What OEE actually measures
Overall Equipment Effectiveness (OEE) is a single percentage that tells you what fraction of your planned production time was truly productive. It was developed by Seiichi Nakajima in Japan in the 1960s as part of Total Productive Maintenance (TPM) and remains the standard shop-floor metric across Indian manufacturing today.
The formula is straightforward: OEE = Availability × Performance × Quality. Each of the three components is also a percentage between 0% and 100%, and OEE is the product of all three. A bay running 90% Availability × 85% Performance × 98% Quality has an OEE of 74.97%.
OEE rolls three categories of waste into one number. Availability captures unplanned downtime, set-up changeovers, and breakdowns. Performance captures slower-than-rated running and minor stops. Quality captures rejections and rework. A single OEE number tells the plant head where the worst loss sits this month - which then dictates where the improvement effort goes.
World-class OEE benchmarks (and what is realistic in Indian factories)
The textbook world-class OEE benchmark is 85%, originally derived from Japanese discrete manufacturing data. The components for world-class are typically Availability 90%, Performance 95%, Quality 99.9%. This is genuinely achievable but uncommon outside Japanese-owned plants and a few Indian Tier-1 OEM suppliers we have measured.
Realistic Indian benchmarks from data across the 60+ shop floors Anand has implemented:
- Greenfield Indian factory, first 12 months: 35-50% OEE typical.
- Established Indian SME manufacturer with no measurement: 40-55% OEE typical.
- Indian factory with active TPM/OEE program for 1-2 years: 60-72% OEE typical.
- Indian Tier-1 auto supplier with IATF 16949 discipline: 72-82% OEE typical.
- Indian world-class (rare): 82-88% OEE.
How to calculate Availability
Availability = Run Time / Planned Production Time
Planned Production Time is your shift hours minus planned stops (lunch breaks, scheduled maintenance, planned changeovers). Run Time is Planned Production Time minus unplanned stops (breakdowns, unplanned set-up changes, material shortages, operator absence).
Worked example: a single-shift bay with 8 hours of scheduled production, 30 minutes of lunch (a planned stop), and 60 minutes of unplanned breakdown. Planned Production Time = 480 - 30 = 450 minutes. Run Time = 450 - 60 = 390 minutes. Availability = 390 / 450 = 86.7%.
The trap with Availability is what you treat as 'planned' versus 'unplanned'. Some factories classify all changeovers as planned (which inflates Availability) and others classify the same changeovers as unplanned (which depresses it). The IATF 16949 standard treats only the minimum reasonable changeover time as planned; anything beyond that is an unplanned stop. We use the same convention.
How to calculate Performance
Performance = (Ideal Cycle Time × Total Count) / Run Time
Ideal Cycle Time is the rated cycle time per piece from the equipment manufacturer or the engineering standard. Total Count is the number of pieces produced (good plus rejects). Run Time is what we calculated for Availability.
Worked example: same bay, Run Time of 390 minutes, Ideal Cycle Time of 2 minutes per piece, Total Count of 175 pieces. Performance = (2 × 175) / 390 = 89.7%.
Performance below 100% means the bay is producing slower than rated. Causes are usually: operator unfamiliarity, worn tooling, sub-optimal cutting parameters, micro-stops (small interruptions of less than a few minutes that the operator does not log), or running at a deliberately slower speed because the operator is uncertain about the part.
How to calculate Quality
Quality = Good Count / Total Count
Good Count is pieces that pass first-time inspection (no rework counted). Total Count is everything produced (good + rejects + rework).
Worked example: 175 pieces produced, 7 rejected, 4 reworked. Good Count is 175 - 7 - 4 = 164 (rework is excluded from Good Count because it required extra work). Quality = 164 / 175 = 93.7%.
A common Indian-factory mistake is counting reworked pieces as good. This inflates the Quality component but hides the real cost of poor quality (rework time, rework material, rework planning overhead). Strict OEE always treats first-pass yield as the Quality input.
Combining the three: Total OEE
Combining the three worked examples: OEE = 86.7% × 89.7% × 93.7% = 72.9%.
This bay's OEE of 72.9% places it in the 'established Indian SME with active OEE program' tier. To improve from 72.9% to 80%, the bay needs to lift one or more of: Availability above 90%, Performance above 92%, Quality above 95%. The single biggest gap in this example is Availability - so the improvement focus should be on the 60 minutes of unplanned breakdown. Was it a tool break, a power outage, an operator-induced jam? The OEE number itself does not answer that. The supporting reason-code log does.
The Six Big Losses (Nakajima's taxonomy)
Every reason your OEE is below 100% maps to one of six categories Nakajima identified. Naming the losses correctly is half the battle.
- Loss 1: Equipment breakdowns - unplanned mechanical, electrical, or process failures. Affects Availability.
- Loss 2: Set-up and adjustment time - changeover time between parts beyond the engineered minimum. Affects Availability.
- Loss 3: Idling and minor stops - small interruptions under 5 minutes (jams, sensor faults, material refills). Affects Performance.
- Loss 4: Reduced running speed - operating below the rated cycle time due to wear, conservatism, or material issues. Affects Performance.
- Loss 5: Defects and rework - pieces that fail first-pass inspection. Affects Quality.
- Loss 6: Start-up rejections - pieces rejected during warm-up after a changeover or restart. Affects Quality.
How to actually improve OEE
Knowing your OEE is 47% does nothing. You have to act on it. From customers who have moved OEE 15-25 percentage points in 12-18 months, the pattern is consistent:
- Measure daily, review weekly. Hourly OEE on a wall display is noise; daily totals reviewed every Friday with operators is the right rhythm.
- Pareto the losses. The 80/20 rule applies: roughly 80% of your OEE loss comes from 20% of the reason codes. Focus there first.
- Tie incentives to OEE. A small per-bay monthly bonus (INR 500-1,500 per operator) when OEE crosses a target shifts ownership from 'head office software' to 'our bay's number'.
- Fix root causes, not symptoms. If breakdowns dominate, do machine-hour-based PM (not calendar-PM). If set-up dominates, do SMED. If quality dominates, do FMEA on the failing operation.
- Resist software-only solutions. An OEE dashboard is necessary but never sufficient. The improvement happens at the bay, not in the system.
Frequently Asked Questions
What is a good OEE score for an Indian manufacturing factory?
World-class OEE is 85%, but realistic Indian factory benchmarks are: 35-50% for greenfield first-year operations, 40-55% for established SMEs with no measurement, 60-72% for SMEs with an active OEE program, and 72-82% for IATF 16949 Tier-1 auto suppliers. Above 82% is considered Indian world-class and is rare outside Japanese-owned plants.
How is OEE calculated?
OEE = Availability × Performance × Quality. Availability = Run Time / Planned Production Time. Performance = (Ideal Cycle Time × Total Count) / Run Time. Quality = Good Count / Total Count. Each component is a percentage between 0% and 100%, and OEE is their product. For example, 90% × 85% × 98% = 75% OEE.
What are the Six Big Losses in OEE?
Nakajima's Six Big Losses categorize every reason OEE falls below 100%: (1) Equipment breakdowns, (2) Set-up and adjustment time, (3) Idling and minor stops, (4) Reduced running speed, (5) Defects and rework, (6) Start-up rejections. Losses 1-2 affect Availability, 3-4 affect Performance, and 5-6 affect Quality. Mapping every operator-logged stop to one of these six categories is the foundation of OEE-driven improvement.
Should rework be counted as good production in OEE?
No. The Quality component of OEE should use first-pass yield only. Reworked pieces required extra work and should be excluded from Good Count. Counting rework as good inflates Quality and hides the true cost of poor quality. This is the standard convention in IATF 16949 and TPM methodology.
How long does it take to improve OEE in an Indian factory?
From data across the Indian factories we have implemented, expect 6-9 months to see initial OEE lift from baseline (typically 10-15 percentage points), and 12-18 months to reach a sustained higher level. The improvement curve flattens after 24 months unless capital is invested in better equipment. The fastest gains come from changeover-time reduction (SMED) and operator-incentive alignment with OEE targets.
Can OEE be measured manually without software?
Yes, but it does not scale beyond 2-3 machines. Manual OEE measurement requires operators to log start time, end time, stop reasons, and good/reject counts on a paper sheet, which then gets transcribed to Excel. The data quality drops within a few weeks because operators dislike paperwork. Automated machine connectivity (signals from the PLC or a simple start/stop tablet) captures Run Time accurately and lets operators log only reason codes and reject counts. ERPDrive's OEE module supports both manual and machine-connected modes.
Sources & References
This article cites the following primary sources. Click through for the official source documents: