How we saved €4,000 in our first year

Years of guessing

We were doing about €35,000 a month in revenue. That is a real restaurant, with a real kitchen and a real team, and for a long time it felt like plenty. Busy Friday, good week. Quiet Tuesday, bad week. That was the entire reporting system.

At the end of every month there was less left than we expected, and we could not say why. Not roughly why — at all. We could tell you exactly what we had sold. We could not tell you what it had cost us to sell it.

The uncomfortable part is that we were not being lazy. We were in the building six days a week. We knew the guests, we knew the menu, we knew which supplier would be late. We just had no idea what any of it was doing to the money, and somewhere along the way we had built the habit of not looking.

Delivery notes went in a drawer. Invoices went to the accountant. The rota was whatever it had been last week, because last week had gone fine.

What we tried

Excel first. A friend built us a spreadsheet with tabs for purchases, sales and waste, and for about six weeks it was beautiful. Then somebody dragged a cell, a formula broke quietly, and the numbers stopped matching reality. We did not notice for a month. After that nobody trusted it — and a spreadsheet nobody trusts is worse than no spreadsheet at all, because it still gives you an answer and you still act on it.

Then the accountant. Ours is good at the job; the problem was the shape of the work. The reports arrived weeks after the month had closed, organised for the tax office rather than for us. They told us accurately what had already happened. By then the food was eaten, the waste was in the bin, and the month we could have fixed was closed.

So we went back to gut feeling, which in practice meant explanations that all pointed away from us. Suppliers had put prices up. The season was slow. Energy. Every one of those was partly true, and every one of them was outside our control — which, we worked out much later, is usually how you can tell you are looking at an excuse rather than at a cause.

What we built

We are not a software company. We are operators who got tired of not knowing, so we built the smallest thing that would answer one question: where does the money go between the till and the bank account?

It turned into a dashboard. Once a week we enter the same short list of figures and it gives us food cost, labour cost and break-even. There is no POS integration, and that is deliberate. Typing the numbers in yourself takes twenty minutes and keeps you close to them. A number that arrives automatically is a number nobody reads.

• Revenue for the week, split between food and drink
• Purchases, entered as the invoices arrive instead of in a monthly pile
• An inventory count on the same weekday, before the first delivery
• Payroll hours actually worked, not hours rostered
• Waste and staff meals, logged when they happen rather than from memory

The first number that hurt

It took two weeks to get our first real insight, and it was not subtle. Food cost was running at 36%. Payroll was at 42%. Seventy-eight cents out of every euro were gone before rent, energy, insurance, or a single repair.

We had been living inside those numbers for years without ever seeing them written down next to each other. Two weeks of honest data entry did what several years of working hard had not: it told us which problem to pick up first.

What changed

Nothing dramatic happened next. There was no single leak, no theft, no villain. There were about a dozen small things, and not one of them would have justified a meeting on its own.

On food, portions were the biggest of them — a heavy hand on the protein does not look like anything on the pass. Under that: ordering for a week we assumed would be busy and then was not, prep for a Friday that never came, staff meals nobody had ever costed, and two dishes that had been losing money on every plate for the better part of a year. We started weighing portions. We ordered against the inventory count instead of against a feeling. We reworked the two dishes and dropped one of them. Food cost went from 36% to 28%.

On labour, we had been running more or less the same rota every week regardless of what the week actually looked like. Once we could see hours worked against sales by day, the overstaffed shifts were obvious — and so were two that had quietly been understaffed, which was costing us in a different way. Nobody was let go. We moved hours to where the guests were. Payroll went from 42% to 34%.

Added up, the dashboard identified about €1,800 a month in savings.

The first year did not return €1,800 twelve times over. It returned about €4,000, and that gap is the honest part of this story. Changes take months to land. We implemented some of them badly the first time. A couple did not survive a busy summer and had to be put back in place in September. €4,000 in a year is not a transformation — it is a small salary we had been throwing away, and the only reason we found it is that we started counting.

Why we made it available

Once we had the thing, we started noticing the same drawer of delivery notes in every kitchen we walked into. Every operator we spoke to knew last month’s revenue. Almost none of them could tell us last week’s food cost.

So we cleaned it up and made it available. We are not consultants and this is not a turnaround programme. There is nothing in here that a competent operator could not work out alone with a calculator and one fixed hour every Monday. What the system does is make sure that hour actually happens, and that the number in front of you is the same number, calculated the same way, week after week.

That is the whole method. Count the same things on the same day. Watch the direction of travel rather than the level. Fix the biggest gap first, then look again. The uncomfortable truth is that we could have done all of this years earlier with a notebook. We did not, because nothing was forcing us to look — and that, more than any feature, is what we ended up building.