Where 85% comes from
The 85% OEE world-class benchmark was established in the context of automotive and discrete manufacturing, where Seiichi Nakajima developed the Total Productive Maintenance framework in the 1960s. It reflects what is achievable on a stable, high-volume production line running a single product with predictable cycle times, minimal regulatory constraints, and no mandatory cleaning downtime.
In that context, 85% OEE — derived from approximately 90% availability, 95% performance, and 99.9% quality — is a meaningful and achievable target. The losses it tolerates are equipment breakdowns, minor stops, speed losses, and startup defects. All of these are, in principle, reducible through better maintenance, better standard work, and better process control.
Food manufacturing shares none of these foundational assumptions. A food plant running 10 SKUs per day across an allergen transition will experience CIP downtime, sanitation changeovers, regulatory hold events, and biological raw material variability that are not reducible to zero regardless of how well the plant is managed. Applying 85% as the target ignores this structural reality.
The losses that food OEE carries by design
Three categories of loss in food manufacturing are structural — they exist because of the nature of food production, not because of management failure.
CIP and sanitation downtime is the most significant. A food plant running daily CIP on a wet processing line loses between 45 and 120 minutes of planned production time per CIP cycle, depending on the system size, the CIP sequence, and the cleaning validation requirements. This is planned downtime in OEE terms — it is excluded from the availability calculation because the line was not intended to run during that time. But it consumes calendar time that a discrete manufacturer would use for production, and it caps the maximum throughput the line can achieve regardless of OEE.
Allergen changeovers add a further layer. A line running five allergen transitions per week at 90 minutes each loses 7.5 hours of production time weekly to changeovers alone. If the line runs two 8-hour shifts, that is nearly 10% of available production time consumed by allergen management. This shows up in OEE as availability loss, and it is not fully recoverable — SMED can compress it, but the validated cleaning time is a floor below which it cannot go.
Regulatory hold times are the third structural loss. Microbiological release testing, allergen swab results, and first article quality approval all introduce hold periods between cleaning completion and line restart. A plant waiting 24 hours for microbiological results before releasing a batch is losing availability to a regulatory requirement, not to an engineering failure.
Estimate changeover duration and lost production value for your specific line type and changeover complexity.
Open changeover estimator →Realistic OEE benchmarks by food segment
Industry data from multi-site food manufacturing deployments suggests that realistic world-class OEE targets in food are significantly lower than 85%, and vary considerably by segment.
High-speed beverage filling lines, running a narrow product range with infrequent changeovers and automated CIP, can achieve 70 to 80% OEE. These lines are the closest food equivalent to the discrete manufacturing context in which 85% was established, and they face the fewest structural constraints.
Dairy processing lines typically achieve 55 to 70% OEE. Mandatory 7-step CIP, strict microbiological release requirements, and product variability from biological raw materials all impose availability and quality losses that are not present in other segments.
Wet food processing lines — meat, fish, ready meals — typically achieve 50 to 65% OEE. High allergen complexity, frequent changeovers, intensive sanitation requirements, and raw material variability in weight and composition combine to create availability and quality losses well above those seen in beverages.
Dry food processing — bakery, snack, powder — can achieve higher availability than wet lines because CIP is less frequent, but quality losses from recipe sensitivity and startup defects often offset this. Realistic targets are 60 to 75% OEE depending on product complexity.
What to measure instead of chasing 85%
The most useful OEE metric for a food plant is not the absolute score — it is the trend and the loss distribution. A plant improving from 52% to 58% OEE over 12 months is making meaningful progress regardless of how far it sits from 85%. A plant stuck at 68% OEE for three years with no understanding of where the losses come from is in a worse position even though its absolute score is higher.
Breaking OEE into its three components — availability, performance, and quality — and tracking each separately against segment-appropriate targets reveals where improvement effort should go. In most food plants, availability losses dominate because of CIP, changeovers, and hold times. Performance losses from speed reductions are the second largest category. Quality losses are typically the smallest category but carry the highest cost when they occur because food safety incidents have consequences far beyond scrap rate.
The Six Big Losses framework, adapted for food manufacturing, gives a more useful picture than the composite OEE score alone. CIP cycles, allergen changeovers, and regulatory holds should each be tracked separately as planned downtime categories, distinct from unplanned downtime from equipment breakdowns. This separation makes clear which losses are structural and which are improvable, and prevents management from setting unrealistic improvement targets based on misclassified data.
When a high OEE score is a warning sign
In food manufacturing, an OEE score above the realistic segment ceiling — a wet processing line reporting 82% OEE, for example — should prompt scrutiny of the measurement methodology before celebration.
The most common cause of artificially high OEE in food plants is incorrect ideal cycle time setting. If the theoretical maximum speed used in the performance calculation is set below the line's actual design speed, the performance component will read above 100%, inflating the composite score. This is sometimes called sandbagging — setting conservative targets that the line can easily meet — and it renders the OEE data useless for improvement purposes.
The second common cause is incorrect classification of downtime. If CIP cycles are classified as scheduled breaks rather than planned downtime, they are excluded from the planned production time denominator, making the availability component appear higher than it actually is. The line looks more efficient than it is because the measurement excludes a major source of lost time.
A food plant with genuinely high OEE has typically invested heavily in changeover reduction, predictive maintenance, and real-time performance monitoring. It has accurate ideal cycle times, honest downtime classification, and a documented improvement history. That combination is rare and worth recognising. An unexpectedly high OEE score without that supporting evidence is almost always a measurement artefact.
The right frame for food OEE
OEE in food manufacturing is most useful as a loss accounting tool rather than a benchmark comparison. It answers the question: of the time we planned to produce, how much was truly productive, and where did we lose the rest? That question has value regardless of whether the answer is 52% or 72%.
The improvement target should be set relative to the realistic ceiling for the segment and the specific line — not relative to an automotive benchmark that was never designed for food. A wet food processing line improving its structural floor OEE by 8 percentage points through changeover reduction and unplanned downtime elimination has delivered significant value. Measuring it against 85% obscures that achievement and misdirects the next improvement effort.