Adventures in Coffee Processing

Adventures in Coffee Processing

Coffee Process Experiments drying on screen tables, Finca Manantiales del Frontino.

On the vast majority of coffee farms, no matter where they are in the world, coffee farmers use variations of the same 3 processes to transform fresh coffee cherries into dry coffee beans ready for sale: the natural, honey and washed processes. Winemakers and beer brewers, on the other hand, are constantly experimenting and pushing the boundaries of their craft- resulting in a steady stream of new and exciting flavors in those beverages. Coffee famously has more flavor compounds than wine- so why aren’t more coffee producers experimenting with them?

We wondered whether there were some inherent properties of coffee cherries that would limit the range of processing options- or whether uninventive coffee processing could be better explained through cultural convention, lack of resources, or simple historic momentum.  We decided to find out.

Luisa surveying our coffee experiments as they dry


Drying coffee in our improvised "drying silo"- a small food dehyrdator.


Sudden rainstorms meant babysitting the drying tables and rushing out with plastic a couple times per day.

Checking moisture levels on experimental coffee lot

In the fall of 2013, Luisa and I spent a month at Finca Manantiales del Frontino, a 170 hectare coffee farm that grows specialty varieties with a focus on quality in Colombia’s Valle de Cauca. We were there primarily to explore three questions:

  1. Why do coffee farmers use the same three processes worldwide?
  2. What other processes are possible in the context of a tropical coffee farm, and what kind of flavors would they create?
  3. Is there a way to add value for coffee farmers by engaging with coffee processing?

Manantiales was a beautiful spot, and everyone we met there was incredibly welcoming, supportive and friendly. We owe a whole-hearted thank-you to the entire team. Please pay them a visit if you’re ever in their neighborhood and try their amazing coffee.

Manantiales set us up with a small group of their best pickers to supply us with the highest quality, uniformly-ripe cherries. We planned experiments using 5kg lots of geisha and red bourbon cherries. Over the course of several weeks, we completed 10 different process experiments with geisha, and another 9 with bourbon. We divided the experiments into two intervention periods: maceration, or any processing while the coffee was still wet, and drying.

The processes were:


1. Refrigerated macerated mashed cherries
2. Macerated mashed cherries with wine yeast
3. Macerated ambient temp. no yeast
4. Honey Bucket (whole cherries left to age 2 days before pulping)
5. Mashed cherries macerated with milk and pectic enzyme
6. Frozen macerated mashed cherries
7. Standard raised bed natural process (Control)
8. Whole cherries simmered at 50c for 30 minutes before following process #7
9. Washed frozen (only macerated geisha lot not sulfited to control bacterial process)
10. Washed refrigerated


    A. Whole cherries frozen
    B. Whole cherries submerged in water with 75ppm SO2 (anoxic honey bucket)
    C. Macerated cherries refrigerated
    D. Macerated cherries with beer yeast ambient temp.
    E. Whole cherries simmered at 50c for 30 minutes before following geisha process #7
    F. Washed with citric acid solution
    H. Washed and frozen
    I. Mixed cacao bean fermentation (after sulfite wash on coffee cherries)

          Ia. Shorter cacao fermentation

    Measuring brix in a bucket of macerating whole coffee cherries.

    In general, we were using what tools we could find to manipulate time, temperature and biotic variables. (We also attempted to manipulate pressure with a stainless steel whip cream maker, but unfortunately the one we bought in Bogotá (and didn’t test until we were at the finca) turned out not to hold pressure. More on pressure experiments soon…)

    When we were finished and had all of our experimental lots dried, hulled and ready for roasting, we took them to Manantiales’ cupping lab in the nearby city of Armenia. The finca has two licensed Q graders on staff who helped us to sort the coffee, roast samples and cup them.

    We cupped each coffee blind, and scored them quantitatively and qualitatively, giving each a numeric score and set of flavor and aroma descriptors. Our lowest score was 8 out of 100! (Which was a joke, really- the coffee was basically ruined and tasted horrible.) Our highest score was an 89, which was a marked improvement over the control.  The complete results are below.

    Roasting samples of the experimental coffees at the cupping lab in Armenia.

    For our first foray into experimenting with coffee, and given the improvised tools and equipment we had access to, we feel like the project was successful. We’ve been told that coffee will humble anyone who starts trying to understand and manipulate it- and we didn’t have very high expectations for our first try.

    More than anything, we generated meaningful information about where to go with future experiments. Our results told us that time, temperature and bacteria all have interesting and varied impacts on flavor development.   Our initial results seem to suggest that colder temperatures during fermentation result in greener, herbal aromas, while warmer temperatures create stronger fruity aromas. Changing the bacterial environment may impact coffee’s acidity and secondary aromas.  

    It’s clear from our results that faster processes yield better results: coffee doesn’t hold up to extended macerations the way that grapes would, for instance. With our two experiments combining freshly hulled cacao beans with freshly pulped coffee, the shorter fermentation time yielded much better results. Our over fermented or over-macerated coffee developed oniony, wine-like, moldy or earthy flavors that are all pretty unpleasant.  2 days or less at full moisture levels is a rough guide we might follow for future rounds.

    All of us smelling each ground sample for dry fragrance before brewing.  Even really horrible tasting coffees sometimes had some great fragrances...

    Smelling the wet aroma of the brewing samples before breaking the crust.

    Recording tasting notes for each coffee sample

    More than these immediate conclusions, we learned that doing anything out of the ordinary with coffee on coffee farms in the developing world is difficult because of lack of hygiene, lack of infrastructure, lack of capital and lack of education. Exploring the nuance of each variable impacting the flavor of our experimental lots would require the ability to create a controlled, lab-like environment. Coffee farms don’t look like wineries or breweries with their shiny stainless equipment, impermeable flooring and ample supplies of disinfectant cleaners. Instead there are dirty boots walking everywhere, unwashed hands, insects, compost piles nearby… and that’s in the best of cases.

    At issue is whether the increased value of coffee from applying a more refined version of one of our experimental processes would justify the investment necessary to achieve successful results at scale. We wonder whether there may be a sweet spot in working with farmers whose land happens to be at lower altitudes and therefore who can’t grow traditional specialty coffees- where the opportunity to increase value is high, since you’re starting from the commodity price. These farms are also those that would require the highest initial investment to get a project like this off the ground, however.

    For farms that already produce specialty grade coffee, the benefit is harder to justify. Where buyers already pay high prices for quality coffee, the marginal quality or increased interest may not sufficiently increase margins. The market for specialty coffee is also complex: few roasters and coffee consumers are willing to spend the kind of money to support the type of unique specialty coffee we’re contemplating here- and the luxury segment is only so big.

    Regardless, this subject is fascinating to discuss and we do feel that new processes should certainly become part of the coffee world in the future- whether along the lines we’ve explored or otherwise. We’re happy to continue the conversation with whoever is interested- and we do hope to take our results and use them to inform more advanced experiments in the future.

    Full tasting notes for each experimental lot:

    (You’ll notice a few are missing- for some reason we don’t have data for them…)

    Red Bourbon:

    A. Aroma: Straw, dried corn, clove, spiced: cedar, pepper. Taste: Long, agreeable aftertaste, bitter chocolate, pepper, cloves, exotic, inconsistent. 85.5
    B. Aroma: Chocolate and fruity, light peach, dry, peanuts, blackberries, figs Taste: almonds and figs, velvety, creamy body, balanced fruity acidity, long and clean aftertaste 85.5
    C. Aroma: Buttery peanuts, toasted hazelnuts, sweet herbal background, chocolate, almonds. Taste: Flat, watery, lacks sweetness and complexity, dry and insipid. 77
    D. Aroma: Onion, malty butter, vinous-liquor, ripe fruit Taste: Light acidity, medium body, buttery cedar, prolonged chocolaty aftertaste, stinging background, onion. 82
    E. Aroma: Chocolate, spices, orange peel, earthy, phenol. Taste: Moderate acidity, rough, heavy, light earth, lacks sweetness and cleanness, astringent, disagreeable. 75
    F. Aroma: Herbal, dry, peanuts, caramel, dried fruits, herbal-chocolate, light vanilla, buttery. Taste: Medium acidity, juicy, pleasant dried fruits. 83
    H. (one cup was oniony, two cups were earthy) Aroma: Red wine, dried fruits, bitter chocolate, raisins. Taste: clean, gripping acidity, medium bodied, dried fruits, long aftertaste, lacks sweetness. 8

      Ia. Aroma: Potato, nuts, currents, pepper, pico de gallo. Taste: Vinous acidity, dried bitter fruits, light astringency. 82.5


      1. Aroma: Chocolate, hazelnut, herbal background, caramel and vanilla. Taste: Mold, silky buttery body, high acid, non-uniform. 79
      2. Aroma: Cedar, citronella, citrus, molasses, tobacco, malt, blackberries. Taste: Astringent, dry, non-uniform, juicy acidity, heavy body, short aftertaste, under-ripe background. 76
      3. Aroma: Bitter chocolate, nuts, straw, pepper, caramel, light herbs, white chocolate. Taste: Dirty background, heavy, astringent. 70
      4. Aroma: chocolate, dried fruits, butter Taste: chocolate, light spices, fresh wood, juicy acidity, sugar cane, molasses, spices, almonds, floral, tea. 85
      5. Aroma: Fruity, chocolate, dry almonds, sweet wood, cilantro. Taste: Astringent, acidic light body, lacks sweetness, rough. 79
      6. Aroma: Fruity, guava, spices, rubber/ woody background, liquor. Taste: Under-ripe background, green, astringent straw, dry, lack of sweetness. 75.
      7. Aroma: Peach, vanilla, caramel, hazelnut, jasmine, rose tea, honey. Taste: Fruity acidity, creamy and silky, peach, short and dry aftertaste, sweet honey. 86.5
      8. Aroma: Jasmine, honey, apricot, plum, blackberry Taste: Nice floral acidity, silky body, complex flavor, short but clean aftertaste with a fresh background, consistent and uniform. 89
      9. Aroma: Strong almonds, mint, honey. Taste: Astringent, under-ripe background, lacks sweetness, lacks defined acidity, very flat. 78
      10. Aroma: Hazelnut, white chocolate, spiced background, molasses, blackberries. Taste: Fruity acidity, rounded body, blackberries, spices, short and slightly astringent, herbal background. 83
        Back to blog

        Leave a comment

        Please note, comments need to be approved before they are published.