Three losses, one after the other
Sixteen percent off in the drum, two percent off at the sorting table, one and a half percent gone between the cooler and the sealed bag. Add those and you get 19.5 percent. Multiply the survival fractions — 0.84 times 0.98 times 0.985 — and you get 0.8109, which is 18.91 percent lost. The gap is 0.59 percent, and on a 132 pound lot that is three quarters of a pound. Small here. It stops being small the moment any of the three gets larger.
The order matters because each step only ever works on what the step before it left behind. The sorting table never sees the moisture that left in the drum, so taking two percent off the green would be counting weight that was already gone.
The mistake that costs you green
Sixteen percent is lost in the roast, so a roaster wanting one pound of roasted coffee charges 1.16 pounds of green. That is wrong, and it is wrong in the direction that leaves you short. Losing sixteen percent means keeping 0.84, and one pound out of a process that keeps 0.84 needs 1 divided by 0.84 going in, which is 1.190 pounds. The error is 0.030 pounds per roasted pound — on 110.88 pounds of roasted coffee, three and a third pounds of green that was never charged.
At eight percent the same error is 0.007 pounds and nobody would notice. At twenty-five percent it is 0.083. It grows with the square of the loss, roughly, which is why it stays invisible until the day somebody takes a roast further than usual.
What the surviving pound is carrying
Green at 4.40 with 55 cents of freight is 4.95 a pound landed. A 132 pound lot is 653.40 of green. Six batches at 14 dollars is 84 dollars of gas, power and labour — six rather than five and a half, because a roaster does not run a half cycle. Total 737.40.
Divide by the green and it is 5.59 a pound. Divide by what came off the cooler and it is 6.65. Divide by what reached a bag and it is 6.89. That last figure is the only one of the three you can put a retail price against, and it is 1.30 a pound higher than the number on the green invoice. Nothing was wasted to make that happen. It is the same money spread over less coffee.
Where this page is guessing and where it is not
The arithmetic is exact. The inputs are not, and the roast loss is the one that decides the answer. It moves with the moisture the green came in at — a fresh crop and the same coffee a year later do not behave the same — and it moves with how far you take the roast, which is a decision rather than a property. There is no published figure that will do. Weigh a few batches.
The sorting percentage is easier to get: sort a batch, weigh what you pulled out, divide by the batch weight. Most roasteries have never done it and are surprised by the answer in one direction or the other.
Questions people ask
How much weight does coffee lose when roasted?
Yours will not match anyone else. It depends on the moisture the green arrived with and on how far the roast is taken, and both of those change from lot to lot and from profile to profile. The page asks rather than tells for that reason. Weigh a charge of green and weigh the same batch cooled, on the same scale, for a handful of batches, and use your own average.
Can I just add the roast loss back on to get the green I need?
No, and the error runs against you. A loss of sixteen percent means keeping 0.84 of what went in, so a pound out needs 1 divided by 0.84 going in, which is 1.190 pounds, not 1.16. The difference is 0.030 pounds per roasted pound. On a hundred pounds of roasted coffee that is three pounds of green you thought you had and did not.
Do the roast loss, the sorting loss and the packing loss add up?
They do not, because they happen one after another and each takes its share of whatever is left rather than of the original. Sixteen, two and one and a half add to 19.5 percent but compound to 18.91. The page multiplies the survival fractions instead, which is the arithmetic that matches the scale at the end.
What does a pound of roasted coffee actually cost me?
More than the green invoice, and the gap is bigger than most people expect. With the defaults here, 4.95 a pound landed becomes 6.89 a sellable pound once the losses and the batch running cost are spread over what survived. That is the figure a retail price has to cover, along with everything the page deliberately leaves out: packaging, labels, labour past the roaster, and the rest of the business.
Should staling loss be in this chain?
Not as a percentage of a lot, which is why it is not here. Coffee going past what you are willing to sell is a real cost, but it is driven by how fast the shelf turns rather than by the lot size, so folding it in as a fixed percentage would give a number that looks precise and is not. Track it against your own returns instead.