To determine a standard cycle time, time the job across several cycles, adjust that observed time for the operator's pace to get a normal time, then add an allowance for personal needs, fatigue, and unavoidable delays. In formula terms: standard time = observed time × performance rating × (1 + allowance). The result is the time a trained operator should take, sustainably, at a normal pace.
Standard cycle time is the number half your plant argues about. Set it too tight and people ignore it; set it too loose and it hides losses and inflates OEE. The goal is a standard that is both defensible and trusted, built the same way every time, from real observation, with the allowance stated openly. This guide covers the classic time-study method, the data-derived alternative, the critical difference between a standard time and OEE's ideal cycle time, and how to keep the number honest.
What is standard cycle time?
Standard cycle time is the target time to complete one unit of work, set for a trained operator working at a normal pace under normal conditions, including reasonable allowances. It is not the fastest time anyone ever hit, and it is not a raw average of whatever the stopwatch showed. It is a rated, adjusted figure meant to be repeatable day after day, the basis for scheduling, costing, staffing, and line balancing.
The word "standard" carries weight: it means agreed, documented, and stable. A standard cycle time that quietly changes whenever a number looks bad is not a standard, it is a fudge factor. The whole value of the figure comes from setting it with a defensible method and then defending it, so that when actual time drifts away from standard, everyone treats the gap as a signal to investigate rather than an excuse to move the goalposts. Distinguish it from raw cycle time which is simply the observed time between units with no rating or allowance applied.
How do you run a time study?
A time study builds the standard from stopwatch observation in three moves: observe, rate, and allow. Each move corrects for something the raw stopwatch misses.
- Break the job into elements. Split the cycle into a few distinct, timeable steps with clear start and stop points. Elements make the study repeatable and let you spot which step varies most.
- Time enough cycles. Record the time for each element across multiple cycles, commonly 10 or more, to average out normal variation. Drop obvious anomalies (a dropped tool, a phone call) and note why.
- Compute the observed time. Average the cycles to get the observed (or "select") time per element, then sum the elements. This is what actually happened, at whatever pace the operator worked.
- Apply a performance rating. Judge the operator's pace against a normal pace, where 1.00 is normal. Normal time = observed time × rating. A rating of 1.10 means they worked 10% faster than normal, so normal time is higher than observed; 0.90 means slower.
- Add the allowance. Multiply normal time by (1 + allowance) to cover personal time, fatigue, and unavoidable delay. Standard time = normal time × (1 + allowance).
The two judgment calls, the rating and the allowance, are where studies go wrong, so both should be set by a consistent method and written down, not eyeballed differently each time.
What is a normal allowance to add?
Allowances typically fall in the 10–20% range, covering three things: personal time, fatigue, and unavoidable delay, together called PF&D. Personal time is for basic needs; fatigue accounts for the pace drop over a shift; delay covers short, unavoidable interruptions like waiting briefly for material. The heavier and hotter the work, the higher the fatigue component, which is why a light bench-assembly allowance sits near the bottom of the range and hot, physical work sits higher.
The allowance is not padding, it is the difference between a sprint time and a sustainable time. A standard with no allowance describes a burst nobody can hold for eight hours; the same job run all shift will always miss it, and people will stop believing the number. The U.S. Department of Labor's guidance on incorporating PF&D allowances into work standards treats these allowances as a normal, expected part of a fair standard, not a concession (Fact Sheet #39D). Set the allowance by policy, one documented figure per work type, so it is applied the same way for every study and cannot be quietly dialed up or down to hit a target.
What are the reference formulas and figures?
The method rests on a small set of standard formulas and one government reference for the allowance, all worth stating plainly.
| Quantity | Formula or figure | Basis |
|---|---|---|
| Normal time | observed time × performance rating | Standard time-study method |
| Standard time | normal time × (1 + allowance) | Standard time-study method |
| Typical PF&D allowance | ~10–20% | Industrial engineering practice |
| PF&D allowance in wage standards | Personal, fatigue, delay | U.S. DOL Fact Sheet #39D |
The U.S. Department of Labor's Fact Sheet #39D is a primary source that treats personal, fatigue, and delay allowances as a normal component of a fair work standard, useful when someone claims an allowance is padding. The two formulas above are the entire arithmetic of a time study; everything else is judgment about the rating and the allowance. Keep the figures on one page so any standard on the floor can be traced back to how it was built, the same transparency that keeps a loss analysis honest.
Can you derive a standard from machine data instead?
Yes, where machines already log cycle times, you can derive a standard from the data instead of a stopwatch. This is often more honest, because it uses hundreds or thousands of real cycles rather than a 20-minute study, and it removes the observer effect where operators speed up while being timed. The method is to pull the distribution of actual cycle times and choose a defensible point on it, commonly a low percentile or the repeatable mode, as the standard, rather than the raw average, which is dragged up by stoppages.
The trade-off is that machine data captures machine time cleanly but says little about manual pace, ratings, or why the slow cycles were slow. So the strongest approach is usually a blend: derive the machine-paced portion from cycle logs, and use a short time study to set the manual elements and the allowance. Data tells you what the machine repeatedly does; the study tells you what a person can sustain and why. For any operation that is mostly machine-paced, the data-derived standard is both easier to maintain and harder to argue with, it is a record, not an opinion.
How is standard cycle time different from ideal cycle time?
They are different numbers built for different jobs, and mixing them corrupts OEE. Ideal cycle time, the one OEE's Performance factor uses, is the fastest repeatable speed the process can run, with no allowance for breaks or fatigue. Standard cycle time includes the allowance and a normal (not maximum) pace, because it is built for scheduling and costing where sustainability matters. Standard time is therefore always longer than ideal time.
The trap is plugging the standard cycle time into OEE as the ideal. Because the standard is padded with allowance, using it as the ideal makes the line look like it is running near 100% Performance when it has real speed losses hiding inside the allowance. Performance can even read above 100% if the standard is loose, which is the classic sign the wrong cycle time is in the formula. Keep the two separate: ideal cycle time for the OEE Performance factor, standard cycle time for planning, staffing, and cost. If you need to reconcile them for a machine-paced line, see flow rate vs cycle time which unpacks the reciprocal between rate and time.
How do you set a standard people trust?
A trusted standard comes from a transparent method, applied consistently, and revisited on a schedule. The steps that build trust are less about the arithmetic and more about the process: involve the operator, show the rating and allowance openly, and never move the number to make a metric look good. When people can see how the standard was built, these cycles, this rating, this documented allowance, they stop treating it as a weapon and start treating it as a shared reference.
Then keep it current. Re-study when the method, tooling, or product changes, not when someone dislikes the result, and log the reason for every revision so the history is visible. A stable standard you trust beats a precise one you game, exactly as with OEE itself. The same logic that makes machine-derived counts more honest applies here: standards built from real cycle data on the floor, the way Harmony captures it (see the platform), are a record rather than an argument, and the CLS case study shows the payoff. Once your standard is set, put the line's ideal cycle time and losses through the OEE calculator and see how the standard supports adherence to plan and honest scheduling.