Answer
Usually both. Use direct (receipts and disbursements) for same-day through ~13-week views where timing and funding decisions matter. Use indirect (P&L and balance-sheet bridge) for monthly and longer planning. Many treasuries run a hybrid in the middle: direct for known cash items, indirect for the residual working-capital bridge.
Why
The methods answer different questions:
- Direct asks when cash moves — collections, payroll, tax, debt, CapEx. It is the right input for sweeps, overdraft avoidance, and short-term investment under liquidity management.
- Indirect asks how much cash the business will generate from earnings and working-capital change. It aligns with FP&A and board packs, but it is weak for tomorrow’s funding gap.
- Data reality — bank and ERP actuals support direct short horizons; reliable monthly close supports indirect longer horizons. Forcing one method across all horizons usually produces a forecast nobody trusts.
- Structures downstream — cash pooling and payment release calendars need direct near-term accuracy; strategic buffers and line sizing need the longer indirect view.
Pick by horizon and decision, not by preference for one spreadsheet style. If the team will not maintain daily driver updates, do not promise a daily direct forecast — shorten the horizon or automate bank/ERP feeds first. Full mechanics: cash forecasting.
Next
Method and design: cash forecasting. Umbrella: liquidity management. Concentration that consumes the short-term view: cash pooling.