How many MC's

[ChrisSchooley]

by MC's I mean more coffees, and by how many I mean How many coffees do you think can go into a blend without the blend becoming muddled and cacophonous.

Secondary question, what would you consider too little of a particular coffee in a blend to really make a consistent impact. For example, I've often felt that having less than 20% of anything in a blend (especially espresso) is pointless as there is no way that you can tell me that 15% of any coffee makes it into every shot.

[ChrisSchooley]

29 views? No one has any thoughts on this?

Maybe if I gave my answer: 3

I think that once you get more than 3 coffees in there, it starts to get muddled and lead to inconsistencies.
Also, as I stated above, I think that using less than 20% of something in a blend really reduces the input of the coffee to the profile.

I would like to hear what others think though.

[AndrewT]

Chris i don't have any absolutes but i do have a few thoughts. First i think blending ~15% in a blend is intended to be more of an influence than something that stands out. If you blend 15% of a color with another color it doesn't necesarily change the color but the shading (the best analogy i can come up with right now). we have a blend that has 13% sumatra and we have tried taking it out of the blend but the dynamic of the coffee changes, so it remains. I cannot gurantee that it ends up in every shot, so i grant you that. But I wonder if you are making your calculations based on 100% of each bean getting incorporated in the ground product? i'm not sure you can predict exactly what happens between the hopper and the grinding burrs. I do agree with your rationale but it contradicts my experience.

What tends to get lost in a forum like this is common experience, which allows us to have a more substantial discussion. It would be fun to come up with an experiement based on this and other discussions in order to extend the conversation. If we come up with a simple experiment, maybe a green broker would provide samples in order to test our theories? what do you think?

[AndrewT]

So this is what i did. I created two blends the first with 15% Sumatra; 25% Brazil; 20% Salvador Pulped Nat; 20% Colombia Washed; and Ethiopia Natural (aleta wondo) the second: 29% Brazil; 23.8% Salavdor; 23.6% Colombia and 23.6% Ethiopia. I roasted both of them and this friday the roasters and our barista will do a triangulation cupping and see what we find out. The roast were 12min roasts; Gold at 6:30, First at 10:00 and 13% moisture loss, if you want agtron i can give you an approx. We are going to do two sets of the same cupping. the first set will be set-up weighing out the whole bean and grinding individually. the second, we will grind before then weigh out the ground coffee. I'm not sure if this will do anything, but i wonder if it will incorporate all coffees in there correct percentages so we can discern the 15% sumatra? This of course does not help us determine how many coffees one can use in a blend without it getting muddled, but it will be fun!

[rojo]

I go with 4...

I've done blends with a 1/4 split between two coffees (12.5% of each) and it seems noticeable and helpful...
I've only done that with espresso looking for just a wee more crema and/or brightness in the blend... Any blend is probably subject to the chance of not containing all it's components proportional to one's intentions, that increases when the percentage of a component is low. The variation seems very slight to me though... not enough to discount it as totally inconsistent. I guess if I were to hone in on a lowest possible percentage, it'd be that 12.5% because we've had 10% of stuff in blends and it isn't as consistent. But that's espresso speaking....

For drip maybe 10% could work out, depending on what you're doing. Airpots and such with a good 3-4OZ of coffee could fare better since there's more coffee in there... I agree that low percentages are for hinting at things in a broader sense rather than standing out specifically. Like trying to borrow a coffee's body without imparting so much of it's flavor or using the darker roast in a little bit to balance some high notes throughout...

It'd be fun to try and look more indepth at this though... Pulling twenty summ' shots and seeing if they taste the same and all that. ...

PS. Chris-
Now I got Black Moon stuck in my head
-Thanks....

[AndrewT]

i guess if i have to answer the question and not ramble on i would say 3 unless an element of the taste profile of the blend is lacking then i would say 4. But i am going to ramble anyway...

Ryan i think even the best barista would have difficulty producing the same shot 20 times in a row even with a single origin. I truly commend your effort though and i would love to hear what you learn in the process. Some may protest, but espresso is so finicky that it can be like trying to line up the cosmos (does that make sense?). I like how you said it, "But that's espresso speaking...." so it makes it difficult to make claims about consistency or inconsistencies. So, word.

I was sample roasting and thinking about this post, that is when i thew together the blends i mentioned below. I guess the question i was trying to get at was how consistently can 15% component be distinguished in a "muddled" (what is muddled anyway?) blend, thus the 5 part blend. Chris, i totally agree with you, but what is fun about coffee is it tends to mess with you in unexpected ways. So with the blend created and roasted i wanted to find out how the other roasters and barista here would respond in a blind triangulation cupping. So if the results are everywhere then we might be able to assume that the 5 part blend is "inconsistent" but if the 15% coffee is consistently picked out then maybe it has something to consistently offer? what do you think?

[ckornman]

A few of our in-house studies have indicated to us that anything less than 10% in a blend may not be noticeable. I'd wager it could be as much as 15%, maybe 20% for an espresso. In an exercise designed out of pure academic curiosity, we were actually able to hide up to 10% past crop dark roast in a fresh crop light roast Kenya (and I'm talking 10% of the beans in each cupping glass) from reasonable detection. Sure, it tasted different, but not significant, particularly compared to higher percentages: when on the table with 25%, 50%, and 75% increments of the same coffees, the 10% & 0% were indistinguishable. When it was just 10% & 0%, we could all tell a difference.

For my part, I've never met a 4-part blend that held a candle to a well-crafted 3-or-less in terms of clarity & articulation. That's what I think of when I say "muddled," a coffee that has lost its sense of self. You can't pick out those hints of origin, the distinctiveness of individual flavor profiles. Now, if you want a blend that just tastes like diner coffee, blend it up - the more the better. You can certainly achieve balance with more parts, smaller percentages, etc. My favorite blends seem to consistently be those 2 & 3-component iterations, where one coffee functions as an anchor for body & sweetness, another provides distinctive flavor, and maybe a third elevates the acidity and subtler flavor elements, or augments the sweetness, or perhaps accents the finish.

Regarding Andrew T's assertion that

...i think even the best barista would have difficulty producing the same shot 20 times in a row even with a single origin...

I agree... but...

I think that the goal of consistency, while lofty, is also admirable and achievable. Much of the role of the pro barista is to mitigate inconsistencies. Consistent dosing, tamping, pressure, temperature, etc., etc., are all part of this process, and if there's one less thing to worry about (i.e., an inconsistent blend), it's that much easier to achieve that ideal.

My thoughts...

[c.hallien]

I saw a coffee here in Toronto claiming to have 7 different countries of origin ... why?
I think Kornman's observations and figures are very accurate. In my experience you can blend 5-20% without the average consumer noticing the difference, so would this mean that those are the minimum %'s to make an impact? Obviously there are differences in origins/varietals/processing methods and their overall impact on the blend; the contribution of 20% Washed Costa Rica vs 20% Sun-Dried Harrar is disproportionate so in that respect it would depend on the components under consideration. Theorectically coffees in blends should work much like spices in a dish, sometimes a dash of salt or sugar are not perceptible at all as sweet or salty but they do work to enhance the other flavors, other times one strong spice can overpower the rest. All that being said, anything more than 3 makes me personally suspicious that
1. the blend is a dumping ground for old and / or slow moving coffee
2. a claim / cert requirement is really the motive
3. some lofty/wordy/descriptive person has infultrated the place and convinced everyone that more components make for a more complex coffee, then I get a visual of some crazed coffee fanatic
All reasons, in my opinion, to be catious.

[ChrisSchooley]

Great replies. Kornman, thanks for the figures. I've found similar results again and again. Also thanks for backing up my consistency argument.

Much of the role of the pro barista is to mitigate inconsistencies

I kind of feel like blending per dose or throw, or at the very least per 1lb, would be a pretty solid way to mitigate inconsistencies. I realize that it may not be realistic in a retail setting, but not impossible.

Ryan, good looking out on the Black Moon, Sh** is mad real.

[AndrewT]

Long story short with our blind cupping (explained above) there was no consistent distinction/identification with a 15% portion of Sumatra, which i think is one of those coffees that is "disproportionate" in its taste effect. With ten tasters tasting ten sets there was only a 38% rate of identifying the correct coffee, and you have a 30% chance of getting it right anyway. And there was no significant difference between the set that was weighed per cup vs. the set that was preweighed and ground before portioning in the cup. So i think this supports the general consensus of the group. more later...