A 13-week cash forecast survives when it costs less to maintain than it is worth to the person reading it. That means a forecast built on actual cash, week by week, with the largest customers and payments shown individually and everything else grouped. It is updated in a fixed fifteen-minute routine each week rather than rebuilt. Forecasts are often abandoned because they were built too detailed to keep current, so by week three the numbers are stale and nobody trusts them. Build the forecast around the update, not the first version.
Why most 13-week forecasts die
The pattern is familiar. A founder, a lender or an investor asks for a cash forecast. Someone builds a detailed model, often driven from the profit and loss forecast, with every expense line and every customer. It is accurate on day one.
Then the first week passes. Actual receipts differ from the plan, two suppliers are paid late, and a new contract starts. Updating the model means touching dozens of lines and reconciling it back to the bank. That takes two hours, and the founder does not have two hours. The model is skipped one week, then the next. By the third week it describes a business that no longer exists.
The problem is not discipline. The maintenance cost exceeded the perceived value. The fix is to cut the maintenance cost until the update fits in fifteen minutes.
Get the granularity right
Three choices set the maintenance cost.
Weekly, direct, cash only. Each column is a week. Each row is cash actually expected to arrive or leave the bank in that week. It is not revenue, not expenses and not profit. Do not derive it from the profit and loss forecast. Forecast receipts and payments directly. Accruals, depreciation and other non-cash items have no place in it.
Individual where it matters, grouped where it does not. List individually the customers and payments that could move the cash position materially on their own: typically the top five to ten customers, payroll, rent, tax payments, loan repayments and the few suppliers that matter. Group everything else into one or two run-rate lines, such as "All other customers" and "Other operating payments", forecast from recent weekly averages.
A threshold for detail. Set an amount below which nothing gets its own row. For a small business, an item that does not change the closing balance by more than a few percent of the minimum cash buffer is noise. Leave it in the run-rate lines.
A structure you can copy
Each week's column runs top to bottom like this:
- Opening cash: all bank accounts combined. For week 1, this is the actual bank balance.
- Receipts: one row per named customer; one row for all other customers; one row for other receipts, such as tax refunds, loans drawn or equity.
- Total receipts.
- Payments: payroll, including payroll taxes; rent; named key suppliers; one row for other operating payments; tax payments, such as GST, VAT or sales tax and advance tax; debt service; capital expenditure.
- Total payments.
- Net cash flow: total receipts less total payments.
- Closing cash: opening cash plus net cash flow. This becomes the next week's opening cash.
- Minimum cash buffer: the level the business should not go below, set by the founder.
- Headroom: closing cash less the buffer. This is the row people actually read.
A worked week. Opening cash is 120,000. Receipts: Customer A 18,000, Customer B 12,500, all other customers 16,000, for total receipts of 46,500. Payments: payroll 38,000, rent 9,500, key suppliers 11,200, cards and subscriptions 2,500, for total payments of 61,200. Net cash flow is minus 14,700, and closing cash is 105,300. Against a buffer of 75,000, headroom is 30,300.
Thirteen columns of that, with one line of notes per week for anything unusual, is the whole forecast. It fits on one screen.
The weekly update: fifteen minutes, same time each week
Pick a fixed slot, such as Monday morning, and do the same five steps each time.
- Replace last week's forecast with actuals. Take receipts and payments from the bank, by row. Closing cash must agree to the bank balance. If it does not, the forecast is not anchored and nothing else matters until it is.
- Note the biggest variances, in a line or two. For example, forecast receipts were 46,500 and actual receipts 39,000. The 7,500 shortfall is Customer B's payment moving to next week. The variance note is where the learning happens.
- Update the next four weeks for known changes: a payment promised, an invoice raised, a hire starting, a supplier pressing for payment.
- Roll the horizon. Drop the completed week and add week 13, using run-rates and anything already known.
- Read the headroom row. If any week falls below the buffer, that is the agenda for the founder's next conversation, whether that means chasing collections, deferring a payment or drawing on a facility.
The update is a roll-forward, not a rebuild. The structure does not change from week to week. Only the numbers do.
What to leave out deliberately
- Profit and loss detail. The forecast answers when cash moves, not whether the business is profitable. Keep the two models separate.
- Daily timing. Weekly is precise enough for decisions and far cheaper to maintain.
- Every small supplier. Group them. Itemising them adds work without changing any decision.
- Scenarios, at first. Run a base case until it has survived a few weeks of updates. Add a downside case later, by adjusting a few named rows, not by building a second model.
- Spreadsheet complexity. No macros, no external links and no hidden sheets. Anyone who has to update the forecast should be able to see how it works.
Signs the forecast is working
After a month, the forecast should agree to the bank every week. The variance notes should be getting shorter as the run-rates settle. And the founder should be using the headroom row to make at least one decision a month: when to chase, when to pay, when to draw. If those three things are true, the forecast will survive, because it is now cheaper to keep than to lose.
Where this fits
A maintained 13-week forecast is usually the first deliverable in a Virtual CFO engagement, because it feeds directly into decisions. It works best on books that are reconciled every month. If yours are not, start with the close, then build the forecast on top of it.
Wissen Solutions builds and maintains 13-week cash forecasts, budgets and board reporting through our Virtual CFO services.
Note on the worked example: the figures are illustrative and were checked by calculation. Total receipts 46,500, total payments 61,200, net cash flow minus 14,700, closing cash 105,300, and headroom 30,300 against a 75,000 buffer.