Part I: Waste, Variation, and Other Productionally-Transmitted Diseases

(excerpted from The Pipeline: A Picture of Homebuilding Production, originally posted on Escape from Averageness in May 2010, updated and reposted here)

“I want to talk about variation and uncertainty”, said the intrepid, results-based consultant.  “It was mentioned earlier that variation is a form of waste, because, to the extent that it does not add value, variation is wasteful.  That thinking largely comes from the Toyota Production System and the Lean Production methodology that TPS later spawned.  But – Taiichi Ohno did not include variation as a form of waste (muda) in the TPS, opting instead to use the term for unevenness (mura), and associate it with the separate principle of stability;  production processes need stability, and variation causes instability.

“In both Lean Production and the TPS, the principle of stability sits between the principle of waste and the principle of standardization, which includes the concept of standards, visual management, and problem-consciousness, which are linked to PDCA problem-solving.  I think there is a reason Mr. Ohno chose to not place waste and variation in the same category.  Given their inherent characteristics, a desire to remove waste in all of its forms is more reasonable than a desire to eliminate variation.”

Turning to the board, the intrepid, results-based consultant wrote:



“Beyond that distinction, there is the series of basic laws of production physics that directly or indirectly form our understanding of variation;  two laws in particular”, she said.  “The first law, called the Law of Variability, states that higher levels of variation degrade the performance of the production system.

“The second law, the Law of Variability Buffering, says that variation will always be buffered by some combination of inventory, capacity utilization, or time.”



“If all RB Builders does is attack variation by attacking waste (in the form of errors, rework, etc.), and it fails to directly attack the variation that causes instability, then it will have to live with a production system that protects itself with some combination – buffers itself with some level – of additional work-in-process, longer-than-necessary cycle times, or wasted capacity.

“Our production system will default to longer durations.  Time is the self-determining buffer, the ‘buffer of last resort’, so to speak.  If we do nothing about variation, yet limit work-in-process and capacity, the result will be long cycle times.


“Buffers – high levels of work-in-process, long cycle times, and unused capacity – allow the system to compensate for variation, but, they result in lost Throughput.  The true cost of variation is the financial throughput – the Gross Income – that RB Builders surrenders to that variation.

“You begin to get a sense that variability is a very big deal”, said the intrepid, results-based consultant.  “Jack Welch used to say, ‘Variation is evil’.  Some variation and uncertainty is natural, and some of it is necessary and planned.  But, the instability that variation and uncertainty cause is a decidedly evil form of waste.

“In process religion, the gods of production will not be mocked”, she said.  “If we fail to directly attack the variation that causes instability, we will have to accept a production system that protects itself with some combination of too much work-in-process, long cycle times, or reduced throughput.”

“But – isn’t protection a good thing?”, deadpanned a superintendent.  “Shouldn’t we be practicing safe production?”

“Yes, protection would be a good idea”, responded the intrepid, results-based consultant, equally deadpan.  “Especially for boys like you, who should be worried about contracting a PTD.

“Put it this way”, she said.  “Which one of the Productionally-Transmitted Diseases would you like to have?  Which combination of longer-than-necessary cycle times, higher-than-necessary levels of work-in-process, and lower-than-possible rates of throughput (because of excess and unused capacity) do you really want?

“Some level of variation and uncertainty is natural, inevitable, and unavoidable”, she said.  “Some variation is necessary, just to protect ourselves in the marketplace;  we don’t offer only one floorplan and elevation.  Whenever we are protecting the output of the system from variation and uncertainty, some level of protective capacity or buffering is needed.

“But – protecting a system from variation comes at a cost, and to the extent that the variation that necessitates the buffering is unnecessary, avoidable, excessive, or uncontrollable, it is a very bad thing.”


(The Pipeline: A Picture of Homebuilding Production is available on the publisher website (, and the author website (, as well as,, and



  1. Ping from Steven G. Twiss:

    Builders with little to no variation are destined to fail in the US as they are quickly identified by potential customers as “them, again”, and ignored.

    The problem is solved when the process, starting with the CAD system through manufacturing, is adept and accepts a large amount of variation without slowing production either in the factory or the field.

  2. Ping from Fletcher Groves:

    Steve — variation is not the same as product variety. The only way a production system can accept a large amount of variation is by having excessively long process duration, higher levels of work-in-process, or excess unused capacity. None of these alternatives are positive or desirable.