Start with the real question: if 30% of invoices slip by two weeks, can you still make it through…

That is the right question, not “what is my bank balance today?”. An SME can look healthy on paper and still hit a liquidity wall next week if several expected payments arrive late. The European Commission notes that late payment damages cash flow and can push smaller firms to rely more heavily on overdrafts or short-term credit. In France, the commercial-payment framework sets rules between businesses, while Service-Public explains that late-payment penalties may apply when those rules are breached.

In practice, cash planning should be treated as a map of pressure points, not as a monthly average. The operational angle is to combine three layers: fixed due dates, probable inflows, and historical lateness by customer or segment. Bpifrance still highlights customer-payment timing as a central cash issue for French SMEs.

  • Useful question: “On what exact date does cash fall below my alert threshold?”
  • Poor question: “What is my average cash position for the month?”
  • Goal: spot risky weeks before they become emergencies

The method: a 13-week mini cash forecast in three columns, then three scenarios

Build a rolling 13-week table and refresh it every week. Line by line: expected inflows, firm outflows, and probable outflows. For inflows, use the estimated payment date, not the invoice date, ideally customer by customer. For outflows, keep the known due dates: payroll, taxes, VAT, rent, purchases, loan repayments, subscriptions. The point is not prediction theater; it is to expose timing gaps.

Then add three scenarios. Central scenario: customers pay according to recent behavior. Cautious scenario: add a delay to weaker accounts. Stressed scenario: worsen the two or three biggest inflows again. This is more robust than a single forecast number because it shows where cash actually breaks. Bpifrance explicitly promotes working-capital management to optimize cash; that principle is highly practical here.

Rule of thumb: if a customer usually pays at day 47, do not schedule that inflow at day 30 just because the invoice “should” be paid. Cash does not need optimism; it needs discipline.

  • Column 1: due date or expected inflow
  • Column 2: estimated actual payment date
  • Column 3: timing gap and cash impact

Fictional numerical example: the company that thought it had €60,000… until Thursday

Fictional example. Atlas Services starts the week with €60,000 in cash. It expects €85,000 in inflows over the next 15 days, but its history shows that 40% of customer payments arrive on average ten days later than planned. Its firm outflows over the same period are €92,000: payroll, social charges, a critical supplier, partial VAT, and rent. If Atlas books all inflows on invoice date, it wrongly concludes it will still have €53,000 after two weeks. In reality, once the delay pattern is applied, it drops to €15,000 before two large payments land.

So the real diagnosis is not “do I have enough customers?” but “when does the cash actually arrive?”. In this example, Atlas must either accelerate collection on two invoices, secure a short-term facility, or postpone a non-essential expense. The decision may vary, but the alert is the same: the cash peak problem is invisible in revenue; it appears only in the mismatch between due dates and delays. European analyses also stress the liquidity harm caused by late payment.

  • Opening cash: €60,000
  • Theoretical inflows: €85,000
  • Firm outflows: €92,000
  • Cash before delay correction: dangerously overstated

Know the limits: your model is only as good as your data

A cash forecast becomes fragile if the underlying data is dirty. Unmatched invoices, inconsistent due dates, merged customer names, forgotten credit notes, and partial payments not tracked correctly will all distort the curve. Even a good spreadsheet becomes misleading if customers, periods, and flow types are mixed together. Data quality is therefore not a luxury for controllers; it is the precondition for management.

You must also separate three things: contractual terms, actual behavior, and dispute risk. A customer can be “on time” under the contract while still paying late because of a delivery dispute. In that case, your forecast must include the chance of blocking, not just the nominal delay. And not all delays mean the same thing: a strategic large account in arrears is handled differently from a small scattered one.

  • Check contractual due dates and real payment dates separately
  • Clean duplicates and invoice status errors
  • Segment customers by behavior, not just by amount

A seven-day action plan: who does what, and what result should be visible

Day 1: finance prepares the list of the 20 largest expected inflows and 20 largest outflows. Observable result: one view of the next 13 weeks of cash.

Day 2: accounting or AR corrects wrong due dates and marks paid invoices. Result: fewer gaps between invoice reality and forecast.

Day 3: sales classifies customers into three payment-behavior groups. Result: a delay pattern by group, not by gut feeling.

  • Day 4 — CFO/owner: set the cash alert threshold and the response plan; observable result: a decision trigger that does not need debate.
  • Day 5 — collections: call or follow up the 5 most sensitive invoices; observable result: written payment confirmations.
  • Day 6 — procurement/ops: defer non-critical spending if needed; observable result: less cash strain in the tightest week.
  • Day 7 — owner: compare the central and cautious scenarios and refresh the rolling forecast; observable result: a living weekly cash model.