Cash Forecast Error Attribution

A business forecast $500,000 of Friday cash receipts but collected $430,000. The $70,000 difference matters, but the useful question is why: a late customer, an invoice dispute, a missing transaction or a faulty bank feed each requires a different response.

Palm describes cash variance analysis as comparing forecast and actual cash and tracing differences, and Intuit's cash-flow variance guide also separates operating drivers, while error attribution adds a documented category and owner to the reconciled gap. Freeze the forecast version, using the estimate issued before the cash arrived and not a later update made with hindsight.

Choose the horizon, since a same-day error differs from a weekly or monthly error and a payment shifted from Tuesday to Thursday can disappear in the weekly aggregate. Reconcile actual cash by confirming bank transactions, currency, value date and transfers before blaming the forecasting model, because duplicate or delayed feeds create false variances.

Define the sign, stating whether positive means actual cash exceeded the forecast or vice versa, because inconsistent signs make a bridge hard to read. Break the movement into streams, since customer receipts, supplier payments, payroll, tax and financing have different owners and forecast drivers.

Separate timing from amount: a receipt expected Friday that comes Monday is a timing miss, while an invoice reduced by a credit note is an amount change. Track omitted events, because an unplanned tax payment or asset purchase can dominate a week and the team should find out why it was absent from the input list.

Trace customer receipts through invoice dates, disputes, promised payment dates and actual history, since "customer says soon" is not the same as a scheduled transfer, and trace payments through purchase-order commitments, invoice approvals and payment runs. A valid new sales plan may change the next forecast, but it does not erase the error in the frozen prior one.

Apply cause codes detailed enough to drive a fix but few enough that staff can classify reliably, and allow a pending category so that an open investigation is marked unresolved rather than given an invented cause. Avoid double counting, because a delayed customer payment should not be classified as both collection delay and missing revenue unless separate amounts are involved.

Build a bridge that starts with forecast movement, adds or subtracts classified differences and reconciles to actual movement, leaving any unexplained remainder visible. Look for systematic bias, since a team that consistently expects collections earlier than they arrive may need a timing adjustment across customers, and prioritise large and recurring misses over one rare acquisition payment.

Review materiality, because a $5,000 miss can matter when payroll is due in a small account. Assign owners, verify that corrective actions improve subsequent frozen forecasts, keep an audit trail and show uncertainty as a range, so that for an owner attribution becomes a learning loop that helps the next cash decision arrive earlier and with clearer evidence.