Takt time, cycle time and lead time, with one example

Takt time is the pace customer demand requires, cycle time is how long one cycle actually takes at the machine, and lead time is how long an order or a part takes from start to finish. Takt is calculated from demand and available time. Cycle time is measured on your station. Lead time is measured across the whole flow, waiting included, and usually runs in days, not seconds.

This guide works one hypothetical machining center through all three, computes each figure two ways, and ends with what to do when cycle time runs past takt. How to measure cycle time covers the timing itself.

What are takt time, cycle time and lead time?#

Each one answers a different question about the same station: how fast it must go, how fast it does go, and how long work takes to get through.

The Lean Enterprise Institute (LEI) gives a plain example: a plant that runs 480 minutes a day for a demand of 240 widgets has a takt time of 2 minutes.^[1] Takt is German for a precise interval of time, like a musical meter, and the idea was first used in German aircraft production in the 1930s.^[1]

LEI’s condition is that cycle time is timed by actual measurement.^[2] The glossary entry for cycle time separates the ideal you enter from the actual the machine reports.

LEI defines production lead time, also called throughput time, as the time a product takes to move all the way through a process or value stream. Order lead time adds the time to get the product to the customer and the delays in entering the order, so it is how long the customer waits.^[2]

Three mix-ups to avoid#

• Takt is not the time a unit takes to make. That is cycle time. Takt doesn’t depend on the machine at all: double the demand and takt halves, on the same machine.
• A longer takt means a slower pace. Demand of 900 in 900 available minutes is a 60-second takt. Halve the demand to 450 and takt is 120 seconds, so the station can go slower, not faster.
• Cycle time is not raw material to finished product. That span is production lead time. A housing that cycles in 108 seconds can spend two days in the cell.

Takt, cycle and lead time side by side#

LEI ranks the measured times: value-creating time is usually less than cycle time, which is less than production lead time.^[2] Takt sits apart from both, because it comes from the customer, not from the process.

How do you calculate takt time for a station?#

You divide the time the station is available to run by the demand it must meet over the same period. Use one period for both: a day’s demand against a day’s available time.

Available production time is the scheduled time less planned breaks. Don’t subtract breakdowns or slow cycles, and don’t add scrap to the demand. Those are the station’s losses, and they belong on the measured side of the comparison.

Takt changes only when demand or available time changes. Win a new order or add a Saturday shift, and recalculate. A breakdown leaves it alone.

Does the station keep up with takt?#

The station keeps up when its cycle time per part is shorter than takt and its good output still covers demand once stops and scrap are counted. Check the cycle first, because it sets the ceiling.

The VF-2’s timed median is 108 seconds a cycle. Keep the figure per cycle, and divide by the 2 parts per cycle only to compare it with takt, which is per housing. Use this measured cycle, not the ideal cycle time, which is the target OEE performance is scored against.

The ceiling assumes the VF-2 never stops and never scraps a part. On a real day it stopped for 54 minutes and scrapped 10 housings. It ran 846 minutes, or 50,760 seconds: 470 cycles, 940 housings, 930 of them good.

Planned time per good housing is 54,000 ÷ 930 = 58.1 seconds. That is still inside takt, with 1.9 seconds to spare instead of 6. The flow KPIs explain why this is the figure to hold against takt, not the cycle.

You probably don’t need takt on a job-shop machine that runs one-off orders with no steady daily demand. Compare each machine’s load against its available hours instead.

How do you measure lead time?#

You measure production lead time by following work from release to finish, or by dividing the work in process by the rate the cell finishes it. Both give an average, and on a stable week they agree.

The second method is Little’s Law. In John Little’s proof, L is the mean number of units in a system, W is the mean time a unit spends in it, λ is the arrival rate, and L = λW, provided the averages hold steady over the period.^[3] On a shop floor, read it as work in process = throughput × lead time.

Thirty working hours is 108,000 seconds. Each housing spends 108 of them in a machining cycle on the VF-2, 0.1% of its time in the cell. The rest is queues, moves and the other operations.

Order lead time is longer again. For example, add a day for order entry, two days waiting for release and a day to pack and ship, and the customer waits 6 working days for a housing whose cycle takes under 2 minutes.

Lead time follows the queue, not the cycle. Cut the work in process from 1,800 to 900 housings at the same 900 a day, and lead time falls to 1 working day with no change to any cycle time.

What happens when demand rises past the cycle?#

When takt falls below the cycle time per part, the station falls behind every day, whatever its uptime. Take the VF-2 again, with demand up to 1,080 housings a day.

The gap doesn’t stay at the station. If you keep releasing 1,080 a day, 150 more housings join the queue in front of the VF-2 each day. After five days, that is 750 housings waiting, or 750 ÷ 930 = 0.8 days of the VF-2’s work added to every order behind them.

Measure cycle time from the machine’s own cycle#

IoTFlows supplies the measured side of the comparison. A SenseAi on each machine reads vibration and sound from outside, with no PLC connection, and reports part counts and cycle times derived from the machine’s cycle. The actual cycle time is the running time between counted parts, so stops stay out of it, as the metrics reference defines.

Each operation keeps its cycle time per cycle, with parts per cycle in a separate quantity per cycle field, the same convention as the VF-2’s two-housing fixture. Shift Production shows the ideal, goal and actual cycle side by side for every operation, with good parts against the shift goal.

Divide your available time by your demand, and hold that takt against the actual cycle. See production monitoring.

What to do when cycle time exceeds takt#

When cycle time per part exceeds takt, no downtime project closes the gap: the station is short even at zero stops. Change the cycle, the hours or the number of machines. The table works each option through on the VF-2 at 1,080 a day.

The cycle target needs room for today’s losses. The VF-2 turns 930 of its 1,000-housing ceiling into good parts, or 93%, so the cycle per housing must be 50 × 0.93 = 46.5 seconds or less: 93 seconds per two-housing cycle. Check it the other way: 54,000 ÷ 93 = 580.6 cycles, or 1,161 housings at zero stops, and 93% of that is 1,080.

The other rows follow from the takt formula. A third 450-minute shift makes 1,350 available minutes, so takt is 81,000 ÷ 1,080 = 75 seconds. A second identical machine halves each machine’s demand to 540, a takt of 54,000 ÷ 540 = 100 seconds. Building ahead works only for a short spike: today’s 30 spare housings a day take 25 days to stock the 750 that five days at 1,080 need.

On a manual station, the cycle is the operator’s work loop. Move work elements to a neighboring station until each loop fits inside takt.

Don’t speed up a station that is already well inside takt. Faster cycles there build housings ahead of demand, and the extra work in process lengthens lead time without shipping a single extra part.