After a number of email exchanges with +Andy Carmichael triggered by his original post The difference between Cycle Time and Lead Time… and why not to use Cycle Time in Kanban, and follow up post More Musings on Little’s Law, I will repost here a reply I sent to the Kanbandev list.
First: we should really drop the usage of Cycle Time, because it’s usage is ambiguous. There are currently two definitions of Cycle Time. No one is intrinsically wrong or right. Their validity depends on the background context.
The people from operations management typically refer to Cycle Time as the time from start to finish of operational work. This usage has a stronger background in engineering and physics. Donald Reinertsen uses this Cycle Time in his Flow book.
The people from lean manufacturing typically refer to Cycle Time as the average time between two successive units exiting the work process. This is the Cycle Time used by Taichi Ohno. Notice well that this Cycle Time is really a measurement of Throughput (items per period of time), because it is the reciprocal of throughput. It is similar to Takt Time with respect to Customer Demand rate.
Funny but true:
If you divide the Reinertsen Cycle Time by the Ohno Cycle Time, then you get WIP!
(This must be the most ambiguous way you can ever imagine to expresss Little’s Law!)
Since in the Kanban community we have a great deal of respect for both of these two schools of thought, it is hard to keep one consistent understanding of the term “Cycle Time.”
I used to use the term as Donald Reinertsen; but after long discussions with Andy Carmichael, I decided to drop the term in favor of:
Flow Time was one of the terms originally used to by Dr. Little in describing his law, when applying it to operations. Of course, I have a biased preference for this term because of my Tame the Flow.
Here are the terms I use in Tame the Flow (self-citing):
- Flow Time (FT), or “Process Time” or “Time in Process” (TIP): The time from the moment work is started to the moment work is finished; in other words the time that a work item spends as WIP (Work in Process). Flow time is used to compute throughput in terms of units of productivity. This is sometimes known as Process Time. We will use Flow Time to compute production throughput or operational throughput and to exploit Little’s Law. Note that in the Kanban Community, flow time is often referred to as lead time (and in particular as “system lead time”); though we will use lead time in a different sense as explained below. Other synonyms for this concepts are: cycle time, throughput time, and sojourn time; though they might be ambiguous.
- Lead Time (LT): The time from the moment a customer makes an order to the moment the customer has fully paid. We use LT to compute financial throughput performance (in the throughput accounting sense of TOC, i.e. money per period of time). Note that this is different from the that used in conventional business, where it relates to the period between order and delivery. We could refer to the latter as “Order Lead Time” signifying the amount of time an order has to be placed ahead of when its delivery is needed; but, as we will see below, we prefer to use Service Time instead.
- Response Time (RT), or “Reaction Time”: The time between the moment a customer makes an order and the moment when that order is actually acknowledged as a work item that will be handled by the service organization or team. In an Agile/Scrum setting, this would be the moment that the work item is added to the “backlog.”
- Service Time (ST), or “Resolution Time”: The time from the moment a customer makes an order to the moment work is delivered back to the customer. This is often known as the (Order) Lead Time. It is used mainly to compute throughput performance in terms of customer service (even when manufacturing and delivering tangible goods). While (Order) Lead Time might be what is considered when negotiating terms with customers, Service Time puts the emphasis on your own organization’s responsibility in servicing a customer need.
So let’s all stop using Cycle Time, and use something we all understand to mean the same thing. I strongly suggest Flow Time when referring to the Reinertsen Cycle Time; and if we ever need to refer to the Ohno Cycle Time, then just use (Operational) Throughput (though Andy suggests Delivery Rate) instead.
In the next revision of Tame the Flow I will replace all “Cycle Time” with “Flow Time.”