Published March 29, 2026 · 10 min read
"What's a good OEE score?" is the most common question plant managers ask after they start measuring. The answer depends entirely on your industry, your equipment, and what you're making. Here's what the real numbers look like across manufacturing sectors — and how to use them without fooling yourself.
The widely cited benchmark is 85% OEE for world-class manufacturing. That number comes from Seiichi Nakajima's original TPM research and has been repeated in every manufacturing textbook since. But here's the problem: most plants aren't even close.
The average OEE across all manufacturing is around 60-65%. That means roughly 35-40% of your planned production time is lost to downtime, slow cycles, or quality defects. The gap between 60% and 85% represents an enormous amount of hidden capacity.
OEE is the product of three independent factors: Availability (are you running?), Performance (are you running fast enough?), and Quality (are you making good parts?). If you haven't measured these before, start with our complete guide to calculating OEE.
But comparing your plastics extrusion line to an automotive assembly robot is meaningless. A pharma filling line that runs at 50% OEE might be perfectly well-run — because half its time goes to mandated cleaning and validation. An automotive stamping press at 50% OEE is in serious trouble.
Context matters more than the number.
The table below shows typical OEE ranges based on published industry data, manufacturing efficiency benchmarks from consulting firms, and our own observations across customer plants. These are guidelines, not absolute standards — your specific equipment, product mix, and operating conditions will shift these numbers.
| Industry | Typical OEE | World-Class | Biggest Loss Category |
|---|---|---|---|
| Automotive / Assembly | 75 – 85% | 90%+ | Minor stops & speed losses |
| Chemical / Continuous | 80 – 92% | 95%+ | Unplanned downtime |
| Plastics / Extrusion | 65 – 80% | 85%+ | Startup scrap & changeover |
| Food & Beverage | 55 – 70% | 80%+ | Changeover & cleaning |
| Cable & Wire | 60 – 75% | 82%+ | Speed losses & breaks |
| Packaging | 50 – 65% | 78%+ | Minor stops & jams |
| Metals / Machining | 55 – 70% | 80%+ | Setup time & tool changes |
| Pharmaceutical | 45 – 65% | 75%+ | Cleaning validation & changeover |
Notice that chemical and continuous processes score highest. That's because they run long campaigns with minimal changeover. Pharmaceutical scores lowest — not because pharma plants are poorly run, but because GMP cleaning and batch record requirements consume a huge portion of available time. OEE benchmark comparison only makes sense within your own industry.
Several structural factors explain why a "good" OEE in one industry would be considered terrible in another:
Continuous processes (chemical plants, refineries, paper mills) run for days or weeks between shutdowns. Every changeover avoided is availability gained. Batch processes (pharma, food, specialty chemicals) stop and start frequently by design. Comparing OEE across these two manufacturing types without accounting for this difference is meaningless.
A packaging line running 15 SKUs per shift will have a fundamentally lower OEE than one running a single product 24/7. The changeover time is real — but it's a business decision, not a failure. If you want to know how to track this properly, see our guide on manufacturing downtime tracking.
Pharmaceutical and food plants lose significant time to mandated cleaning, sanitization, and validation between batches. A pharma clean-in-place (CIP) cycle can take 2-4 hours. That's not downtime you can eliminate — it's the cost of compliance. Some plants exclude regulated cleaning from OEE calculations entirely and measure "productive OEE" separately.
A new high-speed bottling line will naturally run at higher OEE than a 25-year-old CNC lathe with worn ballscrews. Equipment age, product complexity, tolerance requirements, and operator skill all play a role. The most useful OEE benchmark is against your own historical performance, not someone else's plant.
Forget industry averages for a moment. Here's how to establish meaningful OEE benchmarks for your specific operation:
You need at least 30 days of continuous data to establish a baseline. A single shift or a single week is too noisy — you'll catch one good week or one bad week and draw the wrong conclusions. Ideally, collect data automatically from machine signals rather than relying on operator logs. Manual data is always optimistic — operators round up runtime and round down scrap.
A plant-wide OEE average hides everything useful. If you have 10 machines and one is running at 35% OEE while the others are at 75%, the average looks fine at 71%. But that one machine is your bottleneck. Always benchmark per machine and per product.
Break your OEE into the three components and find which one is dragging you down:
A 5% OEE improvement per quarter is ambitious but achievable for most plants just starting to measure. Don't target 85% world-class OEE when you're at 55% — you'll demoralize the team. Instead, set incremental goals: 55% → 60% in Q1, focus on the #1 loss, then 60% → 65% in Q2. Each 5-point improvement compounds. A real-time production dashboard makes these trends visible to everyone on the floor.
OEE is a powerful metric, but it's also easy to game. Here are the traps that undermine real improvement:
The easiest way to inflate OEE is to run longer campaigns with fewer product switches. OEE goes up — but your inventory carrying costs explode, lead times lengthen, and customer responsiveness drops. If your sales team is promising 48-hour delivery on 50 SKUs, running 3-day campaigns to boost OEE is the wrong tradeoff.
Some plants exclude planned maintenance, breaks, or meetings from the OEE calculation. This inflates the number but hides real opportunities. If your planned maintenance takes 4 hours when it should take 2, that's a loss worth seeing. The most honest approach is to measure OEE against total available time and track planned vs. unplanned separately.
A machine running at 95% OEE making a product nobody ordered is worthless. OEE measures equipment effectiveness, not business effectiveness. The goal is profitable output, not a high score. The best plants use OEE to find losses, not as a KPI to be maximized at all costs.
The right approach: Use OEE to identify your biggest losses. Fix those losses. Let the number go up as a consequence, not as a goal.
Most plants that fail at OEE don't fail at the math — they fail at data collection. Operators filling out paper forms or spreadsheets at the end of a shift will always have gaps, rounding errors, and optimistic estimates. The numbers look good on paper but don't match reality.
PulseMQ connects directly to your PLC via MQTT and Sparkplug B and calculates OEE automatically from live machine signals:
No manual data entry. No end-of-shift guessing. The dashboard updates every few seconds, and shift comparison and historical trending let you see whether your improvements are sticking over weeks and months.
PulseMQ calculates OEE automatically from your machine signals. No clipboards, no spreadsheets, no end-of-shift data entry. Live in days, not months.
Get a Free Analysis