OEE vs TEEP vs Utilization: What Each Metric Measures

Three questions, three denominators

• Utilization (also called loading) = Planned Production Time / Calendar Time. Calendar time is all of it — 24 hours a day, 7 days a week, 8,760 hours a year. Utilization measures a scheduling decision: one 8-hour shift five days a week is about 24% utilization before the machine loses a single minute to anything.
• OEE = Availability x Performance x Quality, measured against planned production time only. Time you never scheduled doesn't count against it. OEE measures execution: given the hours we committed, how much fully productive output did we get?
• TEEP (Total Effective Equipment Performance) = OEE x Utilization, equivalently measured against calendar time. TEEP measures the machine as an asset: of everything this machine could theoretically ever produce, what fraction are we getting?

The denominators are the whole story. OEE deliberately forgives unscheduled time — that's a feature, because the crew running the shift doesn't control whether a second shift exists. TEEP deliberately forgives nothing, because the capital tied up in the machine doesn't care about your shift pattern.

Worked example: one machine, one week

A machining center runs one 8-hour shift, five days a week. Calendar time for the week is 168 hours.

1. Planned production time: 5 shifts x 8 hours = 40 hours, minus 2.5 hours of scheduled breaks = 37.5 hours. Utilization = 37.5 / 168 = 22.3%.
2. During those 37.5 hours: 5.5 hours of unplanned stops (breakdowns, changeovers, waiting). Availability = 32 / 37.5 = 85.3%.
3. Output at ideal rate should have taken 27 of those 32 run hours. Performance = 27 / 32 = 84.4%.
4. 97% of parts were good first time. Quality = 97.0%.
5. OEE = 0.853 x 0.844 x 0.970 = 69.8%.
6. TEEP = OEE x Utilization = 0.698 x 0.223 = 15.6%.

Same machine, same week: 70% OEE, 16% TEEP. Neither number is wrong and neither is 'the real one' — they're answers to different questions. The crew earned a respectable 70% on the time they were given. The asset delivered 16% of its theoretical maximum, which means the biggest capacity reserve isn't on the shop floor at all — it's the 128 unscheduled hours.

Tip This is the classic pre-capex check. If TEEP is 16%, 'we're out of capacity' usually means 'we're out of first-shift capacity.' A second shift, weekend running, or lights-out operation on proven jobs is almost always cheaper than a new machine.

Which metric for which decision

• Improving shop-floor execution: OEE, broken into its three factors. It's the only one of the three the crew on shift can actually move, which also makes it the only fair one to put in front of them.
• Deciding whether to buy another machine: TEEP first. Low TEEP with decent OEE means buy hours (shifts), not iron. Low TEEP with low OEE means fix execution before spending anything.
• Quoting lead times and loading the schedule: utilization, projected forward — but discount scheduled hours by demonstrated OEE. A 37.5-hour week at 70% OEE yields about 26 hours of good output at ideal rate; quoting as if all 37.5 will materialize is how backlogs are born.
• Justifying capital already spent, or comparing sites' use of similar assets: TEEP, because it's the metric denominated in the same units the capital was — total time the asset exists.

One fairness rule worth enforcing: never target the shift crew on TEEP or utilization. Both are dominated by scheduling and demand decisions made above their pay grade. Handing a crew a 16% TEEP as if it were their score is how metric programs lose the floor.

Common mix-ups and how to spot them

• 'Our OEE is 24% because we only run one shift.' That's utilization leaking into OEE. Unscheduled time is excluded from OEE by definition; if it's in your number, your planned-time definition is wrong.
• 'The machine is 95% utilized, we need another one.' Utilization here usually means 'scheduled nearly every first-shift hour' — check OEE before signing anything. A machine scheduled 95% of one shift at 60% OEE has a third of its scheduled capacity recoverable for free.
• Comparing your OEE to a competitor's TEEP (or vice versa). Cross-company numbers rarely come with their denominators attached; a plant quoting 85% against planned time and one quoting 55% against the calendar may be identical operations.
• Moving time between 'planned' and 'unplanned' to dress up whichever metric is being reviewed this quarter. Fixed written definitions are the only defense — and the same shift data should produce all three metrics, so they can't drift independently.

The healthiest setup is to compute all three from the same shift log and read them together: utilization tells you how much calendar you bought, OEE tells you how well you spent it, TEEP tells you what's left in the asset. Any one of them alone can be gamed; the trio keeps each other honest.

Common questions

Related guides