Section 46. TimeGlobal Process Cycle Time

46. TimeGlobal Process Cycle Time

Note

As with other definitions described in "TimeIndividual Step Cycle Time," "TimeProcess Lead Time," "TimeReplenishment Time," and "TimeTakt Time" in this chapter, there is disagreement between different Process Improvement camps. To be frank, it really doesn't matter which naming convention you select, as long as you measure the right thing from a project perspective and do the right things based on it. In the subsequent description, focus on what to measure, as opposed to what it's called.

The naming convention used here is chosen for its clarity and practicality. If a naming convention is already established in your business, then use that; however, a little digging usually uncovers a lot of inconsistency and misuse.

In reports, always list the metric and how you define it; that way confusion and inevitable fruitless debate are reduced.

It is useful to read "TimeIndividual Step Cycle Time," "TimeProcess Lead Time," "TimeReplenishment Time," and "TimeTakt Time" in this chapter prior to proceeding with data capture to ensure comprehension.

Overview

(See also "TimeIndividual Step Cycle Time" in this chapter.)

Cycle Time, when applied to the whole process, is the time between entities being processed (not the time taken for a single entity to traverse the whole process, which is known as the Process Lead Time).

Thus, the Global Process Cycle Time is measured by standing at the end of the process and timing between entities as they exit the process, as shown in Figure 7.45.1.

Figure 7.45.1. Graphical representation of a Test of Equal Variance.

Figure 7.45.2. Test of Equal Variances for two operators: Bob and Jane (output from Minitab v14).

Global Process Cycle Time and Process Lead Time can be the same if the process is always empty and processes only one entity at a time (i.e., to get an entity to fall out of the end of the process, that entity has to traverse the process from beginning to end).

Clearly the Global Process Cycle Time is driven both by the Individual Step Cycle Times and the positioning of inventory (Work in Process) in the process. The Global Process Cycle Time can be reduced significantly by having the correct Work in Progress (WIP) in the system.

There is also an obvious relationship here between Global Process Cycle Time and Takt Time (the pace of customer demand). Global Process Cycle Time should be balanced with Takt Time to go quickly enough to meet customer demand and yet not so quickly that you generate excess inventory. Ideally, the Global Process Cycle Time is set up to be slightly less than Takt Time, typically 95% (i.e., quicker) to retain flexibility and robustness to variation in demand.

Logistics

Calculating the average Global Process Cycle Time can usually be done from data captured over a period of a week or so. To calculate variation in the Cycle Time, a longer study is needed to allow normal variation to appear in the process; for example, if there is weekly variability, then data has to be captured over a few weeks to be able to see it.

It takes approximately one hour of a preparatory meeting to structure a data capture method. It is always useful to do a dry run data capture, to ensure the correct data is being captured. For more detail regarding the data capture, see "KPOVs and Data" in this chapter.

Global Process Cycle Time inevitably varies by entity type. It is useful, therefore, to capture either data for just one entity type or preferably data for all (or multiple) entity types and then stratifying the Time by type during the data analysis.

Roadmap

It is possible to calculate the Global Process Cycle Time from as few as 710 data points. By taking the truncated mean (discarding the highest and lowest points and averaging the rest) a good approximation can be made. If understanding the variation in the Global Process Cycle Time is required, then 2530 data points are required as a minimum.

If multiple entity types are examined, seven data points per type allow calculation of an average for each type as well as the whole. If an understanding of variation is required by type, then 2530 data points per type are required.

Step 1.	Determine the timing point for an entity. Usually this is the exact time at which an entity exits the process, but there should be a clear operational definition for this, so that timing is consistent during the data capture. At this point for an entity, the clock starts and when the subsequent entity meets the same timing point the clock stops.
Step 2.	If the Global Process Cycle Time is less than one minute, then it is probably best to use a stopwatch, starting it when the first entity exits the process and stopping it when the second entity exits the process. This is the Global Process Cycle Time for the second entity, so if the entity type varies, then record the entity type of the second entity (not the first). If the step takes significantly longer than a minute then it probably is best to write down the start and stop times and then calculate the Global Process Cycle Time later. Record the times and also the entity type being processed to allow stratification by type later during the data analysis. It is sometimes useful, if the Global Process Cycle Time is short, to time three or more consecutive entities passing through the step to get a better reading on the average time.
Step 3.	From the captured data, calculate mean and variation (standard deviation) in the Global Process Cycle Time.

If data for multiple entity types was captured, analyze the data as a whole and then conduct separate analyses stratifying by entity type.

Interpreting the Output

Interpreting the Global Process Cycle Time should not be done in isolation from other elements. The trick is to understand the relationship between the internal times relating to our process:

• Process Lead Time Total time for an entity to progress through the process

• Replenishment Time to Customer Time to fill the Customer's "in-box" when they request it be refilled

• Global Process Cycle Time Rate (expressed in time) at which the process can generate entities

And to understand the relationship with the external Customer or Market related times:

• Requested Delivery TimeHow quickly the Customer needs delivery of an entity when they request it

• Takt TimeThe pace of usage of entities by the Customer, expressed in time

If the Global Process Cycle Time is greater than the Takt Time, then the process cannot meet Customer demand and options are

• Shedding load from the process to speed it up

• Eliminating work content (specifically NVA)

• Providing the correct Work In Process inventory to ensure no steps are starved of entities to work on

• Adding more resources to the process, specifically on the bottleneck steps

• Using multiple lines in parallel