There are two ways to forecast cash.
Count the money: list every receipt and payment you expect, by date, the way it will hit the bank.
Or derive it: start from the P&L, add back the non-cash items, adjust for working capital, and arrive at cash flow the way the accounts do.
Direct and indirect. Neither is better. They answer different questions, and most forecasting arguments are really two people answering different questions at each other.
Direct: the operational truth
The direct method is the granular view, and it suits shorter horizons. Transaction level forecasting, contracted cashflows. Payroll on the 25th, the VAT payment, the customer who pays on the 10th of each month whatever their terms say. It reflects reality, line-by-line. Its weakness is reach: beyond a quarter, listing individual receipts becomes guesswork. Best used for liquidity management, when ensuring you have enough cash on a given day is the goal.
Indirect: the planner's view
The indirect method rides on the budget: start with revenue projections, margin assumptions, capex and financing assumptions, and work your way back to cash. It reaches as far as the budgeting horizon does, ties to the numbers the board already tracks, and answers the multi-year-scale questions: funding gaps, covenant trajectories, how much FX exposure next year holds. Its weakness is the mirror image: it has no idea whether Friday's payroll clears, because it doesn't know what a Friday is.
You need both, and they need to meet
Which do you need? The practical answer is both. Direct for the next 13 weeks, indirect beyond, with the crossover zone treated with suspicion because it's where both methods are weakest. Reconcile them regularly: look at both over the same quarter and compare. If they disagree materially, one or both needs to be adjusted. Persistent gaps usually mean the indirect model's assumptions aren't tying with the reality of the business.
The same job at three sizes
Start-up. Direct will ensure you have the money to keep the lights on. … read more show less
A runway model needs to be as accurate as possible, but a 13-week forecast doesn't help plan a fundraising cycle. Start with money in the bank, AP & AR, receipts you'd bet on, payments you can't avoid. Derive from the P&L cautiously when your direct forecasts can't take you any further.
Established mid-market. Both, for separate tasks. … read more show less
A 13-week direct forecast connected to bank data, an indirect model riding the budget, and the quarterly reconciliation that keeps them honest with each other. The 13-week forecast is focused on liquidity management, ensuring there are sufficient funds in the accounts to keep things moving, and planning buffers or when facilities need to be drawn. Specific teams should be given responsibility for the different inputs (payroll from HR, revenue from Sales, debt from Treasury/finance, etc.) with a process to make information gathering as smooth as possible. The longer forecast is for investment/facility planning, scenario planning, strategic decision-making and managing exposures.
Large corporate. A systematic approach to both, with buy-in from group stakeholders and a reconciliation process with ownership. … read more show less
The residual risk is organisational: two teams, two forecasts, and a difference nobody is responsible for explaining.