Two horizons, two methods

Practitioners split cash forecasting by horizon. An AFP conference session describes a direct model for the near term, built from open receivables and payables, payment terms and a payment calendar, and refreshed weekly with month-to-date actuals. Beyond that, an indirect model derives cash from the P&L and balance sheet forecast [1]. A treasury software vendor describes the same practice: mature teams run the direct method for near-term liquidity and the indirect method for the longer-term plan, and reconcile the two on the same actuals [2].

In K-Φ the near-term side is the GL run-off and the longer-term side is the budget run-off.

The cost of keeping them apart

Nilly Essaides notes that treasury and FP&A produce short-, medium- and long-term cash forecasts that are often out of sync, which creates confusion at planning time [3]. The seam is usually the culprit. Someone decides where actuals stop and the plan starts, pastes one file into the other, and repeats the exercise every month.

Each half also fails on its own. A ledger-only forecast runs out of road: open items unwind, and then it has nothing more to say. A budget-only forecast is confident about a plan the ledger may already have contradicted. You need both, and you need to know exactly where one hands over to the other.

The two run-offs

GL run-off: what is already on the ledger. Open receivables, payables and debt unwind on their own terms, from the ledger itself. This is where the forecast is most grounded, because every number starts from a booked balance.

Budget run-off: what the plan says next. Past the ledger, the budget carries the horizon, translated into cash flows by timing rules. It is re-forecast each quarter and hardened by whatever closes beneath it.

Closer to today, bank actuals from a treasury management system answer the first few days. K-Φ answers the two later feeds on the same axis.

The junction: a cutoff period

The cutoff is the period where the forecast stops using the ledger and starts using the budget. Everything before it comes from GL run-off, everything after from budget run-off. By default it sits three months past the last actual period in the ledger: two closing months as a floor, plus one month of headroom, so that a close running late does not push actuals into the budget zone.

The cutoff is anchored on the data, not on the calendar. If the ledger falls behind, the cutoff falls behind with it, which is visible, instead of silently replacing actuals with budget. When judgment is needed, the cutoff can be dragged on the forecast chart, at month, quarter, week or day granularity.

Under the hood, the ledger side is computed by a GL engine and the budget side by a separate budget engine, and the two are composed at the cutoff.

Automatic by design

Each month a new ledger arrives, through an import, a QuickBooks or Xero connection, or an API feed. Actuals lock in and replace budget assumptions. The cutoff slides forward. The GL run-off is recomputed from the new balances, and the forecast is versioned: a new version publishes and the previous one locks. Nobody moves the seam.

That is what turns a forecast from a document into a living process, and it is why the junction is the core of K-Φ rather than one feature on a list. You can see it in the Forecast section of k-phi.com.

What you get