Roasting a new green coffee for the first time
A new green coffee is a small unknown. You don't know how fast it will take heat, where first crack will land or what it'll taste like at different roast levels. You can learn it in three or four batches if you go about it in an orderly way.
Start from what you know
Check the sample's details first: the origin, variety, process and altitude. They tell you what to expect. As a very rough guide:
- Dense, high-grown coffees (often washed) need a bit more energy early on and tolerate longer development.
- Softer, lower-grown ones take heat quickly and can scorch.
- Naturals tend to be sweeter and a little less forgiving at the end.
Then use a profile you already trust on a similar coffee as your first batch. Don't invent something new for an unknown coffee.
Roast a test batch smaller than you'd normally
Use maybe 70% of your usual load. It's cheaper if it goes wrong, and it responds more quickly to changes. When you know the coffee, move up to your normal load.
Change one thing at a time
After the first batch, ask what you'd most like to be different. Too sour? Add time after first crack. Too flat? Shorten the roast or raise the heat a little. Choose one lever and move it.
If you change the charge temperature, the heat and the airflow in the same batch, you'll never know which one worked. It feels slow and it's the quickest way to learn.
Log it as you go
Use the habits in the roast log guide. For a new coffee it's worth noting the time to each landmark and the development time especially, since it's the number that tells you how far you pushed the finish.
Cup the batches together
Wait a few days, then cup all of the test batches together, blind if you can. See cupping notes that teach you something. Choose the best one and write its numbers at the top of your notes for that coffee. That's your starting point next time.
When to stop
Three or four batches is usually enough to land somewhere good. After that you're chasing small gains. Lock in a recipe, roast it a few times, and see how the numbers hold. If a batch comes out different, you'll know straight away that something changed, because you have something to compare it to.