Cash Flow Forecasting Becomes the Most Used ERP Report

Ask a finance team which report they ran most often last year and the answer is usually the profit and loss by department, or aged receivables. Ask them this week and the answer is cash. After a fortnight in which equity markets fell sharply, factory output in China stalled and shipping schedules became guesswork, cash flow forecasting has gone from a quarterly treasury exercise to the number the managing director asks for before anything else. Most enterprise resource planning systems cannot produce it. Not because they lack the data, but because the thing being asked for is not a report. It is a weekly operating ritual that happens to use data from the system, and organisations that try to solve it by finding the right menu item will still be looking in May.

Why the standard reports do not answer the question

An ERP is an accrual machine. It records revenue when it is earned and cost when it is incurred, which is the correct basis for reporting profit and the wrong basis for predicting the bank balance. Aged receivables tells you what is overdue against agreed terms. It does not tell you when money will actually arrive, because it assumes customers pay on their due date, and the ones that matter do not. Aged payables has the same flaw in reverse, with the added problem that it excludes everything committed but not yet invoiced: purchase orders raised, contracts signed, capital projects approved, the retention on a job that becomes payable at handover. Then there is the set of outflows that rarely sit in the accounting subledgers at all. Payroll and its statutory companions. Loan repayments and finance leases. Tax payments. Licence and permit renewals. Insurance premiums. Intercompany settlements that somebody moves when they remember. In a typical mid-sized group these items are between a third and a half of total cash out, and they live in a spreadsheet on one person's laptop. So the honest starting point is that a cash forecast is assembled from the system rather than produced by it.

The thirteen-week direct forecast

The format that works is unglamorous and nearly a century old: a direct forecast, weekly, thirteen weeks out, built from receipts and payments rather than from adjusted profit. Why thirteen weeks? Long enough to see a quarter-end, a payroll cycle, a tax payment and a facility review; short enough that the numbers are constructed from actual invoices and commitments instead of assumptions. Anything beyond a quarter is a scenario, not a forecast, and should be labelled as one. The build has four inputs. Opening bank position by entity and by account, taken daily, not from the ledger but from the bank. Expected receipts, derived from open invoices and adjusted for how each customer actually pays. Committed outflows, including everything in the paragraph above. And discretionary outflows, separated deliberately, because that separation is the entire point of the exercise: it shows management what it still controls. One person owns the forecast. One day of the week it is published. Every line has a source and a name against it. That is the ritual, and its value compounds.

Payment behaviour is already in your system

The single largest improvement available to most finance teams costs nothing, and it is not a forecasting tool. It is computing how long each customer actually takes to pay. The data is in the ERP: invoice date, due date, cash application date, by customer, for several years. Calculate the median days from invoice to receipt per customer, not the average, because one dispute distorts a mean badly. Then apply that median instead of the contractual terms. Do it for the twenty customers who represent the largest share of your receivables and you will improve forecast accuracy more than any model change. You will also discover, usually to some discomfort, that several major accounts have never paid on terms and that nobody has ever raised it, because the sales relationship is more visible than the cash consequence.

Levers, in order of speed

When the forecast shows a gap, the instinct is to cut spending. That is rarely the fastest lever. Invoice hygiene is. In every organisation I have examined there is revenue earned and not invoiced, invoices rejected for a missing purchase order reference, invoices issued from the wrong entity, disputed lines holding up otherwise payable invoices, and approvals sitting in someone's inbox. This is money you have already earned, and it is usually recoverable in days rather than weeks. Then collections focus, applied to the twenty accounts that matter rather than uniformly. Then payment term negotiation in both directions, done openly, because the counterparty is doing the same arithmetic. Then inventory release, which is slower than it looks. Then capital expenditure deferral, which is immediate but political. Then facility headroom, which is a conversation with a bank that goes better with a thirteen-week forecast in hand than without.

Practical Guidance for Cash Forecasting Setup

• Build a direct weekly forecast, thirteen weeks, one owner, one publication day. Not an adjusted profit forecast, and not monthly. The cadence is what makes it useful.
• Take opening balances from the bank, by entity and account, every day. Ledger cash is a reconciliation artefact and lags reality by exactly as long as your reconciliation does.
• Replace contractual payment terms with observed median days to pay. Compute it from your own history for the top twenty accounts. This is the highest-return hour you will spend.
• Capture committed outflows that are not yet invoices. Open purchase orders, approved capital projects, retention becoming due, loan and lease schedules, statutory payments, licence renewals, insurance.
• Separate discretionary from committed on the face of the forecast. Management needs to see what it still controls; a single net number hides the only actionable part.
• Run the invoice hygiene sweep before cutting costs. Uninvoiced revenue, rejected invoices, disputed lines, missing references, unapproved approvals. Fastest cash in the building.
• Compare forecast to actual weekly, line by line, and correct for bias. Persistent optimism in receipts is the normal failure. Adjust the assumption, not the format.
• Give the bank the same forecast you give the board. Two versions of the truth will be discovered at the worst possible moment, and the credibility cost exceeds any negotiating advantage.

The Regional Angle

Four features of regional trading structure make the current situation more dangerous than a European or American cash forecast would suggest. The first is cash locked in goods that have not arrived. Gulf importers and distributors routinely pay in advance or against documents, and with Chinese production disrupted and sailings cancelled, a significant amount of working capital is currently sitting in prepayments and in-transit inventory with no reliable arrival date. That money is neither a receivable nor stock you can sell. Add a line to the forecast for advance payments against undelivered orders, with an expected arrival column, and watch it rather than averaging it into inventory. The second is the financing cycle that sits on top of those shipments. Where trade finance is structured around letters of credit and trust receipts, the facility repayment is tied to a shipment and sale sequence that assumed predictable transit times. When shipments slip by six weeks, trust receipt maturities arrive before the goods can be sold, limits stay utilised, and new orders cannot be opened. This is the mechanism by which a supply delay becomes a liquidity event, and it is invisible in any forecast built purely from the ledger. Map your facility maturities against expected shipment arrivals now, and talk to the bank before the first one bites. The third is that a meaningful share of outflow here is already fixed by paper handed to somebody else. Where rent, supplier settlements and instalments have been paid by cheques issued in advance and dated forward, those obligations cannot be quietly deferred by an email; the instrument is in the counterparty's drawer. Layer on the annual lump sums that characterise operating here, trade licence and establishment card renewals, visa and medical processing, insurance, and, if headcount reductions are being contemplated, end-of-service entitlements that have been accrued but never funded. A group that reduces staff to save cash can create the largest single outflow of its quarter in the act of doing so. Model it before deciding it. The fourth is structural: most regional groups have no cash pooling. Balances sit across many entities, free zones and banks, some of them trapped by licence restrictions, minority shareholders or simple inertia, and are moved by instruction when somebody notices. Consolidated cash looks comfortable while a specific entity cannot meet its payroll. Forecast by entity as well as in aggregate, and if inter-entity funding is going to be part of the answer, involve your tax adviser early, because those transfers are loans with documentation, pricing and substance consequences rather than internal bookkeeping.

The objection worth taking seriously

The first objection is capacity. A weekly thirteen-week forecast, properly maintained, is a couple of days a week for somebody competent. Small finance teams already closing the month, chasing collections and answering the auditor do not have that person free, and told to do it anyway they will produce a forecast that is merely a spreadsheet with last month's numbers extrapolated, which is worse than nothing because it will be believed. The second is political. Cash forecasts are behaviourally corrupt in most organisations. Sales inflates collection dates to protect relationships, operations understates spend to protect budgets, and the finance director quietly haircuts the whole thing before it reaches the board. Everyone knows the number is wrong, and the reconciliation each week becomes an argument about blame rather than an improvement to the assumptions. The practical answer to both is to reduce scope rather than ambition. Forecast the twenty largest customers, the ten largest suppliers, payroll, statutory payments and facilities; put everything else in a single behavioural average line. That version can be maintained in half a day a week and is usually within a few per cent. On the political problem, the only fix I have seen work is to publish last week's forecast against actual every week, unedited, with the variance attributed to the assumption rather than to the person. Once the team sees that the exercise is about calibrating estimates rather than allocating blame, the sandbagging subsides. It takes about two months.

Common Questions

Should we buy a treasury or cash forecasting tool?

Not yet. Build the discipline in a spreadsheet fed by system extracts for one quarter. If the ritual holds and the constraint becomes consolidation across many entities and banks, then buy something. Tools do not create the discipline.

Weekly or daily?

Weekly forecast, daily bank balance. Daily forecasting is for organisations genuinely operating close to their limits, and it consumes a person entirely.

How accurate should we expect to be?

Within about five per cent on total receipts at four weeks out is good for a mid-sized business after a couple of months of calibration. Precision matters less than consistency of bias.

What should we expect over the next twelve months?

Expect supply disruption to show up in receipts with a lag, as delayed shipments become delayed sales and then delayed collections, most likely from April onwards. Expect banks to ask for more frequent and more detailed information, and to look closely at covenant headroom. Expect customers to extend their own payment behaviour quietly rather than by asking. And expect the cash forecast, once it is running properly, to become a permanent board item, because no management team that has seen thirteen weeks of visibility agrees to go back to a monthly balance.

---

Cash Forecasting Setup — we build the thirteen-week direct forecast from your own system data, calibrate it against how your customers actually pay, and leave you with a weekly ritual rather than another report.