As many of you know, cupping is a key function of our daily work at Intelligentsia. We use it to decide which coffees to buy, dialing in roasts, quality assurance, blend constitution, and research and development.
Until recently, we’ve used a form that was slightly varied from the Specialty Coffee Association of America’s version. These forms, and most of the other forms out there, use a format where the user can take notes on different sensory elements, and then assign a score at the end.
Our QC and buying teams found limitations with this, and decided to sit down and make a new form. Here’s a link to that form, and to help better explain the features of this form, our QC manager Chris Kornman has the following.
Our new cupping form was built to address a few common problems in coffee evaluation. Some forms tend to rely heavily on an overabundance of attributes and prohibitively complicated mathematics, while others may score too simplistically without giving full attention to readily measurable sensory experiences. Additionally, experienced cuppers and new cuppers alike tend to use a fairly compressed range – just a 10 to 12 point spread on a 100 point scale — to score nearly every coffee they taste. The added difficulty of qualitative distinction (i.e., every 85-point coffee is not the same) is also a problem that lacks a specific solution in many available evaluation forms.
Functionally, our forms are constructed to maximize data capture during cuppings (which, even when performed without a sheet, requires an immense amount of coordination and attention). Sensory evaluation follows the same order as the sheet, so aromatics occur first, followed by a quantitative analysis of sweetness and acidity (the two easiest attributes to evaluate while the coffee is still relatively hot). These first steps are followed by qualitative evaluations of flavor, aftertaste, and complexity. Each of these last three categories is scored on both positive and negative characteristics by the cupper. This is relatively unique attribute of our new cupping form, and a pretty important one: rather than asking the cupper to summarize the score of a flavor in one number, he or she scores the same attribute (e.g., flavor) separately for simultaneous positive (e.g., caramel & cherry) and negative (e.g., bitter & woody) perceptions, resulting in a more holistic record of the experience.
Scoring is performed in 1-point increments, and simple addition/subtraction from a starting number of 80 (the entry level score for a specialty coffee) results in the final score. We’ve found after months of testing that the interaction of positive and negative values allows for both a statistically significant separation between “good” and “exceptional” coffees, and an easily identifiable justification for the final score.
Because cuppings rarely occur in a vacuum, we’ve included a space for notes and scores from the rest of the cupping panel, as well as sufficient real estate for taking notes on data relating to the coffee itself, including important signifiers like origin and roast specs.
Additionally, our Quality database, Cropster, has been customized to accept the values from our new forms. This allows us to preserve our cupping records and share specific cupping trends and valuable feedback with both our roasters and our producing partners across the world. The level of specificity inherent in the cupping form provides more detailed insights into areas for future improvement at the roaster, the storage facility, the dry mill, the washing station, and the farm.