Five Big Mistakes to Avoid in Sample Roasting

A few days ago, I recorded a video about sample roasting. In the video, I shared our “profile” for sample roasting here at Mill City, and I discussed the common mistakes that keep people from sample roasting successfully. This is a topic we get a lot of questions about, so I wanted to reiterate these points here on the blog because we think it’s pretty important. Anything worth mentioning is worth repeating, right? Right.

What is a sample roast?

To start, let’s just clarify what a sample roast is. A sample roast is a short and light roast profile which we apply to every new coffee we’re considering for purchase. It’s a controlled standard specifically designed to produce a roast that’s underdeveloped to highlight any possible defect in the green coffee. We don’t alter anything about the sample roast based on a coffee’s density, moisture content, age, varietal, or processing method. It’s crucial that this roast be the same for every coffee that enters our roastery. 

The SCA has guidelines for sample roasts, which we follow. Their guidelines state that a sample roast should be completed in 8-12 minutes and be cooled immediately. The coffee should be roasted to a 63 on the Agtron scale (the roast level should be measured from ground coffee, not whole bean). The roast should not have any scorching or tipping, and the sample should rest for 8-24 hours before being cupped. 

Our sample protocol

We sample roast on our 500g sample roaster, which is connected to natural gas and has a maximum available gas pressure of 3.8kPa. This is a lot of power for a small roaster. When sampling, we start with a 200g charge weight so we need to reduce our fuel input considerably in order to reach our target roast landmarks.

If you’re roasting a lot of samples, which we occasionally are, and you want to minimize variance in the roast, it’s best to find a set of roaster inputs that can remain static. This means that you won’t have to make any adjustments during the roasting process, so it’s less likely you’ll miss a fuel or an air change. We worked to find the settings which would allow us to reach our target roast color in the recommended amount of time. Here are those settings:

• Charge weight: 200g
• Charge temperature: 360° Fahrenheit
• Fuel setting: 0.8kPa (21% of our maximum gas pressure)
• Air setting: Medium (that’s a 55 on our dial)
• Drum speed: 60RPM (an 80 on our drum speed knob)

This helps us achieve a sample roast profile that looks like this:

• Dry End: 3:30-4:00
• First Crack: 6:30-7:00
• Drop: 8:30-9:00

We cool the coffee quickly (our standard for cooling time on all of our roasters is under 5 minutes) and we bag it in a kraft tin-tie bag or leave it in a coffee tray to off-gas until cupping it the next day. We’re diligent about a 24 hour rest time for roasted samples; if we miss the 24 hour window the next day we’ll re-roast the sample. 

What people get wrong

There are a few things that people sometimes miss when it comes to sample roasting. These aren’t nitpicky problems; these are issues that are seriously affecting people’s ability to sample roast or evaluate greens. These mistakes will keep you from purchasing the right coffee for them or recognizing a high quality sample when it pops up on their cupping table. 

MISTAKE #1: Believing that importers inflate their cupping scores

There are a lot of rumors about cupping scores and marketing. The idea that an importer would artificially inflate the score of a green coffee to sell more bags is silly. Cupping is, by definition, an objective process and anyone who’s calibrated with other cuppers should be able to tell the difference between a run-of-the-mill offering and something truly special. Reputable importers carry a wide range of greens from everyday sippers to once-in-a-lifetime coffees, and they don’t need to tell you that an 82 coffee is an 85 because there are plenty of people looking to buy that 82. 

The reason this rumor survives is because roasters either don’t sample roast well or don’t cup coffee well (or both). They get a great sample in their hands and aren’t able to maintain the quality all the way to the cup. Then, they decry score inflation instead of looking inward and thinking that the problem may be with them. Don’t be like them. Scores are a small part of what you should consider when you’re buying a new green coffee. If the score matters to you, take it at face value and move forward in confidence. 

MISTAKE #2: Using the wrong sample size for your machine

Can you roast a 300g sample in a 12kg roaster? Maybe. Can you do it within the standard (8-12 minutes to a 63 on the Agron scale)? Not likely. Will that tiny amount of coffee roasted in a massive drum turn out scorched, scalded, and acrid? Probably. Coffee that’s not roasted to within the standard for sample roasting can’t be used for a sample evaluation, the flavor notes simply won’t be helpful in making a purchasing decision. 

Maybe your roaster can effectively roast a 600g or 1.2kg sample to the aforementioned landmarks. Great news! Unfortunately, most importers will only send you 450g of green or less as a sample.

As a side note, some people can roast really small batches on really big roasters. How do they do it? We suspect witchcraft. 

MISTAKE #3: Using the wrong roaster to sample roast

If you don’t have a dedicated sample roaster or a production roaster that’s small enough to sample roast, it will be very difficult to successfully roast samples. Even if you can find the magical window of settings that helps you hit the targets on a sample roast one time, it’s not likely you’ll be able to do it 1,000 times. The most important thing when it comes to sample roasting is consistency. 

MISTAKE #4: Not being able to taste the difference

You have to be able to taste coffee well or none of this other stuff matters. Every single time we teach our roasting classes, we ask people what they want to learn more about. Without fail, one of the most common responses is developing better sensory skills. There’s no secret to becoming a better coffee taster. You hone your palate by tasting a lot of coffee. Preferably, taste it with other people so you can calibrate on flavor notes. Ideally, make sure one or two of those people has some experience so you can learn the terminology and vocabulary of specialty coffee. 

All of the most successful sample roasts in the world won’t matter if you can’t tell the difference between citric acidity and ferment. Learn to distinguish between sweetness, levels and types of acidity, and identify balance in coffee. Develop a lexicon of flavor descriptors that helps you communicate with both your importer and your customers. Until you do this, you’re better off letting someone else decide what green coffee to buy.

MISTAKE #5: Having a production roaster that isn’t up to the task

If your sample roaster is great, you might be able to profile small batches of coffee on it. If you get to the point where those profiles really knock your socks off, you should be able to record the phase events and then translate them to your production machine. The times and temperatures won’t be identical, but the percentage of time spent in each phase should be similar, and the production profile should look like a stretched out version of what you did on the smaller roaster. 

If your production roaster, however, lacks the thermal stability or sophisticated control system required to roast high-grade specialty coffee, it’s a lot of work for nothing. Part of running a prosperous company is the ability to scale, but if your coffee quality declines at higher volumes your customers will notice. You need a production roaster that can hit targets as exactly and repeatedly as your sample roaster.

The payoff

Sample roasting, cupping, production roasting, brewing, and selling great coffee are skills that can be practiced and continuously improved. Looking for small changes you can make to increase your consistency, efficiency, and quality is worth the effort. Good coffee is the reward.