Process Wednesday: taste as a specification

From Derek Walker's really fun book "The Management of Chemical Process Development in the Pharmaceutical Industry", a hilarious story about fuzzy specifications with Bristol-Myers' antibiotic Cephalexin:

Similarly, Banyu in Japan bought Cephalexin from Bristol-Myers' facility in Latina, Italy. In the early days Banyu complained that the taste and odor of Latina's Cephalexin batches varied considerably, with many batches failing to meet Banyu' organoleptic criteria, particularly for taste. Since taste and odor are not quantifiable parameters, the dilemma persisted for some time. During a visit to Latina, Banyu's President, Dr. Iwadare, politely explained that Japanese children actually retched when obliged to take Banyu's Cephalexin oral suspension.... 

...In an effort to resolve the problem, Banyu sent their President's son and their Director of R&D to conduct taste tests on batches of Cephalexin and to "train" Latina people to carry out the taste test to identify Cephalexin batches suitable for shipment to Japan.... 

The taste test itself was a ceremony to behold... It seemed to take on the aura of a religious occasion as each taster sniffed his freshly opened bottle and then carefully spooned a very small amount of powder onto a clean plate. After a debate about the powder's appearance, they lifted their plate to their mouths in unison and licked the sample off. A pause followed with exquisite facial expressions registering their reactions to their palates. The mouth was rinsed and the results of the tasting were written down. Only six licks per day could be undertaken such that it took the Japanese several days to lick their way through the assembled batches. The ruling on acceptable batches and unacceptable batches was handed down and arrangements made to ship the acceptable batches to Japan.  

The use of Latina staff to routinely taste test the Cephalexin batches never happened. Quite apart from the concern that the unions might react to people being used as "tasting guinea pigs." it was agreed that a scientific solution was needed. Indeed, Dr. Visibelli and his staff, working with Latina analysts and production people, found that the presence of acetone in the last process step led to the formation of traces of an unstable Schiff base with Cephalexin which degraded to give unidentified products causing the odor and taste problems. 

I am beginning to learn that this is a classic problem in chemical manufacturing. The customer doesn't like the material, but hasn't developed a chemical test to determine what exactly they don't like about it (or that the test is difficult to master, or that the test is subjective (color, taste, odor.) The manufacturer will say, "if you can't define it/specify, then we can't fix it." Rinse, repeat, until someone actually figures out exactly what is going on and addresses the problem directly. Would that we were all Dr. Visibelli and his team, the true heroes in this funny little story.