--- title: "The board asks: can we afford it? The answer should already exist" url: "https://obserbi.com/blog/can-we-afford-it-answer-should-already-exist" description: "Whether a company can afford a machine, a hire or a new site usually takes a week of spreadsheet work. With a standing model of cash, commitments and scenarios it is a query." --- # The board asks: can we afford it? The answer should already exist October 22, 2026·4 min read·Editorial team A brass cash box standing open with its tray lifted, a brass gauge with a plain dial propped in front, three brass keys on a ring and a closed ledger, in warm amber light against black Picture a board meeting at a metal-processing company in the second half of the year. The operations director has a proposal: a second machining centre, because the current one is the bottleneck and the order book says it will stay that way. The chair turns to the CFO and asks the only question that matters. Can we afford it? The CFO gives the honest answer, which is that she will come back in two weeks. In those two weeks a controller pulls the cash forecast, open purchase orders, the loan schedule, capex approved but not yet invoiced, and the hiring plan from HR. Some of it is in the ERP, some in the treasury spreadsheet, some in an email from the bank. The pieces are reconciled by hand, a deposit and a lead time are added, and three versions of the next eighteen months are built. The answer, when it arrives, is a good one. It is also three weeks old by the time the board sits again, built on a cash position that has since moved, and it took the finance team’s best person out of everything else for a month. The next proposal, a regional sales hire, will start the process again. ## The real cause: the model is rebuilt for each question instead of maintained The question “can we afford X” is not a hard question. It is the same question every time, with a different X. What makes it slow is that its inputs are assembled only when someone asks. There is no standing picture of cash, of what the company has already committed to, and of how the next eighteen months look under a few plausible futures, so every proposal pays the full cost of building one. A standing model turns the question into a query. The proposal becomes one more line of commitments laid over a picture that already exists, and the answer is a comparison of scenarios that are already running, not a project. ## Where the standing model goes wrong **Receivables are forecast at face value.** The cash line starts from the receivables ledger as if every customer will pay on the due date. They will not, and the gap is largest in exactly the quarters when a big outflow is being considered. Receipts belong on the dates payment history predicts, customer by customer, as described in [the article on 13-week cash forecasts](https://obserbi.com/blog/why-13-week-cash-forecast-misses). **Commitments are counted only when they become invoices.** A signed supplier contract, an approved capex not yet ordered, a lease renewal, severance under a restructuring already announced: none of these are liabilities in the ledger yet, and all of them are cash the company has already spent. A model that only sees payables sees the company as richer than it is. **Covenants are left outside the model.** Affordability is not just whether cash goes below zero. It is whether leverage, interest cover or a minimum liquidity clause holds in the worst month of the plan. A purchase the company can pay for and still breach a covenant on is not affordable, and the bank will notice first. The mechanics are in [the article on covenants](https://obserbi.com/blog/covenant-breach-early-warning). **The base case is the only case.** One line of future cash is a hope, not a model. The question the board needs answered is: under which conditions does this proposal stop being affordable? That requires at least a downside and an upside, with the assumptions that distinguish them stated. **Second-order effects are ignored.** A new machine brings an operator, tooling, energy, maintenance and working capital for the extra output before the extra output is paid for. A new hire brings a ramp-up period and a pipeline that takes two quarters to convert. The proposal’s own number is usually the smallest number in the decision. **The model has no memory.** When last year’s affordable proposal turned out to be tight, nobody can say which assumption failed, because the model that approved it was overwritten. Every answer should be kept with the scenario it came from. ## What a standing model needs 1. **A cash forecast that updates itself.** Receipts from payment behaviour, outflows from the payment-run calendar, statutory items on their dates, rebuilt from the ERP and bank data every night rather than every Friday. 2. **A register of commitments that is not the payables ledger.** Contracts, approved capex, leases, planned headcount, each with its expected cash dates, assembled from procurement, HR and a few documents, and kept. 3. **The financing terms as constraints.** Covenant formulas, facility limits and repayment schedules, computed against every month of every scenario, not checked quarterly. 4. **Two or three scenarios that run continuously.** A base case, a downside with named triggers, perhaps an upside. The point is not precision; it is knowing which assumptions the answer depends on. 5. **A way to drop a proposal in.** The machine, with its deposit, lead time, financing, operator and working capital, as a set of dated cash lines overlaid on every scenario. The answer is the lowest cash month and the tightest covenant month in each. ## A test you can run this week Take the last proposal the board approved after an affordability analysis. Find the forecast used to approve it, if it was kept. Then list the commitments the company had at that date that were not yet invoices: contracts, approved capex, leases, announced hires. Check whether each was in the forecast the board saw. Then compare the cash position the forecast predicted for the month the proposal landed with the actual bank balance. If commitments were missing and the cash came in below forecast, the board approved a decision on a picture that was knowably too optimistic. That is not a judgment failure; it is a missing register and a missing scenario, both of which can be built. How a board should read the answer is worth stating plainly. A scenario is not a prediction. It says: if receipts, orders and costs behave like this, the lowest cash month is here and the tightest covenant is this one. The board’s job is to decide whether it believes those conditions and whether it can live with that month. That remains judgment, and no model removes it. What a model removes is the three weeks and the hand-built spreadsheet in between. Keeping that picture current every night, on the company’s own hardware, with the commitments and the covenants in it, is what [ObserBI](https://obserbi.com/) is being built to do. One input worth adding is the cost of serving the customers the proposal is meant to win, the subject of [the article on customer profitability](https://obserbi.com/blog/customer-profitability-no-report-shows). - scenario planning - capital decisions - cash - commitments - board reporting ## Keep reading - [Where to run the finance model so the data stays home](https://obserbi.com/blog/where-to-run-the-finance-model) - [The customer profitability no report shows](https://obserbi.com/blog/customer-profitability-no-report-shows)