The Machine Runs, and It Runs Slowly, and It Makes Rejects

Three Ways to Lose Output

A production line fails to produce good units at full rate for three distinct reasons, and measuring only one of them gives a badly misleading picture.

Availability losses occur when the machine is not running: breakdowns, changeovers, material shortages, and waiting for an operator.

Performance losses occur when the machine runs below its rated speed: minor stoppages, slow running, jams cleared in seconds that never appear in any log.

Quality losses occur when the machine produces units that are scrapped or reworked, including startup rejects after every changeover.

Overall equipment effectiveness multiplies the three into a single figure.

The Multiplication Is the Insight

Three dimensions each looking respectable produce a combined figure that is not. That multiplication is why the measure exists and why plants calculating it for the first time frequently find results in the forties or fifties on lines everybody considered to be running well.

Benchmarks generally place world class performance for discrete manufacturing around eighty five percent, which requires roughly ninety percent availability, ninety five percent performance, and ninety nine percent quality.

Where the Hidden Losses Are

The performance component is where most organisations discover something they did not know.

Breakdowns are logged because somebody calls maintenance. Minor stoppages, meaning a jam cleared by an operator in thirty seconds, are not logged anywhere and can consume a substantial share of running time.

Similarly, a line running at ninety percent of rated speed because an operator turned it down to reduce jams looks entirely normal and loses a tenth of capacity permanently.

Those two categories are why automated data collection changes the picture. A machine reporting its own state at high frequency captures stoppages measured in seconds that no manual system records.

What the Number Is For

The measure is most useful applied to a constraint, meaning the process step limiting the output of the whole line.

Improving effectiveness on a non constraint produces no additional output. The line is still limited by the bottleneck, and the extra capacity accumulates as inventory in front of it.

That is the connection between this measure and constraint based management. Measuring everything produces a lot of numbers; measuring the constraint produces a decision.

A related caution is that the measure can be gamed. Running a machine to build inventory nobody needs improves availability and performance while destroying working capital, which is why it should be read alongside inventory and demand rather than alone.

Where the Definition Gets Argued

Several definitional choices materially change the result, and organisations comparing themselves against benchmarks frequently discover they were measuring different things.

Planned downtime, including scheduled maintenance and periods with no demand, is conventionally excluded from the availability calculation, and some organisations include it. Including it produces a lower number that answers a different question, sometimes called total effective equipment performance.

Ideal cycle time determines the performance denominator, and using an optimistic theoretical rate rather than a demonstrated one flatters or penalises the result depending on which was chosen.

Rework may be counted as a quality loss or not, depending on whether the reworked unit is eventually sold.

Comparing effectiveness figures across companies is therefore close to meaningless. Comparing a plant against its own history using a fixed definition is where the value is.

The Behavioural Effect

The measure changes conversations, which is arguably its main contribution.

A maintenance discussion focused on uptime treats a slow running machine as fine. A discussion focused on effectiveness treats it as a loss equivalent to being stopped.

It also makes changeover time visible as a cost rather than as a fixed necessity, which is what drives investment in reducing it, and it connects quality directly to capacity rather than treating scrap as a materials issue.

The Bottom Line

Overall equipment effectiveness multiplies availability, speed, and quality, which produces a number substantially lower than any single dimension suggests and frequently lower than anyone in the plant expected. Its value is in exposing minor stoppages and slow running that no uptime measure captures, and in treating scrap as a capacity loss rather than a materials cost. It should be applied to the constraint rather than everywhere, and compared against a plant own history rather than against a benchmark computed under a different definition.