Tuesday, April 9, 2013

Some thoughts on degrees of roast

Today I received this comment from a reader on one of my earlier posts:

I'm enjoying your post Kevin. I don't mean to move the topic away from drip brew but since you were talking about roasting profiles I thought I would chime in.

I started out at Starbucks and then moved into the "third wave" coffee movement as a roaster. I have found it extremely challenging to grow my knowledge when most of the "respected" roasters are roasting just into first crack and I still insist on taking the roast to the brink of, (or right into), second crack. That means to the rest of the coffee world I roast "dark" and that this style of roasting is completely disregarded as a way of presenting coffee. Just wanted to say thanks for providing some wisdom and giving roasters like me a different influence to turn to.

Thank you!

I appreciate Landon's comments and it prompts me to share a few thoughts. 

One of the basic tasting exercises I've found useful over the past 30 years is what I call tasting a progressive roast of almost every coffee you have. This means roasting on either a sample or production roaster and pulling out samples of a coffee at every degree of roast from City to French. In a production roasting context (I'm assuming a drum roaster with a tryer) you can pull several tryer's worth of City, Full City, Vienna, Italian and French - enough not just to cup, but to brew, laying them out in the the cooling tray or on a cookie sheet. Obviously it's not a useful (or safe!) exercise with soft, low-grown coffees like Brazils or Hawaiians, but otherwise it's something very worthwhile to experience with as many coffee origins as possible. 

It should be fairly obvious that the degree of roast chosen for a coffee ought ideally to be correlated to the potential of the particular green coffee, where that coffee is in its life span, and the brewing method to be used, but equally important and less obvious is what the local water supply is like. Some specifics on all of these, in order:

1. Green coffee: hard, dense, high-altitude coffees of the highest quality are delicious at a wide range of roasts. Classic examples would include top Guatemalans from Huehuetenango or Antigua, Kenya auction lots, and on the natural side of things Yemen Mochas and excellent dry-processed Ethiopians. These coffees are way too acid to be palatable below a classic City roast, come into their own at Full City (chestnut brown, no second pop, no oils on the bean surface) and have a Port-like richness and depth without taking on a burnt character well into the Vienna roast range and all the way up to what in Starbucks or Peet's terms we'd call an espresso roast (Agtron in the 30's, still well shy of Italian or West Coast French). These kinds of coffees are becoming harder and harder to find, and what's happened at Starbucks over the years is the heavy roast lives on while the quality of the coffees it is being applied to has declined dramatically, resulting in an ocean of flat, carbonized coffee fit only for burying in syrups and milk froth. Peet's has done a far better job of buying green coffee that can handle deep roasting with a lot of varietal character expression, but I suspect that global warming will force some lightening of roast intonation even there, as coffees grown at 5500 feet are starting to cup like they were grown at 4000. 

2. Seasonality: by this I emphatically do not mean the bogus "seasonality" of many third wave roasters who use the term to justify offering a ridiculously small selection of washed single origin coffees. Instead what I'm referring to is that from a roasting perspective there should be a range of roast intonations applied to a coffee over its useful lifespan.

Taking a top new crop Guatemalan as an example, it'll arrive and clear customs in, say, May or June, and the coffee will never be better than it is at that point (assuming proper harvesting, drying and reposo). Acidity and aroma are at their peak, and its optimum use (assuming the roaster has a range of roast profiles and end users) would be moderately (City to Full City) roasted, ideally unblended or otherwise combined with coffees of the same caliber, in a blend intended for drip or (better still, however unlikely) vacuum-pot brewing. 

Assuming one is tasting one's production roasts on a daily basis, by October or so (depending on warehouse conditions) this coffee will have faded enough to have acquired a slight woody note, which can and should be "roasted out" by taking the coffee a few degrees darker. Depending on the broader coffee lineup, maybe it's relegated to blends and replaced by fresher single-origins, or perhaps used judiciously to add chocolate notes and some acid snap to a Northern Italian espresso blend that's otherwise composed of softer coffees such as dry-processed Brazils and the like. 

By January or thereabouts this coffee is tasting seriously past crop, and finds its best use in seriously dark Italian or French Roasts, while one eagerly awaits new crop. 

Semi-washed and dry-processed coffees are another animal entirely. Softer ones like Brazils are generally best suited to espresso brewing at roasts in a fairly narrow range that starts at Full City and goes not much past Vienna. Sumatras and other classic Indonesians are tricky to roast and have a tan or blond cast even when objectively quite far along in roast, but they need Full City+ to begin to blossom and earthier examples can go quite far into second pop. Dry-processed Harrars and Yirgacheffees and the rare great lot of Yemen Mocha are among the most forgiving of coffees, needing at least Full City to begin to show their best but also making for remarkable espresso late into second pop. 

3. Brewing method: the paper filters used in drip brewing mute the perceived acidity of a coffee, so the ideal roast for this method will taste a bit shrill in a cupping cup but just right when brewed in a commercial drip brewer. I should also point out that regardless of whether you rinse your paper filters or not the ratio of paper to coffee matters, which is why you'll never get the kind of stellar flavor of a just-brewed batch in a 1-3 gallon commercial brewer from a single-cup pour-over unit, no matter how much weighing of water, pouring through tiny spouts or other such geekery is involved. 

The vacuum pot remains, as Corby Kummer once put it, "the CD player of coffee makers - because all you taste is the coffee," offering a transparent experience of aroma and flavor that's a clear notch better than even the best drip units. Sadly there's been no rival of the sturdy, stainless-steel vacuum pots fit for commercial use from many decades ago and Bodum killed off their promising electric vacuum pots just when they were starting to get the hang of making them, so it's a fragile and arcane relic of a brewing method at this point. 

In general the best roasts for drip for washed coffees are in the City to Full City range, but there are certainly some exquisite exceptions for those who prefer body and lushness to varietal nuance. Some of the most memorable cups of drip coffee I've ever had have been out of 3 gallon urns at Peet's, Starbucks (back in the day) and Spinellli's (R.I.P.), with oceanic body, surprisingly high supporting (rather than leading) acidity and great aroma. 

Pressurized brewing methods increase the perceived acidity of coffee. The obvious case is espresso, where any coffee used needs to be custom-roasted for the purpose. Clearly there are a wide range of palatable roasts, but they begin with the coffee well established in second pop (Vienna roast territory) and end well before Italian roast no matter how hard and acid the bean. 

The French Press and the Aeropress are also pressurized brewing methods, albeit at trivial levels of pressurization relative to espresso. Nonetheless, there's enough pressure here to call for full city roasts at a minimum, and to make these methods particularly good showcases for coffees such as semi-washed Indonesians (Sumatra, Sulawesi) or full city+ dry-processed African coffees for whom body and depth are more significant calling cards than refined acidity. 

During my years at Starbucks the only brewing method used at the roasting plant besides espresso was the plunger pot. Meanwhile a senior executive at Peet's who was deeply involved with their coffee once told me he'd spent an entire year drinking their coffee only as straight shots of espresso - in consequence of which he might have been the only person in Peet's history to complain that the coffees were a bit too acidic!

 A steady diet of nothing but pressurized brewing methods trains the palate to value body over other aspects of a coffee, while the increase in perceived acidity afforded by the pressure makes a coffee that would taste flat brewed drip taste fairly lively. I argued unsuccessfully for years at Starbucks that our roasts ought to be lightened up to suit the drip brewing method that they were being used for in our stores and in most customer's homes - or, alternatively, that if we believed in the plunger pot so much we owed it to our customers to only brew coffee in-store using that method. "Roastmasters" (whatever that means) and buyers in newer shops today who taste nothing but cupping room roasts and then inflict them on their customers are living in the same kind of coffee bubble, and in either case the coffee and the customer (and ultimately the farmer) pays the price. 

4. Water: I don't think it's any accident that the places where the classic Full City roast flourished in the early days of good coffee in the U.S. were locales with naturally-soft [low mineral content] water that was generally free from off tastes and odors. On the East Coast Manhattan was (and still is) legendary for its soft water, while in Chicago or Milwaukee the high mineral content dictated much lighter roasts that would still show some life after being blunted by the water. On the West coast Peet's had good soft water in Berkeley, and the same held true in Seattle and Portland.

When we at Starbucks started expanding beyond these favored areas it was either change our roast or change the water, and I ended up specifying water filtration and treatment systems for places like L.A. and San Diego that cost a small fortune, taking rock-hard water with sulfur and salt water intrusion and running it through softeners, reverse osmosis and remineralization treatments to effectively duplicate the neutral ~3 grain hardness water needed for the coffee to taste the way it should. 

Cupping room roasts at retail and other modern phenomena

Getting back to the note from Landon that provoked this post, what I would call a cupping room roast - light City, first pop just ending - is indeed on offer for both retail consumption and, unbelievably enough, in more than a few espresso doser-grinders, in some of today's Third Wave cafés. Considering that most of these establishments are run by people with no actual coffee training perhaps it's not surprising that the range of pejorative terms that apply to such roasts (underdeveloped, cereal-like, green, bland and so on) are unknown to them, but such roasts represent an extreme over-reaction to Starbucks and its many imitators. 

What would be nice to see in coffee is the kind of stylistic diversity one sees in craft beers, with one roaster specializing in the coffee equivalent of Belgian lambics (which I guess would be nothing but edgy, wildly-aromatic Ethiopian naturals!) and another into Porters and Stouts (the Peet's universe) and many other options in between, but a lot of what I see at retail, at least in the Pacific Northwest, are roast-alike and taste-alike clones of Intelligentsia and Stumptown with the same half-dozen origin countries all represented at a very narrow range of roasts from light City to light Full City. 30 years ago we had Peet's and Starbucks pushing the limits on the dark side while Pannikin in San Diego, Kobos in Portland, Schapira's in upstate New York, Freed, Teller & Freed in San Francisco and (best of all, IMHO) The Coffee Connection in Boston offered superb true Full City roasts. Many of us hoped that that level of choice was a harbinger of a greater range of good choices to come, but it looks like we were overly optimistic. 


  1. Kevin, this is an amazingly information-dense post, and I enjoyed reading every word of it. A few random comments:

    As you may know, James Hoffman's written about the effects of water on coffee, but he also mentions the interesting possibility that local water quality affects how roasters roast their coffee: http://www.jimseven.com/2012/02/05/scandinavian-travels/

    For someone who tastes coffee through a press pot or a paper-filtered method, the coffee's acidity should be muted, so the cupping roast (and brew) for that coffee should be very acidic if a drip is going to taste normal. Is this correct? If so, then one would think that 3rd wave acid bombs (as tasted in home brews) should be pretty rare. Yet, in my experience, that is one of the stereotypical characteristics of a 3rd wave roast.

    How much does the low pressure extraction of an Aeropress affect the brew? I've done a brew of the same coffee ground to the same size in a CCD and Aeropress with equivalent coffee/water ratios of 60g/L, with similar steep times, water temps, and stirring, and yet the Aeropress always yields a brew that is much lighter than the CCD. My only guess was that the Aeropress's paper filters filter a lot more stuff. I wish I had a refractometer to measure these things. Any thoughts?

    By vacuum pot, do you mean siphon pot? For example: http://www.sweetmarias.com/sweetmarias/index.php/coffee-brewers/vacuum-brewers.html

  2. Hi Andre - thanks for your thoughtful comments and questions.

    I enjoyed James Hoffman's post. He reports that the cupping room in Copenhagen used RO water for brewing, which is always a disaster and does indeed make any coffee taste flat. You need some mineral content in the water, and realistically speaking the roasts on offer ought to be adapted to the local tap water unless you're in a truly toxic, my-water-glows-in-the-dark situation as for example in Southern California or Texas where it's realistic to expect customers to use bottled spring water for brewing and a sophisticated in-store filtration system is assumed.

    In my experience paper filters mute perceived acidity and the plunger pot slightly increases perceived acidity. What I see a lot of the 3rd wave folks doing is selling and brewing roasts that are either fit only for cupping (~ light City range) or too underdeveloped even to cup (not finished with first crack/true institutiuonal or cinnamon). There is a lot of roasting to eye, rather than to taste, going on, it seems to me.

    What is a CCD brewer? Clever Coffee Dripper (a guess on my part). I like the Clever but in my experience you're always battling rapid cooling of coffee in the plastic. The new Bonavita porcelain Clever clone (preheated) looks like a promising improvment, but it'll be heavy and fragile. I use a quite fine (~ Melitta) grind in the Aeropress and for all of these methods I prefer coffee at the upper end of the strength range: 65-70 grams per liter. The darker the roast the heavier the dose and the higher the brew temp. needs to be due to less actual soluble material/less coffee in the coffee. Assuming you rinse both the Aeropress and the Clever filters I would think you're going to have far less coffee absorbed and retained by the tiny disc in the Aeropress vs. the comparatively massive filter in the Clever. Rinsed or not I don't routinely detect a paper taste in Aeropress coffee, whereas I usually do in any of the small capacity drip brewers.

    Finally, yes, vacuum pot = siphon pot.

    Thanks again for your comments.

  3. Kevin,

    Once again, thanks for the very insightful post. It's funny, I have lived my entire life going back and forth from Southern California and Texas, and I just kind of assumed everyone used bottled water for home brewing.

    So I am curious, what is driving the uber-light roasts of the Third Wave roasters? Is it purely reactionary to Starbucks and Peet's? Is it economics since lightly roasted beans are more dense and therefore yield a greater volume of roasted beans compared to darker roasts? Is it because the lighter roasts taste less like "coffee" and are lacking in bitterness (which the American and Western European palates do not typically enjoy)? This is the part that eludes me. I don't understand why the Third-Wave is roasting so lightly and selling underveloped coffees. Moreover, I don't understand why they are having so much success doing so unless nobody wants to point out that the emperor is not wearing any clothes.

    I would think that roast level is just a means to an end: great tasting coffee. But, it seems for some roasters that the roast level is an end unto itself. It also seems like many Third-Wave roasters care more about origin and farms than what is in the cup. Read through the descriptions on their sites and 80-90%, if not more, of the copy is about the farm/farmer with very little detail paid to the actual coffee and even then the descriptors used rarely ever show up in the cup.

  4. Aquabrew was the pioneer manufacturer of really good commercial brewers that could accommodate a 5 gallon bottle of water. I think they might have been bough out by Newco, but in any case for lower-volume cafés and OCS use their brewers are great.

    Not all of the 3rd wave folks roast ultra-light, but many do. I certainly don't think it's economics, since the few percent they save with lower shrink rates during roasting are irrelvant compared to the huge sums squandered buying "microlots" whose quality is often no better than that of coffee available in full containers for "C" plus a differnential, plus the huge expense of ceaseless (and largely unnecessary and unjustified) travel to farms, splashy self-congratulatory websites, etc.

    As for how much success they're having, that's clearly a relative thing. Espresso-based milk drinks and wholesale trade in coffees far less glamorous than a few highly-touted Cup of Excellence lots still pay the bills most places, and in moset cities the volume of coffee sold by all of the 3rd wave places put together probably doesn't equal what Starbucks moves in one neighborhood (leaving out venues like supermarkets and Costco which is where purchases of actual whole bean coffee occur).

    I agree with you about many roasters caring more about origin and farms than what's in the cup, and the customer's interest in value for money spent, quality, a choice of origins and flavors and consistency doesn't seem to factor in at all. At its worst the 3rd wave model is solipsism and narcissism dressed up as "direct trade." We've gone from the old colonial model of roasters paying as little as possible for coffees, using them in proprietary blends and scrrewing both the producer and the consumer to a new model of screwing the consumer in order to get as much money as possible to the farmer and above all the roaster, and that apparently is called progress (or "direct" trade).

  5. Hi Kevin, thanks for the useful information. I'm always surprised and learning something new when I read you --- for example, I didn't know that the water should have some minerality (TDS?) for brewing coffee. I've been using RO or distilled water all along.

    The CCD is indeed the Clever, and I've heard other people say that it loses too much heat, so I decided to measure it. At an ambient temperature of around 69F, this morning's brew went from 208F (water in the measuring cup out of the microwave) to 198 (after 360 g water poured into a CCD with 22 g of coffee) to 183 (after 3 minutes and some stirring).

    Is a 15F degree drop important? I don't know. I suppose it depends on how much of the extraction is done up front when the water first hits the grounds compared to later on in the steeping. One indirect way to test this might be to wrap the CCD in neoprene, measure the temperature loss, and see if there is any taste difference for presumably lower temp loss.

    I've also accidentally discovered that darker roast coffee (eg. Peet's Ethiopian Fancy) needs more extraction to have its full flavors come out. I was so surprised the first time I had strong, acidic berry flavors come out of one of Peet's coffees. I will try out the higher dosing, though I've found that with higher dosing, there is a tendency to underextract if all else is equal. I usually need to stir more or steep longer when I increase the amount of coffee relative to water.

  6. Hi Andre -

    About the only thing worse than RO water for brewing is softened water.

    That's a pretty precipitous tempuerature drop in the Clever, but I think you'd see similar numbers in a Hario or Chemex. It just underlines the fact that drip coffee is by nature a batch (not a single serve or by-the-cup) brewing method, and said batch needs to be at least a quart (but a half gallon is better and a gallon or two better still) for optimum results.

    The Clever is of course nor really a drip brewer but an immersion one, which is why it kinda works - and certainly works better than Harios and the like. But contrast all of these with brewing in a real electric drip brewer - say a Fetco on the commercial side or a Bonavita, Technivorm or Brazen for home use. With any of these you can preheat the filter cone with 200 degree water, and in the case of the commercial brewers the filter basket is made of stainless steel and retains a lot of that heat. At sea level you'd set the water temperature to 200-205 F coming out of the spray head, which means proper extraction and even on a cold morning brewed coffee coming out of the filter at at least 185 F.

    Personally I think that a great coffee is going to be delicious at a pretty broad range of extraction temperatures, but it's sure nice to be in control of that variable as much as possible. The ideal in that regard is clearly the vacuum or siphon pot, where the water in the lower chamber is actually boiling throughout the brew cycle and the drop to ideal brew temp is consistent.

  7. Yes, a siphon pot is now on my shopping list. I wish I drank enough coffee to justify one of the big brewers!

    I'm curious if you've seen the Espro press pot. It has a double-wall vacuum-insulated steel pot and the plunger has dual metal filters so it filters better than a traditional French press. I have the little one but haven't used it much because it's kind of messy and retains too much coffee below the filter. Perhaps I will have to give it a go again now.

  8. Yes saw the Espro at SCAA a couple of years ago. Elegant design and does what they say it does, but I'm not sure that some sediment isn't part of the charm of the French Press. The Bodum Columbia series which is also double-wall stainless but doesn't have the microfiltration and a similar pot by Nissan are also appealing, but I'm pretty happy with the Aeropress for small-scale brewing and a Nissan thermos with #6 filter cone for quart-at-a-time drip coffee for two.

  9. Hello Kevin,

    I'm really enjoying this blog. There is much food for thought and you write well - so hope the muses continues to inspire you to write on all things coffee.

    Regarding degree of roast, it's possible to go quiet light on the roast but still produce something that tastes like coffee should taste, but I have tasted coffee from some of these 3rd wave roasters that is almost still green in appearance and tastes of tea rather than coffee. I'm an old fashioned roaster (though relativey new in the business) but I believe that roasted coffee should taste like coffee rather than lemon grass or green tea.

    It seems to me that one of the problems with these really light roasts is that the beans have not been allowed to develop enough after first crack - and that what is being offered is something closer to a cupping roast. The problem is then compounded when these light roasts are used for espresso extraction - producing an espresso which is like sucking on a lemon. The thinking from some of these roasters seems to be that there is only one roast profile suitable to bring out the essence of these beans. For washed coffees in particular a different roast profile should be used for coffee to be used for espresso extraction - generally a longer roast development time with a higher finish temperature and consequently darker degree of roast is required to dampen the acidity. Naturals and pulped naturals also taste more balanced in an espresso when pushed to a darker roast degree but generally don't need to be pushed to the same degree as a washed coffee.

  10. Thanks for your comments. I agree completely about the roasting of course, and would only add that progress in coffee, as in wine or food, comes from careful study of the best of what already exists as the basis for real innovation. In the case of espresso, many in the business today clearly haven't learned from the vast amount of experimentation and experience that the Italians offer so freely. There's a reason for the range of roasts they use, for the kinds of coffees selected for blends, and for the dosages, contact times and brew temperatures of classic espresso. There are even good reasons why Italian espresso blends are forumulated with the expectation that straight shots will be consumed with sugar added.

    Espresso should be about subtlety, nuance, balance and above all a perfume that lingers on the palate for many minutes after the last sip. Espresso that makes you think you're sucking on a lemon, or, alterantively, eating burnt toast, or that absolutely MUST be buried in steamed milk in order to be drinkable, are the kind of "innovations" we could do without.

  11. I completely agree with you that I wish there was a wider variety of roasting styles out there. As it seems to be a lot of roasters out there that have no idea what they are doing like you state, whether they be roasting too dark or too light doesn't matter because it stinks either way.

    Since you stated you liked Coffee Connection, what do you think of George Howell's Terrior? Personally, I have not had it enough to emphatically state that he roasts light all the time, but the ones I have had were very much on the light side. Is he roasting too light? Or does he have the experience in coffee to adequately roast light?

    What are good examples of roasters out there who are doing a good job?

  12. A friend who just did a pretty major tour of newer roaster-retailers on the West Coast, from L.A. to Vancouver, reports much the same thing: lots of folks roasting without a clue and buying & roasting coffee with no background other than pulling shots on an espresso machine.

    That said, there certainly are roasters out there paying careful attention to roast degree, but I hesitate to offer a lot of names because there are so many new players and the market is so dynamic that I'd probably end up leaving a lot of worthy roasters out.

    As for Terroir, it's been a couple of years since I've drunk their coffee regularly but the roasts there do seem to be lighter than at the Coffee Connection - erring in the direction of too light - more towards the City or American rather than classic full city range, and the Agtron scores I've seen in Coffee Review tend to bear this out. This may well be George Howell's preference these days, and Lord knows he knows what he's doing and the coffees being bought are sure to be stellar, albeit in a very narrow range of taste.

    As far as nationally-available coffees that are roasted sensitively and in a reasonable range I would suggest checkout out Allegro Coffee (either at Whole Foods or online). My successor there, Christy Thorns, does an outstanding job.

  13. Kevin,

    As a young student of coffee (10 years experience and still feeling novice) this was just pure pleasure reading. I work today at Coffee-Tech Engineering, a manufacturer of roasters from Israel, and so many things you wrote here fit our ways of thinking, teaching and believing. I might add that we are also firm believers in emphasizing conduction rather than convection and this too I find to have many things in common with your post.
    It's funny, and sometimes frustrating, but we find ourselves many times telling others that with all due respect to sophism of various kinds, the result of a roast is in one main place - in the cup.
    Thank you for making me smile :)

  14. Hi Kevin. I enjoy your blog however I must take issue with the idea that there is a degree of roast that any coffee "should" or "should not" be roasted to. Roast preference is as personal as the degree to which a person likes their toast or their meat done. For every person who thinks that a cupping roast or 'blond' roast is far too light there is someone who cherishes it. The very light Scandinavian roasts are named for just that reason..many in Scandinavia prefer an extremely light roast and the resulting high acidity.They are not all wrong or incorrect, they just prefer what they like. If there was a correct roast for Yemeni or Sumatran coffee, then after hundreds of years we would all be doing it the same. I have been told that Kona 'should' be roasted light..but travel to Kona and find even a French or Italian roast is offered..because some prefer that. They are not wrong.. they know what they like.
    When I opened my shop 17 years ago I offered 32 different varietals at 11 different roasts, done while you wait (7-10 minutes total) with a free cup of coffee to make it enjoyable.I had customers who enjoyed a Cinnamon Roast of Kenyan (drop temperature of around 420) and I had a New Guinea Kamarl that was most popular at a Finnish Roast (drop temperature of 403). I also had customers who enjoyed an Italian Roast (drop temperature of 475) of the very same coffees. To each his own.
    Currently I pick a varietal of the week and roast light to dark Monday to Friday and I do not sell the beans past a day old. If the customer discovers a certain coffee that they love they can get it custom roasted the way they like any time they want. After all these years, the only rule I have learned is that there is no optimum roast for any coffee.. or rather that the optimum roast is whatever the individual customer prefers and none of them totally agree concerning each varietal.