Sitting in on (and lending a hand to) the Sensory Workshop for craft breweries. It’s an all day event, and right now Annette Fritsch (Boston Brewing) is giving an overview introduction.
More updates to follow as I think of them.
Update: Annette is explaining the importance of separating your preference for a beer from whether or not it tastes like it should. As a panelist in a production facility, you will likely be tasting some beers that you dislike (I know I do), but you must resist the urge to mark the sample down and instead focus on how close the sample is to what it should taste like. Very difficult.
Update 2: Teri Horner (MillerCoors) is covering raw material sensory analysis. They check their dilution water (apearance, taste, flavor) every FOUR hours!!
Update 3: Just before lunch, Amanda Benson (Deschutes) introduced a small handfull of flavor standards and explained their origins and characteristics. Later, they will pass out 3 blinded samples and quiz the attendees as to their identity.
Now, Annette is back and discussing sensory methods for QA.
In/Out method: testing whether samples are in or out of your normal variation (go/no go). Hard to do statistical analysis on this simple data. Intensive for panel leader, as identification of issues can be subjective. +: short, easy to train, simple. -: no descriptive information, subjective, need lots of data, directly impacts the release of products so it can lead to bias (“we need to ship this beer!”, or “I will have lots more work to do if we destroy this beer”).
Degree of Difference:
You know samples are different (via diff tests), degree of difference tests can tell you how different.
+: less time and resources than full descriptive profiling. -: useless unless attributes are specified and panelists are trained, cutoffs must be established.
Some discussion of Difference from Control (an article on this already exists on this site; I’ll skip it here).
Update 4: Next up, Gwen Conley from Port Brewing/Lost Abbey to discuss intermediate/advanced training.
Panel types: acceptance, difference/descrimination, production release, profiling. Most of this talk on descriptive profing. She has little information on her slides, and her oratory is fast and all over the place so we’ll see how much of this I can record…
Now she gives examples of scales and intensity rating for basic tastes. They’ve passed out more tasting samples for a demonstration: ketchup, for example, where flavors are broken apart and rated on their intensities.
Shelf life: hold beer at cold, room temp, and hot (partially for micro reasons, not absolutely necessary). I think hot (85-90F) is useful to an extent, as it can give you a rough idea of how products behave differently which can be useful for solving issues further upstream.
Package/product interactions: can liner material can pick up aromas, so take care where they are stored. “Garbage truck caught fire outside Ball plant, and cans smelled like burnt electrical wire, fish, diapers, etc”.
Update 5: Now welcome Lindsay Guerdrum, from New Belgium, discussing in-process tasting. What to do if malts are mixed in a silo? If a fermentation is moving too slow? Without in-process tasting, you may not catch this until it gets too expensive to fix. It saves money and time and resources to have this type of tasting procedure. It keeps products consistent for you consumers. Validate your panelists continually with spiked samples to ensure they are trained and sensitive. Who to put on this panel? Validated, knowledgable of your products, consistent, unbiased (yeah right), unemotional, communicative people. Define and establish: anomalies, who monitors and communicates issues, where and when in the process it is tasted. Make sure you have support from management and tasters; get them to buy-in. Build a profile of the flavor of your beers at different stages of production for comparison. Write decision trees and SOP’s for different scenarios. Hold a blending panel to determine acceptable blending ratios for anomalous beers. Comunicate results. Track anomalies to watch for patterns and pre-empt further issues.
Update last: the round-table expert panel, with the above folks and Lauren Salazar (New Belgium). I think I’ll just leave it here, since this will be all over the place. If something noteworthy comes up, I’ll mention it.