Phnom Penh, Cambodia · June 2022

Juniper Contamination

The blast radius was not determined by the severity of the failure. It was determined by how long the detection gap allowed it to run.

In June 2022, I received a 50kg juniper lot from a Vietnam-based supplier, and it went straight into production without any quality gate at receipt. The lot was used across four consecutive batches between July and August 2022, spanning three products. All four batches initially passed post-production quality control. The contamination was real, but the protocol had no way to detect the same contaminated lot moving through consecutive batches and different products. Detection came 56 days after the lot was received, through an unplanned batch comparison during a routine tasting session. By then, 715 bottles had been filled and distributed.

The investigation confirmed the contamination was isolated to a single lot. I quarantined all four batches on the same day the root cause was confirmed. I did not act on initial suspicion. I acted on confirmed evidence. The structural fix was to test every new botanical lot through single-botanical distillation before it entered production. This moved detection upstream and added one to two days to the cycle, which was the right trade.