Short answer
Key facts
- Formula
- Availability × Performance × Quality
- Computed per
- Machine · job card · part · operator · shift · day
- Availability from
- Logged run vs setup, pause and breakdown time
- Performance from
- Pieces counted vs the part's ideal cycle time
- Quality from
- Rejects logged with a reason at the machine
- Benchmark
- 85% world-class · ~60% typical · ~40% first measurement (OEE.com)
What is OEE and why do job shops get it wrong?
OEE is three questions multiplied together. Was the machine running when it was supposed to be (availability)? When it ran, did it run at the speed the part allows (performance)? Of what it made, how much was good (quality)? The standard definitions and benchmarks are documented at OEE.com; the six big losses behind them come from TPM.
Job shops get it wrong for a specific reason: the inputs live in three places. Run time is in the operator's head, piece counts in a register, rejects at the inspection table. By the time someone types them into Excel, the stops under ten minutes are gone and setup has been booked as production. The OEE that comes out is flattering and useless.
How does OpeEye compute OEE?
| Factor | What it needs | Where OpeEye gets it |
|---|---|---|
| Availability | Planned time, and every minute the machine was not running | Operator taps setup / pause / breakdown with a reason; auto-count box or controller reports run state automatically |
| Performance | Ideal cycle time and pieces made | Ideal cycle from the part master; pieces from operator taps, the auto-count box or the control's counter |
| Quality | Good pieces vs rejects | Rejects logged at the machine with a reason (dimension, tool mark, burr…) |
Because every event carries a job card, an operator and a timestamp, OEE can be sliced the way you actually ask the question: this machine today, this part across the month, this operator this shift, the night shift against the day shift. And because the same events feed the downtime report, the OEE number comes with its reasons attached.
A worked example: one machine, one shift
A turning centre on an 8-hour shift with a 30-minute planned break: 450 minutes planned. The operator logged 62 minutes of stops (a 35-minute setup, a 15-minute tool breakage, 12 minutes waiting for material) — 388 minutes running. Availability = 388 ÷ 450 = 86.2%.
The part's ideal cycle is 48 seconds, so 388 minutes could yield 485 pieces. The machine produced 410. Performance = 410 ÷ 485 = 84.5%. Nine pieces were rejected for a dimension issue: Quality = 401 ÷ 410 = 97.8%.
OEE = 0.862 × 0.845 × 0.978 = 71.2%. That looks respectable — until you notice the 62 minutes of stops alone were worth 77 pieces at the ideal cycle, and the 35-minute setup was planned for 20. Try your own numbers in the OEE calculator.
What the OEE screens show
OEE per machine
OEE per job and part
Shift comparison
Loss breakdown
Daily & shift reports
Wall TV mode
What to do with the first number
Don't chase 85%. Chase the biggest loss. In most job shops that is availability — setups running over plan and short stops nobody used to write down — and it is the cheapest to fix because it needs no new machine. Fix the top reason, watch the number for two weeks, then take the next one.
See your real OEE by the end of the first shift.
Tablets on your machines in an afternoon. Free 15-day trial, on-site across Gujarat, remote anywhere in India.
Questions shop owners ask
What is OEE, in one line?
+
OEE (Overall Equipment Effectiveness) is the percentage of planned production time that a machine spends making good parts at its ideal speed. It is availability × performance × quality: 100% means no stops, no slow cycles and no rejects.
What is a good OEE for a CNC job shop?
+
The common benchmark (OEE.com) is 85% for world-class, around 60% for a typical manufacturer, and around 40% for plants measuring for the first time. Job shops with frequent setups usually sit below high-volume lines, which is exactly why setup time should be tracked separately.
Where does OpeEye get the numbers for OEE?
+
From the floor as it happens: availability from the logged run, setup, pause and breakdown time; performance from pieces counted against the part's ideal cycle; quality from rejects logged with reasons. Nothing is estimated at month-end.
Can OEE be computed per job, not just per machine?
+
Yes. Because every piece, stop and reject is attached to a job card, OpeEye shows OEE per job and per part as well as per machine, shift and day — so you can see which parts run well and which ones quietly lose money.
Do I need sensors to measure OEE?
+
No. Operator tablets give you all three factors from day one. An optional auto-count box or a direct controller connection makes the piece count and run state automatic on the machines where you want that.
Why does OEE from the register always look better than the real number?
+
Because registers round: stops under ten minutes are not written, setup is booked as running, and rejects are counted at inspection rather than at the machine. Logging each event at the moment it happens, with a reason, is what makes the number honest — and useful.
Related
- How to calculate OEEThe formula, a worked example on a turning centre, and a calculator.
- Downtime trackingOne-tap stop reasons, breakdown alerts in seconds, hours lost by cause.
- Production trackingPieces vs target per machine, operator, shift and job — without registers.
- CNC machine monitoringWhat it does, how data gets off the machine, what it costs.