What I Learned From Cupping Every Coffee We Considered Stocking This Year
I've cupped more coffee this year than I can count. Dozens of samples from importers, samples from friends in the industry, samples we roasted and tested ourselves before deciding to add them to the catalog. Each cupping taught me something. Some of the lessons are obvious in hindsight. Some of them I'm still working through.
This is what the year of cupping taught me about choosing coffee — six lessons that will shape what we stock going forward.
Lesson 1: The cup score is less predictive than I thought
I mentioned this in the cup score post, but it's worth repeating from a sourcing perspective. A high cup score tells me a coffee is clean, balanced, and defect-free. It does not tell me whether I'll like it.
I've passed on 87-point coffees that customers loved and stocked 86-point coffees that customers asked about for months. The difference wasn't quality — both were good. The difference was flavor character, which the score doesn't capture.
Going forward, I'll still look at the score as a floor. But I won't use it as the deciding factor. The cup in front of me matters more than the number on the sheet.
Lesson 2: Process matters more than origin for flavor personality
Two Ethiopian Yirgacheffes from the same washing station can taste completely different if one is washed and the other is natural. Two Costa Rican Tarrazús from neighboring farms can taste almost identical if both are washed.
Process is the bigger lever for flavor character. Origin provides the base. Processing determines how loud the base gets.
This changed how I think about diversifying the catalog. Adding more origins gives us coverage. Adding more processes gives us variety.
Lesson 3: The "wow" coffees don't always sell
The most exciting coffee I cupped this year was a natural-process Yemeni Mocca with strawberry-jam aromatics and a long, sweet finish. It scored 88.5. I loved it. I ordered it.
It didn't sell well. Customers found it unusual, in a way they didn't love. The same customers who would have rated a washed Ethiopian 9/10 rated this one 6/10. Not because it was bad — because it was different in a way they weren't expecting.
The lesson: my excitement about a coffee doesn't translate to customer excitement. The cup has to match what customers want, not what excites me as a buyer. I've gotten better at predicting which "wow" coffees will translate and which won't.
Lesson 4: Consistency matters more than peak quality
The best cupping score I've recorded this year was a 91-point microlot from a specific farm in Bolivia. It was extraordinary. We didn't stock it.
Why? Because it was a one-time lot. We'd have 30 pounds. After that, we'd never get the same coffee again. Customers who tried it would have loved it. Customers who missed it would have asked why we don't have it.
Consistency — being able to offer the same coffee year after year, or at least season after season — is more valuable to our customers than a one-time peak experience. We prioritize coffees with reliable supply over exceptional one-off lots.
Lesson 5: The 80/20 rule applies to green coffee sourcing
About 20 percent of the coffees I've cupped this year are ones I'd be excited to stock. The other 80 percent are good but unremarkable — clean, balanced, correct, but not exciting.
The 20 percent is hard to source. Every roaster is looking for the same coffees. The competition drives prices up and makes it harder to build long-term relationships with producers.
The 80 percent is easy to source and easy to roast. It's where most "specialty" coffee lives.
The challenge: find the 20 percent, build a relationship with the producer, lock in supply, and offer it at a price that's fair but sustainable. This is the work of running a roastery.
Lesson 6: Cupping fatigue is real
After cupping 8 to 10 coffees in a session, my palate loses sensitivity. The 11th coffee tastes more like the 10th. The 15th coffee tastes like every other coffee. By the 20th, I can't tell the difference between a 84 and an 88 — they're all just "coffee."
The professional fix: cup no more than 6 to 8 coffees per session, take breaks, eat unsalted crackers or drink water between cups, and don't cup more than two sessions per day.
The reality: I try to follow this and often fail. When importers send me 12 samples to evaluate, I cup all 12. When I'm trying to lock in next quarter's lineup, I cup 20 in a session. The cupping scores from late in the session are less reliable than the scores from early.
Going forward, I'm going to be more disciplined. The fatigue effect means I've made sourcing decisions on cuppings that weren't as accurate as I thought they were.
What this changes about how we stock coffee
Three concrete changes:
- More washed coffees, fewer naturals. Most customers prefer the cleaner profile. We still want some naturals in the lineup for variety, but the bulk of what we offer will be washed.
- Fewer one-off microlots. We'll keep some rotating single-origin lots for variety, but the core lineup will be coffees we can re-order.
- More communication about why a coffee is on the shelf. When a coffee is unusual, customers need to know before they buy. When a coffee is reliable, they need to know that too. The catalog page should explain.
We've written about how a new coffee earns a spot and how we decide which coffees make our lineup every quarter. The cupping lessons are how those decisions get made.
What I'd recommend to anyone starting a roastery
If you're starting out and wondering how to source coffee:
- Cup everything. You can't make good decisions without tasting. Even bad cuppings teach you something.
- Cup with other people. Comparing notes is the fastest way to calibrate your palate.
- Keep records. What did this coffee taste like? Why did you pass on it? Six months later, your notes will teach you things you didn't see at the time.
- Be honest about your biases. I like bright, clean coffees. My catalog reflects that. If you want a balanced catalog, you need to actively source for the things you don't naturally gravitate toward.
A year of cupping is enough to teach you the basics. Three years is enough to teach you what you don't know yet. Five years is enough to start trusting your palate.
I'm at year two. I'm still learning. The coffee I'm most excited about next year is one I haven't cupped yet.
Brew slow. Sip slow.
— The Cozy Sips team
P.S. We cup every batch twice before it ships — once on the day we roast, once a week later after the coffee has rested. The second cupping is what tells us whether the batch is ready to ship. We've written about why we cup every coffee twice — the operational side of the same practice.