Cycle Time Analysis

What Cycle Time Analysis Actually Measures

Cycle time is the elapsed time from the start of one work cycle to the start of the next at a defined observation point — usually the end-of-line unit.

Cycle time analysis breaks that elapsed time into elements: pick, place, fasten, inspect, walk, wait.

Element-level breakdown is what turns a number into a plan — an average cycle time tells you nothing about which element to attack, the element breakdown tells you exactly.

• Value-added elements — the customer would pay for these
• Necessary non-value-added — required today, target for elimination or offload
• Waste — the seven wastes, direct target for removal
• Wait — imbalance signal, feeds line-balancing work

How to Do a Cycle Time Analysis

Follow a repeatable sequence.

Rushing this produces numbers no one trusts.

• 1. Define the work elements with the operator, not for them.
• 2. Choose an observation window that covers normal variation — different operators, shifts, and part variants.
• 3. Observe 30+ cycles per operator (stopwatch) or the full recording window (AI workstation analysis).
• 4. Tabulate element times per cycle. Keep raw data — do not go straight to averages.
• 5. Plot the distribution: median, 5th and 95th percentile, variance. The tails are where the story lives.
• 6. Compare against takt. Any element above takt is a bottleneck candidate.
• 7. Update standard work with the operator, based on the observed best method, and audit its use.

Why the Average Is a Lie

Average cycle time is the number that gets reported and the number that misleads.

A workstation with a median of 47 seconds and a 95th percentile of 78 seconds behaves nothing like a workstation with a median of 55 seconds and a 95th percentile of 58 seconds — even though the averages might be similar.

The tail is what causes the missed schedule, the overtime and the WIP buildup.

Any cycle time analysis that reports only an average is failing the plant that commissioned it.

AI-Accelerated Cycle Time Analysis

AI workstation analysis extracts element-level cycle times from video across every observed cycle — not a sample of 30.

The engineer defines elements once, validates the model's classification on a sample, and gets full distribution across every operator and shift.

Time from recording to first defensible analysis drops from days to hours.

The full method is described on the AI Workstation Analysis page.

What to Do With the Result

A cycle time analysis is useful only if it changes what happens at the workstation.

Every completed analysis produces: an updated operator standard work sheet posted at the station, a yamazumi that feeds the next line-balance conversation, and a top-3 improvement list that enters the daily management cadence with a named owner and a due date.

Analysis without a floor-level installation is a report — not an improvement.

Frequently Asked Questions