OEE Calculator (Availability x Performance x Quality, TEEP, and ISO 22400 Loss Accounting)

Industrial engineering methodology and OEE loss-accounting workflow

This calculator converts a shift record into Overall Equipment Effectiveness, the single ratio that tells you what share of planned production time actually produced good parts at full speed. OEE is the product of three independent ratios, Availability, Performance, and Quality, and its power is that it decomposes a vague sense of “the line is slow” into three measurable loss buckets that point at different countermeasures. The operational objective is to convert raw shift data into an honest effectiveness number and, more importantly, into the split between time lost to stops, speed, and scrap, so a maintenance, process, or quality action can be aimed at the largest loss rather than guessed at.

The data workflow runs end to end. You enter the planned production time (shift length net of planned stops such as breaks and scheduled meetings), the unplanned down time, the ideal cycle time for the part, the total count produced, and the good count. The tool computes run time, applies the three governing ratios, and returns Availability, Performance, Quality, OEE, and the fully productive time in the same units you entered, plus TEEP when calendar time is supplied. It also flags which of the three factors is dragging the number down, because a 75 percent OEE built from a 70 percent Availability is a different problem from the same 75 percent built from a 78 percent Performance.

A naive reading treats OEE as one opaque percentage and chases it directly. The shop floor rewards the opposite: OEE is an accounting identity for the six big losses, and the calculation only earns trust when the inputs are measured against the correct time base. Load time versus calendar time, planned versus unplanned stops, ideal versus nameplate cycle time, and total versus good count each have a precise definition below, because getting any one of them wrong silently moves the loss from the bucket that owns it into one that does not.

Governing equations: OEE equals Availability times Performance times Quality

The core model is a product of three dimensionless ratios, each bounded between 0 and 1 (0 to 100 percent).

OEE = Availability (A) x Performance (P) x Quality (Q)

Each factor is defined against an explicit time or count base:

• Availability (A) = Run Time / Planned Production Time. Run Time = Planned Production Time minus unplanned Down Time. Planned Production Time is shift length net of planned stops (breaks, planned maintenance). Units: minutes or seconds; the ratio is dimensionless.
• Performance (P) = (Ideal Cycle Time x Total Count) / Run Time. Ideal Cycle Time is the theoretical fastest time per part (the nameplate or best-demonstrated rate), in seconds per unit. P is capped at 1.0; a value above 1 means the ideal cycle time is set too slow.
• Quality (Q) = Good Count / Total Count. Good Count excludes rejects, rework, and startup scrap.

Two derived metrics extend the model. Fully Productive Time = Planned Production Time x OEE, the time that yielded saleable output. TEEP (Total Effective Equipment Performance) = OEE x Utilization, where Utilization = Planned Production Time / All Time (the full calendar, typically 1440 minutes per day). TEEP answers a different question than OEE: OEE measures how well you run when you are scheduled to run; TEEP measures how much of the entire clock, including unscheduled hours, becomes good output, which is the metric for capacity and capital decisions. There is no Imperial versus SI distinction for the ratios; only keep cycle time and time bases in the same unit before dividing.

Applicable standards and testing frameworks: ISO 22400, SEMI E10, and JIPM TPM

OEE is defined slightly differently across frameworks, so a defensible number states which one it follows.

Compliance sets the boundaries. ISO 22400 dictates that planned maintenance is removed from planned production time (it is not an availability loss), while unplanned breakdowns are. SEMI E10 separates non-scheduled time from downtime, which is why OEE (scheduled base) can read 75 percent while TEEP (calendar base) reads 22 percent on the same equipment. Where a definition is contested, for example whether a planned changeover is a planned stop or an availability loss, state the convention and apply it identically across every asset so the numbers are comparable.

Key input variables and the six-big-loss classification

Every OEE input maps to one of the six big losses, and the value of the tool is that it routes each measured minute or reject into the factor that owns it. The table is the reference map; measure your own losses against it rather than assuming a distribution.

Deration factors: how the six losses erode nominal capacity

Nameplate capacity assumes the machine runs every scheduled second at full speed with zero defects. The six losses derate that ideal, and because OEE multiplies the three factors, they compound: a machine that is 90 percent available, 90 percent performing, and 90 percent quality is not 90 percent effective but 72.9 percent. Small, individually acceptable losses stack into a large capacity gap.

Availability losses: breakdowns, micro-stoppages, and changeover

Every unplanned stop and every changeover minute counted as down time reduces Run Time and therefore Availability. Breakdowns are attacked with preventive and predictive maintenance; setup time is attacked with SMED (single-minute exchange of dies). A plant running many short production runs is usually availability-limited by changeover, not by breakdowns, and the calculator’s Availability figure separates the two so the right program is funded.

Performance losses: minor stops and reduced speed

While running, a machine loses throughput to idling, minor stops absorbed into the cycle, and speed below nameplate. These are the hardest losses to see because the machine looks like it is working. Performance is computed from the gap between the ideal cycle time and the actual output over Run Time, which surfaces speed loss that a stopwatch on a single cycle would miss.

Quality losses: process defects and startup rejects

Scrap and rework consume capacity twice: the time to make the bad part and the time to remake it. Quality separates defects produced during stable running from startup and post-changeover rejects, because the first points at process control and the second at setup standardization. First-pass yield, not final yield after rework, is the correct input.

Nominal OEE versus safe operating capacity and TEEP

Peak nameplate output and realistic sustainable output are far apart, and OEE quantifies the gap. The JIPM world-class benchmark is 85 percent OEE, but that figure is a target for a mature discrete line, not a universal pass mark: a high-changeover job shop and a continuous process plant have structurally different ceilings. Set the target from your own loss structure, not a headline number. TEEP then reframes the ceiling for capacity planning: with Utilization = Planned Production Time / All Time, a line at 75 percent OEE run two shifts of five days has far more hidden capacity (higher achievable TEEP) than the OEE alone suggests, which is the number to bring to a capital or outsourcing decision before buying another machine.

Reverse-engineering the factors from a target OEE

The identity inverts, which turns OEE from a scorecard into a design tool. Given a target and two known factors, solve for the third; given a target output, solve for the OEE it demands.

• Required Availability: A = OEE / (P x Q).
• Required Performance: P = OEE / (A x Q).
• Required Quality: Q = OEE / (A x P).
• Good parts from an OEE: Good Count = OEE x Planned Production Time / Ideal Cycle Time.
• OEE demanded by a target output: Required OEE = (Target Good Count x Ideal Cycle Time) / Planned Production Time.

For example, to reach 85 percent OEE on a line already holding 90 percent Availability and 99 percent Quality, required Performance = 0.85 / (0.90 x 0.99) = 95.4 percent, which tells you the improvement must come from speed and minor stops, not from more maintenance.

Five OEE case studies and worked calculations

Case 1: baseline discrete line, standard conditions

Shift 480 min, breaks 60 min, so Planned Production Time = 420 min = 25,200 s. Unplanned Down Time = 47 min, so Run Time = 373 min = 22,380 s. Availability = 373 / 420 = 88.8 percent. Ideal Cycle Time = 1.0 s/part, Total Count = 19,271. Performance = (1.0 x 19,271) / 22,380 = 86.1 percent. Good Count = 18,848, so Quality = 18,848 / 19,271 = 97.8 percent. OEE = 0.888 x 0.861 x 0.978 = 74.8 percent, and Fully Productive Time = 25,200 x 0.748 = 18,850 s = 314 min.

Case 2: high-changeover job shop, availability-limited

Same 420 min planned, but six changeovers of 20 min each add 120 min of setup counted as down time, so Run Time = 300 min and Availability = 300 / 420 = 71.4 percent. With Performance 90 percent and Quality 98 percent, OEE = 0.714 x 0.90 x 0.98 = 63.0 percent. The dominant loss is setup, so a SMED program that halves changeover to 10 min lifts Availability to 85.7 percent and OEE to 75.6 percent, a bigger gain than any speed or quality action here.

Case 3: constrained line, only the bottleneck OEE moves throughput

A three-station line has OEEs of 82, 65, and 88 percent; the 65 percent constraint governs line output. Raising the first station from 82 to 92 percent adds zero saleable parts because it was never the limit. Raising the constraint from 65 to 75 percent increases line throughput by 75/65 = 15.4 percent. The lesson the calculator makes concrete: measure OEE per asset, then improve the lowest OEE on the critical path first.

Case 4: high-speed packaging, performance-limited

On the same 22,380 s of Run Time with an Ideal Cycle Time of 0.5 s/unit, a line producing 35,000 units has Performance = (0.5 x 35,000) / 22,380 = 78.2 percent. Availability is a healthy 95 percent and Quality 99.5 percent, giving OEE = 0.95 x 0.782 x 0.995 = 73.9 percent. The loss is minor stops and speed, invisible on a walk-by, so the countermeasure is a short-stop study and infeed reliability, not maintenance.

Case 5: reverse calculation, target OEE to required speed

A plant must reach 85 percent OEE and already holds Availability 90 percent and Quality 99 percent. Required Performance = 0.85 / (0.90 x 0.99) = 95.4 percent. Converting to output: at Planned Production Time 25,200 s and Ideal Cycle Time 1.0 s, the target Good Count = 0.85 x 25,200 / 1.0 = 21,420 parts per shift. The reverse pass converts a KPI goal into a concrete speed target and a parts-per-shift commitment.

Boundary conditions, mathematical limits, and model assumptions

The model assumes a single, well-defined ideal cycle time and a countable good/total split over one time base. It strains at the edges. On a continuous or batch process (chemicals, paper, heat treat) there is no discrete part count, so OEE is adapted with a rate-based Performance and a yield-based Quality, and the naive part-count form does not apply. When the product mix changes within the shift, a single ideal cycle time misrepresents the run, and OEE must be computed per product or with a weighted ideal. Performance above 100 percent is a definitional error, not a real result, and signals a mislabeled ideal cycle time. The three factors are treated as independent, which is a modeling convenience: in practice a breakdown often causes startup scrap, coupling Availability and Quality, so a single root cause can appear in two buckets. Finally, OEE measures effectiveness of a machine, not profitability; a high OEE producing unsold inventory is overproduction, and takt, not OEE, governs how fast the line should run.

Common OEE mistakes and data interpretation pitfalls

• Using calendar time as the base. Dividing by 24 hours instead of planned production time collapses OEE and TEEP into one wrong number; keep the scheduled base for OEE.
• Padding the ideal cycle time. Setting a slow ideal inflates Performance and hides the largest hidden loss on most lines.
• Counting rework as good. Quality must use first-pass good count; counting reworked parts as good erases the quality loss and its cost.
• Chasing a single OEE number. The same OEE can come from very different factor splits; without the A, P, Q breakdown the improvement target is a guess.
• Improving a non-bottleneck. Raising OEE on an asset that is not the constraint adds cost and no throughput; improve the constraint first.

Integration into MES, ERP, CMMS, and value stream mapping

OEE is a hub metric that feeds the wider stack. In an MES it is computed in real time from machine states and counts, driving andon, loss Pareto charts, and shift reporting. In a CMMS the Availability loss and its breakdown detail prioritize the preventive maintenance plan and feed reliability metrics such as MTBF and MTTR. In ERP and capacity requirements planning, effective capacity equals nameplate capacity times OEE (or TEEP for the calendar view), so the OEE figure directly sizes how much a work center can actually deliver against the master production schedule. In value stream mapping, per-process OEE identifies which step erodes flow and whether the constraint is a stop, a speed, or a quality problem. Because the same inputs feed cost of poor quality and downtime cost, a consistent OEE keeps maintenance, quality, scheduling, and finance arguing from one set of numbers.

OEE frequently asked questions

Sources, disclaimer, and editorial transparency

The OEE definitions, formulas, the Six Big Losses, and the world-class benchmark used here follow recognized sources, including Vorne / OEE.com, the Lean Enterprise Institute, ASQ, and Seiichi Nakajima’s foundational work on TPM. Definitions follow ISO 22400 for manufacturing KPIs, SEMI E10 for equipment states, and the JIPM TPM six-big-losses framework. This calculator and guide were built by Zeeshan Abbas and technically reviewed by Rimsha Nadeem Anwar, a Six Sigma Black Belt industrial engineer; see our Editorial Policy for how each tool is researched, built, and tested.

Results are accurate estimates for planning and education, not certified engineering advice. Validate outputs against your own measured data and engineering judgment before changing a line, committing capital, or making staffing decisions. See our full Disclaimer. OpsCalculators.com is operated by MAFHH INTERNATIONAL LTD. Your inputs are processed in your browser and are never stored; see our Privacy Policy.