EQUIPMENT LOSS

OEE: Availability, Performance, and Quality

Build an OEE percentage that points to the loss to investigate instead of becoming an unexamined score.

Published
Reviewed

DIRECT ANSWER

The calculation in one paragraph

Overall equipment effectiveness (OEE) equals availability × performance × quality. Availability compares run time with planned production time; performance compares actual output with the theoretical output at the ideal cycle time; quality compares good count with total count. Multiply decimal factors, then diagnose the weakest component.

Open the OEE calculator

Define the three loss buckets

Availability captures stops during planned production time: breakdowns, adjustments, and other events that reduce run time. Performance captures speed losses and minor stops while the equipment is considered running. Quality captures rejected output, including units that require rework if they were not good on the first pass.

The Lean Enterprise Institute describes the same three-part structure and multiplies the factors. Multiplication matters: a strong quality rate cannot cancel a serious availability loss.

Use traceable boundaries and an achievable ideal rate

Agree when planned production time starts, what marks a stop, which count is total count, and what qualifies as good. A shift that excludes changeovers cannot be compared directly with one that treats them as availability loss. Preserve event logs and counter sources with the calculation.

Ideal cycle time should be the fastest sustainable cycle for the defined product and equipment under correct conditions—not a budget rate padded with assumed loss and not a one-off record that the process cannot repeat. For mixed products, calculate performance from product-specific ideal times or equivalent standard units.

Treat OEE as a loss map, not a league table

A single OEE number is useful for trending one constrained asset under stable definitions. It is weak evidence for ranking unrelated machines, plants, or product families because their loading rules, ideal rates, product mixes, and quality definitions differ.

Start improvement with the factor and event category that represents the largest recoverable loss. A lower percentage with accurate loss capture is more actionable than a higher percentage produced by excluding inconvenient time.

WORKED EXAMPLE

Shift-level OEE example

  • Planned production: 450 min
  • Stop time: 45 min
  • Ideal cycle: 0.50 min/unit
  • Total count: 760
  • Good count: 744
Formula(405 ÷ 450) × ((0.50 × 760) ÷ 405) × (744 ÷ 760)82.7% OEE (90.0% × 93.8% × 97.9%)

Performance is the largest factor loss after availability. The factor view tells the team where to open the event data; 82.7% alone does not.

OEE factor reference

FactorFormulaTypical losses
AvailabilityRun time ÷ planned production timeBreakdowns, setup, adjustment
Performance(Ideal cycle × total count) ÷ run timeMinor stops, reduced speed
QualityGood count ÷ total countScrap, start-up rejects, rework
Calculate with decimals, then present each factor and OEE as percentages.

A practical workflow

  1. Fix the asset, product scope, and planned production window.
  2. Calculate run time as planned production time minus stop time.
  3. Calculate availability, performance, and quality from source records.
  4. Investigate any factor above 100%; it usually signals a bad ideal rate, unit mismatch, or count boundary.
  5. Trend the factors and Pareto the underlying losses before selecting corrective work.

EDGE CASES

Frequently asked questions

Can OEE be greater than 100%?

A valid factor should not exceed 100%. A higher result usually means ideal cycle time is too slow, time units do not match, overlapping counts were used, or run time was understated. Correct the data instead of simply capping the published result.

Should planned maintenance be included?

If the asset was not scheduled to produce, it is commonly outside planned production time. If maintenance occurs inside a committed production window, excluding it hides availability loss. Define the calendar rule once and apply it consistently.

Is 85% always world class?

No universal target makes unlike operations comparable. Use a stable baseline, accurate loss definitions, customer requirements, and the economics of the constrained process to set an improvement target.