After that, hot water is poured over the grounds in the cup and allowed to brew a few minutes. This leaves a crust of coffee grounds floating at the top of the cup which I break with a spoon while moving my nose close to the surface to evaluate the aroma as I break through that crust. Here I found cups 3-7 most interesting. The first cup has some unpleasant attributes and the second cup was just uninteresting. Darker cups weren't bad, but not characteristic of customer expectations.

Then it's time to taste the coffees, lightest to darkest. My favorite on the first sip was cup 2 with some very nice complex fruitiness, but it's not enough for a coffee to only be great on the first sip.

The 2nd tasting goes in the opposite direction: darkest to lightest. Here I started to pick up on the characteristics I'm looking for at cup 7, but 6 was better than 7 and 5 was better than 6. This let me narrow the range of coffees I'm considering to the 2-5 range, all of which for this flight is something that I'd consider a light roast (cup number to darkness isn't fixed between flights. I pull more heavily in whichever regions I most expect to have what I want to get out of that coffee).

From there, I focus on tasting within that range until the coffee is cool and consider whether I want to pick something directly to attempt to replicate on the big roaster or if I think changing how much time I'm spending in key temperature ranges (roast faster or slower) might improve on what I'm tasting. Sometimes I need to go back to the little roaster to try something else, but I'm usually pretty confident that I can make an adjustment if needed and move directly into production testing.

I've been doing this with every coffee I roast for over a quarter century now so I've gotten decently good at reasoning about what I can do to achieve the cup characteristics I want.