Overall equipment effectiveness
OEE calculation and interpretation: time boundaries, ideal cycles, first-pass quality, reconciled losses, aggregation and improvement decisions.
Tags: production, planning, oee, productivity
On this page
The worked OEE shift excludes 30 min from 480 scheduled min, leaving 450 planned min. 60 stop min leave 390 running min. At an assumed 0.5 min/piece ideal cycle, 600 total and 570 first-pass good pieces give 86.67% availability, 76.92% performance, 95.00% quality and 63.33% OEE using unrounded factors. The direct check is 570 × 0.5 ÷ 450. Retain the same planned-time and good-count boundaries when comparing shifts.
OEE expresses good output at the ideal rate as a share of planned production time. Vorne separates availability, performance and quality to locate losses. Vorne, Calculate OEE
Changes to planned time or ideal cycle can raise reported OEE with no additional conforming output.
Establish a consistent calculation basis
For a single product with one ideal time per piece, use the following relationships:
| Quantity | Calculation |
|---|---|
| Run time | Planned production time − recorded stop time |
| Availability | Run time / planned production time |
| Performance | Ideal cycle time × total count / run time |
| Quality | First-pass good count / total count |
| OEE | Availability × performance × quality |
The factors cancel algebraically to:
OEE = ideal cycle time × first-pass good count / planned production time
Vorne defines ideal cycle time as the fastest cycle achievable under optimal conditions, and good output as passing the process the first time without rework. It identifies a performance value above 100% as a reason to investigate an incorrectly high ideal cycle time. Vorne, definitions and formulas
Keep time units consistent. If the period is measured in minutes and the cycle in seconds per piece, convert one before division. Confirm whether the counter represents cycles, individual pieces, cavities, batches or another quantity. A four-cavity cycle does not produce only one piece merely because the machine emitted one cycle pulse.
Decide what time is included
Document the reporting period, time zone, scheduled production intervals, excluded intervals and stop categories. Planned downtime is not automatically excluded from OEE: a setup inside time intended for production can be an availability loss. Vorne's FAQ specifically includes changeovers and cautions that excluding them hides improvement opportunities. It discusses breaks and maintenance in relation to whether they displace intended production. Vorne, OEE FAQ
Define a stop-duration threshold and keep its treatment consistent. Vorne's Quick Start distinguishes logged stop time, which affects availability, from shorter interruptions captured through performance loss. It also requires an explicit changeover start/end convention. Its suggested thresholds are examples, not universal machine settings. Vorne, Start to Measure OEE, copyright 2016
Prevent overlap and double counting. A recorded stoppage should not occupy two simultaneous loss categories in the time balance. Changing the stop threshold may redistribute loss between availability and performance; with the same planned time, ideal cycle and good count, it should not change the direct overall score.
Original reconciled shift example
Assume an invented 480-minute reporting shift. An agreed 30-minute nonproduction interval is excluded, leaving 450 minutes planned production. During that planned time, recorded stops total 60 minutes. A single product has an assumed 0.5 min/piece ideal cycle. Total output is 600 pieces, including 570 first-pass good pieces and 30 pieces requiring rejection or rework.
- Run time = 450 − 60 = 390 min.
- Availability = 390 / 450 = 86.67%.
- Performance = (0.5 × 600) / 390 = 76.92%.
- Quality = 570 / 600 = 95.00%.
- OEE = 0.866666… × 0.769230… × 0.95 = 63.33%.
The direct check gives 570 × 0.5 / 450 = 63.33%. Do not multiply prematurely rounded factors when reconciling the result.
Expressed as equivalent time, the planned-time balance is:
| Component | Time |
|---|---|
| Availability loss | 60 min |
| Performance loss: 390 − 600 × 0.5 | 90 min |
| Quality loss: 30 × 0.5 | 15 min |
| Fully productive equivalent: 570 × 0.5 | 285 min |
| Total planned production time | 450 min |
The equivalent losses reconcile the calculation. The 90-minute performance residual can represent several causes.
If the same output is reported after incorrectly excluding an additional 60 minutes of setup from planned time, the direct score becomes 285 / 390 = 73.08%. Output has not changed. This original sensitivity example illustrates why a definition change must be disclosed, not treated as physical improvement.
Collect counts and interpret quality carefully
Define where counts occur and which production interval they belong to. Reconcile total, good and rejected quantities, including counter resets, partial batches, startup output and delayed inspection. Do not count a repaired piece as new first-pass good output while also retaining its original rejected production event.
Separate OEE's quality factor from final shipped yield and material yield. The loss basis concerns output at the chosen process boundary. A later inspection finding may require correcting the earlier quality record; document the period and revision rather than silently changing a historical score.
Siemens' Insights Hub OEE documentation explicitly ties downtime totals to machine states and its configured time model. Its theoretical output also depends on the products and their individual design speeds within the selected time ranges. This illustrates why a dashboard number must be interpreted with its configured model, not just its displayed label. Siemens, OEE Standard Formulas
Aggregate without hiding the basis
Do not take a simple average of percentages from unequal periods. For compatible single-product periods, sum fully productive equivalent time and divide by summed planned production time. For example, 100 productive-equivalent minutes / 200 planned minutes and 90 / 100 combine to 190 / 300 = 63.33%, rather than the unweighted 70% average of 50% and 90%.
Mixed products require an explicit rate and aggregation method. Preserve each product's ideal cycle, count and time allocation. Do not multiply one averaged cycle by an unrelated total count or assume that count-based quality weighting equals ideal-time weighting when products have different cycles. State the implemented method before comparing systems.
For a line, define the asset or process boundary whose losses matter to throughput. Vorne's FAQ recommends measuring at the constraint and includes starvation and blocking among relevant external losses. This is guidance for locating improvement effort, not a reason to maximize every machine's output independently. Vorne, measurement location
Use the metric to choose improvements
Retain stop reasons and the three factors alongside OEE. Investigate changes using comparable products, equipment, shifts and definitions. Pair the score with actual good output, demand, due-date performance, safety and cost; producing unwanted inventory to lift utilization is not necessarily useful improvement.
Vorne's benchmark discussion explicitly rejects one universal ideal score and cautions against fixation on the commonly quoted 85% figure. Product mix and changeover frequency affect comparability. Set targets from the process and its opportunities rather than treating a headline benchmark as an acceptance standard. Vorne, World-Class OEE
Record asset, period, time exclusions, stop rules, count sources, first-pass rule, ideal-cycle revision and aggregation. Keep the raw values for reconciliation.
Source versions
Reviewed 2026-11-01. The Vorne Quick Start PDF carries © 2016 in its actual footer; no newer issue date is inferred. The cited Vorne live pages and Siemens rolling manual expose no separately verified technical issue date in the retrieved content. The numerical examples, reconciliation and aggregation cautions are original analytical guidance, not company reporting policy or a universal target.