The Six Big Losses in Manufacturing: Identifying and Eliminating Hidden Costs

Your machines are running, your operators are on the floor, and your OEE still lands well below target. The gap hides in the six big losses: equipment failure, setup and adjustments, idling and minor stops, reduced speed, process defects, and reduced yield. Each one consumes capacity you have already paid for, and most never make it to a shift report. Here is what each one looks like on the shop floor and what it takes to remove it.

Key points

• Six categories account for almost every hour between the capacity you pay for and the output you ship.
• Two losses sit in availability, two in performance, and two in quality, so each maps to a number you can move.
• Idling and minor stops are underreported because operators clear them faster than they can log them.
• Manual shift logs catch breakdowns but miss stoppages under five minutes, whereas automated capture identifies unbudgeted costs.
• On high-mix lines, reducing the number of changeovers is normally the fastest way to recover hours.

What are the six big losses?

The six big losses are the categories of waste that sit between what a machine could produce and what it delivers on the day. Seiichi Nakajima defined them in 1984 as part of Total Productive Maintenance, and forty years on, the six big losses in lean manufacturing still carry the same names (1). Every hour that fails to become a good part belongs in one of them.

The framework earns its place because it is exhaustive. Instead of chasing a vague efficiency problem, you sort a measured gap into six buckets, each with a different owner and a different fix.

How do the six big losses map to OEE?

OEE multiplies availability by performance by quality, and each factor accounts for exactly two losses. If the calculation is new to you, our guide to what OEE is and how to calculate it works through the formula with examples.

That pairing is what makes the six big losses in OEE practical rather than academic. A score of 72% on its own tells you nothing you can act on. The same 72% built from 88% availability, 85% performance, and 96% quality points straight at minor stops and speed loss, and tells your maintenance team where not to spend the week. 

The widely cited world-class target is 85% OEE, built from 90% availability, 95% performance, and 99% quality (1). Very few plants reach it. What matters more is the direction your own score moves over a quarter.

Which losses take machine time away?

Availability losses are the visible ones. Equipment failure covers any unplanned stop: a tooling break, a motor fault, a line starved by upstream equipment, or a line blocked by downstream equipment. Most plants already record these reasonably well, because someone has to call maintenance.

Setup and adjustments cover planned stops: changeovers, warm-up cycles, cleaning, tooling adjustments, and in-process quality checks. Changeovers are the items most plants can compress, which is why the SMED method was built around them (2). The category also grows quietly as product variety increases, so a plant that keeps adding SKUs loses availability without anything breaking.

The fix is rarely faster hands. It is fewer changeovers and better sequencing, grouping jobs by tool, color, or material, so the line moves between similar orders. That is a planning decision, which is why advanced planning and scheduling tends to recover more hours than a stopwatch exercise on the changeover itself.

Why does a running machine still lose output?

Performance losses are quieter and usually larger. Idling and minor stops mean the equipment halts for a minute or two: a misfeed, a jam, a blocked sensor, a quick clean. Nobody calls maintenance, so nothing gets logged, and a stop that happens forty times a shift disappears from every report you read.

Reduced speed is the other half of the pair. The machine runs, but below its ideal cycle time, and the causes are ordinary:

• Worn, dirty, or poorly lubricated equipment.
• Material sitting outside specification.
• Ambient conditions such as heat or humidity.
• Operators running a machine they rarely work on.
• Deliberate slow running to avoid jams that were never fixed.

That last cause deserves attention. Operators often slow a line to compensate for jams that no one fixed, turning a visible loss into an invisible one. Matching people to equipment using a competence matrix removes a fair share of the speed loss that is written off as inexperience.

What do process defects and reduced yield cost?

Process defects are the scrap and rework produced while the process is stable: incorrect settings, handling errors, expired material. Reduced yield covers everything defective from startup until the process settles, including the parts you throw away as routine after every changeover.

Reduced yield ties quality straight back to availability. Each changeover carries a scrap tail, so a line running eight changeovers a week pays that cost eight times, and cutting the changeover count reduces two of the six losses at once.

Quality loss also costs more per unit than any other category, because a rejected part has already consumed material, machine time, and labor. A minute lost to a jam costs you a minute. A part scrapped at final inspection costs you everything spent on it upstream.

How do you find losses that nobody is recording?

Ask three supervisors what their biggest loss is, and you will get three answers, all of them availability losses, because those are the ones people remember. Manual logs capture breakdowns reliably and micro-stoppages almost never, so the categories that hurt most stay invisible.

Automated capture changes that. When machine signals feed the calculation directly, stops under five minutes are counted, speed is measured against the ideal cycle time rather than the estimated one, and every category carries a figure somebody can be held to. OEE monitoring software runs that capture continuously and splits the result by machine, shift, and operator. For a rough baseline, the free OEE calculator works from data you already keep.

Book a 15-minute demo and see your own six big losses broken out by machine and shift, including the downtime reasons your current reports leave out.

Sources

(1) Nakajima, S. (1988). Introduction to TPM: Total Productive Maintenance. Productivity Press. Japanese original published in 1984 by JIPM.

(2) Shingo, S. (1985). A Revolution in Manufacturing: The SMED System. Productivity Press.