A short field note on the single most useful piece of finance infrastructure for owner-operated businesses: the 13-week rolling cash forecast.
Nearly every engagement we begin includes some version of this work. By the third or fourth week of an embedded fractional CFO retainer, the 13-week cash forecast is usually the artifact that has shifted the most behavior — more than the closing process, more than the KPI deck, more than the budget. We thought it was worth writing down what we have learned about building one that actually gets used.
Why 13 weeks
One quarter of weekly granularity is roughly the right horizon for an operating business. It is long enough to surface payroll cycles, sales tax remittances, debt service, insurance renewals, and seasonal AR/AP swings — but short enough that the line items still mean something. Twenty-six weeks becomes guesswork. Six weeks misses the next obligation. Thirteen is the practical floor for “I can see what is coming.”
Build it from the bank, not the books
The forecast should reconcile, week one, to the actual operating bank balance. Not the GL cash account, not the trial balance, not the bank reconciliation report — the live balance. If the model and the bank do not tie, the model is wrong. Tie-out happens every Monday, before anything else.
Three columns, not one
Each week should carry three columns: forecast, actual, and variance. The forecast freezes at week-start. Actuals fill in as the week closes. Variance is the conversation. Without the variance column, the forecast becomes a wishlist; with it, the forecast becomes a learning loop.
Receipts and disbursements, not P&L
The forecast is a cash document, not a P&L document. Revenue is replaced with expected receipts by customer and aging bucket. Expenses are replaced with scheduled disbursements by category and due date. Depreciation does not appear. Accrued bonuses appear in the week they are paid. The discipline of forecasting cash, not earnings, is itself the lesson.
One owner
The model needs a single named owner — usually the controller or fractional CFO. Models with shared ownership decay. Models with named ownership compound. The owner sends the weekly variance note to the principal and the bank. The note is two paragraphs, never three.
What it changes
Within a quarter, the same model that started as a forecasting exercise becomes the basis for AR collection cadence, vendor payment policy, credit line draw decisions, and owner distribution timing. By the second quarter it has usually replaced two or three legacy reports that no one was reading anyway. By the third quarter, the owner stops asking for it because it shows up on their desk on Monday morning without being asked.
If you would like to talk about cash forecasting in your own business, we are happy to do that without a meeting. Email andrew@denniscapital.net with a sentence or two about your situation and we will reply.