OEE for Meat Plants: Why the Textbook Formula Doesn't Fit
OEE in its original habitat
Overall Equipment Effectiveness — OEE — was invented for discrete manufacturing. A car assembly line is the canonical example. The formula:
OEE = Availability × Performance × Quality
Availability is uptime as a percentage of scheduled time. Performance is actual throughput as a percentage of design throughput. Quality is good units as a percentage of total units. Multiply them together and you get a single percentage that, in theory, tells you how well the line is running.
Why it falls apart in protein
The meat plant manager who tries to compute OEE for a fabrication line runs into three immediate problems:
- Quality is non-binary. An assembly line either makes a good widget or a defective one. A beef plant produces a continuous distribution of cuts at different grades, with different yields, against orders that demand specific grades. "Quality = 98%" is meaningless when the question is "98% of what — what should we have cut, or what we did cut?"
- Performance depends on raw material. A car line is fed the same chassis every time. A fab line is fed carcasses whose grade, weight, conformation, and ribeye area vary continuously. If today's cattle yield 2% less 112A Ribeye than yesterday's, that's not a performance miss — that's a different problem entirely.
- Availability is confounded with cleanup. USDA-required cleaning, knife sanitation, blade changes — these are not "downtime" in the OEE sense. They're regulatory floor.
What works instead
Most operationally mature meat plants have abandoned classical OEE in favor of three separate measures, tracked daily:
- Yield variance vs. matrix. Actual yields against the yield matrix, by product and carcass class. A −2.1% yield variance on 189A Tenderloin from PRI carcasses is a specific, investigable signal. A composite OEE number is not.
- Throughput vs. line speed × scheduled time. Actual carcasses processed against what the chain speed and schedule would predict. Differences here are real availability/performance issues — usually mechanical or staffing.
- Spec compliance. Percent of fabricated product that meets the buyer's spec (weight range, trim, fat thickness) without rework. This is the real-world equivalent of OEE's quality term, but tied to actual customer-facing acceptance.
How MakeSheet computes them
Yield variance is automatic — the LP knows what it expected; the production write-back knows what was actually produced. Throughput is computed from the carcass count and the day's scheduled hours. Spec compliance comes from the QC log, joined to the IMPS-spec table for each shipped item.
Each of the three has its own dashboard. They're tracked independently because they reveal different problems. Collapsing them into a single number — OEE-style — would hide more than it shows.
The deeper point
Operations metrics work when they tell you what to do next. "OEE = 73%" doesn't tell anyone what to do. "Yield variance on 112A is −2.1% from PRI carcasses for the third day running" tells the QC supervisor exactly where to look. The formula you choose isn't about following a textbook — it's about whether the number leads to action.
