Run rate is the speed at which a line is producing, expressed as units per unit of time, parts per minute, cases per hour, tons per shift. Instantaneous run rate is the speed right now, while the line is running; average run rate is total output divided by total time, including every slowdown and stop. The gap between them is performance loss.

Ask two operators for the line's run rate and you will often get two very different numbers. One quotes the nameplate speed, what the machine does when it is humming. The other quotes what actually landed on the pallet by end of shift. Both are "run rate," and the distance between them is the most useful number neither of them said out loud. This post separates the three run rates that get confused, ideal, instantaneous, and average, shows why the gap is pure performance loss and gives you a clean way to measure all three.

What is run rate in manufacturing?

Run rate is a production speed. It is the rate at which a process turns out units, and it is the reciprocal of cycle time: if a machine completes one part every 3 seconds, its run rate is 20 parts per minute, and its ideal cycle time is 3 seconds. The two are the same fact stated two ways, rate looks at throughput, cycle time looks at the interval between units, and any conversation about run rate is really a conversation about which cycle time you mean.

There are three run rates worth naming, and keeping them separate is most of the battle:

Ideal, instantaneous, and average run rate across one shift (illustrative)Three run rates, one shift (illustrative)ratetime in shiftideal run rateaverage run rateinstantaneousperformance loss band
Illustrative trace. The jagged line is instantaneous run rate, near ideal when running, zero during stops. Averaging it across the whole shift, including the stops, drops it to the average run rate. The shaded band is performance loss.

What is the difference between instantaneous run rate and average throughput?

Instantaneous run rate measures speed while running; average throughput measures speed including everything that stopped the running. A filler rated at 400 cans a minute might genuinely hit 395 cans a minute every second it is moving, excellent instantaneous run rate, and still deliver an average of 260 cans a minute across the shift because it jammed nine times, ran slow after each restart, and waited twice on an empty infeed. The equipment is fast. The process is not.

This is why "the line runs at 400 a minute" is one of the most expensive sentences on a plant floor. It is true and useless at the same time. Nobody schedules against instantaneous run rate; the warehouse fills at the average. When the two are quoted interchangeably, planning off nameplate speed, then acting surprised at end-of-shift counts, the plant chronically over-promises and under-ships. Separating the two turns a vague "the line is slow" into a measurable gap you can attack.

How do you calculate run rate?

Each run rate has its own simple formula; the discipline is using the right denominator:

Run rateFormulaWhat the denominator counts
Ideal1 ÷ ideal cycle timeNothing lost, design speed
Instantaneousunits ÷ running time onlyTime the line was actually moving
Averagegood units ÷ scheduled timeAll scheduled time, stops included
The only difference between the three is what the denominator forgives. Ideal forgives everything, average forgives nothing.

A worked example with hypothetical numbers. A packaging line is scheduled for a 480-minute shift, rated at 400 cans/minute (ideal). It runs for 410 minutes and is stopped for 70. During running time it produces 155,800 cans. Instantaneous run rate = 155,800 ÷ 410 = 380 cans/minute, 95% of ideal, so it holds speed well when moving. Average run rate = 155,800 ÷ 480 = 325 cans/minute, only 81% of ideal. The 55-can-per-minute gap between instantaneous and average is stopped time; the 20-can gap between ideal and instantaneous is slow running. Both are performance loss, and they have different fixes.

Run rate waterfall: ideal to instantaneous to average (illustrative)From ideal to actual: where the cans go (illustrative)Ideal 400/minInstant 380/min-20 speedAverage 325/min-55 stopsSlow running costs 20/min; stops cost another 55/min. Different losses, different fixes.
Illustrative. The step from ideal to instantaneous is reduced-speed loss; the step from instantaneous to average is stopped-time loss. Naming which one you have tells you where to work.

Why is the gap between run rate and throughput pure performance loss?

Because every unit of that gap is capacity the line was built to deliver and didn't. In OEE terms, the ratio of average output to ideal output is the Performance factor, and it is fed by exactly two of the six big losses: reduced speed (running below ideal) and minor stops and idling (the short, uncounted stoppages that murder the average). Neither shows up as a breakdown on the maintenance log, which is why both are chronically underestimated. A line that never has a "real" downtime event can still bleed a fifth of its capacity to speed and micro-stops.

The two losses behave differently on the floor. Minor stops are individually trivial, a misfeed cleared in fifteen seconds, a sensor blocked, a starved infeed, and collectively enormous, precisely because they are too short to log by hand. Reduced speed is quieter still: the line "always runs at 80% because it jams at full speed," a workaround that hardens into the assumed rate until someone measures the instantaneous run rate against nameplate and asks why. Both are performance loss, both live in the gap, and both are invisible unless run rate is measured continuously rather than reconstructed from shift totals.

By the numbers. Performance loss is rarely a small tail. The six-loss framework built into OEE attributes speed and minor-stop losses to the Performance factor, and on many lines that factor sits in the 80–90% range even when Availability looks healthy, meaning 10–20% of running capacity vanishes to a gap most shift reports never name. For context on how these losses roll into a plant-level number, see the U.S. Federal Reserve's industrial production and capacity utilization series, which tracks how much of installed capacity U.S. manufacturers actually use.

How do you measure run rate reliably?

Measure the running time and the stopped time separately, or the two performance losses collapse into one unusable average. The sequence:

  1. Fix the ideal run rate first, in writing. Use the nameplate or a demonstrated best sustained rate, per product, and defend it, a soft ideal makes every downstream loss look smaller than it is.
  2. Timestamp starts and stops, not just end-of-shift counts. You cannot separate reduced speed from micro-stops without knowing how many minutes the line was truly moving.
  3. Compute instantaneous run rate = units ÷ running minutes, and compare it to ideal. The gap here is reduced-speed loss.
  4. Compute average run rate = good units ÷ scheduled minutes, and compare it to instantaneous. The gap here is stopped-time and minor-stop loss.
  5. Trend both against ideal every shift, per product and per line. A single blended number hides which loss is growing.
  6. Attack the larger gap first SMED and infeed reliability for stops, tooling and setpoint work for speed, then re-measure. Chasing the smaller loss is how improvement projects stall.

How do you improve run rate?

Improve the two gaps separately, because they answer to different tools. To close the stopped-time gap between instantaneous and average run rate, attack changeovers and micro-stops: SMED to shrink the time a line sits idle during product changes, infeed and material-presentation fixes to stop the starving and blocking, and short-interval root-cause work on the top three recurring jams. Most plants find their biggest single win here, because minor stops are both the largest hidden loss and the cheapest to remove once they are finally counted.

To close the reduced-speed gap between ideal and instantaneous run rate, work the physics: worn tooling, conservative setpoints left over from an old quality scare, lubrication and alignment that let the machine hold its rated speed, and a real challenge to any "we always run it slow" rule that no one can defend with data. Both efforts feed the same scoreboard, every point of recovered run rate lifts the Performance factor and, through it, the whole OEE improvement effort. The discipline is to name which gap you are closing before you start, so you can prove the fix worked when you re-measure.

How does run rate connect to OEE and the plant KPIs?

Run rate is the raw material of the OEE Performance factor, and it is the bridge between a machine's speed and the plant's promises. Average run rate drives real throughput which drives whether the manufacturing KPI stack, schedule attainment, on-time shipment, cost per unit, comes out green or red. When planners schedule against ideal or instantaneous run rate instead of the average the line actually holds, the miss is baked in before the shift starts. The honest planning number is the average, adjusted for the mix you are about to run.

The catch is that average run rate is only trustworthy if the stops and slow patches underneath it were captured as they happened. Hand-tallied shift totals can tell you the average landed low; they cannot tell you whether the culprit was speed or micro-stops, because the fifteen-second jams never made the clipboard. Plants that detect stops and counts straight off the equipment, the way Harmony feeds live run-rate and downtime signals into dashboards operators actually watch (see the platform or the CLS field story), can finally see the two performance losses as separate, fixable problems instead of one disappointing number at the end of the day.