A factory owner in Rajkot told me his machines run 24 hours a day, 6 days a week. Business was good, he said, but margins were thin and deliveries were always late. I asked him one question: what is your OEE? He looked at me like I had asked him to recite the periodic table.
His machines were running 24 hours. They were not producing for 24 hours. There is an enormous difference. Between tool changes, material waiting, quality rejections, unplanned breakdowns, and speed losses that nobody tracked, his actual productive output was roughly 38% of what those machines were theoretically capable of. He did not have a capacity problem. He had a visibility problem.
Overall Equipment Effectiveness — OEE — is the single most useful metric that most Indian SMEs do not track. Not because it is complicated, but because it has been wrapped in so much consulting jargon and expensive MES software that small manufacturers assume it is not for them. It is. And you can start measuring it tomorrow with a phone app and a notebook.
What OEE actually is
OEE is a single number that tells you how effectively a machine converts available time into good parts. It combines three factors:
OEE = Availability × Performance × Quality
Each factor captures a different type of loss:
- Availability measures how much of the planned production time the machine was actually running (as opposed to being down for breakdowns, changeovers, or waiting for material).
- Performance measures how fast the machine ran compared to its ideal speed (accounting for slow cycles, minor stops, and idling).
- Quality measures how many of the parts produced were actually good (accounting for scrap, rework, and startup rejects).
Each factor is expressed as a percentage. Multiply all three, and you get OEE as a percentage.
Why all three factors matter
Tracking only one factor gives you a misleading picture:
- A machine with 95% Availability but 60% Performance is available but slow — you are wasting capacity through speed losses.
- A machine with 90% Performance but 70% Quality is running fast but producing scrap — you are wasting material and time.
- A machine with 95% Quality but 50% Availability makes good parts when it runs, but it barely runs — breakdowns or changeovers are eating your time.
OEE forces you to see all three dimensions simultaneously. A factory that only tracks "uptime" can have 90% uptime and still have a terrible OEE of 45% because performance and quality are silently bleeding capacity.
How to calculate each component — with real numbers
Let's use a concrete example from an Indian factory context. We will track a CNC lathe in a Rajkot job shop for one 10-hour shift.
Step 1: Calculate Availability
Availability = Operating Time ÷ Planned Production Time
Planned Production Time is the total shift time minus any planned stops (scheduled maintenance, planned breaks that are formally excluded from production expectations).
| Item | Time |
|---|---|
| Shift duration | 10 hours (600 minutes) |
| Planned lunch break | 30 minutes |
| Planned maintenance (scheduled) | 0 minutes |
| Planned Production Time | 570 minutes |
Now subtract unplanned downtime:
| Downtime event | Duration |
|---|---|
| Breakdown (hydraulic leak repair) | 45 minutes |
| Waiting for material from store | 25 minutes |
| Changeover between jobs | 40 minutes (2 changeovers × 20 min each) |
| Tool change (insert breakage) | 15 minutes |
| Total Unplanned Downtime | 125 minutes |
Operating Time = 570 - 125 = 445 minutes
Availability = 445 ÷ 570 = 78.1%
Step 2: Calculate Performance
Performance = (Ideal Cycle Time × Total Parts Produced) ÷ Operating Time
Or equivalently: Performance = Actual Output ÷ Theoretical Output (at ideal speed during operating time).
The CNC lathe has an ideal cycle time of 4 minutes per part for the product being run.
- Theoretical output in 445 minutes at ideal speed: 445 ÷ 4 = 111.25 → 111 parts
- Actual parts produced: 88 parts
Performance = 88 ÷ 111 = 79.3%
The gap (23 parts) comes from minor stops (operator stepping away, measurement pauses), reduced speed (running at conservative feed rates because of material hardness), and small interruptions that do not count as formal downtime.
Step 3: Calculate Quality
Quality = Good Parts ÷ Total Parts Produced
- Total parts produced: 88
- Rejected parts (dimensional out of tolerance): 4
- Rework parts (needed deburring redo): 2
- Good parts first time: 82
Quality = 82 ÷ 88 = 93.2%
Step 4: Calculate OEE
OEE = 78.1% × 79.3% × 93.2% = 57.7%
The Rajkot CNC Lathe — Complete OEE calculation table
| Parameter | Value | How measured |
|---|---|---|
| Shift duration | 600 min | Fixed |
| Planned breaks | 30 min | Fixed |
| Planned Production Time | 570 min | Shift - breaks |
| Breakdown time | 45 min | Logged by operator |
| Material waiting time | 25 min | Logged by operator |
| Changeover time | 40 min | Logged by operator |
| Tool change (unplanned) | 15 min | Logged by operator |
| Total downtime | 125 min | Sum of above |
| Operating Time | 445 min | PPT - downtime |
| Availability | 78.1% | OT ÷ PPT |
| Ideal cycle time | 4 min/part | From machine spec / time study |
| Theoretical output | 111 parts | OT ÷ ideal cycle time |
| Actual output | 88 parts | Counted |
| Performance | 79.3% | Actual ÷ Theoretical |
| Total parts produced | 88 | Counted |
| Good parts (first pass) | 82 | Inspected |
| Rejected/rework parts | 6 | Inspected |
| Quality | 93.2% | Good ÷ Total |
| OEE | 57.7% | A × P × Q |
This is a realistic number. Not a textbook number — a real one. And it tells the factory owner something paper records never could: his machine is productive only 57.7% of the time it is supposed to be running.
What "good" OEE looks like — and why you should not aim for 85% on day one
The world-class OEE benchmark is 85%. You will see this number in every manufacturing textbook and every consultant's slide deck. It is a useful target for a Toyota or a Bosch plant with decades of continuous improvement, dedicated maintenance teams, and standardised products.
For an Indian SME job shop running custom work with frequent changeovers, an OEE of 85% is not a starting target. It is a fantasy that will demoralise your team if you set it as the goal.
Here is a more realistic progression:
| Stage | Typical OEE | Timeframe | What it means |
|---|---|---|---|
| First measurement (reality check) | 30-45% | Day 1 | This is where most Indian SMEs actually are. Don't panic. |
| Basic fixes applied | 45-55% | Month 1-3 | Low-hanging fruit: better material staging, basic PM, reduced changeover |
| Sustained improvement | 55-65% | Month 3-9 | Systematic attack on top losses, operator involvement |
| Good for a job shop | 65-75% | Month 9-18 | Competitive and profitable for high-mix work |
| Excellent | 75-85% | Year 2+ | Requires disciplined TPM, standardised work, skilled team |
The key insight: going from 35% to 55% OEE is worth far more in rupees than going from 75% to 85%. The first improvement is low-cost and high-impact. The last improvement is expensive and incremental. Focus where the money is.
The six big losses
OEE losses fall into six categories, grouped under the three OEE factors. Understanding these helps you target improvements.
Availability losses
1. Equipment failure (breakdowns). Unplanned stops due to mechanical, electrical, or hydraulic failures. In Indian SMEs, this is often the largest single loss — particularly on older machines where preventive maintenance is reactive ("fix it when it breaks").
2. Setup and changeover. Time lost switching from one job to another — changing fixtures, loading programs, adjusting settings, first-article inspection. In a job shop doing 4-8 changeovers per shift, this can consume 15-25% of available time.
Performance losses
3. Idling and minor stops. Brief interruptions — operator checking dimensions, clearing chips, adjusting coolant, fetching tools. Each stop is small (30 seconds to 5 minutes), but they add up to significant time over a shift.
4. Reduced speed. Running the machine below its rated speed. This happens when operators reduce feed rates because of tool wear, material variation, vibration, or simply because they were trained on conservative parameters and nobody updated them.
Quality losses
5. Process defects. Scrap or rework during normal production. Dimensional errors, surface finish problems, material defects discovered during machining.
6. Startup rejects. Defective parts produced during warmup, after a changeover, or at the start of a new batch. The first 2-3 parts after a setup change are often scrap — and in a job shop with frequent changeovers, this adds up.
How to start measuring OEE without expensive sensors
You do not need IoT sensors, machine connectivity, or a ₹30 lakh MES system to start measuring OEE. Here is the minimum viable approach:
Method 1: Manual logging with a structured form
Create a simple A4 form for each machine per shift with four sections:
- Downtime log (reason + duration for every stop over 5 minutes)
- Parts count (total produced, total good)
- Ideal cycle time (pre-printed for common products)
- Shift hours and planned breaks
The operator fills this in during the shift. The supervisor reviews it at shift end. Someone enters it into a spreadsheet. You have OEE data.
Pros: Zero cost, can start tomorrow. Cons: Relies on operator discipline, data entry delay, same handwriting/accuracy issues as paper job cards.
Method 2: Phone app with manual input
Operators use a phone app to log start/stop times, downtime reasons, and part counts. The app calculates OEE automatically. No sensors needed — the operator is the sensor.
Pros: Real-time data, automatic calculations, no re-entry needed, timestamps are precise. Cons: Requires phones and Wi-Fi on the floor, 10-20 seconds of operator time per event.
This is the method most Indian SMEs should start with. It balances accuracy with practicality. The data is not as precise as machine-connected sensors, but it is 90% accurate — and 90% accurate data that you actually have is infinitely more useful than 99% accurate data from sensors you have not installed.
Method 3: Machine connectivity (future state)
Connecting machines directly to capture cycle times, spindle load, and alarm data. This is the gold standard but requires investment in hardware (IoT gateways, PLCs with OPC-UA capability, or retrofit sensors) and software. Budget ₹15,000-50,000 per machine depending on the machine age and controller type.
Most SMEs should start with Method 2 and upgrade specific bottleneck machines to Method 3 once the OEE measurement habit is established.
Which machine to measure first
Do not start OEE tracking on every machine. Start with one. The right one. Here is how to choose:
Pick your bottleneck machine. The machine with the longest queue of waiting jobs. The one that every other operation waits for. The one that determines your factory's throughput.
If you are not sure which machine is the bottleneck, look for these signs:
- Work-in-progress (WIP) piles up in front of it
- Jobs frequently wait for it while other machines sit idle
- Expediting and overtime are concentrated on this machine
- Delivery delays trace back to this machine's capacity
Start measuring OEE on this machine because every percentage point of improvement here translates directly into more factory throughput. Improving OEE on a non-bottleneck machine is mathematically interesting but commercially useless — the factory output does not change.
Once you have 2-4 weeks of OEE data on the bottleneck, expand to the next 2-3 critical machines. Within 3 months, you should be tracking OEE on your top 5-8 machines — the ones that determine your capacity and delivery performance.
OEE and the six big losses — a practical attack plan
Once you have data, you need to act on it. Here is the prioritisation framework:
First: Attack availability losses
Availability losses are usually the biggest component of poor OEE in Indian SMEs. Two specific targets:
Reduce breakdown time with basic preventive maintenance (PM). You do not need a full TPM programme. Start with:
- Daily 10-minute operator checks (oil levels, coolant condition, unusual sounds, chip buildup)
- Weekly lubrication schedule (written, assigned, checked)
- Monthly critical-component inspection (belts, bearings, hydraulic hoses, way wipers)
A Pune CNC shop reduced breakdown time by 40% in 3 months just by implementing a daily checklist and a weekly lubrication round. No new parts, no new machines. Just consistency.
Reduce changeover time with basic SMED principles. SMED (Single-Minute Exchange of Die) sounds complicated. The core idea is simple: separate what you can do while the machine is still running (external setup) from what you must do with the machine stopped (internal setup).
Example: On a CNC lathe, loading the next program, preparing the next fixture, and staging the raw material can all happen while the current job is still cutting. If your operator waits until the last part is done and then starts preparing for the next job, you lose 15-30 minutes per changeover that could be eliminated.
Second: Attack performance losses
Update cutting parameters. Many Indian job shops run on cutting speeds that were conservative estimates set years ago. A time study often reveals that feed rates and spindle speeds can be increased by 10-20% without quality issues — particularly on newer inserts and with proper coolant.
Reduce minor stops. Track what operators do during those 2-3 minute interruptions. Common culprits: searching for tools (fix with shadow boards), measuring parts (fix with go/no-go gauges at the machine), waiting for crane/material handling (fix with scheduling).
Third: Attack quality losses
First-article inspection discipline. Many startup rejects happen because the operator runs 5-10 parts before checking dimensions. If the first part is wrong, all 10 are scrap. Enforce a mandatory first-article check after every changeover.
Tool wear monitoring. Track the number of parts per insert and change proactively before the tool wears past tolerance. This converts random quality failures into predictable tool changes.
How OEE connects to quoting accuracy and delivery promises
Here is where OEE stops being an academic metric and becomes a commercial tool.
Quoting accuracy
When you quote a job, you estimate machine hours. That estimate is based on ideal cycle times — the time the machine takes per part when everything goes perfectly. But everything never goes perfectly. Your OEE tells you exactly how imperfect reality is.
If your CNC lathe has an OEE of 58%, and the ideal cycle time is 4 minutes per part, the effective cycle time is:
Effective cycle time = Ideal cycle time ÷ OEE = 4 ÷ 0.58 = 6.9 minutes per part
If you quote using the 4-minute ideal time, you underestimate the actual hours by 42%. Your quote is 42% too cheap on the machining component. Your margin evaporates.
If you quote using the 6.9-minute effective time, you price correctly. You might lose a few price-sensitive orders, but you do not lose money on the orders you win. This is the connection between shop-floor metrics and commercial success.
Delivery promises
The same logic applies to delivery dates. If you promise delivery based on ideal cycle times, you commit to a schedule your factory cannot meet. If you use OEE-adjusted cycle times, your delivery promises are realistic — and you hit them consistently.
A Coimbatore precision machining shop improved their on-time delivery from 62% to 84% in four months — not by working faster, but by quoting and scheduling with OEE-adjusted times. They did not produce more. They promised less. And they kept every promise.
| Quoting method | Estimated hours for 500 parts | Actual hours needed | Margin impact |
|---|---|---|---|
| Ideal cycle time (4 min/part) | 33.3 hours | 57.5 hours | 42% cost underestimate |
| OEE-adjusted (6.9 min/part) | 57.5 hours | 57.5 hours | Accurate pricing |
Skipping the theatre
OEE measurement becomes theatre when it stops driving action. Watch for these signs:
- Dashboard tourism. People look at the OEE dashboard, nod thoughtfully, and change nothing. Data without action is entertainment.
- Gaming the numbers. Operators reclassify "unplanned downtime" as "planned maintenance" to boost the Availability number. The OEE improves on paper. Nothing changes on the floor.
- Measuring everything, acting on nothing. You track OEE on 30 machines but have no improvement plan for any of them. Measurement without improvement is bureaucracy.
- Chasing the 85% number. The goal is not a number. The goal is more throughput, better margins, and reliable delivery. If your OEE is 52% and your factory is profitable with on-time delivery, you do not have an OEE problem.
The antidote to theatre is simple: every OEE review must end with one specific action. Not a strategy. Not a discussion. One change that will be made this week on one machine. The change can be small — move the tool cabinet closer, schedule a PM task, update the cutting parameters. If you do one small improvement per week for a year, you will have made 52 changes. Your OEE will be unrecognisable.
Getting started — this week
- Pick one machine. Your bottleneck. The one everyone waits for.
- Define the ideal cycle time for the most common product that runs on it. Check with the operator — they know.
- Track for one week. Use a phone app or a paper form. Log every downtime event over 5 minutes, total parts produced, and good parts.
- Calculate OEE at the end of the week. Do not panic at the number.
- Identify the top loss. Breakdowns? Changeovers? Slow speed? Rejects?
- Fix one thing. The biggest, easiest-to-address loss. Implement the fix.
- Measure again next week. See if the number moved.
That is it. No consulting engagement. No ₹25 lakh software purchase. No six-month implementation timeline. One machine, one week, one number, one action.
Manufacturing productivity in India does not improve through grand programmes. It improves through small, measured, weekly changes on the machines that matter most. OEE is simply the tool that tells you where to look and whether your changes are working.
QuoteERP tracks OEE data from your shop floor and feeds it directly into your quoting engine — so your quotes reflect real machine performance, not theoretical cycle times. Stop guessing. Start measuring. See how it works →