OEE in one line: Availability x Performance x Quality
OEE multiplies three factors into a single percentage. Availability is the share of planned production time the machine was actually running, so it captures unplanned downtime and changeovers. Performance is how fast it ran against its ideal cycle time, so it captures slow running and minor stops. Quality is the share of good parts out of total parts produced, so it captures defects and rework.
Multiply them and you get the headline number. An 85 percent OEE is widely cited as world-class, which is roughly 90 percent availability, 95 percent performance, and 99 percent quality. Most plants start far lower, around 60 percent, which means there is real room to recover. Use the OEE calculator to baseline your line before you start, then attack the weakest factor first.
Cut unplanned downtime with preventive and predictive maintenance
Unplanned breakdowns are the single biggest drain on Availability. A reactive, run-to-failure culture guarantees that machines stop at the worst possible moment. Shifting to a planned regime, time-based preventive maintenance for wear parts and condition-based predictive maintenance for the rest, moves stoppages into scheduled windows where they cost far less.
Start by ranking assets on how often they fail and how much each stop hurts throughput, then build maintenance plans for the worst offenders first. Track mean time between failures so you can prove the regime is working and tune the intervals over time.
Speed up changeovers with SMED
Every changeover is downtime, so it eats Availability directly. Single-Minute Exchange of Die, the method Shigeo Shingo developed at Toyota, targets changeovers under ten minutes by separating internal setup, which must happen while the machine is stopped, from external setup, which can be prepared while it still runs.
Convert as much internal work to external as possible: stage tooling and materials in advance, pre-heat or pre-position dies, and standardize fasteners. Teams that apply SMED routinely cut changeover time by half or more, which frees capacity without buying a single new machine.
Eliminate minor stops and idling
Small, frequent stoppages, a jam, a misfeed, a sensor fault cleared by an operator in under a couple of minutes, rarely get logged, yet they quietly erode Performance. Because each one is short, they hide in the gap between recorded run time and actual output, and they add up to hours across a shift.
Hunt them by watching the line, not the reports. Tally every micro-stop for a week, find the recurring root causes, and engineer them out with better guarding, guides, or feeder tuning. Eliminating chronic minor stops is often the fastest single Performance gain available.
Run at the ideal cycle time, stop slow running
A machine that runs but runs slow loses Performance just as surely as one that stops. Speed loss creeps in through worn components, conservative operator settings, poor material flow, or a line that has quietly drifted below its rated cycle time and nobody reset it.
Define the ideal cycle time for each product on each machine, then measure actual speed against it continuously. When the gap appears, treat it as a defect to be diagnosed, not a number to be accepted. Restoring rated speed often recovers double-digit OEE points with no capital spend.
Reduce defects and rework
Every scrapped or reworked part is time and material the line spent producing nothing sellable, which is exactly what the Quality factor penalizes. Startup rejects after a changeover and steady-state defects during a run both count, so both deserve attention.
Use mistake-proofing (poka-yoke) to make defects impossible rather than merely detectable, apply statistical process control to catch drift before it produces scrap, and run structured root-cause analysis on the defects you do see. Stabilizing the process at startup alone can lift Quality noticeably.
Standardize work and train operators
Inconsistent methods produce inconsistent OEE. When every operator runs the line a little differently, speed, stop frequency, and defect rates swing from shift to shift, and improvements made on one shift evaporate on the next. Standardized work, the documented best-known method for each task, locks in gains across the whole team.
Capture the best current method, train every operator to it, and make the standard visible at the machine. Then improve the standard, not the individual: each refinement becomes the new baseline for everyone, so OEE ratchets up instead of bouncing around.
Add condition monitoring and IoT sensors
You cannot predict a failure you cannot see coming. Low-cost sensors for vibration, temperature, current draw, and cycle counts turn maintenance from a calendar guess into a data-driven decision, which is what makes predictive maintenance possible at scale. They also catch speed loss and developing faults long before an operator would notice.
Instrument the assets that hurt most when they fail, stream the readings to a central system, and set alert thresholds tied to real failure modes. The same telemetry that warns you of a bearing about to seize also feeds the automatic OEE measurement in the next step.
Measure and visualize OEE continuously
You cannot improve what you do not measure. Manual, end-of-shift OEE logs are late, incomplete, and easy to game, which is why micro-stops and speed loss stay invisible. Automatic, real-time measurement from machine signals gives you an honest number and, more usefully, the breakdown of where the losses sit.
Put a live OEE display on the floor so the team sees the score and the top losses as they happen, and review the same data in management dashboards for trends. Continuous measurement closes the loop on every other lever on this list: it tells you which factor to attack next and proves whether the last change worked.
Tracking OEE in Odoo
Most of these levers need the same thing: trustworthy production data in one place. Odoo Manufacturing (MRP) captures work-order durations, planned versus actual run time, and scrap, while Odoo Maintenance manages preventive plans and logs equipment downtime against the same assets. Together they give you the raw inputs for Availability, Performance, and Quality without a separate system. Octura configures work centers, capacities, ideal cycle times, and maintenance triggers so your OEE number reflects what is really happening on the floor.
Explore Odoo Manufacturing (MRP) →References
- OEE.com, the OEE primer on Availability, Performance, and Quality. Definitions, the world-class 85 percent benchmark, and the six big losses. oee.com
- Seiichi Nakajima, Introduction to TPM: Total Productive Maintenance (Productivity Press). The foundational text that defined OEE and the Total Productive Maintenance framework. productivitypress.com
- Lean Enterprise Institute, lean manufacturing and SMED resources. Standardized work, quick changeover, and continuous improvement methods. lean.org