Why Performance should never legitimately exceed 100% in an OEE calculation, what causes it to happen anyway, and how to correctly determine a true ideal cycle time.
Why Performance Can Never Legitimately Exceed 100%
Performance = (Ideal Cycle Time × Total Count) ÷ Run Time, and by definition, "ideal" cycle time represents the theoretical fastest possible cycle the equipment is actually capable of achieving — its true design rate, not merely a typical or conservative rate. Since nothing can run faster than its own theoretical maximum, a correctly calculated Performance factor can approach but never legitimately exceed 100%. A Performance calculation returning a result above 100% isn't evidence of unusually excellent performance — it's a direct mathematical signal that the "ideal cycle time" input is wrong.
What Actually Causes This to Happen
A Performance value over 100% occurs specifically when the ideal cycle time entered into the calculation is slower (larger) than the equipment's true design rate — if the actual production data shows the equipment genuinely running faster than the stated "ideal" figure, the math reports an impossible result. This is a data-quality problem, not a calculation error in the formula itself — the formula is working correctly; it's being fed an incorrect input.
Why This Mistake Happens So Commonly
Ideal cycle time is often mistakenly set to a "typical," "average," or deliberately conservative rate rather than the equipment's true theoretical maximum — sometimes because the true design-rate specification isn't readily available or documented, sometimes because a conservative figure is used out of habit or an assumption that it's "safer," and sometimes because the equipment's actual achievable rate has genuinely improved since the ideal cycle time figure was originally established (through maintenance, tooling upgrades, or process improvements) without the reference ideal cycle time being updated to match.
Where the True Ideal Cycle Time Actually Comes From
The correct source for ideal cycle time is the equipment manufacturer's design/nameplate rate specification, or — when nameplate data isn't available or the process has been modified since original installation — the fastest sustained cycle time actually observed under genuinely ideal conditions (no minor stops, no quality issues, running at full design speed) during a careful, deliberate observation period specifically intended to establish this reference figure. This is not the same as an average cycle time across a normal shift (which already includes the minor stops and slowdowns Performance is specifically meant to capture as losses) — using an average cycle time as if it were the ideal cycle time defeats the purpose of measuring Performance loss at all, since it bakes existing losses into the baseline itself.
Why an Inflated Ideal Cycle Time Doesn't Just Cause a Math Error
Beyond producing a nonsensical over-100% Performance readout, using an inflated (too-slow) ideal cycle time systematically understates real Performance losses across every OEE calculation using that same reference figure — even calculations that happen to come out below 100% are still artificially inflated relative to what a correct ideal cycle time would show, masking genuine speed-loss opportunities that a correctly calibrated ideal cycle time would reveal. This is why establishing an accurate ideal cycle time is treated as a prerequisite to meaningful OEE tracking, not a detail that can be approximated loosely without consequence.
How to Correct an Established Ideal Cycle Time Figure
If historical OEE tracking has been using an incorrect (too slow) ideal cycle time, correcting it going forward will produce a step-change drop in reported Performance and OEE scores compared to prior periods — this isn't a sign that equipment performance actually got worse; it's the calculation becoming more accurate. This is worth communicating clearly to stakeholders reviewing OEE trends when a cycle time correction is made, so the resulting score change is understood as a measurement correction rather than a genuine performance regression.