The six big losses aren’t a theory. They’re happening on your lines right now. This post gives you a practical breakdown of each one: what causes it, how to spot it, and why some of them stay hidden for months despite costing you more than the obvious stuff.
Introduction
Most plants lose on production capacity daily. Not to power cuts or supply chain disasters, or to major breakdowns. Instead, it’s six predictable, recurring losses that contribute.
What’s more is that most of those losses are happening right now, on your lines, and your current tracking system probably isn’t catching all of them.
The six big losses are a framework given by TPM to manufacturers, so that they stop flying blind. Each one targets a specific part of your OEE score: Availability, Quality, or Performance. Identify the loss, find where it occurs, measure it honestly, and you finally have something you can act on.
This post goes through all six. What each loss looks like on the floor, why some are more obvious while others stay hidden for months, and how to detect them before they quietly drain your ROI, shift after shift.
Where the six big losses come from
The concept comes from Total Productive Maintenance, or TPM, developed in Japan in the 1970s. The idea was simple: before you can eliminate waste, you need a common language for it. Seiichi Nakajima, one of the architects of TPM, identified six recurring ways equipment fails to produce at its maximum potential. Those six categories became the foundation of OEE as we know it today.
Each loss attacks one of the three OEE pillars: Availability, Performance, or Quality. Some are obvious. Some are quiet, which usually prove to be more expensive.
The three pillars and their losses
| OEE Pillar | Loss Category | Loss Name |
|---|---|---|
| Availability | Unplanned Stops | Equipment Failure |
| Availability | Planned Stops | Setup and Adjustment |
| Performance | Small Stops | Idling and Minor Stoppages |
| Performance | Slow Cycles | Reduced Speed |
| Quality | Startup Waste | Start-up Rejects |
| Quality | Production Waste | Process Defects |
Now let’s go through each one.
Loss 1: Equipment failure
Pillar attacked: Availability
This is the one everyone notices. A machine goes down mid-shift. Production stops. People stand around. The maintenance team gets called.
Equipment failure is any unplanned stop lasting long enough to log. The threshold varies by plant, but a common rule is anything over five minutes. Below that, it usually gets recorded as a minor stoppage (Loss 3).
The examples are familiar: a bearing seizes, a conveyor belt snaps, a sensor gives a false reading and trips the line, a motor overheats. The failure itself is often the last event in a longer story. Most catastrophic breakdowns are preceded by weeks of warning signs that weren’t acted on.
How to detect it: Availability loss is the easiest to measure because the line is stopped. The harder question is what caused it. Track failure modes, not just downtime minutes. If your records say “mechanical failure” every time, you’re logging, not learning. OmniOEE captures the timestamp and duration automatically, but the failure category needs to come from the operator or maintenance technician at the time of the event.
Loss 2: Setup and adjustment
Pillar attacked: Availability
Every time you change a product, a format, a mould, a batch or a configuration, the line isn’t producing. That time counts against you.
This is where a lot of plants argue with their OEE numbers. “But we have to do changeovers. That’s just manufacturing.” True. But the question isn’t whether changeovers happen. It’s how long they take and whether that time is being tracked honestly.
Setup and adjustment losses include the changeover itself and everything after: waiting for first-off approval, trimming parameters, warming up to stable output. Some of that tail can last longer than the changeover.
At Emirates Macaroni Factory, Roberto Traversay and his team were initially surprised by how much planned stop time was being absorbed by post-changeover adjustments they hadn’t been tracking separately. The changeover was logged as 20 minutes. The actual return to rated output was closer to 45.
How to detect it: Compare your planned changeover time against actual time to first good unit. The gap is your adjustment loss. If you’re not measuring time-to-first-good-unit, you’re underreporting Loss 2.
Loss 3: Idling and minor stoppages
Pillar attacked: Performance
This is the most underestimated loss on the list. A minor stoppage is a brief interruption, typically under five minutes, where the machine pauses but doesn’t require maintenance. The operator clears it and restarts. Nothing is logged. Repeat this 40 times in a shift and you’ve lost an hour of output that never shows up anywhere.
Examples: a label jams and needs to be pulled, a product falls off a conveyor and triggers a sensor, a fill nozzle drips and gets wiped down, a cap doesn’t seat properly and the reject gate opens. Each one is seconds to a couple of minutes. Individually, they feel negligible. Cumulatively, they’re often the single biggest performance killer on a line.
The reason they go undetected is cultural as much as technical. Operators are trained to keep the line running. Stopping to log a two-minute interruption feels like more effort than it’s worth. So they clear it and move on, and the loss disappears.
How to detect it: Manual tracking almost never captures Loss 3 accurately. You need automated cycle time monitoring. OmniOEE flags any cycle that exceeds the ideal cycle time, which surfaces micro-stoppages that operators wouldn’t log. On some lines, fixing this category alone moves OEE by 8 to 12 percentage points.
Loss 4: Reduced speed
Pillar attacked: Performance
The machine is running. Nothing is stopped. But it’s not running at its designed speed.
This one is subtle because it’s invisible without a baseline. If a line’s rated speed is 120 units per minute and it’s running at 100, nobody necessarily notices. The line is moving. Product is coming out. But you’re leaving 20 units per minute on the floor.
Reduced speed happens for several reasons. Sometimes it’s deliberate: operators slow the line to reduce defects or avoid jams. Sometimes it’s gradual: wear and tear over months causes a slow drift in output that never triggers an alarm. Sometimes the theoretical maximum was set optimistically during installation and the line has never actually hit it.
The “ideal cycle time” problem is real. If your benchmark is wrong, every OEE calculation built on it is wrong. Before you chase Loss 4, make sure your ideal cycle time reflects what the machine can actually do at full capacity, not what it was rated for on a spec sheet ten years ago.
How to detect it: Compare actual units produced per hour against theoretical maximum units per hour. Any gap that isn’t explained by stoppages is speed loss. This requires accurate ideal cycle time data and automated counter readings. Manual tallying at the end of a shift doesn’t give you the granularity to catch this.
Loss 5: Startup rejects
Pillar attacked: Quality
After a changeover, a startup or a long stoppage, the line runs but the first units off don’t meet specification. They’re scrapped or reworked. You produced them, you spent the materials and the energy, but they don’t count.
Startup rejects are highest when parameters are still stabilising: temperature, pressure, viscosity, tension, speed. In food manufacturing, the first run after a product change may involve several batches of product that fail QC checks. In injection moulding, the first shots after a mould change are often dimensionally inconsistent until the tool reaches operating temperature.
Plants that track quality well often separate startup rejects from in-process defects specifically because the root cause and the fix are different. Startup rejects are a process design and changeover quality problem. Process defects are a running stability problem.
How to detect it: Log quality separately by production phase: startup, steady state, end of run. If you only track total shift rejects, you can’t tell whether your quality problem lives in the first 20 minutes or throughout the day.
Loss 6: Process defects
Pillar attacked: Quality
This is defects during normal production. The line is running at speed. No stoppages. But some percentage of output is off-spec.
Process defects include units that are scrapped outright and units that get sent for rework. Both count. Rework still consumed production time, labour and materials to produce a unit that couldn’t ship on first pass.
Common causes: raw material variation, tooling wear, environmental changes (humidity, temperature), operator variation across shifts, or a process that was never as stable as it looked. In FMCG, a common culprit is filling weights drifting outside tolerance. In food, it’s foreign body checks triggering rejects. In packaging, it’s print registration errors.
Quality losses are often treated as a QC department problem. They’re actually a production floor problem. By the time a defect reaches inspection, the loss has already happened.
How to detect it: Defect rate per hour, not just per shift. If your defect rate spikes at shift handover or after a material lot change, you’ll only see it with hourly data. Shift-end totals hide the pattern.
Prioritising which loss to attack first
Not all six losses are equal on every line. Most plants have one or two dominant losses that explain the majority of their OEE gap.
A rough prioritisation framework:
- Start with Loss 1 (equipment failure) if unplanned downtime is regularly disrupting production schedules. This is usually the most visible and the one most likely to get maintenance budget approved.
- Start with Loss 3 (minor stoppages) if your line looks like it’s running but OEE is still low. This is the most common hidden loss in FMCG and food manufacturing.
- Start with Loss 4 (reduced speed) if your line is running without stoppages but throughput is consistently below target. This often surfaces after Losses 1 and 3 are addressed and you can finally see what’s underneath.
- Losses 5 and 6 (quality) deserve separate attention if your first-pass yield is below 98%. Quality losses compound: the defect cost isn’t just the scrapped unit, it’s the production time spent making it.
How OmniOEE helps you tackle all six losses
Most plants that struggle with OEE aren’t struggling because they don’t care. They’re struggling because they can’t see clearly enough to act.
Manual tracking gives you a partial picture. Operators log what they notice. Changeovers get recorded. Major breakdowns get a timestamp. But minor stoppages go unlogged because they clear too fast. Speed drift goes unnoticed because there’s no automated baseline to compare against. Quality defects sit in a separate spreadsheet that never gets connected to production data.
OmniOEE is built to close those gaps across all six categories.
- For Loss 1 and Loss 2, the platform captures unplanned downtime and planned stop events automatically, with timestamps and durations that don’t depend on an operator remembering to log them. Maintenance teams get real-time alerts when a line goes down, so response time drops.
- For Loss 3 and Loss 4, which are the two losses most likely to be invisible in your current system, OmniOEE monitors cycle times continuously against your ideal cycle time baseline. Every deviation gets flagged. Minor stoppages that used to vanish into the shift now show up as a pattern you can analyse and address.
- For Loss 5 and Loss 6, OmniOEE connects quality data directly to production events. If your defect rate spikes in the first 20 minutes after a changeover, you’ll see it. If a specific shift consistently produces more rejects than others, the data will show you that too.
The result is an OEE number you can actually trust, and a loss breakdown specific enough to tell you where to focus.
The bottom line
OEE is only useful if you know what’s pulling it down. The six big losses give you the framework to find out. Equipment failure and setup time are the ones most plants already track. Minor stoppages, reduced speed, startup rejects and process defects are the ones that tend to stay hidden, and they’re often the bigger problem.
Map your losses before you try to fix them. If you want to see how the platform maps to your lines, speak to our experts.