Table of Contents
84
16 min
14.11.2026
Flour mill: two points of extraction cost half a million
Kateryna Gulchenko
Senior Customer Support Manager
“For ten years I chased extraction. We celebrated every point. Then I did the sums and saw that two points of extraction had cost me half a million. More flour, less money.”
Pavlo Danylchuk runs a flour mill in Myrhorod. Seventeen years in grain processing, ten of them as the owner. Eighteen thousand tonnes of wheat a year, three grades of flour and bran. Twenty-four people and a lab of his own.
The year closes with 4 065 600 ₴ of profit on revenue of 192 888 000 ₴. Two percent and a bit. In milling that is normal, and that is exactly why a mistake here surfaces only when somebody goes looking for it.
Pavlo was looking in the wrong place. For ten years he measured his mill by one figure, and the figure was correct. It simply counted the wrong thing.
Where a mill keeps its money
Almost nowhere, and that is the nature of the trade.
Eighteen thousand tonnes of grain turn into a hundred and ninety-three million of revenue, and the owner is left with two percent and a bit. You get to be wrong here exactly once.
| Item | For the year |
|---|---|
| Revenue | 192 888 000 ₴ |
| Grain | 154 800 000 ₴ |
| Electricity | 7 560 000 ₴ |
| Packaging | 2 462 400 ₴ |
| Payroll | 9 600 000 ₴ |
| Fixed costs | 14 400 000 ₴ |
| Profit | 4 065 600 ₴ (2.11%) |
Why the margin here is always thin
A mill earns on the gap between what it paid for grain and what it sold the output for. Everything else is the technology of that gap.
Grain is 80.3% of revenue. The mill does not set its price: there is the exchange, the elevator, the season. What you actually steer is about one fifth of the revenue.
When four fifths of the budget are set from outside, a mistake in the remaining fifth eats a share of the profit, not a share of the revenue. And the profit here is two percent.
What comes out of a tonne of grain
A mill does not make flour. A mill splits grain into fractions.
A tonne of wheat yields four different goods, and the price between the extremes differs by more than three times. Which share lands in which is the miller's main decision.
| What comes out | Share of grain | Price per tonne |
|---|---|---|
| Top grade | 45% | 14 200 ₴ |
| First grade | 22% | 11 800 ₴ |
| Second grade | 9% | 9 200 ₴ |
| Bran | 22% | 4 100 ₴ |
| Shrinkage and losses | 2% | — |
Total flour extraction is 76%. That is 13 680 tonnes a year.
The rest of the grain physically cannot become flour: the coat, the germ and what evaporates away. The ceiling is close, and every mill in the world lives in a narrow corridor around those same seventy-odd percent.
Everyone in milling knows the number 76. It stands in the reports, mills are compared by it, the technologist is praised for it.
The other 24% is bran and shrinkage. Bran sells too, just over three times cheaper than flour, and in the owner's head it lives as waste rather than as a product.
What one grade costs
This is where the whole arithmetic of the article lives.
Between top grade and first there is 2 400 ₴ a tonne. Between top and second, 5 000 ₴. Bran plays in another price league altogether: 4 100 ₴ against 14 200 ₴.
Grade is set by the ash content of the flour, and everything that is not pure endosperm drags that ash up:
- the coat that did not come away cleanly in the grind
- the germ with its fat, which spoils both the colour and the keeping
- field impurities that arrived together with the grain
- small, shrivelled kernels with more coat in them than core
Why this weighs more than tonnes
Moving a hundred tonnes from top grade down into first is minus 240 000 ₴. To win that back in tonnage you would have to add more than sixteen tonnes of top grade. And there is nowhere to add them from: the grain is the same.
So in milling the money moves sideways across the grades, not up and down the extraction rate. Yet extraction is what everybody watches.
That is nobody's stupidity. Extraction really was the main indicator for decades, back when flour was one product and the state took it by the bulk. Grades five thousand hryvnias apart are a far younger thing than the habit.
Why extraction became the metric
Because it is simple and you see it every day.
Weigh the grain in, weigh the flour out, divide. One figure, one percentage, no arguments. You can put it on the board in the mill and watch it climb.
“We had extraction up on the board. Every shift. The technologist would come in in the morning and look there first, not at the price list.”
There is no arguing with extraction: the scales show what they show. That finality is exactly what makes it so convenient.
Grade is another matter. Ash content comes back from the lab two hours later, it has to be read, and it does not add up into one neat figure for the board. It can also be contested: it is argued over with the buyer.
A metric people argue about loses to a metric nobody argues about, even when the second one measures the wrong thing.
What was tried before
Pavlo decided to drive extraction from 76% up to 78%.
Technically that is easy: a harder milling regime, tighter rolls, deeper recovery out of the bran. Two points means 360 extra tonnes of flour a year, and on paper it looked like clean money.
Why it is easy to believe
Because extraction is the only part of milling the owner steers directly. The price of grain is dictated by the market, the price of flour too. The roll setting is yours.
Extraction really did climb. So did revenue.
“I worked it out on a napkin myself: three hundred and sixty tonnes at fourteen two hundred. Five million. I showed that napkin to everybody.”
How much revenue actually rose
By 72 000 ₴ for the year.
Not five million and not half a million. Seventy-two thousand, on revenue of almost a hundred and ninety-three million.
That is less than four hundredths of a percent. If revenue swung with the weather or the exchange rate, a change like that would have gone unnoticed entirely.
Where the effect went
The napkin assumed the extra 360 tonnes would come out as top grade. They came out as second. And what used to be top grade travelled down along with them.
A harder grind pulls the coat into the flour, ash content rises, and what passed as top grade yesterday is first grade today. Top grade became 42% instead of 45%, second grade 12% instead of 9%.
In tonnes it looks like this: 7 560 of top grade instead of 8 100, which is 540 tonnes less. Every one of those tonnes lost at least 2 400 ₴, and some of them dropped two steps at once.
What the board did not show
Two lines that arrived together with the new regime.
- flour went to 14 040 tonnes instead of 13 680, and packaging cost 64 800 ₴ more
- a harder grind eats more power: 448 ₴ per tonne of grain instead of 420
- that is 106.67% of the base consumption, so 8 064 000 ₴ instead of 7 560 000 ₴
- electricity added 504 000 ₴ in all
Add it up: revenue plus 72 000 ₴, packaging minus 64 800 ₴, power minus 504 000 ₴.
Power is no small change here. A harder regime means more passes, a tighter gap and a higher load on every system. Power per tonne of grain rose by six point seven percent — on a monthly bill that is invisible, and over a year it is 504 000 ₴ — more than the whole bottom-line minus.
What it cost
Minus 496 800 ₴ for the year.
That is 12.2% of the mill's annual profit. Handed over for two points of a metric that was drawn on the board every single day.
There is more flour and less money, and between those two facts there is no contradiction.
“The worst part is that for a year we counted it a victory. Extraction seventy-eight, everybody pleased. I even wrote the technologist a bonus.”
Why the mistake is so hard to catch
Because it does not produce a loss.
“I asked the bookkeepers to count it again. They counted. I asked a third time. The figure did not move.”
The mill is still in the black. Revenue even rose a little. No report flashes red, because the one figure that would flash is counted nowhere: how many hryvnias a tonne of grain yields once it is split into grades.
Where the substitution happens
Extraction answers the question “how much flour”. Money answers the question “which flour”. Those are different questions, and the second one never stood on the board.
Mistakes that produce a minus show up within a month. Mistakes that produce a plus of seventy-two thousand instead of five million live for years, because formally they work.
The move that works
Pavlo took the milling regime back to where it had been — the same 76% extraction. And then he took on what happens before the rolls.
The logic is simple: if grade is spoiled by everything that is not endosperm, it is cheaper to take that out before the grind than to separate it afterwards.
Grain arrives from the field with everything that grew there: weed seeds, fine impurities, damaged kernels. All of it carries higher ash content than endosperm, and all of it drags grade down before the grind has even started.
- one more separation stage before conditioning
- a longer rest after conditioning, so the coat comes away more evenly
- the fine fraction taken off as its own stream instead of going with the main one
- lab checks on ash content at each of the first three break systems
What it costs
Thirty-four hryvnias per tonne of grain, which is 612 000 ₴ a year. No new equipment and no new people: a different order of operations and one extra cleaning stage.
The dearest part turned out to be not the cleaning but the patience. A longer rest means grain standing in a bin instead of going to the grind, and for the first few weeks that looks like lost pace.
What the preparation gave
Extraction is 76% again.
The same 13 680 tonnes of flour. The same packaging. The same power. The only thing that changed was how those tonnes fell across the grades: top grade 48%, first 20%, second 8%.
Revenue rose by 1 764 000 ₴. Take off the 612 000 ₴ the preparation itself costs and the net effect is 1 152 000 ₴, or 28.3% of annual profit.
Bran neither grew nor shrank: the same 22%. The preparation does not pull flour out of the bran, it pulls out of the flour what should never have got in there.
“We did not take the board down. We just wrote a different figure on it.”
Three columns side by side
It shows best when all three regimes stand in one table.
| Indicator | Before | Harder grind | Better preparation |
|---|---|---|---|
| Flour extraction | 76% | 78% | 76% |
| Top grade | 45% | 42% | 48% |
| First grade | 22% | 24% | 20% |
| Second grade | 9% | 12% | 8% |
| Bran | 22% | 20% | 22% |
| Tonnes of flour | 13 680 | 14 040 | 13 680 |
| Effect for the year | — | −496 800 ₴ | 1 152 000 ₴ |
The middle column added tonnes and took away grade. The right one added not a single tonne and took away nothing.
Look at the first row. By the metric drawn on that board for ten years, the middle column is the best of the three, and the right one is no different from the left at all.
The gap between the two moves is 1 648 800 ₴, and neither of them touched the price of grain.
Why this is not a universal rule
Grade mix is not always dearer than extraction.
If a mill makes one flour with no grades at all, this whole arithmetic disappears: there extraction is the money. The same goes for feed milling, where the fractions are not separated at all.
When extraction really does rule
The difference appears where your product has grades at different prices. The wider the gap between top and second, the dearer every point pushed downwards. When the gap is narrow, chasing extraction becomes sensible.
At Pavlo's mill the gap is 5 000 ₴ a tonne. That is what decided everything.
You can test it at your own mill in five minutes: take the difference between the top grade and the second and divide it by the price of the top one. If it comes out above a third, your money is in the grade mix.
“The board now carries hryvnias per tonne of grain. The technologist was indignant at first that it was not his number. A quarter later he asked to add the top grade share himself.”
What to watch every month
Three figures are enough, and extraction is not among them.
- hryvnias per tonne of grain after the split into grades, as a single line
- the share of top grade in the flour produced
- power per tonne of grain, because the milling regime shows up right there
If the top grade share falls while extraction climbs, you have just traded money for tonnes. If both fall, the trouble is in the grain, not in the regime.
None of the three asks for a new accounting system. Revenue by grade is already on the invoices, tonnes of grain in the intake log, power on the bill. All that is missing is one division a month.
Where to start
Take last month and work out one figure: divide the revenue from all output by the tonnes of grain processed.
Then take the month when extraction was the highest of the year and work out the same thing.
If the second figure turns out to be lower, you have found your money — and found it where the prize for the best result has hung for years.
The money does not vanish — you simply do not see it
Milling looks like a business about grain and rolls. It is really about which grade every tonne you have already bought ends up in.
Pavlo did not have a technology problem. He had a board in the mill carrying a figure that measures tonnes rather than hryvnias.
Two points of extraction cost 496 800 ₴. Three points of top grade at the same extraction brought 1 152 000 ₴. Both figures are true, and between them lies the whole difference between a tonne and a grade.
How to see this at your own mill
All the arithmetic above rests on one row the board in the mill does not have: how many hryvnias a tonne of grain yields once it is split into grades. While the mill is measured by extraction, neither the 496 800 ₴ that two points cost nor the 1 152 000 ₴ that the same preparation brought without a single extra tonne is visible. When the same data is collected automatically, the hryvnia per tonne of grain is visible every month — which is exactly why owners of small and medium businesses use Finmap: to see the financial picture rather than assemble it at the end of the month.
Read also
Topic foundation: Management Accounting Explained: What It Is and Why Owners Need It
Share valuable content — become a source of insights
Questions that come up right away
Why did raising extraction give a minus rather than a plus?
Because the grade balance shifted along with extraction. Top grade fell from 45% to 42%, second grade rose from 9% to 12%. Revenue climbed by only 72 000 ₴, while packaging added 64 800 ₴ and power 504 000 ₴. That is minus 496 800 ₴ in all, or 12.2% of annual profit.
How much is one grade worth per tonne?
Between top grade and first the difference is 2 400 ₴, between top and second it is 5 000 ₴. Bran plays in another price league altogether: 4 100 ₴ against 14 200 ₴ for a tonne of top grade. That difference is what decides where a mill keeps its money.
Why did extraction become the main metric at all?
Because you see it every day and there is no arguing with it: weigh the grain, weigh the flour, divide. Ash content comes back from the lab two hours later and does not add up into one neat figure for the board. Extraction measures tonnes, while a mill sells grades.
What did better grain preparation give?
Top grade rose from 45% to 48%, first grade fell to 20%, second to 8%. Extraction stayed where it was at 76%, so the same 13 680 tonnes of flour. Revenue rose by 1 764 000 ₴ against costs of 612 000 ₴, a net effect of 1 152 000 ₴, or 28.3% of profit.
What does the preparation itself cost?
Thirty-four hryvnias per tonne of grain, which is 612 000 ₴ a year. That buys one more separation stage, a longer rest after conditioning, the fine fraction taken off as its own stream, and ash content checks on the first three break systems. No new equipment is needed.
Is grade mix always more important than extraction?
No. If a mill makes one flour with no grades, extraction is the money — the same goes for feed milling. The difference appears where grades carry different prices, and the wider the gap between top and second, the dearer every point pushed downwards.
Why does the report not show the mistake?
Because it does not produce a loss. The mill stays in the black, revenue even rises a little, and no line flashes red. The one figure that would show the problem is counted nowhere: how many hryvnias a tonne of grain yields once it is split into grades.
Which figures are worth watching every month?
Hryvnias per tonne of grain after the split into grades, the share of top grade in the flour produced, and power per tonne of grain. If the top grade share falls while extraction climbs, you have just traded money for tonnes. All three are already in the invoices, the intake log and the electricity bill.
Any questions left?
We are ready to answer them.
Finmap support
