Building a finance function from scratch

The business was selling and growing. Everything behind that was improvised: no month-end, no closed books, no management accounts, no plan written down anywhere. Bills arrived from every direction, and cash was burning with no line of sight on when it would run out.

Nothing that existed could be used for a decision. Gross margin swung from negative to fifty percent between consecutive months — not because the business changed that much, but because it was not being measured consistently. Costs were not allocated. There was no standard costing, so nobody could say what a unit actually cost to sell.

And the most serious problem was the one that looked like an asset: tens of thousands sitting in receivables, counted as revenue, that was never going to arrive. The founder was showing those numbers to prospective investors.

• Revenue recognition that holds. Revenue allocated to the periods it belongs in, bad debt cleaned out, and one definition of revenue that stayed the same every month.
• Costing, properly. A full costing sheet with accurate inventory costs and supplier terms behind it, so gross margin became a number that meant something rather than an accident of timing.
• AP and AR processes that actually existed — bills captured, approved and paid on a schedule, and receivables chased against terms instead of hopefully.
• Cash flow management with a real horizon. Forecast far enough ahead that the business knowingly operated for two months on effectively no cash on hand while an investment closed. That was survivable because it was planned, not discovered.
• Automated sales reporting that splits discounts, refunds, cancellations and replacements apart, so the top line reflects what was actually earned rather than what was ordered.
• Unit economics and the KPIs above them: average order value, customer acquisition cost, units ordered against units delivered, and stock-outs — the handful of numbers that told the founder what to do next.
• A management pack a non-finance founder could actually read, used to run the business monthly and to raise investment.
• Tax recovered rather than left on the table: relief reclaimed on the written-off debt, and a six-figure R&D tax credit that carried payroll through several lean months.

• Revenue roughly 4x over the two years, from a low five-figure monthly run rate to a consistent six-figure one — measured the same way throughout, which is the only reason the comparison means anything.
• Gross margin is stable and moves only when something real moves. It can be planned against.
• Contribution turned positive for the first time in the business’s life.
• The books close every month, and the founder can answer an investor’s question without going to find out.

• No inventory management system. Cash was too tight to justify one, so I built a costing sheet instead. A costing sheet that is right beats a system the business cannot afford — and it kept the option open for later.
• No data pipelines, no warehouse, no BI stack. The sales reporting was automated inside the tools already being paid for. The business needed numbers it could trust by Friday, not infrastructure — and I can build infrastructure, which is exactly why it was my call to make and not to reach for.
• No forecasting model more elaborate than the decisions required. Cash flow was forecast to the horizon that mattered for survival and fundraising, and no further.

Two years, and it was hard in the way this work usually is: long nights, disagreements, and a lot of arguing about numbers that people did not want to be true. Standing up a finance function inside a business that has never had one is not primarily a technical problem. It is persuasion, repeated monthly, until the numbers become the thing everyone refers to rather than the thing someone produces.