Production Time: Definition, Formula & the OEE Time Cascade

What is production time?

Production time is the total time a plant or line spends actually producing output — measured as the portion of the planned shift during which the equipment runs and converts input into product. It excludes scheduled shutdowns, planned breaks, and by definition unplanned stoppages. It is not the time from order to delivery; that is lead time, and it is not the time between two finished parts; that is cycle time. Production time sits in the middle: longer than a cycle, shorter than a lead time, and — critically — measurable in real time at the machine.

Across the 15,000+ machines SYMESTIC has connected, the single most consistent finding about production time is that it is the most overestimated number in the factory. Ask a shift supervisor how many hours their line produced last week, and the answer is usually 10–20 % too high. Not because anyone is dishonest — because manual estimation systematically forgets micro-stops, forgets the last fifteen minutes before shift change, and rounds in the direction of what the schedule said should have happened. Accurate production time is the first number an instrumented plant usually sees correctly, and it is often the first uncomfortable one.

Production time vs. cycle time vs. lead time vs. takt time

Six time concepts get used interchangeably in production conversations and mean substantially different things. The table below separates them.

Two relationships from this table are worth burning into memory. Cycle time and takt time govern the line's rhythm. When cycle time is faster than takt time, the line is capable of meeting demand; when slower, it isn't, and no amount of overtime on adjacent equipment will fix it. Production time and throughput time sit in different universes. Production time is a plant metric ("how much did the line run?"); throughput time is a unit metric ("how long did this specific part spend in the factory?"). You can have high production time and terrible throughput time at the same time — that's what happens when work piles up between value-adding steps.

The OEE time cascade — the canonical structure

The most useful way to understand production time is the Nakajima six-big-losses time hierarchy that underlies OEE. It decomposes total calendar time into progressively narrower buckets, and each transition corresponds to a specific class of loss.

In most practical conversations, "production time" means either Operating Time or Net Operating Time — the time the machine was genuinely making product, not being set up, repaired, or idle. The distinction matters, because it determines what question the number answers. Operating Time tells you whether availability problems are hurting you; Net Operating Time tells you whether the machine ran fast enough when it was operating. Both are needed; they are not the same.

How to calculate production time

The base formula is direct:

Production Time = Planned Production Time − Unplanned Downtime − Setup/Changeover Time

Which, in the OEE cascade, is Operating Time. For Net Operating Time, subtract performance losses as well; for Fully Productive Time, subtract quality losses on top. The manufacturing-lead-time formula — relevant when planning orders rather than measuring a machine — is structurally different:

Manufacturing Lead Time = Setup Time + Run Time + Move Time + Queue Time + Wait Time

The key point: these are not two formulas for the same thing. The first answers "how much did the machine actually run?". The second answers "how long does one order take to get through the factory?". Mixing them up — and estimating production time from order-level planning data — is one of the fastest ways to produce numbers that look reasonable and mean nothing.

What shortens production time (and what quietly steals it)

When a machine's production time falls short of its planned production time, the culprit falls into one of four categories. Each has its own signature in the data and its own fix.

• Breakdowns and unplanned stops — the most visible loss. Usually well-documented in paper logs, often systematically under-reported in duration (operators restart the machine before logging, so the timer starts later than the stop). Automatically-captured PLC signals eliminate this bias.
• Setup and changeover time — the second-largest loss in high-mix environments. Often confused with production time because the machine is "occupied", but by convention setup is excluded. SMED methods reduce it; accurate measurement is the precondition for SMED being targeted at the right changeovers.
• Micro-stops and speed losses — the most expensive class, because it is the least visible. A machine running at 85 % of nameplate speed for a full shift loses more production time than a 30-minute breakdown, and nobody will notice without automated cycle counting.
• "Phantom production time" — time logged as running that wasn't — the category nobody wants to discuss. Shift handover gaps, "running but not producing" states, end-of-shift cleanup booked as production. Typically adds up to 30–60 minutes per shift in a plant without automated capture. When the MDE/BDE gets installed, this time appears as a sudden "drop" that management interprets as the system being wrong. It isn't.

How production time is actually measured on the shopfloor

In practice, production time can be captured three ways, and they produce materially different numbers:

1. Manual booking (paper, tablet, or BDE terminal at shift end). Operator logs "ran from 06:30 to 14:00, 30 minutes break". Easy, cheap, and systematically optimistic. Typical error: 15–20 % overstatement.
2. Semi-automatic booking (operator acknowledges stops as they happen). Gateway detects the stop; operator classifies it. More accurate for classification, still depends on operator discipline for duration. Typical error: 5–10 %.
3. Fully automatic capture (PLC signals or digital-I/O gateway). The machine tells the system when it is running. Sub-second accuracy. No operator interpretation required for duration; the operator's role is to add a reason code when prompted.

The Klocke pharma-packaging implementation illustrates the gap. At the Weingarten site, fully-automatic capture via digital-I/O gateways delivered 7 additional hours of production time per week within the first three weeks — not because the plant started running more, but because the previously-invisible stops and speed losses became visible and addressable. The 7 hours didn't appear; they had been there all along, mis-booked as production. This is the consistent pattern across every MDE rollout I have personally worked on for more than two decades.

FAQ

What's the difference between production time and cycle time?
Production time is the total time the line is producing, measured at plant or shift level (hours per shift). Cycle time is the time between two consecutive finished parts, measured at the unit level (seconds per part). Production time ÷ cycle time ≈ number of parts produced, if the machine ran at nominal speed without losses.

Is production time the same as uptime?
Closely related but not identical. Uptime is the portion of scheduled time the machine was operational (available); production time is the portion during which it was actually producing output. A machine can be "up" but idle, or "up" in setup — both count as uptime but neither as production time.

How accurate is manually-recorded production time?
In the plants SYMESTIC has benchmarked, manual booking overstates production time by 10–20 % on average. The error is structural, not dishonest: operators reliably forget short stops, round shift boundaries, and book setup time as production. Automatic capture closes this gap and is typically the first measurable gain of an MDE deployment.

What's a good production-time utilisation?
Expressed as (Operating Time ÷ Planned Production Time), typical discrete-manufacturing plants run 65–80 %. Above 85 % is strong; below 60 % indicates significant availability losses (breakdowns, long changeovers) or planning problems (line starved of orders, material, or people). The number is most useful when broken down by loss reason, not as a single aggregate.

How do I increase production time?
In this order: (1) measure it accurately — manual estimates will hide the biggest opportunities; (2) reduce unplanned downtime through TPM and structured maintenance; (3) cut changeover time via SMED; (4) address speed losses and micro-stops, which usually turn out to be the single largest invisible loss; (5) only then revisit the shift plan. Trying to "add hours" through overtime before these steps is almost always the worst choice.

Related: OEE · Production Efficiency · Cycle Time · Takt Time · Production Optimization · Machine Data Capture (MDE)

About the author

Martin Brandel

MES Consultant at SYMESTIC. 30+ years of industrial automation and machine connectivity — from Simatic S5 in 1991 to OPC UA and IoT gateways today. Dipl.-Ing. Nachrichtentechnik. Specialist in brownfield MDE/BDE, retrofit projects, and capturing accurate production time from mixed machine parks without PLC intervention or production interruption. · LinkedIn