own forecast fails test three and cannot be fixed. A spreadsheet with a list of dated committed payments beats an automatic forecast built on a wrong payment assumption, and it costs nothing. The wider selection questions sit in the guide to accounting software.

Which test do small business packages most often fail?

Test two, where card payments are involved, and test three. Both are workable around. Both are worth knowing before you rely on the numbers.

Should the accountant run the tests?

They will spot different failures, particularly on the repayment split and on closed periods. Running them together is an hour well spent, because they inherit whatever the tool cannot do. The routine in the bookkeeping foundations guide is a fair measure of what the remaining work actually is.