Production Metrics: Cycle Time, Takt Time, and OEE

A production floor runs on a handful of numbers. They tell you whether the line can keep up with orders, where it is losing time, and how much of what it makes is good. You do not need a math background to use them. You need to know what goes into each number, because the most common mistakes come from counting the wrong thing.

What You Will Be Able to Do

• Tell cycle time, takt time, and lead time apart.
• Calculate takt time from available time and customer demand.
• Find a line's bottleneck and its maximum output.
• Calculate OEE and say what each of its three factors shows.
• Calculate first pass yield and read a machine repeatability spec correctly.

Three Times That Get Confused

Cycle Time

Cycle time is the total time from the start of one production cycle to the start of the next. It covers everything that repeats, not only the time the machine is working.

A press runs for 18 seconds. Loading and unloading take 6 seconds, and a quick check takes 4 seconds. The cycle time is 18 + 6 + 4 = 28 seconds. Hourly output is 3,600 seconds ÷ 28 = about 128 parts.

To measure it, time at least 10 cycles from the same point in each cycle, then take the average. One cycle can fool you. A common mistake is reading the run timer on the control and calling that the cycle time, which leaves out loading, unloading, and checks.

Takt Time

Takt comes from a German word for the beat in music. Takt time is the pace at which you must produce to meet customer demand.

Takt time = available working time ÷ customer demand

Available time means planned working time. Take out breaks, lunch, and planned meetings. Do not take out breakdowns, because those are the losses you are trying to find.

Worked example:

1. An 8-hour shift is 480 minutes.
2. Subtract a 30-minute lunch and two 10-minute breaks: 480 - 50 = 430 minutes.
3. Convert to seconds: 430 x 60 = 25,800 seconds.
4. The customer needs 600 units per shift.
5. Takt time = 25,800 ÷ 600 = 43 seconds.

One unit must come off the line every 43 seconds. Takt comes from the customer, and cycle time comes from the process. Comparing the two tells you where you stand.

Bottlenecks and Line Output

A line cannot run faster than its slowest station. That station is the bottleneck, and its cycle time sets the pace for the whole line.

Worked example, for a line with four stations:

The slowest is station 2 at 48 seconds. Maximum output is 3,600 ÷ 48 = 75 units an hour. The other stations are faster, but they cannot push more through than station 2 lets pass.

Speeding up station 1 gains nothing. Parts just pile up in front of station 2. Cut station 2 to 40 seconds, and it now ties with station 3. Output rises to 3,600 ÷ 40 = 90 units an hour, and both are now the bottleneck. When two stations tie for slowest, improving just one changes nothing. You have to improve both to go any faster.

OEE: Overall Equipment Effectiveness

OEE stands for Overall Equipment Effectiveness. It combines three factors into one percentage that shows how well a machine is used: of the time it was scheduled to run, how much made good parts at full speed. Managers track it because it shows where the lost time went.

OEE = Availability x Performance x Quality

Worked example, for one shift:

1. Planned production time is 450 minutes.
2. The machine was down 45 minutes for a changeover and a jam, so run time is 405 minutes.
3. Availability = 405 ÷ 450 = 90 percent.
4. The ideal cycle time is 0.5 minutes per part, and the machine made 729 parts.
5. Performance = (0.5 x 729) ÷ 405 = 364.5 ÷ 405 = 90 percent.
6. Of those 729 parts, 700 were good.
7. Quality = 700 ÷ 729 = 96 percent.
8. OEE = 0.90 x 0.90 x 0.96 = 0.778, or about 78 percent.

Each factor looks fine by itself, but together they show that more than a fifth of the planned capacity was lost. You can check the answer another way: good parts x ideal cycle ÷ planned time = 700 x 0.5 ÷ 450 = 78 percent.

The 85 Percent Benchmark

An OEE of 85 percent or higher is widely quoted as world-class. The figure comes from Seiichi Nakajima's work on Total Productive Maintenance. It is usually broken down as about 90 percent availability, 95 percent performance, and 99.9 percent quality.

Treat it as a target to work toward, not a pass or fail grade. Many good operations run below it. The most useful comparison is a machine against its own past results, because different products and processes do not compare cleanly. OEE only works when the inputs are honest, so log your stops, reasons, and scrap as they happen. If the ideal cycle time is padded or short stops go unrecorded, the number looks better and the real losses stay hidden.

First Pass Yield

First pass yield is the percentage of units that make it through a process right the first time, with no rework and no scrap.

First pass yield = units good the first time ÷ units started

A line starts 500 units. Twenty need rework and 10 are scrapped, so 470 were good the first time. First pass yield = 470 ÷ 500 = 94 percent.

Final yield would count the 20 reworked units as good, which gives 490 ÷ 500 = 98 percent. That higher number hides the rework, which is why first pass yield is the better measure of the process.

For several steps in a row, multiply the yields. Steps at 98, 95, and 99 percent give 0.98 x 0.95 x 0.99 = 92 percent overall.

Reading a Repeatability Spec

Repeatability describes how closely a machine returns to the same commanded position over repeated cycles under the same conditions. It applies to machine tools, robots, presses, and pick-and-place equipment.

The plus-or-minus sign matters. A number with ± means that much on each side of the target, so the total range is twice the number.

Worked example: a machine has a repeatability of ±0.002 inch. Commanded to 5.000 inches, it can land anywhere from 4.998 to 5.002 inches, a total range of 0.004 inch. If your part tolerance is ±0.003 inch, a total of 0.006 inch, the machine's own spread uses up two-thirds of your tolerance before tool wear or temperature change takes any.

Repeatability is not the same as accuracy. Accuracy is how close you land to the target. Repeatability is how close the results land to each other. A machine that always lands 0.010 inch high is repeatable but not accurate, and an offset can fix it. A machine that scatters cannot be fixed with an offset.

Common Mistakes

Check Your Understanding

1. A plant has two 7-hour shifts of available time and must ship 1,200 units a day. Its slowest station takes 45 seconds. Can it keep up?

Available time is 14 hours, or 50,400 seconds. Takt is 50,400 ÷ 1,200 = 42 seconds. The 45-second station is slower than takt, so the plant cannot keep up. It would make 50,400 ÷ 45 = 1,120 units, which is 80 short.

2. Three stations run at 25, 36, and 30 seconds. A team cuts the first station to 20 seconds. What happens to output?

Nothing. The bottleneck is the 36-second station, so output stays at 3,600 ÷ 36 = 100 units an hour.

3. A machine has 80 percent availability, 95 percent performance, and 98 percent quality. What is its OEE, and where would you look first?

OEE = 0.80 x 0.95 x 0.98 = about 74 percent. Availability is the weakest factor, so start with breakdowns and changeovers.

4. A positioning table is rated at ±0.0002 inch repeatability. What total spread should you expect?

0.0004 inch, because the spec applies to each side of the target.

Key Takeaways

• Cycle time is start of one cycle to start of the next. Takt time is available time divided by demand.
• If cycle time is above takt, you cannot meet demand.
• The slowest station sets line output, so improve the bottleneck first.
• OEE is availability times performance times quality, and 85 percent is the widely quoted world-class benchmark.
• First pass yield counts only units that were right the first time.
• A repeatability of ±X means a total range of 2X.