A while back, a user posted a question about "exploring the extraction space" on Home-Barista.com.
"In various reviews, you speak of varying the basket size, dose, temperature, drink size, and extraction time. Exploring this space with a new coffee seems like a hugely daunting task, particularly for those of us who home-roast and therefore may have at most half a pound of a given blend, with no assurance that the next roast of the same coffee will be exactly the same. Do you have any rules that help you do it?"
At the time I gave the best response I could come up with. But since then I've continued to think about the topic. Actually... it's one of those really, really big questions. And it's something I've practiced without really thinking about it in a systematic way. The question made me realize this - and has made me think about the issue in a more structured way.
I think I'm probably ready to start talking about what this means - about what to do -- basically to try and propose a starting point for further discussion of the problem.
First of all... I've been incredibly lucky. I'm in a unique situation. I have access to a huge number of coffees - and I don't have to pay for them. I have a lot of free time. And I have the prototype La Marzocco GS3.
Most coffee professionals don't have the time to spend doing this sort of experimentation. Most coffee pros don't really have access to the range of coffees from all sorts of roasters. Home enthusiasts tend to have to pay for their coffees. Home enthusiasts really don't have access to a piece of equipment like the GS3. For that matter... very few pros have that luxury!
This is a unique moment in time for me - and I'm trying to milk it for all its worth. It isn't going to last forever.
So... what is the process for exploring the extraction space? What are the goals? What is the methodology?
What is this and how do we do it?
Right now... it's still all in some state of flux.
It's still largely trial and error, guided by some very rough guidelines, leading to some equally rough general rules and structures and patterns.
At it's most extensive, I'll start with a coffee I know nothing about. In other words, I don't know what beans are in the blend. I don't know anything about the style or philosophy of the roaster (i.e. I've never tasted their coffee before). I've not cupped the coffee. I have no input from the roaster on target extraction goals.
In this case, I'll always start with a visual evaluation of the beans. The goal is to try and narrow down a starting point. I'll look for degree of roast, I'll try to identify beans. The idea is to try and recognize commonalities with previously evaluated coffees. As an example, if I see a very light degree of roast and from visual clues see that it may contain Monsooned Malabar I will use what I have learned from working with the Vivace Dolce as a starting point.
I'll then move on to evaluating aroma of the beans. The idea again is to try and identify beans. I don't need to identify the exact beans per se - it's more that I'm again looking for experienced patterns to narrow down a starting point.
As a generic starting point I tend to go with the following parameters:
- LM ridged double basket,
- 200F brew temp,
- 9BAR brew pressure,
- 18 gram dose,
- 2oz extraction in 27 seconds.
Based on the initial visual and aroma evaluations, I will start by making some assumptions on the brew temp. If the degree of roast seems light, I'll adjust the baseline brew temp up to 202F. If the roast is dark, I'll drop that baseline down to 198F. I then tune this based on the beans. If, for example, it looks like the coffee is high-grown washed arabica I'll reduce the brew temp. If I see aged or monsooned coffees I will increase the brew temp (both from the baseline above). So a light roasted coffee with monsooned beans will move up to 203F as a starting point.
At this point, I move on to establishing a dose starting point. To do this I look at the bean composition and the "signature taste" of the roaster and the coffee. If the coffee has a lot of naturals or pulped naturals I will adjust the dose down from the baseline. If the coffee is mostly high-grown arabica I'll increase the dose. So, for example, with the Terroir Daterra Reserve I'd go with a 17.5 gram dose. But with the Stumptown Hairbender I'd go with a 19 gram dose. The "signature taste" is a harder one and requires some knowledge of the roaster. Is the person roasting this coffee a fan of low acidity espresso? Are they a "chocolate bomb" aficionado? If you know what they like out of their espresso you can do minor adjustments to your dose. So, for example, based on this I would actually drop the Terroir Daterra down to a 17 gram dose but would up the Stumptown Hairbender to a 20 gram dose.
Of course, the shortcut for all the above (as implied earlier) is to recognize a common "type" of espresso. For example, from working with the Vivace Dolce I would immediately start with a brew temp of 203F and a slightly down-dosed triple basket pulled ristretto when beginning experimentation with the Victrola Streamline. Or from working with the Stumptown Hairbender, I would start with a brew temp of 197.5F and an updosed double basket pulled short when exploring the Lighthouse Espresso.
In either case, once I have an adjusted temp and dose baseline I'll start experimenting with extraction. I always start by re-evaluating the brew temp. I'll pull a shot and evaluate it. Is it alkaloid? Is it thin? Is it sour? Astringent? Based on the taste, I will alter the temp by small degrees to find the sweet spot. On a side note... many people use visual clues for adjusting temp (looking for crema, colours, patterns, etc in the flow and in the cup). I find that this can be both distracting and often inaccurate. For me - I find it far more worthwhile to focus on the flavour.
Once I've found what I feel is the brew temp sweet spot, I'll start working on adjusting dose from that initial baseline number. I go about this by focusing on two things. First - clarity of flavour and second - roundness and balance. If the cup is "muddied" I'll reduce the dose. If the cup isn't fully developed and sweet and rich I'll up the dose.
It's important to keep in mind that there are some limitations to these guidelines. In particular, there are coffees that benefit dramatically from increased headroom in the basket - and others that are best with restricted headroom. In these cases, the actual dose is less critical than the headroom - in other words, the dose is a way to reach headroom goals. I recently tasted an experimental espresso from Andrew Barnett at Ecco Caffe that required a ton of headroom in the basket. This held true across different baskets and the resulting different doses. With too little headroom, the coffee lost all sweetness.
With a lot of coffees this will get me to the point where I'll have a cup profile I really like. But this probably represents a minority of all coffees.
There are times when I won't be able to get to where I want to be with just these factors. It's at this point that I start looking at changes to extraction volume and basket size. For example, I've found that some lighter roasted delicate coffees tend to end up poorly developed no matter what I do - especially when they are pulped naturals. But if I then swap to a triple basket and deliberately down-dose (19 grams) I "open up" the coffee and it becomes more defined and clear. Or with monsooned coffee I find that the only way I can get the desired sweetness without getting a "wet cardboard" aftertaste is by going with a triple basket, normal dose and then pulling a ristretto shot.
As a general rule, I tend to initially tune extraction volume at this point rather than going to a different basket. Much of the time, tuning this parameter will result in my identifying a desirable profile. To do this I tend to start decreasing the shot volume while keeping extraction time constant. Given that I'm also keeping dose and brew temp constant, this requires a constant fining of the grind. With the exception of certain unusual experiments, I tend to never go below 1.25oz in volume. If I reach that point and still have not found the extraction I desire I will start exploring basket sizing.
With baskets - I'll swap to a triple basket in three cases.
First - if there is a recognized espresso type that I've learned performs best in this basket and I can identify the coffee as seemingly belonging to this type.
Second - if, in experimenting with dose and brew temp I've found that the flavour of the espresso is "constricted" and the coffee seems to me to be a delicate, light bodied one.
Third - if I've gone through all the various permutations above and cannot get the flavour I desire.
The reason I tend to leave this swap to the end is that, once you've switched baskets you really need to re-evaluate all your parameters. Even brew temp might change on you. As a result, I'd rather wait until the end and then treat this as an entirely new process - starting from the beginning again.
A lot of the time I'll find a "sweet spot" that I like but then start wondering about what a coffee would taste like when pulled differently. I generally believe that all great coffees have at least two different "sweet spots" depending upon your own personal taste. As a result, finding an "ideal" extraction should never be the end of your exploration - but merely one more lesson to learn and one more pattern to add to your lexicon.
Finally... the above is, at best, an incredibly rough and inaccurate set of lessons learned. For example... the reality is that there are times when changes to dose are going to require changes in brew temp rather than vice versa. And I have not even gone into brew pressure!
That being said - right now the biggest challenge for me is that I'm doing all of this in a partial vacuum. I would be learning far faster and developing these theories both more quickly and more accurately in a collaborative environment.
I am confident that there are people who have already gone through this - I am sure that I'm repeating the mistakes and successes of others.
I am equally confident that there are others doing exactly the same thing as me in other locations.
My hope is that setting down my (rough and probably at least partially incorrect) ideas is going to result in these people sharing results with me - and this is going to create that collaboration and accelerate this process.