Professional tasting practice, vocabulary and communication of results
This domain tests professional communication: tasting with a repeatable method, using agreed vocabulary, scoring transparently and writing results that others can act on.
Protocol discipline
Standardise what you can: blind coding, 20-22 degrees C samples, 3-5 g pieces, water and neutral cleansers, rest between samples, no strong flavours beforehand. Record sample code, origin, percentage, batch and date before tasting. Smell, snap, melt, taste and finish in the same order each time so sessions remain comparable.
Language
Separate description from judgement. A descriptive note reads: red fruit 3/5, acidity 4/5, astringency 3/5, clean finish. It does not read: delicious, too bitter. Use recognised descriptors (fruity, floral, nutty, caramel, earthy, smoky, vinegary) with intensity and position in the mouth. Reserve hedonic language for a clearly labelled preference section.
Scoring and reports
Use a fixed scorecard with defined anchors for aroma, flavour, texture, balance, finish and defects. State the panel, reference samples and scale version. In reports, give the finding, the evidence and the likely cause; a defect note should point to a process step such as fermentation, drying, roasting, tempering or storage, not to a person.
Calibration and bias
Calibrate regularly with colleagues and reference samples. Watch for halo effects (one strong note colours everything), order effects, expectation bias from labels and price, and adaptation, where sensitivity drops after repeated tasting.
Worked example. Two tasters score the same bar 6 and 9 out of 10 for overall quality. The difference usually sits in defect weighting and in the weight given to finish. Compare attribute sheets first: if the notes agree but the overall score differs, the scorecard anchors, not the palates, are the problem.
Traps. Writing a tasting note from the label; using smooth and rich as if they were attributes; scoring before the finish; presenting one taster's opinion as a panel result; forgetting to record temperature and sample size.