OEE · Overall Equipment Effectiveness
OEE software that measures the machine, not the register
LiveMach OEE software is an industrial IoT tool for Indian manufacturers. Its own sensors, fitted outside the machine, record run time and stoppages. A cloud dashboard turns them into Availability, Performance and Quality, and then OEE, for every shift. It is for plant heads, production managers and factory owners.
OEE is the auditor and consultant word for something you already understand: how much your machines are really producing against what they could produce. LiveMach measures the machine directly, so the number is the machine speaking, not a supervisor remembering.
What is OEE?
OEE stands for Overall Equipment Effectiveness. It is one number, written as a percentage, that tells you how much of your planned production time actually turned into good parts.
It is built from three factors multiplied together: Availability, Performance and Quality.
OEE = Availability × Performance × Quality
Availability
Did the machine run when it was supposed to?
Run time divided by planned production time. Changeovers, breakdowns and no material pull it down.
Performance
Did it run at full speed?
Actual output divided by what the machine should have made at its rated cycle time. Small stops and slow running pull it down.
Quality
Was what it made saleable?
Good parts divided by total parts. Rejects, rework and startup scrap pull it down.
The three multiply, they do not average. This is why OEE drops fast. Three factors that each look respectable on their own can still land you in the fifties.
Try the multiplication yourself
Move the three sliders and watch the shift block shrink. This is the point most people miss about OEE: three good looking numbers multiply into a bad one.
Illustrative calculator
- Planned production time440 min
- After Availability374 min
- After Performance281 min
- Good output269 min
85.0% × 75.0% × 96.0%
OEE61.2%
Availability 85 percent, Performance 75 percent, Quality 96 percent gives an OEE of 61.2 percent. Of 440 planned minutes, 269 minutes produced good parts.
How is OEE calculated? A worked example
Take one CNC lathe on a single 8 hour general shift. These are ordinary numbers from an ordinary day, not a good day.
The shift
- Shift length
- 480 minutes
- Lunch and tea, planned
- 40 minutes
- Planned production time
- 440 minutes
- Changeover for a new job
- 45 minutes
- Breakdown, coolant pump
- 25 minutes
- Waiting for material
- 20 minutes
- Rated cycle time
- 40 seconds per piece
- Pieces produced
- 420
- Rejected at inspection
- 18
The arithmetic
- 1
Run time
440 − 45 − 25 − 20 = 350 minutes
Planned production time minus the three unplanned stops. Lunch and tea are already out.
- 2
Availability
350 ÷ 440 = 0.795, so 79.5%
The machine was available for four fifths of the time you planned for it.
- 3
What it should have made
350 minutes ÷ 40 seconds = 525 pieces
At the rated cycle time, 350 minutes of running is 525 pieces.
- 4
Performance
420 ÷ 525 = 0.80, so 80%
It ran, but slower than rated. Minor stops, a jam, an operator stepping away.
- 5
Quality
(420 − 18) ÷ 420 = 402 ÷ 420 = 0.957, so 95.7%
18 pieces went to the reject bin. 402 were saleable.
- 6
OEE
0.795 × 0.80 × 0.957 = 0.609, so 60.9%
Sixty one percent. Not a disaster, and not unusual.
A cross check that makes it click
There is a second way to reach the same number, and it is the one that makes OEE land. 402 good pieces at 40 seconds each is 268 minutes of genuinely productive work. You planned for 440. 268 divided by 440 is 60.9%, the same answer. The other 172 minutes went somewhere, and OEE is the tool that tells you where.
Illustrative shift, not live data
Where a 480 minute shift actually goes
The same shift from the calculation above, drawn to scale. Every band is real time that was planned for production.
What each band is
- Running350 min
- Changeover45 min
- Breakdown25 min
- No material20 min
- Planned break40 min
480 minutes total
How does LiveMach capture each OEE factor?
We are precise about this, because two of the three come from the machine on their own and the third does not.
| Factor | Where the number comes from | Effort from your team |
|---|---|---|
| Availability | Machine state, read directly from the machine by our sensors. Running, stopped, idle and off, every minute of every shift, night shift included. | None. It records itself. |
| Performance | Production counts from the machine, against the rated cycle time you give us once at setup. | None after setup. |
| Quality | Reject and scrap data. Either the operator enters rejects on the dashboard, or we fit a custom sensor where the process allows it. | This one needs a decision. Operator entry starts quickly. A sensor is fitted per deployment. |
So Availability and Performance are hands off. Quality is not, and we will not pretend otherwise. Most plants start with operator entry, then add a sensor on the lines where rejects matter most.
The reason codes matter as much as the number
An OEE of 61% on its own changes nothing. What changes things is knowing that 45 of the lost minutes were changeover, 25 were the coolant pump and 20 were material. LiveMach records each stoppage with its reason, so the number always comes with the story behind it.
Why does manual OEE tracking almost never work?
Manual OEE fails for one simple reason. The data is written from memory, hours after the event, by somebody whose actual job is running the machine.
The register is filled at the end of the shift
Nobody stops the machine at 11:40 to write down that it stopped at 11:40. The entry is made at 2 pm, from memory, and a 25 minute breakdown becomes "about half an hour".
Reasons get miscoded
When the operator is not sure, everything becomes "breakdown" or "other". Then your Pareto is useless, because the biggest bar means nothing.
Small stops disappear completely
A four minute stop to clear a jam never gets written anywhere. Twelve of those in a shift is 48 minutes, showing up as a Performance loss nobody can explain.
Nobody trusts it, so nobody acts on it
This is the real cost. Everyone in the review meeting knows the numbers are soft, so it becomes an argument about the data instead of a decision about the plant.
OEE makes the register problem sharper, because it multiplies three shaky numbers together and the error compounds.
What are the six big losses in OEE?
The six big losses are the standard breakdown of everything that pulls OEE down. Two hit Availability, two hit Performance and two hit Quality. They come from TPM practice and they map straight onto the three factors.
| Loss | Hits | What it looks like on your floor |
|---|---|---|
| Breakdowns | Availability | Coolant pump fails, belt snaps, motor trips. A fitter is called. |
| Setup and adjustment | Availability | Job change, die or mould change, tool change, first-off approval. |
| Small stops and idling | Performance | Jam cleared in three minutes, sensor blocked, operator away. Never written down. |
| Reduced speed | Performance | Feed dialled down for a worn tool, and never dialled back up. |
| Process defects | Quality | Parts rejected during the run. Dimension drift, surface finish, contamination. |
| Reduced yield at startup | Quality | Scrap between machine start and stable process. Worst on moulding and printing. |
Two of these six usually own most of your lost time. Which two is different in every plant, so guessing does not work.
Is OEE part of TPM?
Yes, OEE is part of TPM. TPM stands for Total Productive Maintenance. OEE is the measurement layer underneath it, the number that tells you whether the TPM work is doing anything.
TPM came out of Japan and reached India through JIPM and TPM Club India. Plenty of Indian plants run it seriously, especially auto component suppliers whose customers ask for it. If your plant has kaizen boards and a TPM steering committee, you are already inside this framework.
The problem is almost always the same. The pillars are running, the boards are updated, and the OEE on the board still comes from a register. So nobody can tell whether last quarter of focused improvement moved anything. A measured OEE gives the kaizen team a before and after they can defend.
We do not sell a TPM programme. We measure the machines so your TPM programme has real numbers to work with.
How do you calculate OEE in Excel?
You can calculate OEE in Excel properly, and if you are starting out you should. Build one row per machine per shift, record nine values, and let three formulas do the rest.
Here is the sheet. Nothing hidden, nothing gated.
Columns to record
- ADate
- BShift
- CMachine
- DShift length in minutes
- EPlanned breaks in minutes
- FUnplanned downtime in minutes
- GRated cycle time in seconds
- HTotal pieces produced
- IPieces rejected
Formulas
- Planned production time
=D2-E2 - Run time
=J2-F2 - Availability
=K2/J2 - Performance
=(G2*H2/60)/K2 - Quality
=(H2-I2)/H2 - OEE
=L2*M2*N2
Column J is planned production time and K is run time, so the later formulas point at them. Format the last four as percentages. Keep a second sheet with one row per downtime event and its reason code, otherwise the number has no story.
Where the spreadsheet stops working
- Every value in column F depends on somebody remembering. That is the weakest link in the sheet.
- There are no timestamps. You know 90 minutes were lost. You do not know when, and without when you cannot find the pattern.
- Reason discipline slips within a month. "Other" becomes the biggest category and the Pareto turns into noise.
- One person maintains it. When they leave the company, it dies.
- It cannot see the night shift honestly, and the night shift is usually where the loss is.
Use the spreadsheet. Learn what the number means on your own machines. The day you start arguing about whether column F is accurate, the spreadsheet has done its job and the machine needs to report itself.
What is a good OEE score?
85% is quoted everywhere as world class. It comes from 90% Availability, 95% Performance and 99.9% Quality multiplied together. Very few plants anywhere sit there for long.
Here is the honest version for an Indian job shop. Plants that measure OEE properly for the first time usually land in the 40s or 50s. That is not a failure. It is the first accurate number they have had, and it is almost always below what the register was showing.
The first win is not the benchmark. It is knowing where the time goes. See that changeover is eating 45 minutes a shift across 12 machines and you have a project. See that one machine has 40 small stops a week and the one next to it has four, and you have a question worth asking.
Reading your own number
Below 50%
Normal for a first honest measurement. Gains here are large and usually cheap.
50% to 65%
Typical for a mixed job shop with frequent changeovers.
65% to 80%
Strong. Changeover and material flow have been worked on seriously.
Above 85%
World class, usually only on a dedicated line running one product.
Compare your machine against itself last month, not against a number from a textbook. That is the comparison that tells you whether anything is improving.
How do you get your real OEE number?
The usual first step is a pilot on one machine or one line. We fit the sensors, you watch the number for two weeks, and then you decide whether it is telling you the truth.
Common questions about OEE
- How is OEE calculated?
- OEE is Availability multiplied by Performance multiplied by Quality. Availability is run time divided by planned production time. Performance is actual output divided by what the machine should have produced in that run time at its rated cycle time. Quality is good pieces divided by total pieces. Worked on the shift above: 79.5% × 80% × 95.7% gives 60.9%.
- What does 85% OEE mean?
- 85% OEE means that out of every 100 minutes of planned production time, 85 minutes worth of good parts came out. It is usually quoted as world class, and it comes from 90% Availability, 95% Performance and 99.9% Quality. For most Indian job shops it is a long term direction, not next quarter target.
- How do I calculate OEE in Excel?
- Record nine values per machine per shift: shift length, planned breaks, unplanned downtime, rated cycle time, pieces produced and pieces rejected, plus date, shift and machine. Then Availability is run time over planned production time, Performance is rated cycle time times pieces over run time, and Quality is good pieces over total pieces. The full column layout and formulas are written out in the Excel section above.
- What are the six big losses in OEE?
- Breakdowns and setup or adjustment losses hit Availability. Small stops or idling and reduced speed hit Performance. Process defects and reduced yield at startup hit Quality. Together they account for everything that pulls OEE below 100%, and most plants find that two of the six own the majority of their lost time.
- Is OEE part of TPM?
- Yes. TPM stands for Total Productive Maintenance, and OEE is the measurement layer under it. TPM gives you the improvement structure through its pillars, and OEE tells you whether that structure is producing results. Indian plants usually meet TPM through JIPM or TPM Club India, and the common gap is that the OEE number on the TPM board still comes from a hand written register.
- Can I get OEE from old machines without changing them?
- Yes. LiveMach reads the machine from outside, using sensors on indicator lamps, contactors, the drive current or a proximity sensor. The machine does not need a network port or any digital output, and nothing inside is cut or rewired. What each machine can give is agreed at a survey before fitting.
- What happens to OEE data if the internet goes down?
- The device keeps logging to its own memory and syncs on its own when the connection returns. The stoppages during the outage are not lost, and the timestamps stay correct.
- How much does OEE software cost?
- There is no public price list. The cost depends on how many machines you fit, which sensors each machine needs and which add ons you want. It is agreed in writing after we understand your machines. The usual first step is a pilot on one machine or one line.
Related reading
What is OEE? The plain guide
The formula, the three factors and a worked example, in simple words.
Machine downtime tracking
Availability is the biggest of the three factors for most plants. Downtime tracking is how you attack it.
CNC machine monitoring
How we read spindle state and production counts off a CNC without touching the controller.
The LiveMach sensor device
What the hardware is, what it reads and how it is fitted outside the machine.
How LiveMach pricing works
Why there is no price list, what decides the cost and how a pilot starts.
About LiveMach
Who builds LiveMach and what it is meant to do for Indian factories.

