OEE Stuck at 60%? —— Six Losses Elimination in Five Steps
Many companies can calculate their OEE and break down the losses, but after half a year, the numbers remain unchanged. The reason is simple: they treat OEE as a "report" rather than an "improvement map." OEE is the product of time utilization rate, performance rate, and quality rate. If any one of these rates is held back, the overall number will not improve. Behind these three rates are six major losses: breakdowns, changeovers, minor stops, reduced speed, defects, and startup losses. The numbers are just a health check; what truly boosts OEE is the closed loop of "loss elimination": accurate calculation, attribution, target selection, root cause analysis, and standardization.
This article uses a case study from a machining workshop to break down how these five steps are executed.
1. Accurate Baseline Calculation: Let the Numbers Speak the Truth
The first step in OEE improvement is not to make changes but to standardize the data. The time utilization rate looks at "how much actual machine running time there is within the planned time," considering changeovers, breakdowns, and downtime as losses. The performance rate examines "how much slower the actual cycle time is compared to the theoretical cycle time," where minor stops and speed losses are hidden. The quality rate focuses on "the proportion of products that are correct the first time," managing defects, rework, and startup losses. These three rates should be calculated separately for each shift and each machine, with continuous recording for more than two weeks before establishing a baseline. Avoid using the best single-day performance as a reference.
Many companies fail at this first step: data is often filled in by the team leader based on memory at the end of the shift, and losses are averaged out, masking the true issues. A changeover that takes 40 minutes once and 5 minutes another time might average to a harmless 22.5 minutes, but these are two entirely different scenarios. Manually estimated running times are often inflated, leading to a baseline that is 10 percentage points higher than reality, which invalidates all subsequent improvements. Therefore, data sources should ideally be the machine's own operation logs, inspection records, and production ledgers. If these are unavailable, it's better to assign a dedicated person to record data for a week rather than rely on estimates. Once the data is standardized, use a Pareto chart to rank the losses by hours. The top three losses typically account for more than 70% of the total, making the primary focus areas clear.
2. Loss Attribution: Break Down the Numbers to Workstations and Time Periods
With the baseline established, the next step is to attribute each of the six losses: which machine experiences breakdowns, how often, and how long each repair takes; which shift and product type are most affected by changeovers; and which time periods see the most minor stops and how long each stop lasts. The more detailed the attribution, the more precise the improvement.
The workshop's baseline OEE was 62%. The breakdown showed that changeovers accounted for the most loss at 38 hours per month—averaging 45 minutes per changeover, with over 50 changeovers each month. Minor stops came second at 25 hours per month—occurring about 20 times a day, each lasting one to two minutes. Defects and rework accounted for 12 hours, while the rest were speed losses and startup losses. During attribution, it's important to distinguish between "equipment issues" and "management issues": long changeover times are often not due to slow equipment but to disorganized tools, lack of standardized procedures, and insufficient preparation. These losses can be reduced without spending money on new equipment.
3. Target Selection: Focus on One or Two Losses at a Time
Trying to tackle too many losses at once will lead to failure. Score and rank the losses based on "loss hours × improvement difficulty × impact area," and focus on one or two primary targets at a time, thoroughly addressing them before moving on to the next. The workshop chose to target minor stops first: each stop lasted only one to two minutes, but with 20 stops a day and a total of 25 hours a month, the improvement was easy, quick, and most likely to build confidence. Changeovers were selected as the second target but with a slower pace to avoid competing for resources in the first round.
The principle is to tackle "high-frequency small losses" first, then move on to "low-frequency major losses." High-frequency small losses occur daily, and even a small improvement can be seen in the numbers immediately. Low-frequency major losses, while significant in single instances, have a longer improvement cycle and slower verification, which can demoralize the team if tackled first. Creating a positive feedback loop in the improvement process helps the team sustain long-term efforts.
4. Root Cause Elimination: Use 5Why to Dig Deep
After selecting the targets, use the 5Why method to trace the root causes, stopping only when you reach the management level of "people, machines, materials, methods, and environment." For example, with "20 minor stops a day," the questions might be: What are the machines doing when they stop? —— Waiting for materials. Why are they waiting for materials? —— The upstream process did not notify of the changeover. Why was there no notification? —— There was no signaling mechanism. Why was there no signaling mechanism? —— The logistics rules were missing. The root cause is identified, and the solution involves two steps: a light signal 10 minutes before the upstream changeover, and a safety stock line on the material rack that triggers automatic material requests when below the line.
Two weeks later, minor stops were reduced from 20 times a day to three, and OEE improved by nine percentage points in a single month. Two key points to note: first, solutions should be implemented through poka-yoke and visualization, relying on lights, lines, and standards rather than "increased responsibility." Second, distinguish between short-term and long-term solutions—light signals are a short-term fix, while integrating changeover times into production plans and establishing pull-based material request rules are long-term defenses. Both should be included in the action plan.
5. Standardization and Review: Prevent Regression
Once improvements are effective, they must be standardized to prevent regression after three months. The new changeover process should be documented in the standard work instructions, the light signal rule included in the shift handover inspection form, and the safety stock line marked on the material rack. New employees should be trained on these procedures. The OEE board should be updated by shift, with immediate response to anomalies, rather than waiting for weekly meetings. Review the losses weekly: re-rank the Pareto chart to see if the targeted losses have truly decreased and if they have shifted to other areas, then select the next round of targets.
Three months later, the workshop's OEE stabilized around 81%, with changeover losses reduced by more than half due to quick changeover improvements, and minor stops not regressing. The significance of the review lies in anchoring the results of each improvement in the process, ensuring that the next round of improvements builds on a new baseline.
The six losses are the targets, and the five-step method is the gun—every OEE improvement is hidden in the closed loop of loss elimination.
Knowledge code: 7.3.1
Version: v20260818
Author: Quality Think Tank
Quality Think Tank is dedicated to providing systematic knowledge, methodologies, and practical tools for quality management professionals, helping companies continuously enhance their quality capabilities.