Beverage scheduling is sequencing chemistry. Diet after regular needs a rinse; dark after light needs more; allergen or flavor carryover risks need a validated washout. None of this is optional, what is optional is how often the week's sequence forces the expensive transitions. The washout count is a decision the schedule makes, usually by accident.
The rules the good boards already know
- Light to dark: within a family, run pale flavors before dark ones, the rinse is shorter in that direction, and some reverse transitions need a full washout the forward direction avoids.
- Cluster families: citrus with citrus, cola with cola, so intra-family changeovers stay minor and the major transitions happen once, between blocks.
- Diet and sugar sequencing: order matters for brix carryover and rework risk, and the direction rule differs by product pair; the matrix should reflect validation, not habit (measure it from machine data).
- Full sanitations at breaks: the unavoidable long washouts belong at shift end or planned downtime, so they overlap time the line was losing anyway.
What a washout actually costs
A CIP cycle costs line time (30 minutes to several hours), water, chemicals, effluent load, and syrup or product loss in the lines. A sequence forcing one avoidable 90-minute washout per line per week costs a shift of capacity per line per quarter, invisible because each individual washout looked necessary given the order, and nobody prices the order itself.
Why the sequence decays by Wednesday
The Monday board is usually decent. Then a rush order lands mid-week, gets slotted where it fits by due date, and breaks a light-to-dark chain, buying a washout nobody prices at insertion time. This is where propose-and-approve sequencing earns its keep: the AI re-proposes the whole remaining week around the insertion, showing the washout cost of each placement, and the scheduler chooses with eyes open (the scheduling architecture; see it in context on beverage bottling floors).
Measure, then trust the count
Washout durations should come from the machines, valve states and flow, not from standards, because real CIP times drift with product, season, and crew. Once measured, the weekly washout count becomes a tracked KPI of scheduling quality: when it drops while volume holds, the schedule got smarter, and the capacity appeared without anyone cleaning faster.