How to Calculate OEE: Overall Equipment Effectiveness Formula & Worked Examples

2026-06-14

OEE = Availability × Performance × Quality. A plain-English guide to the OEE formula with a fully worked example, the Six Big Losses, benchmarks, OEE vs TEEP, and how to track it without an ERP.

Ask three people on a factory floor how their best line is performing and you'll get three answers: "running fine," "we hit target most days," and "ask maintenance." None of them is a number you can act on. OEE — Overall Equipment Effectiveness — turns that vague sense of "fine" into a single percentage, and more importantly, into a breakdown that tells you exactly where the time is going.

This guide covers what OEE is, the exact formula, a fully worked example with real numbers, the Six Big Losses behind every lost point, and the one thing that decides whether your OEE is useful or fiction: where the data comes from.

What OEE actually measures

OEE answers one question: of the time you planned to make good product, how much did you actually spend making good product? A machine scheduled to run for an 8-hour shift could, in theory, produce a fixed number of perfect units. OEE measures how close you got to that theoretical maximum — and splits the gap into three named losses you can attack.

It is the product of three factors:

OEE  =  Availability  ×  Performance  ×  Quality

Availability = run time / planned production time     (downtime losses)
Performance  = actual output / theoretical output     (speed losses)
Quality      = good units / total units               (defect losses)

An OEE of 100% means every planned minute produced a sellable unit at full rated speed. That never happens. What matters is the three-way split: it tells you whether your biggest enemy is downtime, slow running, or scrap — so you stop guessing which problem to fix first.

The three factors, one at a time

Availability captures downtime losses — every minute the line was supposed to run but didn't. Breakdowns, changeovers, waiting for material, no operator, a jammed conveyor. You start from planned production time (the shift, minus genuinely planned stops like a scheduled lunch or a planned-maintenance window) and subtract all the unplanned stops to get run time.

Availability = Run Time ÷ Planned Production Time

Performance captures speed losses — the line ran, but slower than its rated cycle time. Minor stops, idling, worn tooling, a cautious operator, sub-optimal settings. You compare what the line actually produced against what it should have produced at full rated speed for the time it was running.

Performance = (Total Count × Ideal Cycle Time) ÷ Run Time

— or equivalently — Actual Output ÷ Theoretical Output

Quality captures defect losses — units made but not sellable as first-pass good. Rejects, rework, startup scrap.

Quality = Good Count ÷ Total Count

Multiply the three and you have OEE. Each factor is a percentage; so is the result.

A worked example

Take a packing line on a single 8-hour shift.

• Shift length: 480 minutes
• Planned stops: 30 min lunch + 10 min scheduled tea break = 40 min → Planned Production Time = 440 min
• Unplanned downtime: 25 min breakdown + 35 min changeover + 15 min waiting for material = 75 min → Run Time = 365 min
• Ideal cycle time: 1 unit every 2 seconds = 30 units/min
• Total units produced: 9,200
• Rejected units: 200 → Good units = 9,000

Now run the numbers:

Availability = 365 / 440                       = 82.9%
Performance  = (9,200 × (1/30 min)) / 365       = 306.7 / 365  = 84.0%
Quality      = 9,000 / 9,200                     = 97.8%

OEE = 0.829 × 0.840 × 0.978                      = 68.1%

A 68% OEE. The headline isn't the point — the breakdown is. This line's single biggest enemy is downtime (Availability at 83%), with speed losses close behind. Quality is healthy. So the improvement effort belongs on the 75 minutes of breakdowns, changeovers and material waits — not on chasing the 200 rejects. Without the split, a team often spends weeks on the wrong loss.

The Six Big Losses behind every OEE score

OEE's three factors map to six classic loss categories. Naming the loss is what makes it fixable:

Availability losses

• Breakdowns — equipment failure, unplanned stops
• Setup & changeovers — time lost switching products, sizes, or tooling

Performance losses

• Minor stops & idling — short jams, misfeeds, sensor blocks (often under-counted because nobody logs a 40-second stop)
• Reduced speed — running below rated cycle time for any reason

Quality losses

• Startup rejects — scrap during warm-up or after a changeover, before the process stabilises
• Production rejects — defects during steady-state running

The most under-measured of these is almost always minor stops — they don't feel like downtime, so they never get logged, and they quietly eat your Performance number.

What counts as "planned" time — and why it changes everything

The single most common way OEE gets manipulated (deliberately or not) is by moving the boundary of planned production time. Classify a long breakdown as "planned maintenance" and Availability jumps. Exclude a whole product changeover as "not scheduled to run" and the number flatters you.

Pick a definition and hold it constant. The cleanest convention: planned production time = the time you intended to make product. Lunch and a pre-scheduled PM window come out; everything else — including changeovers and breakdowns — stays in and shows up as a loss. If you want to also see the cost of unscheduled time (nights, weekends, holidays), that's a different metric — see TEEP below.

OEE vs TEEP vs OOE — which one do you want?

These three differ only in what they treat as the denominator:

• OEE measures against planned production time — "how well did we run when we meant to run?" This is the day-to-day operating metric.
• TEEP (Total Effective Equipment Performance) measures against all calendar time, 24/7/365 — "how much of our asset's total capacity are we using?" Good for capital and capacity-expansion decisions.
• OOE (Overall Operations Effectiveness) sits between the two, including some unscheduled time.

For shop-floor improvement, use OEE. For "do we need to buy another machine or add a shift?", look at TEEP.

What is a good OEE score?

The widely cited benchmarks:

• 100% — theoretical perfection (every unit, first-pass good, at full speed, zero stops)
• 85% — world-class for discrete manufacturing; a realistic long-term goal
• 60% — typical for manufacturers not actively managing OEE
• 40% — common for a line that's never been measured, and entirely recoverable with basic tracking

Most companies that measure OEE for the first time are shocked to land in the 40–60% band. That gap isn't bad news — it's the easiest capacity you'll ever find, because it needs no capital, just attention pointed at the right loss.

A word of caution: don't chase a single OEE number across different products or lines. A line running a hard-to-make product will show lower OEE than one running an easy SKU, and that comparison tells you nothing. Trend each line against itself over time.

The mistake that makes OEE worthless: guessed data

OEE is only as honest as the data underneath it — planned time, downtime reasons, output counts, and rejects. Guessed OEE is worse than no OEE, because it gives a confident number that points improvement effort in the wrong direction.

The failure mode is almost always the same: downtime is reconstructed from memory at the end of the shift, minor stops are never logged, and reject counts are rounded. By the time the number reaches a Monday meeting it's a week old and nobody trusts it. The fix isn't a better spreadsheet — it's capturing the data at source, on the floor, as it happens: every stop with a reason, every output and reject count, every changeover. Do that and OEE stops being a debate and becomes a dashboard.

This is the same discipline behind tracking changeovers to cut downtime and behind continuous-improvement methods like Kaizen and DMAIC — none of them work on remembered numbers.

How to start tracking OEE (without an ERP rollout)

You don't need an MES or a six-figure system to begin. You need four things captured reliably for each line, each shift:

1. Planned production time — shift length minus planned stops, on a fixed definition

2. Downtime, with a reason — every unplanned stop, tagged (breakdown / changeover / material / no-operator…)

3. Total output count — units off the line

4. Reject count — units that aren't first-pass good

From those four, OEE and its Availability / Performance / Quality split fall out automatically — and the reason codes on downtime instantly tell you which loss to attack. Start with your most important line, get the capture habit solid, then expand.

How Flobri does it

Flobri lets manufacturers capture OEE inputs on the floor and roll them up live — no MES, no code:

• Operators log it where it happens — a shift / production entry records output, rejects, and every stop with a reason code, on a phone or terminal at the line.
• OEE is calculated from real data, not month-end memory — Availability, Performance and Quality are computed from the logged time and counts, so the loss breakdown is always current.
• Leaders see one dashboard — OEE trend by line, the dominant loss, and the worst-performing shift — so the weekly review is about fixing the right loss, not arguing about whose numbers are right.
• You describe the process, it goes live in minutes — no implementation project; you can go from a plain-language description to a working tracker the same day.

Frequently asked questions

How is OEE calculated?

OEE = Availability × Performance × Quality. Availability is run time ÷ planned production time, Performance is actual output ÷ theoretical output at rated speed, and Quality is good units ÷ total units. Multiplying the three gives a single percentage, and the split shows whether your biggest losses are downtime, speed, or defects.

What is a good OEE score?

85% is considered world-class for discrete manufacturing, 60% is typical, and 40% is common for lines that have never been measured. More useful than the absolute number is the trend of each line against itself.

What is the difference between OEE and TEEP?

OEE measures performance against planned production time; TEEP measures against all calendar time (24/7/365). Use OEE for shop-floor improvement and TEEP for capacity and investment decisions.

Why is my OEE different every time I calculate it?

Almost always because the definition of "planned production time" keeps moving, or downtime and rejects are estimated from memory. Fix the planned-time definition and capture downtime, output and rejects at source as they happen.

Can I track OEE without an ERP or MES?

Yes. You only need four data points per line per shift — planned time, downtime with reasons, total output, and rejects. A simple no-code workflow that captures those on the floor is enough to produce a live, trustworthy OEE.

Flobri turns shift-floor data into live OEE — operators log output, rejects and downtime reasons at the line, and Availability/Performance/Quality roll up to one dashboard automatically. See how Flobri turns a process description into a live workflow.