Tasting Coffee – Initial Thoughts

2010 January 4
by Joe Johnston

I consider coffee to be a beverage which should be tasted and evaluated in the same way as other food and beverage. What I look for in any beverage can be distilled into three basic qualities: complexity, balance and being true to character. Complexity is a layering of interesting flavors which make themselves known at different times during the tasting process: the aroma, the initial taste, mid-taste, and lingering aftertaste. One dimensional beverages like Kool-aid are characterized by a sweet, artificial fruit flavor which varies little over the tasting process. Kids like it, but it holds little interest for adults. Balance is the harmonious whole of sweet, bitter, salty, sour and umami that is satisfying. Straight lemon juice lacks balance: it is essentially all sour. One can use lemon juice as an acidifier to balance the flavors of a salad dressing, however. Being true to character simply means that the taste of a beverage is within the normative range for that type of beverage. One expects a pinot noir to taste like a pinot noir not a cabernet, even if they may prefer a cabernet.

It seems as though coffee is judged by coffee professionals through a different lens than other beverages. Often coffee professionals focus more on being true to character than being balanced or complex. Food and beverage professionals who taste a broad range of food, wine, and other beverages will often grade a coffee much higher or lower than a coffee professional might. In my opinion, the current trend in coffee roasting/cupping places too much emphasis on acidity and not on balance. I plan on testing this hypothesis over then next few months.

7 Responses
  1. January 9, 2010

    I think this is a particular line of thought that I would be interested in seeing more experiments on in the future. Perhaps we could run these with professed foodies and espresso junkies in the valley?

    • January 9, 2010

      We should test the hypothesis. I have found this to also be true with beer. Many beer cognoscenti are die hard “hop heads” who value bitterness over balance or complexity. Let us construct an experiment and see if my hypothesis is correct or just a random thought. Consider this entry from Godshot’s blog http://godshot.blogspot.com/2009/11/state-of-sf-coffee_24.html

      • January 9, 2010

        That was a fascinating read. Might deserve a post of its own.

      • Mark permalink
        February 10, 2010

        Two things: There needs to be a testing methodology in place before any roaster prototype cranks out any beans. The ability to detect and measure roasting success will enable you to get to a finished product that much more quickly. I liked the article you linked to and share the author’s bias against the palette’s of coffee ‘professionals.’ This past weekend I got to sample the wares of a local roaster and was overwhelmed with the sourness of the brew as well as the lack of complexity. The brewing method (it was not espresso) yielded an underextracted cup, also. Maybe people who are too ‘into’ coffee can be as dangerous as those audiophiles who believe in exotic power cables and cryogenically frozen interconnects.

        Also, I just discovered the IPA craze, and am therefore something of a hop-head. However, I’ve tried about 15 different IPA’s and have gotten to the point that I can only drink Modus Hopperandi by SKA brewing. It delivers a strong hoppy essence, but it also brings grapefruit, pine and other herbal notes along with a smooth finish (nothing metallic here). It also looks very good and has a wonderful, thick mouthfeel. Please try it if you have not had the chance. A worthy but definitely second place contender in my opinion is Firestone Walker IPA.

        I agree with you about complexity. I want a fireworks show in my mouth, not just one noisy but colorless explosion.

        • February 10, 2010

          Thanks, Mark. I will try your IPA suggestion. We carry Dogfishhead 90 Minute IPA at one of our restaurants and I absolutely love it. Not because of its hops, which are fully present and vibrant, but because of complexity, balance and a bit of the unexpected.

          I have had several impressionable young coffee students bring me under-roasted coffee (you can still see exaggerated wrinkles, under-sizing, and obviously light color) and I have brewed it and also (just to drive home the point) extracted it on the espresso machine. SOUR! übersour! They loved it, but because they were told that it was the most glorious coffee in the world and that this is how quality coffee should taste.

          I also laugh (internally) when they tell me it has notes of gooseberry and a quince finish or something. First of all, most of them have never eaten a gooseberry or a quince. Second, even if they HAD eaten one they would not be able to identify it blindfolded in a pure form. So how can they possibly claim with certitude that any sort of faint flavor component in a black liquid is that of an uncommon fruit? Here’s how: they were told that it is present. It is basically group think. Are there expert tasters who can identify components — yes, undoubtedly. They have taken the time, effort and years (YES years) to develop a palate.

  2. Mark permalink
    February 10, 2010

    Along the lines of sensory testing, here is a panel discussing evaluation of audio and how easily our senses are misled.


    I wonder if a good way of finding a testing methodology for you is to find a coffee you really like, and have a double-blind test with your chosen taste-testing panel. Is the typical slurp-spit tasting really the gold standard? Why not get four Newco OCS-8’s and brew four coffees: a test bean (the one you think is really good), two similar beans of similar quality and freshness and one reference bean (i.e., a bean that is a little stale). The not-so-good bean is there to help give the tasters some scale of reference (Harman does this with speakers – it tends to keep the testers from exaggerating the differences between the other beans). The testing would be double blind, and you’d just be looking for people’s preferences. If you brewed all four coffees at the same time and each person got to taste a couple ounces of each (spitting is ok, too) then maybe you’d get a good picture of what coffee is preferred by a wide variety of people.

    One thing you could also do that Harman does is train amateur tasters. You could train a panel by giving them stale coffee, underroasted and overroasted coffee, so that they could get good at identifying roasting mistakes. I tried to home roast with several different roasters and even bought a 5 lb ambex and ruined hundreds of pounds of beans. That experience made me well acquainted with what failure tastes like. I taste those mistakes all the time when tasting small batch roasted coffee, from roasters all over the country. Anyway, Harman found that they could hone in on errors in prototypes an order of magnitude faster when using trained listeners than non-trained, even though both groups ended up with the same results.

    Just some thoughts.

    • February 10, 2010

      I like your thoughts on standardized tasting. I am not a fan of the standard cupping procedure for selecting green purchases being used a general tasting technique. Personally, I like to use a series of french presses. We are planning on doing a tasting panel with french presses later this month using two groups: coffee professionals vs. culinary professionals. The results may be surprising due to a mental reference standard that is different in the two groups.

      It is wise, as you say, to taste known defects so as to be able to detect them in blind sample.

Comments are closed.