Roast consistency, and what moves it
Updated
On this page7
The green coffee changes
Moisture content, density and bean size all vary by lot and by how long it has been stored, and all three change how a bean takes heat. The same profile on a new lot gives a different result, which is why a lot change is a moment to taste carefully rather than assume.
Batch size changes everything
A half-full drum and a full one behave differently, and a roaster run below its designed capacity often roasts badly rather than simply roasting less. Standardizing batch size is one of the cheapest consistency improvements available.
Ambient conditions matter more than they should
A cold morning, a warm afternoon and a machine that has already run five batches are three different starting points. Back-to-back roasting heats the machine itself, so the fifth batch of the day is not the first batch of the day even with identical settings.
Which is why you log the profile, not the settings
The settings are what you asked for; the profile is what happened. Recording the actual temperature curve is the only way to tell a batch that drifted from one that was run differently, and it is what makes a good roast repeatable by somebody else.
Taste against a retained sample
Keep a small amount from each roast, labeled, and taste a new batch alongside the last good one rather than against memory. Slow drift is invisible batch to batch and obvious against something from a month ago.
Decide what good looks like, in writing
A short description of the correct cup, agreed while a batch is right, is what lets somebody other than you judge the next one. Without it quality control is one person’s palate and it leaves when they do.