Financial projections template: building a forecast you can defend

A financial projections template gives you the rows — revenue, cost of sales, overheads, cash — and that structure is genuinely useful. What it cannot give you is the thing every reader tests first: whether the numbers come from stated assumptions or from typing until the chart looked right. Build the model so the difference is visible, and most of the argument disappears.

The practical shape for a small business is three linked sheets: assumptions, a monthly profit and loss, and a monthly cash flow. Balance sheet if a bank asks. Nothing else, until the business is complicated enough to need it.

What goes in a financial projections template

• An assumptions sheet, first, where every input lives: price, volume, conversion, churn, headcount, salary, rent, payment terms. Nothing is typed anywhere else.
• Revenue built from a driver — customers × price, or units × price — never as a single growing line.
• Cost of sales as a percentage or per-unit cost that traces to the same drivers, so margin moves when volume does.
• Overheads by category and month, with the founder's salary included even if it is not being paid yet.
• Headcount as a small table — role, start month, cost — because people are usually the largest and lumpiest cost.
• Cash flow separate from profit, with payment terms applied: money invoiced in March and paid in May belongs in May.
• VAT or sales tax modelled if you are registered. It is not your money, and it is the most common reason a profitable forecast runs out of cash.
• A closing cash balance per month, which is the single number the whole model exists to produce.

Profit and cash are not the same, and forecasts that ignore the gap are the ones that fail in month seven. A business can be profitable on paper and unable to make payroll because customers pay at 60 days and staff are paid at 30. If you build only one thing carefully, build the cash line.

Building it step by step

1. Open with the assumptions sheet and fill it before touching any other tab. If a number cannot be justified in one sentence, it is a guess — label it as one.
2. Colour inputs differently from formulas so anyone reviewing knows what is safe to change. This single convention prevents most model errors.
3. Build twelve months across, one column per month, and do not summarise into quarters until the monthly version works.
4. Link revenue to drivers with formulas, never hard-coded figures. A model with typed revenue cannot answer any question.
5. Add payment terms to convert invoiced revenue into cash received, and supplier terms to convert costs into cash paid.
6. Sanity-check the outputs against reality: is the implied revenue per employee plausible, is the growth rate one anyone in your sector has achieved.
7. Build two more cases by changing only the assumptions sheet — a slow case and a good case. If changing one input requires editing three tabs, the model is wired wrong.
8. Update actuals monthly beside the forecast. A forecast never compared to what happened teaches nothing.

Three cases, not one

A single forecast invites a single argument about whether it is right. Three — base, slow, and good — shift the conversation to which assumptions drive the difference, which is the conversation worth having. The slow case matters most: it tells you the month you run out of cash if things take twice as long, and that date is usually the most important number in the entire model.

Ettex Sheets handles this comfortably: import an existing XLSX with formulas intact, 150+ functions with autocomplete and instant recalculation, conditional formatting to flag a negative closing balance the moment it appears, charts from any range for the cash curve, pivots when the model outgrows a single view, and cell comments with threads so an accountant can query the churn assumption against the cell it lives in. Version history means a forecast sent to a bank in March is still recoverable, and export to CSV or XLS hands the file to anyone who wants it elsewhere. The written plan around the numbers belongs in Ettex Docs.

Said plainly: Ettex does not supply a ready-made financial model, does not connect to your bank or accounting data, and has no scenario or forecasting engine. This is a spreadsheet with real formulas — the model is yours to build. That is the honest trade: more work up front, and a forecast whose every number you can explain.

Forecasts rarely stand alone: they are the numbers section of something larger. A business plan template gives the surrounding structure — market, model, operations and risks — so the projections have context a reader can judge.

Forecasts sometimes miss money already earned. Research and development tax relief returns a share of qualifying project costs, and loss-making companies can often take it as a cash credit.

Errors that show up in every review

• Revenue typed rather than driven, so no question about volume or price can be answered.
• Founder salaries and payroll taxes left out, understating costs by a third.
• Cash and profit treated as the same line, with payment terms ignored entirely.
• VAT or sales tax counted as revenue, which flatters cash until the return is due.
• Hard-coded numbers hidden inside formulas, which is how models become unauditable.
• One case only, so the reader has no way to see which assumption the whole thing depends on.
• No comparison against actuals, so the same optimistic assumption is repeated every quarter.

A financial projections template gets you the rows. What makes it a forecast is one assumptions sheet, formulas that trace back to it, cash separated from profit, and a slow case that tells you when the money runs out.