Cash flow forecasting that starts from money that already moved.

Money moving in the horizon

EUR 247,740,000

Every projected receipt and payment added together, across the thirteen weeks

Of it, on a dated record

EUR 166,940,000

An open invoice, a payroll run, a loan repayment, a filing — each carried at its own date

Of it, read off closed weeks

EUR 80,800,000

Card settlements, small recurring costs, receipts that arrive without an invoice

Open supplier invoices

Every unpaid item in your ledger, placed on the date its own terms make it due. Not an average delay applied to a total, and not a monthly bucket: the invoice, the entity and the account it leaves from.

Customer invoices, at the date that customer pays

The amount comes off the invoice. The date comes off what that counterparty has actually done, week after week, on statements that were matched and closed — so a customer who runs eleven days past terms sits on the curve eleven days past terms.

Payroll runs

The cycle, the entity and the account each run leaves from. Payroll is the largest single outflow most groups have and the easiest one to place exactly, which is why the tight weeks in a horizon are usually payroll weeks.

The debt schedule

Repayments, interest dates and maturities read off the facilities you already hold. The week a loan comes due is on the curve before anyone builds a slide about it, and it moves when the facility does.

Tax and filing dates

The calendar each entity actually files on, in its own country. A group filing in six countries has six calendars, and the horizon carries all of them rather than a group average.

Weeks that have already been reconciled

What has no record gets its shape from what the group has lived: card settlements, small recurring costs, receipts that arrive without an invoice. It is drawn as that, in its own color and its own line on the split, instead of being blended into the half you can prove.

A week we cannot evidence is named on the screen and left out of the figure it would otherwise sit inside, with the reason attached. A total that quietly absorbs a gap is worse than a total with the gap written across it, because the gap is the part a treasurer needs to be told about.

01How far ahead does the forecast go?

As far as the records do. Day by day across the weeks where dated items dominate, week by week over the thirteen-week horizon this page shows, and month by month beyond it. One set of records read at three resolutions — not three models to keep in step.

02Do we have to keep a forecasting spreadsheet in step with it?

No. The inputs are already here: the statements your banks send, the open items in your ledger, the payroll calendar, the debt schedule and the filing dates. What you maintain is the assumptions you want to override, and every override is recorded with who made it and when.

03What happens to a week we have no evidence for?

It is marked on the surface and left out of the figure it would otherwise sit inside, with the reason beside it. A total that quietly absorbs a week nobody can evidence is worse than a total with a hole named in it, because the hole is the thing worth knowing.

04Can we compare the forecast against what actually happened?

Actual vs Budget is a screen of its own. It puts the closed month beside the plan approved for it, and the actual column is reconciled bank movement rather than an estimate typed in afterwards. Each difference opens onto the movements behind it.

05Can each entity forecast its own cash and still roll up to the group?

Yes, and it is the same curve. An entity treasurer works on their own accounts and their own commitments; the group reads the sum of them with intercompany movement taken out first, so the two never disagree about a payment that has one leg on each side.