The standard way to measure manufacturing equipment efficiency is Overall Equipment Effectiveness, or OEE. It combines three factors into a single percentage: how much of your scheduled time the machine actually runs, how fast it runs compared to its theoretical best, and how many good parts it produces. An OEE score of 100% means you’re making only good parts, as fast as physically possible, with zero downtime. The world-class benchmark is 85%, though most operations fall well below that.
The OEE Formula
OEE is the product of three components, each expressed as a percentage:
- Availability = Run Time / Planned Production Time
- Performance = (Ideal Cycle Time × Total Count) / Run Time
- Quality = Good Count / Total Count
Multiply all three together and you get your OEE score. If your machine has 90% Availability, 95% Performance, and 99% Quality, your OEE is 0.90 × 0.95 × 0.99 = 84.6%. Notice how three individually strong numbers still produce an OEE below the 85% world-class target. That multiplicative effect is exactly what makes OEE useful: it exposes compounding losses that look minor in isolation.
Each component tells you something different. Availability captures time lost to breakdowns, changeovers, and other stops. Performance captures speed losses, meaning periods where the machine runs but slower than its design speed. Quality captures the cost of scrap and rework. When your OEE drops, you can immediately see which of the three factors is dragging it down and direct your attention there.
What Each Component Measures
Availability
Availability answers the question: of the time you planned to run, how much did the machine actually run? Every minute of unplanned downtime (breakdowns, missing materials, waiting on operators) and planned stops during scheduled production (changeovers, cleaning, adjustments) reduces this number. If you scheduled eight hours of production but the machine only ran for six and a half hours after accounting for all stops, your Availability is about 81%.
Performance
Performance compares actual output speed to the machine’s theoretical maximum. Even when a machine is running, it often runs slower than its design speed due to worn components, poor lubrication, substandard materials, operator inexperience, or frequent micro-stops like jams and misfeeds that last only a minute or two. Performance captures all of that lost speed in a single number.
A critical input here is “ideal cycle time,” the fastest time the machine can produce one unit. This is not an average of past performance. It’s the designed maximum, typically drawn from the equipment manufacturer’s specifications for a given product and material. Using averages is a common mistake that masks real losses. If your products follow different process routings or use different materials, each product type needs its own ideal cycle time.
Quality
Quality is the simplest component: good parts divided by total parts. OEE uses a first-pass yield perspective, meaning any part that needs rework counts as a defect even if it’s eventually salvageable. This includes both defects during steady-state production (from incorrect settings, handling errors, or expired materials) and reduced yield during startup, when machines often produce scrap before reaching stable operation.
The Six Big Losses
Behind the three OEE components sit six categories of loss that account for virtually all equipment inefficiency. Mapping your downtime and waste to these categories is what turns OEE from a number on a dashboard into an improvement tool.
Two losses reduce Availability. Equipment failure covers any unplanned stop: breakdowns, tooling failures, unplanned maintenance, or being starved by upstream equipment. Setup and adjustments covers planned stops during scheduled production, including changeovers, cleaning, warmup time, planned maintenance, and quality inspections.
Two losses reduce Performance. Idling and minor stops are short interruptions, typically a minute or two, caused by misfeeds, material jams, blocked sensors, or obstructed product flow. These are resolved by the operator on the spot. Reduced speed captures periods when the machine runs but below its ideal cycle time, often caused by dirty or worn equipment, poor environmental conditions, or operators still learning the process.
Two losses reduce Quality. Process defects are bad parts produced during normal steady-state production. Reduced yield is scrap produced during startup before the machine reaches stable operation, most commonly tracked after changeovers. Equipment that inherently creates waste after startup, like a web press, will always show some reduced yield loss.
OEE Benchmarks by Industry
The 85% world-class benchmark was established in the late 1990s and remains widely cited, but actual targets vary significantly by industry. Automotive manufacturing, with its high levels of automation and robotics, typically achieves 85% to 95%. Food and beverage operations, constrained by strict hygiene standards and process complexity, land between 70% and 80%. Pharmaceutical manufacturing, with its rigorous quality control requirements, generally falls between 65% and 75%. Other discrete production environments with frequent product changeovers and partial automation often sit between 50% and 65%.
If your plant is running at 40% OEE, that’s not unusual for a facility that hasn’t systematically tracked losses before. It does, however, mean there’s enormous untapped capacity. Many manufacturers discover they can significantly increase output without buying new equipment simply by addressing the losses OEE reveals.
OEE vs. OOE vs. TEEP
The only difference between these three metrics is how you define “available time,” which changes the Availability calculation. All three still multiply Availability × Performance × Quality.
OEE uses scheduled production time as its baseline. If a machine isn’t scheduled to run on weekends, those hours don’t count against you. This makes OEE the best metric for measuring how well you use your planned production time.
OOE (Overall Operations Effectiveness) uses total operations time, which includes unscheduled time like weekends when the plant is open but the machine isn’t assigned work. This gives you a broader picture of how well you use the time you could theoretically be running.
TEEP (Total Effective Equipment Performance) uses all available time: 24 hours a day, 365 days a year. It shows the absolute maximum capacity of your equipment and is most useful for long-term capacity planning. If your TEEP is 35%, you know your equipment is only producing 35% of what it physically could if it ran around the clock making perfect parts at full speed. The gap between TEEP and OEE is your “hidden factory,” the untapped capacity that could delay or eliminate the need for capital investment in new equipment.
Collecting the Data
You can calculate OEE with a clipboard and a spreadsheet or with sensors wired into every machine. The formula is identical either way, but the quality of the data changes dramatically.
Manual collection means operators log start times, stop times, downtime reasons, and part counts, usually at the end of a shift. This introduces three structural problems. Data capture is incomplete because operators miss short stops, forget minor issues, or round times while they’re busy managing the line. Timing is poor because events logged from memory hours later can’t reconstruct exactly what happened. And over time, definitions drift as different shifts and different people categorize the same events differently.
Automated collection pulls data directly from machine controllers and sensors, timestamping every stop and speed change as it happens. It catches micro-stops and small speed losses that manual logging misses entirely, and it applies the same logic consistently across shifts and plants. The tradeoff is cost and configuration complexity. If sensors or logic rules are set up incorrectly, the errors are silent and structural, affecting every calculation until someone discovers and fixes them. Automated systems can also make bad data look authoritative, so you still need people reviewing the output and flagging numbers that don’t match reality on the floor.
For most operations starting out, a manual approach is fine. The act of tracking losses consistently, even imperfectly, surfaces problems that were previously invisible. As you scale up, automated collection pays for itself through better accuracy, real-time visibility, and dramatically less labor spent on data entry and consolidation.
Categorizing Downtime Consistently
The biggest accuracy problem in OEE tracking isn’t the math. It’s inconsistent categorization of stops. If one shift logs a material jam as “equipment failure” and another logs the same event as “minor stop,” your data becomes unreliable and your improvement efforts target the wrong problems.
Planned downtime includes preventive maintenance (cleaning, calibrating, replacing parts), changeovers and setup, employee breaks and shift changes, and scheduled training. Unplanned downtime includes equipment breakdowns, material shortages, power outages, human errors, and safety incidents. Building a standardized list of downtime codes, training every operator on when to use each one, and reviewing the data regularly for consistency is essential groundwork. Without it, your OEE number is just noise.
Visualizing Equipment Efficiency
A good OEE dashboard serves three roles: clarity, context, and responsiveness. For clarity, focus on core KPIs (OEE, its three components, downtime breakdown, and scrap rates) in a consistent layout. Resist the urge to pack in every available metric. For context, use trend lines and heatmaps rather than single-point snapshots. A machine running at 72% OEE today means very little without knowing whether it was at 65% last month or 80%. For responsiveness, the dashboard should update live during production and use color-coded indicators (red, yellow, green) to flag deviations without requiring anyone to read through rows of numbers.
Operators on the floor need a simple view: current OEE, active downtime reason, and recent stop history. Supervisors and plant managers need shift-over-shift and line-by-line comparisons. Executives need plant-level trends and capacity utilization. One dashboard rarely serves all three audiences well, so plan for layered views from the start.
Using Value Stream Mapping Alongside OEE
OEE tells you how efficiently individual machines run, but it doesn’t tell you whether those machines are the right constraint to focus on. A machine at 60% OEE that has plenty of excess capacity relative to customer demand may matter far less than a machine at 80% OEE that’s your bottleneck. Value stream mapping is the tool that fills this gap. It maps the full flow of materials and information from raw input to finished product, identifies where work-in-process piles up, and reveals which equipment is capacity-constrained relative to actual demand. Once you know your constraint, you can direct OEE improvement efforts where they’ll have the biggest impact on throughput and on-time delivery rather than optimizing machines that aren’t limiting your output.