Cash forecast or financial model: which one you need first

A cash forecast asks what will be in the bank in the next few weeks. A financial model asks what happens to the business over the next few years if you make certain choices, such as hiring ahead of revenue or raising prices.

The first is built from facts you can see: invoices that are due, bills that will be paid, payroll dates, loan payments, and the tax payments you already know about. The second is built from assumptions you choose. Both are useful. They fail in different ways.

A common version looks thirteen weeks ahead, one column per week. Each week starts with the opening balance, adds expected collections, subtracts payments, and ends with a closing balance that becomes the next week's opening.

For a single month it might read: opening cash $120,000, collections $78,000, payroll $62,000, rent and software $14,000, and a quarterly tax payment of $9,000. The month closes at $113,000. Add the next month's numbers, then the next, and the forecast shows the week cash gets tight, not just the average.

A model connects drivers, such as customers, prices, hiring dates, and churn, to revenue, costs, profit, and the balance sheet, usually by month. It lets you ask what happens if sales come in 20 percent below plan, or if two engineers start in September instead of January.

A model is the right tool for a budget, a fundraising conversation, a lender, or a decision about a large commitment. It is a poor tool for knowing whether Friday's payroll clears, because every figure in it is an estimate.

Build the cash forecast first. It takes a few hours, it uses numbers you already have, and it prevents the problem that ends companies: running out of money while the income statement looks fine. Update it weekly by replacing last week's guesses with what actually happened.

Add a model when you have a decision that depends on the longer view. Start with the lines that change decisions, such as revenue, payroll, and the few large costs. Resist the urge to model every expense. A small model that people trust beats a large one that nobody updates.

The two tools should tell one story. The model's cash balance for the next three months should land close to the cash forecast's. When they differ, find out why. Usually the model assumed customers pay in thirty days and they actually pay in forty-five, or it left out a one-time payment the forecast already knows about. Fixing that gap improves both.

Feed the actual results back after each month's close. A model that is never compared with actuals turns into fiction within a quarter.