Wednesday, 11:17 a.m. Line 2 stops. The operator waits five minutes, expecting it to restart. At 11:24 he calls maintenance. The technician is on another job and arrives at 11:34. A fouled sensor. Restart at 12:00.
That evening, the production report says "target missed". The next day, an extra hour makes up the volume. By month end, nobody knows the stoppage happened, or that it has happened eleven times since January, always on the same sensor.
Those 43 minutes exist nowhere. Which is exactly why they keep happening.
Key points
- OEE (Overall Equipment Effectiveness) is calculated as Availability × Performance × Quality.
- The classic TPM benchmark puts world-class at 85 %. Plants measuring for the first time usually find a number well below it.
- The number itself is useless. What matters is the reason code attached to each stoppage.
- Before investing in another machine, it is worth knowing how much capacity is already asleep inside the one you own.
What is OEE, exactly?
Overall Equipment Effectiveness measures how much of your available time actually turns into good product. It breaks into three multiplied factors:
- Availability: time the machine could run, over time it should have run. Breakdowns and changeovers live here.
- Performance: actual rate over nominal rate. Micro-stops and slow running live here.
- Quality: good parts over parts produced. Scrap and rework live here.
A work centre at 90 % availability, 95 % performance and 98 % quality shows an OEE of 83.8 %. The multiplication is unforgiving: three individually respectable factors produce a mediocre result, and that is precisely the lesson of the metric.
Why 43 minutes leave no trace
A stoppage is not recorded because there is nowhere to record it. The operator has no field to fill, the technician writes the job in a notebook, and production catches up the next day, which erases the symptom.
The result is a structural blind spot. Management sees a volume, sometimes a delay, never the cause or the recurrence. Eleven 43-minute stoppages on one sensor is nearly eight hours of lost production, split into eleven incidents nobody ever added up.
The six big losses
Total productive maintenance has classified them for forty years, and they hold up in every plant:
- Breakdowns: long, visible, usually already tracked in some form.
- Changeovers and setups: often the biggest loss by volume in high-mix SMEs.
- Micro-stops: one to three minutes, dozens of times a day, never declared. Usually the largest hidden loss.
- Slow running: the machine runs, but below nominal rate, and nobody remembers what nominal was.
- Start-up scrap: the first parts after a setup.
- Production scrap: quality losses in steady state.
How Odoo measures OEE without an "Industry 4.0" programme
Many manufacturers postpone this, assuming every machine must be instrumented. That is not needed to start. Odoo computes OEE from work order tracking, which requires a button, not a PLC. At Prism Technology, we start with:
- Configured work centres with capacity, efficiency and setup time. That is the basis for theoretical time.
- Starting and stopping work orders from the shop floor tablet view: two buttons, no keyboard.
- Downtime reason codes selected when the machine is blocked: breakdown, setup, waiting for material, waiting for operator, quality. This is the most valuable data in the whole setup.
- OEE computed natively per work centre, with history, no custom development.
- Preventive and corrective maintenance: maintenance plans per equipment, MTBF and MTTR tracking, and a maintenance request raised with one button from the operator's screen.
- Optional IoT sensors, when a critical work centre justifies automatic cycle time capture rather than manual entry.
The order matters: start with human capture at the bottleneck, prove the value, then instrument what deserves it. The reverse produces long projects and dashboards nobody opens.
Before buying a machine, look at OEE
This is the most profitable use of the metric, and the most commonly skipped. When a plant saturates, the reflex is to buy capacity. If the bottleneck runs at 58 % OEE, there are 27 points between it and the 85 % benchmark — capacity already bought, already installed, already depreciated.
That does not mean never invest. It means an investment decided without knowing the OEE of the work centre concerned is a bet, not a decision.
The reason code matters more than the number
An OEE figure on a screen gains you nothing. What gains you something is the sorted list of this month's downtime reasons. In most plants, that list shows two or three causes concentrating the bulk of losses, and they are usually mundane: tooling stored away from the station, material arriving late from the store, a setup only one person knows how to do.
None of those is solved by buying equipment. All of them first have to become visible.
Frequently asked questions
Should OEE be measured on every machine?
No. Start with the bottleneck, the work centre that sets the plant's throughput. OEE measured everywhere and used nowhere is data entry for nothing.
What OEE should we target?
The classic 85 % benchmark is a reference point, not a universal target: it means little in a high-mix plant where changeovers are structural. The useful target is your own trend. Gaining five points at the bottleneck in six months beats chasing an absolute number.
Will operators actually declare stoppages?
Yes, on one condition: that declaring visibly gets the annoying thing fixed. A system perceived as individual performance monitoring will be worked around within three weeks. A system that finally gets the fouled sensor replaced sustains itself.
Take action
Prism Technology is an official Odoo partner in Belgium, based in Walloon Brabant, focused on manufacturing, inventory and supply chain. Name your bottleneck: in 30 minutes we show you how to measure its OEE with what you already have, before considering any investment.
👉 Book your 30-minute demo — Contact us
