Platform · financial core

What you have, and what is already spoken for

Bank balance is not cash position. Position is balance minus what is committed — payroll, taxes, cheques not yet presented, approved bills, and the subcontractor who will invoice on Friday — and almost nobody can see that number without building it.

$413KCurrent
$186K1–30
$94K31–60
$42K61–90
$29K90+

Collection priority — ranked by recoverability, not by age

Northwind Trading$48,20074 dayspays at 71 avg · low risk · soft reminder
Fulton Systems$31,40096 daysfirst late invoice in 3 years · call, do not dun
Depot Industrial$22,900112 daystwo broken promises · escalate to owner
Harbor Logistics$18,60038 daysrenewal in 14 days · hold all dunning
Position across all accounts13-week forecast from commitmentsCollection behaviour per customer

What it does

Six things, specifically.

Position, not balance

Every account and entity, with committed outflows netted — approved bills, scheduled payroll, tax obligations, and payments in flight but not yet cleared.

Thirteen weeks forward

Built from open receivables, open payables, contracted recurring revenue, and payroll — a projection of known obligations rather than a growth assumption applied to last quarter.

Behaviour per customer

The customer who always pays at day 52 is modelled at day 52, not at their stated 30-day terms. Averaged DSO hides exactly the customers that matter.

By entity, then pooled

Position per legal entity with intercompany visible, so a group with cash in the wrong subsidiary can see that rather than discovering it at a payment run.

Multi-currency

Balances translated at current rate with exposure by currency shown, because a comfortable consolidated position can conceal a shortfall in one of them.

Shortfalls flagged early

Where the projection dips below a threshold you set, the alert comes weeks ahead with the drivers listed rather than the week it happens.

Why the spreadsheet forecast is always wrong

The standard thirteen-week model takes last quarter’s collections, applies a growth rate, subtracts a payroll estimate, and produces a curve. It is wrong in a specific and consistent way: it smooths exactly the lumpiness that causes cash problems.

Real cash trouble is not a gradual decline. It is a quarter where three large customers each slip by two weeks while payroll, a tax payment, and an annual insurance renewal land in the same fortnight. A smoothed model cannot represent that because it averaged away every input that would have shown it.

Cash problems are not gradual. They are three slow payers and a tax date landing in the same fortnight, and averaging is what hides it.

Behaviour, not terms

Payment terms describe an agreement. Payment behaviour describes what happens, and the gap between them is stable enough per customer to be predictive.

Modelling each customer’s actual pattern — including the ones who pay early for a discount and the one who reliably pays at 60 on 30-day terms — produces a materially different curve from one built on terms, and the difference concentrates in your largest accounts.

Commitments are knowable

Most of the outflow side is not a forecast at all. Approved bills have due dates. Payroll is scheduled. Tax obligations are calculable. Subscription renewals are contracted. Purchase orders are committed even where the bill has not arrived.

Treating those as known rather than estimated removes most of the uncertainty from the outflow side, which leaves collection timing as the genuinely variable input — and that is where modelling effort belongs.

Cash in the wrong entity

Groups routinely discover at a payment run that the consolidated position is healthy and the entity that owes the money is not. Moving cash between entities has tax and legal consequences that make it slow.

Position per entity, with intercompany balances visible, turns that into a decision made two weeks ahead rather than a scramble on the day.

Limits

Where this does not help.

It cannot predict a surprise

A customer who has always paid on time and then does not is not predictable from history. The forecast narrows the range; it does not eliminate it.

Not a treasury management system

Investment laddering, hedging, and debt covenant modelling belong in a treasury product. We handle position, forecast, and exposure.

It does not move money

Payment execution runs through your bank and your approvals. We show what should be paid and when, and a person releases it.

Questions

What people ask.

How far ahead does it forecast?
Thirteen weeks at weekly grain by default, extendable. Beyond that, known commitments thin out and it becomes planning rather than forecasting.
Where does collection timing come from?
Each customer’s actual payment history rather than their stated terms. The gap between the two is stable enough per customer to be predictive.
Does it work across entities?
Position per legal entity with intercompany visible, plus a pooled view. Cash in the wrong subsidiary is a common and avoidable surprise.
Can it trigger payments?
No. It shows what should be paid and when; execution runs through your bank under your approvals.
What about foreign currency?
Balances translated at current rate with exposure shown per currency, because a healthy consolidated position can hide a shortfall in one.

See what is actually spoken for.

Tell us how you forecast cash today. We will show you what commitments change.