OEE Benchmarks for Indian Factories: What 'Good' Actually Looks Like

World-class OEE is about 85%. A typical discrete factory runs 55–60%. And only around 6% of plants worldwide ever reach 85% — arguably closer to 3% once you correct for how they measure. Those are the numbers people quote. What almost nobody says out loud: there is no India-specific national OEE dataset to benchmark against, so an Indian plant head comparing to a single global average is comparing to the wrong thing. This page gives you the real distribution, its sources, and a way to judge your own OEE without a national comparator.

This is the numbers page. If you need the definition — what OEE is, the Availability × Performance × Quality formula, and a worked rupee example — start with OEE explained for Indian factories and come back. Here we answer a different question: when your machine scores 62%, is that good?

Key points

• 85% = world-class, 55–60% = typical, 40% = poor. These are the widely-cited TPM bands (oee.com, Evocon).
• World-class 85% is roughly 90% Availability × 95% Performance × 99.9% Quality — three near-perfect factors, not one (Lean Production).
• Only ~6% of plants hit 85%+, likely ~3% after measurement error — across 3,500+ machines in 50+ countries (Evocon). The distribution peaks at 55–60%.
• There is NO India-specific national OEE survey. Vendor pages that publish "India OEE benchmarks" disclose no source or sample — treat them as estimates, not data.
• India's context is structural: an MSME worker produces about 14% of a large-enterprise worker's output (KPMG–CII) — so a mid-size Indian plant's honest OEE often starts low, and that is normal, not shameful.
• The right comparator is you, last month — because the global average was measured on machines nothing like yours.

What "good" actually looks like: the global bands

Every credible source agrees on the same rough ladder. The 85% figure traces to Seiichi Nakajima's TPM work in the 1970s; the companies that won Japan's Distinguished Plant (PM) Prize all scored above 85% (oee.com).

Sources: oee.com, Lean Production, Evocon.

Two numbers anchor the ladder:

• Typical is ~60%, not 85%. "Most manufacturing companies, even today, have OEE scores closer to 60%," and the same source reports seeing more companies below 45% than above 85% (oee.com). A sensor dataset of 3,500+ machines puts the peak of the distribution at 55–60% (Evocon).
• 85% is three near-perfect factors multiplied. World-class breaks down as roughly 90% Availability × 95% Performance × 99.9% Quality ≈ 85% (Lean Production). That 99.9% quality bar is why 85% is so rare — one point of scrap and you're already off it.

How rare is 85%, really?

Rarer than the round number suggests. In a dataset of 3,500+ machines connected across more than 50 countries (May 2023–June 2024), roughly 6% of manufacturing organizations scored 85% or above — and the analysts note the true figure is "likely even lower," nearer 3%, once you strip out plants that over-count because of how they measure (Evocon).

So "world-class" is not a target for next quarter. It's the ceiling perhaps one plant in twenty ever touches. For a mid-size Indian factory, chasing 85% is the wrong goal; moving from 55% to 65% is worth far more money and is actually achievable. (What that 10-point move is worth in rupees: the cost of factory downtime, and how to measure it.)

The honest part: there is no Indian OEE baseline

Here is what separates this page from the vendor blogs. No Indian government body, industry association, or peer-reviewed study publishes a national OEE dataset. The Ministry of MSME, CII and APQC benchmark productivity and competitiveness, but not OEE as a sampled national statistic — competitor-level equipment data is almost never shared openly.

You will find pages titled "India OEE benchmark 2026" giving neat per-sector ranges. We checked one: it lists Automotive Tier-1 at 60–75%, Tier-2/3 at 40–58%, textiles at 40–55%, and so on — but discloses no data source, no sample size, and no methodology. It is vendor-authored estimation, not a survey (TeepTrak). Directionally plausible; not something to quote as fact.

So don't. If someone hands you "the Indian factory OEE average," ask what it was measured on. The honest answer is that the global bands above are the best evidence we have, and they were built mostly on European, American and East Asian plants — not on a Ludhiana forging shop or a Tiruppur knitting unit.

Why an Indian plant's number often starts lower — and why that's normal

The structural backdrop matters, because it means a low baseline is expected, not a verdict on the plant head:

• An MSME worker generates only about 14% of the productivity of a worker in a large enterprise (KPMG–CII, Talent Imperatives for MSMEs).
• Only about 10% of the MSME workforce has formal vocational training, against 50–60% in OECD economies (KPMG–CII).
• Yet the sector still contributes roughly 30.1% of India's GDP and employs about 32.84 crore people (KPMG–CII).

Add high-mix low-volume orders, frequent changeovers, and (very real) grid dips before the DG picks up, and a mid-size Indian discrete plant that honestly measures 50–60% is squarely normal — with a large, recoverable gap. The gap is the opportunity, and it is money: it shows up as what one hour of downtime costs in India and in the downtime, theft and accident statistics for Indian factories.

How to judge your own OEE without a national number

Since there's no Indian yardstick, use these five rules instead of a benchmark:

1. Measure honestly first, compare second. A hand-logged OEE over-reports by roughly 8–15 points because un-logged minor stops inflate the Performance factor (see the OEE explained walkthrough). A "65%" on paper is often really low-60s. Fix the measurement before you trust the comparison.
2. Compare yourself to yourself. The only clean benchmark is your own machine, last month. Beat that. The global 60% was measured on equipment nothing like yours.
3. Benchmark the bottleneck, not the average. A plant-wide OEE average hides the one machine that gates output. A 90% number on a feeder ahead of the constraint is not a win — it's overproduction. Judge the constraint machine.
4. Split the score before you judge it. "62% OEE" tells you nothing on its own. 62% from 70% Availability is a maintenance/changeover problem; 62% from 75% Performance is a minor-stops problem. The factor that's dragging is your target — not the headline.
5. Set a realistic target: +10 points, not 85%. If ~6% of the world hits 85%, it is not your FY goal. Moving your weakest factor up 10 points is achievable, measurable, and pays back.

The recurring theme: OEE benchmarking fails in India not because plants are bad, but because the Performance losses — minor stops and speed loss — are invisible on paper, so the number you compare is fiction. You can't benchmark what you don't measure.

Measure it before you benchmark it

The one factor that quietly wrecks both your score and your ability to compare it is Performance — the two-minute jams and dialled-down speeds no shift book records. Closing that blind spot doesn't require wiring every machine into an MES. A camera already pointed at a cell can tell running from stopped and timestamp every transition, turning "the line felt slow" into a countable record — which is exactly how you spot hidden productivity loss with cameras and run production-line monitoring in India without a PLC on every machine.

That measured downtime is also what makes a benchmark honest: continuous capture counts the stops your logbook rounds off. Want a first, rough number today? Put your machine's downtime minutes and part margin into the downtime calculator and see the rupee gap between your line and a world-class one — then close it a point at a time.

The hard part isn't detecting machine state; it's placement — which camera, at which angle, with a clear sightline to each machine's run-state cue. That's the survey problem Mama solves: record a two-minute phone walk of the floor and get back a plan of which machine each camera should watch, and where a pillar is currently blocking the view.

FAQ

What is a good OEE score for an Indian factory? There is no India-specific official benchmark, so use the global bands: world-class is about 85%, typical is 55–60%, and below 40% is poor (oee.com, Evocon). For a mid-size Indian plant, a measured 50–60% is normal with a large recoverable gap. The most useful target is beating your own last month, not chasing 85%.

Is there an official India OEE benchmark or national dataset? No. No Indian government body, industry association, or peer-reviewed study publishes a sampled national OEE statistic. Pages that advertise "India OEE benchmarks" typically give per-sector ranges with no disclosed source or sample size — treat them as vendor estimates, not data.

What percentage of factories actually reach 85% OEE? Roughly 6% of manufacturing organizations score 85% or above, and the true figure is likely closer to 3% once measurement errors are corrected — across a dataset of 3,500+ machines in 50+ countries (Evocon). World-class is the top few percent, not a routine target.

Why is world-class OEE 85% and not 100%? Because 85% is already three near-perfect factors multiplied: about 90% Availability × 95% Performance × 99.9% Quality (Lean Production). Each factor has irreducible losses — changeovers, minor stops, some scrap — so 100% is a theoretical ceiling no real plant sustains.

My plant scores 55%. Should I be worried? 55% is typical, not alarming — it's near the global peak of the distribution. The useful move is to split it into Availability, Performance and Quality, find the weakest factor, and raise it 10 points. See OEE explained for the split and a rupee example.

How do I benchmark OEE if I can't trust my own number? Fix the measurement first. Hand-logged OEE over-reports by roughly 8–15 points because minor stops go unrecorded, so instrument the Performance factor — with a PLC or a camera watching machine state — before comparing to any benchmark. You cannot benchmark what you do not measure.