Resources / Reference
Manufacturing metrics and lean glossary
Manufacturing runs on a shared vocabulary, and most of it is math. This glossary defines the effectiveness, quality, reliability, and lean terms that show up on every plant dashboard, in plain language, with the formulas behind them. Start here, then go deeper in the guides.
How the metrics connect
Most manufacturing metrics answer one of four questions. This glossary groups the terms that way, so a definition sits next to the others it works with.
- Effectiveness and utilization: how much of your time turned into good product (OEE, TEEP, Availability, Performance, Quality, Utilization).
- Quality and yield: how much of your output met spec the first time (First Pass Yield, Yield, Scrap Rate, SPC, Cp, Cpk, Ppk).
- Reliability and maintenance: how dependable your assets are (MTBF, MTTR, Predictive Maintenance).
- Lean and flow: how smoothly work moves through the line (Takt Time, Cycle Time, Ideal Cycle Time, Bottleneck, Changeover, SMED).
The metrics also stack on each other. Availability, Performance, and Quality multiply into OEE; OEE times Utilization gives TEEP; and first-pass yield rolls up across steps into total yield. For the full worked math behind the effectiveness stack, read the OEE and TEEP guide.
Effectiveness and utilization metrics
OEE (Overall Equipment Effectiveness). The most common effectiveness metric. It is the product of three factors, Availability times Performance times Quality, and expresses the share of planned production time that was truly productive. A score of 85% is often cited as a world-class benchmark for discrete manufacturing.
TEEP (Total Effective Equipment Performance). OEE multiplied by Utilization. Where OEE measures against scheduled time, TEEP measures against all 168 hours in a week, so it exposes the true capacity ceiling of an asset, including the hours you never scheduled.
Availability. The share of scheduled production time an asset was actually running. It is calculated as run time divided by planned production time, and it is reduced by both unplanned stops such as breakdowns and planned stops such as setup.
Performance. How close a running asset came to its ideal speed. It is ideal cycle time times total count, divided by run time. Minor stops, idling, and slow cycles pull it below 100%.
Quality. The share of produced units that meet specification on the first pass, calculated as good count divided by total count. Rejects and units that need rework both count against it.
Utilization. The share of all calendar time that production is actually scheduled to run. It captures the loss from unscheduled hours, whether from no demand, no staffing, or weekends, and it is the factor that separates TEEP from OEE.
Six Big Losses. The standard framework that maps every OEE loss to one of six causes: breakdowns and setup or adjustment (availability), small stops and reduced speed (performance), and startup rejects and production rejects (quality). Naming the loss is the first step to pricing it.
Quality and yield metrics
First Pass Yield (FPY). The percentage of units that move through a step correctly the first time, with no scrap and no rework. Also called first-time yield. Units that are reworked and later pass still count against FPY, which is what makes it stricter than a final-inspection number.
Yield. The ratio of good output to total input. Across a multi-step line, rolled throughput yield multiplies the first-pass yield of every step, so a line of many good-enough stations can still post a low overall yield.
Scrap Rate. The share of produced material that cannot be reworked and is discarded, expressed as a percentage of units or input weight. Scrap is a direct hit to both quality and material cost.
SPC (Statistical Process Control). The practice of watching a process with control charts to separate normal common-cause variation from special-cause variation. It tells an operator when a signal is real and worth acting on, and when a reading is just noise.
Cp (Process Capability). A measure of whether a process spread fits inside its specification limits, assuming the process is perfectly centered. It is the spec width divided by six standard deviations, and it ignores where the process is actually centered.
Cpk (Process Capability Index). Cp adjusted for how centered the process really is between the limits. It uses short-term, within-subgroup variation. A Cpk of 1.33 is a widely used minimum target for a capable process.
Ppk (Process Performance Index). The same idea as Cpk, but computed from long-term, overall variation. Because it reflects how the process behaves across shifts and days, Ppk usually reads lower than Cpk and is the more honest picture of real-world performance.
Reliability and maintenance metrics
MTBF (Mean Time Between Failures). The average operating time between failures of a repairable asset, calculated as total run time divided by the number of failures. Higher MTBF means a more dependable asset, and it is a core input to reliability planning.
MTTR (Mean Time To Repair). The average time to restore a failed asset to service, from the moment it stops to the moment it runs again. It is total repair time divided by the number of repairs. Lower MTTR lifts availability directly.
Predictive Maintenance (PdM). A condition-based strategy that uses sensor data and models to forecast a failure before it happens, so work is scheduled just in time instead of on a fixed calendar or after a breakdown. KaizenFlow's Reliability Forecaster surfaces that forecast; the plant's own maintenance system still executes the work order, because KaizenFlow is an intelligence layer, not a CMMS.
Lean and flow concepts
Takt Time. The rate of production that keeps pace with customer demand, calculated as available production time divided by customer demand. If takt is 60 seconds, one good unit must leave the line every 60 seconds to stay on demand.
Cycle Time. The actual time to complete one unit at a given step, measured start to finish. Compared against takt time, it shows whether a step can keep up with demand.
Ideal Cycle Time. The fastest cycle time a process can achieve under optimal conditions, the theoretical minimum. It is the baseline used to compute the Performance factor in OEE.
Bottleneck. The step with the least capacity, which sets the maximum throughput of the entire line. Improving any non-bottleneck step rarely raises output until the constraint itself moves.
Changeover. The work of switching a line from one product to the next, including teardown, setup, and the run up to the first good part. Changeover time is one of the largest and most controllable availability losses.
SMED (Single-Minute Exchange of Die). A lean method for cutting changeover to under ten minutes, single-digit, by converting internal setup steps done while the machine is stopped into external steps done while it still runs.
From metrics to dollars
A metric is only useful when it points at a decision. On its own, a falling OEE or a rising scrap rate tells you something is wrong, not what to fix first or what it is worth.
KaizenFlow reads these same metrics from the systems a plant already runs, including MES, SCADA, ERP, and historians, through more than 43 connectors. An ensemble of nine AI specialists then ranks every loss behind the numbers by dollar impact and confidence, and the results are reconciled into a verified savings ledger signed by the customer's finance team. You can see the full loop on the platform overview.
In the design-partner program, KaizenFlow models target ranges of 8 to 18% lower unplanned downtime, 5 to 12% less scrap, 4 to 11% more throughput, and 3 to 7% lower energy use. These are modeled ranges from the program, not achieved customer results.
Frequently asked
What is a good OEE score? Many discrete manufacturers treat 85% as a world-class OEE and around 60% as typical, but the honest answer is that a good score is one that is stable and trending up on your own line. The gap between your OEE and 100% is where the recoverable losses live.
What is the difference between OEE and TEEP? OEE measures productivity against scheduled production time. TEEP multiplies OEE by Utilization to measure against all calendar time, so it shows the true capacity ceiling of an asset, including hours you never scheduled.
Are Cpk and Ppk the same thing? No. Both compare process spread to spec limits and account for centering, but Cpk uses short-term within-subgroup variation while Ppk uses long-term overall variation. Ppk usually reads lower and reflects real-world performance over time.
What is the difference between MTBF and MTTR? MTBF is the average run time between failures, so higher is better. MTTR is the average time to repair a failure and return to service, so lower is better. MTBF measures reliability; MTTR measures recovery speed.
Does KaizenFlow replace my MES or CMMS? No. KaizenFlow is a manufacturing intelligence layer that connects on top of the MES, SCADA, ERP, and historians you already run. It surfaces and ranks opportunities; your existing systems still execute the schedule and the work orders.
From metrics to money
Put a dollar figure on every loss.
KaizenFlow ranks the losses behind your OEE, scrap, and downtime by dollar impact, then reconciles the wins into a finance-signed ledger. See it against your own line.