How We Track Roast Loss by Gram During Every Batch (the Number That Tells Us Whether a Roast Went Right)

If you've ever wondered what small-batch roasters actually track during a roast, the answer is mostly weight. Every batch starts as a measured weight of green coffee and ends as a measured weight of roasted coffee. The difference between the two — what we call "roast loss" — is one of the most useful numbers we have for knowing whether a roast went the way we wanted.

Roast loss is the difference between the weight of the green coffee going into the roaster and the weight of the roasted coffee coming out. It's expressed either as a percentage of the green weight or as an absolute mass in grams. For a 1 kg batch that comes out at 850 grams, the roast loss is 150 grams (15%).

This is the inside view of what roast loss actually measures, why it tells us whether a roast went right, the targets we use for different coffees, and what to do when the number doesn't match the target.

What roast loss measures

Roast loss covers two things happening to the bean during roasting:

• Moisture loss. Green coffee is 10-12% water by weight. The drying phase of the roast drives that moisture out. Most of the moisture loss happens in the first 4-5 minutes of the roast.
• Organic loss. Various organic compounds leave the bean during roasting. The Maillard and pyrolysis reactions produce CO2, water vapor (additional to the moisture), and various volatile organic compounds. These compounds leave the bean and end up in the roaster exhaust.

The total mass loss is roughly 12-22%, depending on:

• Roast level. Light roasts lose less mass than dark roasts. Lighter roasts end at 11-15% loss. Dark roasts end at 15-20% loss.
• Origin characteristics. Some origins lose more mass than others. Aged coffees tend to lose more than fresh coffees. Higher-grown coffees sometimes lose slightly more.
• Roasting speed. Slower roasts tend to lose slightly more mass than faster roasts, because the bean spends more time in the development phase.

For a typical light to medium roast of fresh crop coffee, the loss is 13-16%. For a medium-dark to dark roast, the loss is 15-20%.

Why we track it batch-by-batch

The reason roast loss is worth tracking is that it tells us whether the roast went the way we wanted it to. Specifically:

• Roast profile consistency. If the same coffee from the same origin consistently loses 14.5% in our setup, we have a baseline. If a batch comes out at 13.8% or 15.2%, something was different about that roast.
• Charge weight vs charge temperature variance. A lighter green coffee (lower moisture, possibly older crop) loses less mass because there's less water to lose. A higher-moisture green coffee loses more.
• Drop timing variance. A batch that hits first crack later (longer drying phase) tends to lose more mass than a batch that cracks earlier. We aim for consistent first crack timing across batches.

Tracking roast loss helps us answer specific questions:

• Did this batch roast the same as the previous batch? Yes, if the loss is within 0.3% of the previous batch.
• Is the green coffee behaving differently? Maybe, if the loss is consistently off.
• Did the roaster running hotter this morning? Possibly, if the loss is higher than normal at the same roast level.

The number is one diagnostic among several. It's the most useful single number we track.

How we measure it

Our process:

• Pre-roast. We weigh the green coffee on a 0.1-gram scale before charging. We record the charge weight.
• Post-roast. After cooling, we weigh the roasted coffee on the same scale. We record the roasted weight.
• Calculation. Roast loss = (green weight - roasted weight) / green weight × 100.
• Recording. Each batch gets the numbers written on a log sheet — charge weight, roasted weight, drop time, drop temperature, plus a brief note on the cup. We keep the logs in a binder and a digital copy.
• Check. If the loss falls outside the expected range for that coffee and roast level, we note it and adjust the next batch.

The measurement takes 30 seconds at the end of each batch. Over the course of a year, the log data lets us answer questions about specific coffees, specific times of year, and the trajectory of our roasting.

Targets for different coffees

The targets we use for different coffees:

• Light roast Ethiopian, Kenyan, Panamanian. 13-15% loss. Lighter roasts for delicate origins end at the lower end of the loss range.
• Medium roast Colombian, Guatemalan, Costa Rican. 14-16% loss. Medium roasts for balanced origins end in the middle of the range.
• Medium-dark roast Brazilian, low-grown Sumatran. 16-18% loss. Heavier roasts for body-heavy origins end at the higher end.
• Dark roast (espresso blends, dark filter). 18-20% loss. The heaviest roasts end above 18%.

These ranges are starting points. The actual target for a specific coffee gets refined as we cup and dial in.

What to do when the number is off

Common causes of off-target roast loss and what to do:

• Loss is higher than expected. Causes: longer drop time (did the roast run longer than intended?), higher moisture green coffee (maybe a newer crop lot), cooler charge temperature (maybe the roaster was cooler at the start), or different bean density (maybe a different origin lot).
• Loss is lower than expected. Causes: shorter drop time (did the roast end earlier than intended?), lower moisture green coffee (older crop or stale), higher charge temperature (roaster was hotter at the start), or higher density beans (different origin lot).
• Loss is consistent but cup is off. The roast is hitting the right window for weight, but the cup doesn't match. Possible cause: inconsistent roast development (the bean profile was uneven), different green coffee chemistry (the lot changed), or the drop temperature was off.

For any of these, the fix is usually one of three things:

• Adjust charge weight or temperature. If the roaster running temperature is different, adjust the charge temperature to compensate.
• Adjust drop time. If the loss is consistently off, the drop time or drop temperature is probably wrong.
• Cup and adjust. If the loss is right but the cup is off, the development phase probably needs tweaking.

The fix usually shows up in the next batch.

What it doesn't tell you

Roast loss doesn't tell you the cup quality. A batch can be at the target loss and taste flat, or off-target loss and taste great. The cup is the truth. The numbers are diagnostics.

It also doesn't tell you the development time relative to the total roast time. A 14% loss batch with 80% development time was probably developed longer than one with 70% development time — which affects the cup even if the loss is similar.

The most useful additional data points to track alongside roast loss:

• Drop temperature. What's the bean temperature at the drop? Higher drop temperature = more development, more body, more roast character.
• Drop time. How long did the batch take? Longer batches tend to lose more mass per minute in the development phase.
• First crack timing. When did the bean hit first crack? Earlier = lighter roast. Later = darker roast.

These three numbers plus roast loss give a reasonable picture of what happened during the roast.

Our process today

What we actually do, day to day:

• Every batch gets logged. Charge weight, roasted weight, drop time, drop temperature, cup note. In a binder and a spreadsheet.
• Roast loss is checked first. Before the cup note, we check that the loss is in the expected range. If it's off, we cup more carefully and adjust the next batch.
• Trends matter more than individual batches. We look at the loss data over weeks and months, not batch to batch. Seasonal variation and crop change show up in trends.

The number isn't the answer. It's one of the most useful questions. Tracking it consistently gives us a feedback loop that catches drift before customers notice it.