The last of OEE's three factors accounts for a loss the first two miss entirely: output that was made but cannot be sold. A machine can be available and fast and still waste a share of its production on units that fail specification, and quality rate is the factor that captures that waste. This guide defines the OEE quality rate as good units over total units, draws the sometimes-blurred line between quality rate and first-pass yield, and explains how reject counts and analyzer results feed it in a process plant.
Quality Rate (OEE) in one line: The quality rate in OEE is the number of good units divided by the total number of units produced, expressed as a percentage. It measures how much of what the equipment made met specification the first time, so anything scrapped or requiring rework counts as a quality loss. In the Six Big Losses framework it captures startup rejects, the defects made while a process stabilizes, and production rejects made during stable running, and it is what makes OEE a measure of fully good output rather than mere throughput.
Quality rate is the simplest OEE factor to state: good count divided by total count. Total count is everything the equipment produced during the run time; good count is the subset that met specification without rework. If a line produced a thousand units and forty had to be scrapped or reworked, the quality rate is 960 divided by 1000, or 96 percent. The other four percent is quality loss, capacity that was consumed making something that could not be sold as first-quality.
The insistence on good the first time is deliberate and is what distinguishes OEE's quality factor from a simple final-yield number. A unit that was defective and then reworked into an acceptable one still consumed the equipment's capacity to make the original defect, so from the equipment's standpoint that capacity was lost even though the unit was eventually saved. Counting reworked units as losses keeps OEE honest about how much of the asset's true potential was captured.
Because quality is defined over the units produced, it multiplies cleanly with availability and performance to give OEE. Availability captured lost time, performance captured lost speed, and quality captures lost yield, and only the units that were made, made at speed, and made good count toward the final OEE. This is why a strong-looking asset can still have a modest OEE if a persistent defect rate quietly erodes the good count.
Quality rate and first-pass yield are closely related and often equal, but they are not defined identically, and the difference matters in a multi-step process. First-pass yield is typically the fraction of units that pass through a step, or a whole process, correctly the first time without any rework, and across several sequential steps the overall first-pass yield is the product of each step's yield, which can be much lower than any single step suggests. It is a process-quality metric that can span many operations.
OEE's quality rate is scoped to a single piece of equipment or operation and to the units that equipment produced during the measured run. In practice, for a single machine measured on a good-first-time basis, the quality rate and that machine's first-pass yield coincide. The divergence appears when people conflate a machine's OEE quality with a multi-step rolled yield, or when rework is counted as good in one metric but not the other. Being explicit about scope prevents the two from being mixed.
The practical guidance is to treat OEE quality strictly as good-first-time output of the specific asset being measured, and to reserve first-pass and rolled yield for describing how quality flows through a chain of operations. Both are valuable, but they answer different questions: quality rate asks how well this equipment made good product, while rolled yield asks how much good product survives an entire process. Keeping the definitions distinct is what lets each remain meaningful.
In a discrete line, quality rate needs a good count and a reject count, and both can come from the control system. A reject or scrap counter incremented when a unit fails an inspection, together with the total production counter, gives the two numbers the factor requires. A cloud SCADA such as Merobix historizes these counter tags read from the field over Modbus, DNP3, OPC UA, or MQTT, so the good count is simply the total minus the logged rejects over any chosen window, computed from the same live data used to run the line.
A process plant complicates this because quality is often a continuous property of a stream rather than a pass/fail on discrete units. Here the relevant signals are analyzer readings and lab results, such as a moisture analyzer, a gas chromatograph, or a density measurement, that indicate whether product is on-specification. Periods when an on-line analyzer shows the stream off-spec, or when a stream is diverted to slop or reprocessing, represent quality loss, and the historized analyzer trend is what identifies when and how much of the production failed to meet spec.
Because the platform records both the reject counters and the analyzer values continuously, the quality factor can be tied to when defects occurred, which is what separates the two quality losses. Rejects clustered immediately after a start or grade change implicate startup rejects and point at process stabilization, while a steady background of off-spec during normal running implicates production rejects and points at the process or feed. Having the good/total counts and analyzer results in one historized record turns quality rate from an end-of-shift tally into a diagnosable, continuously available metric.
Quality rate is the good count divided by the total count produced during the run, expressed as a percentage. Good count is the units that met specification without rework, so anything scrapped or reworked is a quality loss. For example, if 1000 units were produced and 40 were defective, the quality rate is 960 divided by 1000, or 96 percent.
They are closely related and often equal for a single machine measured on a good-first-time basis, but they are not defined identically. First-pass yield is the fraction passing correctly the first time and can span multiple process steps, where the overall yield is the product of each step's yield. OEE quality rate is scoped to a single asset and the units it produced, so the two diverge when a machine's OEE quality is confused with a multi-step rolled yield or when rework is counted differently.
In a process plant, quality is usually a continuous property of a stream rather than a pass/fail on discrete units, so it is measured from analyzer readings and lab results such as moisture, composition, or density against specification. Time that an on-line analyzer shows the product off-spec, or when a stream is diverted to slop or reprocessing, is treated as quality loss. Historizing the analyzer trend lets the plant quantify how much production failed to meet spec and when, which feeds the quality factor.
This page references the protocol specifications published by the organizations below. Editions, product capabilities, and documentation change over time - confirm current requirements and specifications directly with the source.
Last reviewed: July 27, 2026. Merobix is not affiliated with, endorsed by, or sponsored by these organizations; their names are used only to identify the standards and products discussed.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.