Chapter 25: Roasting on the Numbers, Not Around Them
The Specialist's Action Plan
Most roasters work out their margin once. It happens in a quiet week — a spreadsheet built in an afternoon, a few satisfying numbers, a quiet right, we're profitable. Then the file gets buried under green orders and service calls and the next month's roster, and it's never opened again.
The specialist doesn't do this. The specialist runs the business on the numbers — not because they love spreadsheets, but because they've learned what the numbers do when nobody's watching them.
They drift.
This is the whole problem in one word. Margin erosion never announces itself. There's no alarm, no warning light, no bad day you can point to. Green prices tick up a few cents a kilo across a season. Roast loss creeps from sixteen percent toward eighteen as a drum wears or a profile slips. A wholesale account you priced two years ago quietly slides below the figure that actually makes it worth servicing — and keeps ordering, and keeps feeling like a win, while it slowly costs you money on every bag.
None of this shows up on any given Tuesday. It shows up months later, all at once, as a quarter that should have been profitable and somehow wasn't — and by then you're reverse-engineering where it went instead of catching it while it was cheap to fix.
The four posts before this one handed you the instruments. True COGS per kilo. The roast-loss reality. The capacity-margin link. Pricing built on your real sales mix rather than hope. But an instrument you read once is just a number you wrote down. The difference between a roastery that knows its margin and one that holds it is discipline — a small, fixed set of figures, checked on a cadence, so drift gets caught while it's still a rounding error rather than a reckoning.
That discipline is the Action Plan. Four numbers, run on repeat:
1. Know your true COGS/kg — and re-run it every time green moves.
Not the invoice price. The loss-adjusted, overheads-loaded cost of a finished kilo out the door. This is the number every other number leans on, and it's the one that moves most quietly. When your green price changes, this changes — so it gets re-run, not assumed.
2. Watch utilisation, not just volume.
Your overheads cost the same whether the drum is flat-out or cold. So every additional good kilo through those same fixed costs is almost pure margin, and idle capacity is the single most expensive thing in the building. Volume tells you how busy you are. Utilisation tells you how much margin you're leaving on the floor.
3. Price off your real sales mix — and hold the floor.
A blended number hides the account that's bleeding you. Know your retail-versus-wholesale split per size, know the floor — the price below which an account costs you money to keep — and hold it. Holding the floor is a decision you make in advance, in the calm, so you're not making it under pressure in front of a customer.
4. Check the cadence, not the crisis.
Fifteen minutes a month beats a frantic annual reckoning every single time. The point isn't to obsess — it's to glance. A specialist reads their margin dashboard the way they'd read a roast curve: a quick, practised look that tells them whether anything's drifting before it becomes a problem.
That's the engine. Not a document you build once and bury — a living dashboard you check on a rhythm, so the margin you worked so hard to find in the first four chapters is the margin you actually keep.
You don't have to feel your way through any of it. The dashboard already exists — COGS, capacity, pricing, and annual profit, all modelled on one screen.
The Silverback Tech Note: The ten-year-old spreadsheet
“I once sat in a roastery with an owner who told me, with real pride, that he'd "done his numbers." He had — once. He pulled up the spreadsheet to show me. It was built the better part of ten years earlier, and not a cell had changed since.
In that decade his green had moved through several price cycles. He'd switched origins twice. He'd added a bigger roaster, taken on two staff, moved to a unit with higher rent, and picked up a clutch of wholesale accounts at prices he'd set by feel and never revisited. The spreadsheet knew none of it. It was still confidently reporting a cost per kilo from a business that no longer existed — his own business, ten years younger.
Here's the part that stayed with me: he wasn't careless. He was a good roaster running a busy operation. He'd done exactly what most owners never get around to at all — he'd worked out his margin properly, once. The trouble was the word once. A cost model isn't a monument you build and stand back from. It's an instrument, and an instrument that hasn't been read in ten years isn't telling you about today. It's telling you about a day a decade gone, in a voice convincing enough to plan around.
We re-ran it together that afternoon. His true COGS/kg had climbed by a margin that explained, almost exactly, why a business that felt busier than ever didn't feel any more profitable. Nothing dramatic had happened. Everything small had — green creep, a heavier roast loss on the new drum, two accounts that had quietly drifted under the floor. Each one invisible on its own. All of them together, compounding silently for ten years.
That's the whole case for cadence in a single visit. The number isn't the win. Re-running the number is the win — and the gap between fifteen minutes a month and ten years is the difference between catching drift while it's a rounding error and discovering it as a decade you can't get back.”
Run your numbers free → Pegasus Coffee Calculator.
The Pegasus Coffee Calculator builds retail and wholesale prices straight up from your true cost per kilo, for every bag size, with the margin on a slider you can move — Update your numbers as much as you want, export them as a CSV or PDF as many times as you need.