Cycle time, throughput time and lead time for one part, worked through

The terms, as this page uses them

Plants use these words in different ways, so agree on the definitions before you compare numbers. This page follows the Lean Enterprise Institute's Lean Lexicon where it can and says where it does not.

Cycle time has two common meanings. In lean and on value stream maps it is a step's rhythm: the time to make one part at a station, as measured, which is the Lean Lexicon's definition. In the Factory Physics textbook, and in some plants' reports, cycle time is the average time a job takes from release at the start of its routing to the end of it, which this page calls throughput time. For the same shaft the first meaning gives 84 seconds and the second gives 7.0 days, so ask which one a number is before you act on it. This page uses the first.

Throughput time varies too. The Lean Lexicon treats throughput time and production lead time as one measure, door to door. Many plants use throughput time for the shop floor only, first step to last, and production lead time for dock to dock. This page keeps them apart so you can see what waits outside the machines: raw material and packed parts. Throughput on its own is a rate, not a time: the parts a day that leave the process, 600 here.

Takt time and cycle time: can the line keep up?

Takt time is 54,000 s ÷ 600 = 90 s: the line has to finish a shaft every 90 seconds on average. A step's cycle time is the time on one machine divided by the machines doing that step. Grinding takes 150 s per shaft, but two grinders work side by side, so the step finishes one every 75 s.

• The lathe is the bottleneck. At 84 s it is the slowest step, so the line can make at most 642 shafts a day.
• Every step is under takt, so the line can meet demand, but only just at the lathe: 600 shafts take 600 × 84 = 50,400 s of its 54,000, which leaves 3,600 s, or 60 minutes a day. More than an hour of lathe stops in a day means a short day or overtime.
• Adding the step cycle times gives 326 s, and that sum means nothing. It is not how long a shaft is worked on (401 s, because each grinder spends 150 s on a shaft) and not how long a shaft takes to get through (days).

Production lead time and Little's Law

Production lead time adds the two piles outside the six steps: bar stock waiting at receiving and packed shafts waiting for the truck. A value stream map gets it the way Learning to See teaches: divide each pile by daily demand to turn it into days, add the days, then add the processing time.

Little's Law says that in a stable process, average work in process (WIP) = throughput × average time in the process, so lead time = WIP ÷ throughput. Count everything between the receiving dock and the truck: 7,800 shafts in piles, plus the shafts on machines. A machine holds a shaft only while it works on it, so on average 600 ÷ 54,000 × 401 = 4.5 shafts are on machines. WIP is 7,804.5 shafts, and 7,804.5 ÷ 600 a day = 13.007 days: the same as the waits and processing times added up for the tagged shaft. It has to be, because each pile's days are Little's Law applied to that pile.

That makes Little's Law a check on your own numbers. Count the WIP one morning and divide it by what you ship per day. If tagged parts take much longer or much less than that on average, a pile is growing or shrinking, the count missed something, or the WIP and the shipments were counted between different points.

• Use averages over several weeks. Little's Law is about averages, and one tagged shaft is one sample.
• Divide by what actually leaves the process, not by a forecast. Learning to See divides by daily customer demand, which is the same number only when the line ships what the customer takes.
• Count WIP and throughput between the same two points. Bar stock counts in the dock-to-dock figure; it does not count in the saw-to-pack throughput time.
• It holds whatever order parts are taken in, first in first out or not. It tells you the average wait, not what any single part will wait.

Factory Physics adds a benchmark. Raw process time (401 s here) multiplied by the bottleneck rate (one shaft every 84 s) gives the critical WIP: the parts a line with no variation at all would need to run at full rate, 401 ÷ 84 = 4.8 shafts. A real line needs more, to ride through stops and changeovers, but 4,200 shafts between the saw and the packing bench is far more than that, so most of the WIP is there for other reasons, such as the saw cutting whole bar lots ahead.

Value-added ratio

Value-added time is 280 s. Production lead time in seconds is 13.007 days × 54,000 = 702,401 s. The value-added ratio is 280 ÷ 702,401 = 0.04%. Counting every second of processing, 401 s, it is still only 0.06%. Both say what the timeline shows: the time is in the piles. Some plants divide by calendar time instead of working time, which makes the ratio smaller still, so say which you used.

Measure your own in one shift

1. 1

   Pick one part and the two ends

   One part number or a family that shares the steps. Write down where the clock starts and stops, such as receiving to truck, or first step to last, and keep the same ends next time.

2. 2

   Work out takt

   Available time per day after breaks, divided by the parts the customer takes per day. The takt time calculator does it from shift time, breaks and demand.

3. 3

   Time each step and find the bottleneck

   Time at least ten parts at each step with a time study sheet, divide by the machines doing that step, and compare with takt in the cycle time calculator. The bottleneck calculator adds machine uptime.

4. 4

   Count every pile on one walk

   Count the same morning, from the truck back to receiving, raw material and packed parts included, and divide each pile by daily output. The lead time worksheet does the sums and names the biggest pile.

5. 5

   Check it two ways

   Divide total WIP by what you ship per day in the lead time calculator, then tag five parts over two weeks and add up their waits in the throughput time calculator. If the two disagree badly, recount.

6. 6

   Trace a few orders and draw the map

   Take the dates from recent orders: received, released, started, packed, shipped, delivered. Then draw the flow on the value stream map template and plan the changes with the future state VSM worksheet.