Roast plan calculator: weekly orders to batches and green needed

Short answer. Work from roasted kilos back to green, never the other way round. Green a week = roasted needed ÷ (1 − roast loss) × (1 + buffer). Divide by your batch size for batches, by your roast days for batches a day, and multiply by cycle time for hours at the machine. 220 kg roasted a week at 15% loss with a 10% buffer is 285 kg of green: 24 batches of 12 kg, twelve on each of two roast days, a little over three hours of roasting a day. Thirteen weeks of that is 3,700 kg of green, or 62 bags of 60 kg.

The calculator

Enter the roasted kilos your cafés order in a normal week, not your best week. The defaults are the worked example further down: 220 kg roasted a week, 15% roast loss, 12 kg batches, 16 minutes a batch, two roast days of eight hours, and a 10% buffer.

Batch size is the green weight you load, not what comes out. On a 15 kg roaster, 12 kg of green is a full batch by most roasters' standards: the rated capacity is a maximum for dense coffee, and most roasteries run somewhere between 70 and 85% of it. Cycle time is the whole batch: roast, cooling, dropping and charging the next one, so 15 to 20 minutes for a drum machine, not the 11 minutes on the profile.

The arithmetic, written out

Every line of the calculator is one multiplication or one division. You can keep the whole thing in a spreadsheet, and a lot of roasteries do.

• Roasted a week, with buffer = roasted needed × (1 + buffer)
• Green a week = roasted needed ÷ (1 − roast loss) × (1 + buffer)
• Batches a week = green a week ÷ batch size, rounded up
• Batches a roast day = batches a week ÷ roast days, rounded up
• Roasting hours a roast day = batches a roast day × cycle minutes ÷ 60
• Roasted off a roast day = batches a roast day × batch size × (1 − roast loss)
• Green for n weeks = green a week × n, and ÷ 60 for bags

Two things in that list are worth saying out loud. You divide by one minus the loss, you do not multiply by the loss: 220 kg at 15% loss needs 258.8 kg of green, not 253. And batches round up, because half a batch is not a thing you can roast. Twenty-three and a bit batches is twenty-four batches, and the extra three kilos of green is part of why the buffer exists.

Roasted coffee needed a week

The kilos your cafés will actually take, from confirmed and standing orders, not the total your accounting software reports for last month. More on that below.

Roast loss

Also called shrinkage: the weight the coffee loses as water leaves it. Twelve to eighteen percent covers most specialty roasting, with lighter filter roasts at the bottom of that range and darker espresso at the top. Use your own figure, measured over a few weeks of real batches, and use a different one per coffee if your profiles differ that much.

Batch size

Green in. Rated drum capacity is a ceiling set for dense coffee, so 12 kg on a 15 kg drum and 25 kg on a 30 kg drum are ordinary working batches. Big naturals and low-density high-grown lots take up more room per kilo, and some roasters drop the charge weight for them.

Cycle time

Drop to drop. Roast plus cooling plus emptying the cooling tray plus getting the drum back to charge temperature. If you have never timed it, time a full roast day and divide, rather than guessing from the profile length.

Roast days and hours available

Which days you roast and how long the person at the machine can actually stand there. Eight hours in the field does not mean eight hours of roasting; see the worked example.

Buffer

Samples for prospects, retail bags, the market stall, batches that go wrong, and the café that phones after the cutoff. Ten percent is a reasonable start. If you are running a guest coffee programme or sending a lot of samples, it is higher.

A worked example: 220 kg a week on two roast days

A roastery with about twenty wholesale accounts takes 220 kg of confirmed roasted coffee a week. It roasts on a 15 kg machine in 12 kg batches, 16 minutes a batch, on Mondays and Wednesdays, and adds 10% for samples, retail and waste.

Three hours twelve minutes is the roasting. The roast day is not three hours twelve minutes. Add thirty to forty-five minutes of warm-up before the first charge, the time to weigh out twenty-four batches of green, the profile changes between coffees, emptying the chaff collector, packing, labelling and cleaning down. On this plan a roast day is a comfortable morning, which is the point: the plan should fit inside the day with room for the thing that goes wrong.

The order of the batches matters as much as the count. Group batches of the same coffee together so you change profile as few times as possible, and arrange the day so the charge temperature climbs rather than swings, because a drum heats faster than it cools. Twenty-four batches across four coffees is a calm day. Twenty-four batches across eleven coffees is not, even though the calculator says the same thing.

To turn the same kilos into money, the wholesale coffee price calculator takes green price, roast loss, packaging and labour to a price per kilo, and how many wholesale accounts you need works backwards from a gross profit target to the number of cafés and the kilos they imply.

Plan from confirmed orders, not last month's total

Last month's total is an average of accounts that are each moving in their own direction, and averages hide exactly the thing a roast plan needs to know. A book that did 950 kg last month might be one café up 40 kg and another down 40 kg, or it might be one café that has quietly stopped ordering while a new one started. Roast against the average and you will be short on the week the new café takes its full order and long on the week the quiet one skips.

Plan from three layers, in this order:

1. Confirmed orders for this delivery round. These are certain. They are the plan.
2. Standing orders that have not been skipped. These are nearly certain, and they are what you can also buy green against.
3. The buffer, for everything else: samples, retail, waste, and the account that orders after the cutoff.

This is the practical reason to have an order cutoff. A cutoff is not administrative tidiness; it is the moment the roast plan stops being a forecast and becomes a list. Without one, every roast day starts with a guess and ends with either a shortfall or a shelf of coffee nobody ordered.

Watch the second layer for decay. A standing order that is never skipped but is quietly getting smaller, or an account that has slipped from ordering every week to every nine days, changes your baseline before it changes your total. That is the whole argument for measuring each account against its own rhythm rather than watching the monthly figure.

A roast plan and a roast schedule are not the same thing

The two words get used interchangeably and they answer different questions.

The roast plan

What has to exist, and by when. So many kilos of each coffee, expressed in whole batches, needed for Tuesday's delivery and Thursday's. It comes out of the order book at the cutoff, and it is checked against two things: the green you have, and the delivery days it has to feed.

The roast schedule

The order those batches go through the machine, on which roaster, with whom at the controls. It is made on the roast morning from the plan, and it is checked against the clock and the profile changes. A run sheet is the schedule on paper, with a line per batch to tick and a column for the actual drop weight.

Keeping them separate is what lets you answer a café's question without re-planning the week. "Can you add 6 kg for Thursday?" is a schedule question if the green is already there and the day has room: half a batch, already inside the buffer. It is a plan question if it means an extra batch of a coffee you have two kilos of green left of. The person answering the phone should be able to tell which, which usually means the plan needs to be written down somewhere other than in the roaster's head.

Record the actual drop weight per batch on the run sheet. Over a few weeks that gives you your real roast loss per coffee, which is the input this calculator is most sensitive to and the one most roasteries are guessing at.

What breaks the plan, and how much slack to keep

A roast plan that only works when nothing changes is not a plan. Four things break it regularly:

• One big account moves its order day. An account taking 30 kg that shifts from the Tuesday round to the Thursday round moves 12% of the worked example's week between two roast days: about three batches out of one day and into the other. It does not change the weekly total at all, which is why it is invisible on a monthly report and painful on a Wednesday.
• A guest coffee. A small lot, an unfamiliar profile, an extra changeover, and usually a few sample batches before it is right. Budget it in the buffer and in the day, not just in the green. The guest coffee guide has the green arithmetic for a rotation.
• A machine service. An annual service, a burner fault or a motor takes out a whole roast day. On two roast days a week that is half your capacity for that week, and the coffee still has to go out.
• A sick roaster. On most small roasteries one person roasts. If nobody else can run the machine, one phone call at seven in the morning is the same as a machine failure.

The slack that absorbs all four is not one number, it is two. Keep about 10% of the week's kilos in the buffer, which is what the calculator's buffer input is for, and keep about a third of each roast day unbooked. In the worked example that means planning for twelve batches on a day that could take eighteen or twenty. Then a moved order day is absorbed by the day that has room, and a lost roast day can be caught up on the next one rather than turning into an email to twenty cafés.

The late bulk order is the one to think about before it happens. One roaster in Fresh Cup's August 2026 survey of thirty roasters asked cafés to understand that "if you call in a bulk order and want it the same day, that may be difficult on both sides". Deciding in advance what you will say to that call, and what it costs you, is easier than deciding at half past four.

How much green to hold, and what it costs to hold it

Green coffee is the largest thing a small roastery owns and the slowest to turn into money. It is bought in contracts agreed months before shipment, paid for long before a café orders a bag of it; one roaster told Fresh Cup the money is spent "six months to a year before you ever get your order". Every week of cover on the floor is cash that is not in the bank and cannot be spent on anything else.

Here is the worked example's 284.7 kg a week, at four levels of cover, with the cash at €9.00 a kilo of green:

Green per week = 220 × 1.10 ÷ 0.85 = 284.7 kg. Bags rounded up. €9.00 a kilo is an illustration, not a market price; use your own landed cost.

Six months of cover for this roastery is about €67,000 on the floor. That is why many roasteries have shortened their green contracts, from six months of cover to three or four. Shorter cover is not obviously better: it means buying more often at whatever the market is doing, and it means the lot you liked may be gone. What it does is put less money at risk against a book that can change, which is the fear roasters state plainly — that a café backs out or stops buying and you are left holding green you cannot sell.

Two more things push the same way. Green does age: a lot bought a year ahead of use will be tasting like past crop by the end of the contract, so long cover on a delicate washed coffee costs you in the cup as well as in cash. And the coffee you have bought is already a bet on your book staying the shape it is. If one account is 30% of your wholesale, thirteen weeks of green bought against it is thirteen weeks of concentration risk; the page on account concentration works through what a healthy spread looks like.

Roasted stock is the opposite problem and should not be used to solve this one. Espresso wants a few days of rest and is usually at its best within three or four weeks, so a shelf of roasted coffee ages while you look at it. Hold green, roast to the plan.

How a standing order changes the maths

A standing order is the part of the week you can buy green against, and it is the single biggest improvement you can make to this calculation. It converts a forecast into a baseline.

Take the worked example again and say 150 of the 220 kg comes from standing orders that cafés set themselves and can skip or pause. That is 68% of the week known in advance. The green behind it is 150 × 1.10 ÷ 0.85 = 194.1 kg a week, or 2,524 kg over thirteen weeks: 43 bags you can put on a three-month contract with a straight face. The remaining 70 kg a week is bought shorter, in smaller lots, or from an importer's spot stock where you pay a little more for the flexibility.

A standing order that the café can skip is still a far better forecast than a guess, because a skip is visible before the cutoff rather than as a hole in the week. The roaster's instinct is to ask for a volume commitment instead. A commitment written into terms is one route; a standing order the café controls, with a cutoff, usually gets you the same predictability with none of the resentment, and cafés agree to it because it saves them the weekly reorder.

Order Book does not place a standing order by itself: the café orders from its account each time, with its usual order at the top of the sheet, or you build a draft it confirms with one tap.

Split the plan the way you split the green: contract against the baseline, buy the top layer short. Then a café leaving costs you an account, not a container.

When to add a roast day rather than a longer one

A longer day is cheaper than another day, right up until it is not. Warm-up, weighing, cleaning down and the chaff collector are a fixed cost per roast day, perhaps an hour that produces no coffee, so adding a third day for two extra batches loses money and goodwill. Add hours first.

Add a day when one of these is true:

• The plan needs more than about six hours of actual roasting on one day. That is a nine-hour day once everything around it is counted, and the last batches of a long day are the ones that get away from a tired roaster.
• One delivery day carries more than about half the week's kilos. Splitting the roasting to match the deliveries is what gets fresher coffee into the box, not more total capacity.
• A single fault would take out the week. Three roast days means a lost day costs a third of the week, not half.
• Freshness is a selling point you make. Roasting on Monday for a Thursday delivery is a different product from roasting on Monday for the following Tuesday.

The arithmetic is easy to check either way. In the worked example, 24 batches over two days is twelve a day and 3.2 hours at the machine; over three days it is eight a day and 2.1 hours, with every delivery closer to its roast. Over one day it is 24 batches and 6.4 hours of roasting, which is a full and inflexible day with nowhere to put a machine fault. Put your own numbers in the calculator and read the fit line: it is the only output that tells you plainly that the week does not work.

When you do add a day, add it to the calendar before you need it and tell the cafés on that round. A delivery day that moves twice in a month costs you more trust than a late order costs you money.

Questions roasters ask

How much green coffee do I need for 100 kg of roasted coffee?

About 118 kg at 15% roast loss, because green = roasted ÷ (1 − loss). At 12% loss it is 114 kg and at 18% it is 122 kg. Add a buffer for samples, retail bags, waste and the café that orders late: with 10% on top, 100 kg roasted needs about 129 kg of green.

How many batches an hour can a drum roaster do?

Three or four back to back once the machine is hot. A batch is roughly 10 to 14 minutes of roast plus cooling, dropping and charging the next one, so 15 to 20 minutes a cycle. Warm-up, weighing green, changing profile between coffees and cleaning the chaff collector are not in that number, which is why a roast day is always longer than the roasting.

What batch size should I use on a 15 kg roaster?

Between about 10.5 and 12.75 kg of green, which is 70 to 85% of the drum's rated capacity. The rating is a maximum for dense coffee, not a target. Run much below half and the drum holds too much heat for the amount of coffee in it, which makes the roast hard to control and hard to repeat.

How far ahead should a roastery buy green coffee?

Buy the volume you can see, for as long as you can see it. Many roasteries have shortened their green contracts from six months of cover to three or four, because the cash and the risk of a café leaving both sit with the roastery. Contract against confirmed and standing orders, and buy the rest in shorter lots or from an importer's spot stock.

Should a roastery roast to order or roast to stock?

Roast wholesale to order, against the roast plan the cutoff produces, and hold stock only of the house blend and only a few days of it. Roasted coffee ages in a way green does not: espresso wants a few days of rest and is usually at its best within three or four weeks. Stock that ages on your own shelf is money spent twice.

When should a roastery add a roast day instead of a longer one?

When the plan needs more than about six hours of actual roasting on one day, when one delivery day carries more than half the week's kilos, or when a single machine fault would take out the whole week. A roast day has a fixed cost in warm-up and cleaning, so do not add one for two extra batches.

Where Order Book fits

Order Book does not buy or count your green; this page does the arithmetic and your importer holds the stock. What it does is produce the left-hand side of it. Confirmed orders close at the cutoff and become the roast plan for that round, grouped by coffee and expressed in the kilos you have to roast, then the run sheet for the roast day, the picking list for the boxes and a delivery note per café. Each coffee can carry its green price and roast loss, so insights show the gross margin you are actually making per coffee and per account, and a change in what green costs you shows up where you can see it.

It counts the roasted stock on the shelf, but it keeps no green inventory, and it does not track green contracts or bags in the warehouse. If that is what you need, it is a different tool.

Start free for 30 days, no card, or see what it costs.