OEE Improvement Playbook

A systematic approach to measuring OEE correctly, identifying your biggest losses, and building an improvement program that sticks.

Version 1 · Updated March 2026

Problem

Most plants that measure OEE find they are running between 45% and 65% — which means 35-55% of their available production capacity is being lost to downtime, speed losses, and quality rejects. That is the equivalent of running a second or third shift worth of output and getting nothing for it. The financial impact is direct: unplanned downtime alone costs the average discrete manufacturer an estimated 5-20% of productive capacity annually. When OEE is low, the default response is to buy more equipment or add overtime, both of which are capital-intensive fixes for what is fundamentally a losses problem.

Step-by-step approach

1
Get the OEE calculation right on your constraint
Start with one machine or line — your bottleneck. Calculate OEE as Availability times Performance times Quality, using planned production time as your denominator, not calendar time. Availability is (planned time minus unplanned stops) divided by planned time. Performance is (actual cycle count times ideal cycle time) divided by running time. Quality is good count divided by total count. Do not skip this rigor. Most plants that claim to measure OEE are miscalculating at least one factor, usually Performance, because they do not have an accurate ideal cycle time.
2
Log every stop and categorize by the Six Big Losses
For two weeks, capture every stoppage on your target line — planned and unplanned. Categorize each into one of the Six Big Losses: equipment failure, setup and adjustment, idling and minor stops, reduced speed, process defects, or startup rejects. Use a paper log at the machine if you have to — do not wait for a digital system. At the end of two weeks, build a Pareto of losses in minutes. Your top one or two categories will account for 60-70% of total lost time. That is where you focus first.
3
Launch autonomous maintenance on the target line
Train your operators to perform first-level maintenance: cleaning, inspection, lubrication, and tightening on their own equipment. Start with a team-based deep clean where operators and maintenance work together to bring the machine back to baseline condition. Then create a simple checklist — 10-15 minutes per shift — that operators own. This is not about offloading maintenance work. It is about catching early signs of deterioration before they become breakdowns. Plants that implement autonomous maintenance properly see a 30-40% reduction in unplanned downtime within the first year.
4
Run a changeover reduction event on your longest setup
Identify your longest changeover on the bottleneck line. Video the full changeover and map every step as either internal (machine must be stopped) or external (can be done while running). Move as many internal steps to external as possible — pre-staging tools, pre-heating molds, preparing materials before the changeover starts. This is the SMED methodology and it reliably cuts changeover time by 30-50% on the first pass. Shorter changeovers mean more available production time and smaller batch sizes, which reduces inventory.
5
Build a weekly OEE review cadence with the production team
Post OEE results daily at the line, broken down by Availability, Performance, and Quality — not just the overall number. Review the trend weekly with the production supervisor, maintenance lead, and quality lead. Focus the review on the top loss from the prior week and what specifically will be done differently this week. Set a 13-week rolling target and celebrate improvement, even if the absolute number is still low. The cadence matters more than the target — consistent attention to losses is what drives OEE from 60% to 85% over 18-24 months.

What good looks like

World-class discrete manufacturers sustain OEE above 85%, but the real differentiator is not the number — it is the system behind it. Top performers have operators who own their equipment condition through daily autonomous maintenance routines, real-time loss tracking that triggers immediate response instead of end-of-shift paperwork, and a weekly review rhythm that treats every percentage point of OEE loss as a problem to solve. They run changeovers in single-digit minutes because they have methodically separated internal and external setup tasks.

Industry median: 60%. Top quartile: 72%.

Common failure modes

OEE improvement programs fail most often because plants start measuring OEE but never build the loss-elimination discipline around it — the number becomes a report that goes to management instead of a tool that drives daily action on the floor. The second failure is trying to improve OEE across the entire plant at once instead of piloting on one constrained line, which dilutes focus and produces mediocre results everywhere. Third, skipping autonomous maintenance and jumping straight to advanced maintenance techniques means the equipment never reaches a stable baseline condition, so every improvement is built on a crumbling foundation. Finally, many plants set an OEE target without breaking it into Availability, Performance, and Quality components, which makes it impossible to diagnose which type of loss is actually dragging performance down.

This playbook is based on:

See how you benchmark on OEE