In his 2013 book titled "Gemba Walks", 2nd edition, Dr. Jim Womack makes the statement that "quality is free". The claim is made within the context of a manufacturing environment where rework (the process of repairing defects on products just built) is far more expensive than doing it right the first time. Womack is not alone. Phillip Crosby's 1979 book "Quality is Free" coined the term.

Intuitively, most of us know this. We also assume that doing things right the first time requires a larger upfront investment, more time planning, or both. I've written previously about "good, cheap, fast" and how, in some cases, the "pick two" logic simply doesn't hold true. As one quality assurance executive said: "the cost of good quality is the cost of the process. It is cheap if the process fits the design and people can execute the process." This implies that the product design, production steps, and process executors (i.e., the people) must all work in harmony to keep the cost of good quality down.

Suppose you and I are working on some software together. Should we pair up and work together or code independently to increase our output? The concept of "pair programming" suggests the first option is superior. Some unofficial, unsubstantiated data suggest that programmers are 15% slower when working in pairs, but have 15% fewer defects. So it sounds like an equal trade-off between time and quality, right?

That's superficially true, but this choice is a no-brainer. The cost of producing code that works most of the time (but has bugs) is generally low. The 15% reduction in defects regularly translates into lower total costs, even when accounting for the slower pace of work. Having worked in manufacturing as a young man, I saw highly paid engineers reworking defective, half-built products. I knew their cost to the plant was more than a handful of assembly and test technicians like me. Maybe 15% more line workers contributing to this process would bring smoother flow and fewer defects, reducing the demand for expensive rework.

Make no mistake, the time value of money definitely plays a role here, not to mention the importance of being early to deliver products in emerging markets. I understand all that. However, I cannot recall a single personal example where after-the-fact poor quality costs were less than the "doing it right" costs. I'm also not bashing the concept of "minimum viable products" designed to validate business assumptions. I am simply claiming that rework is expensive and that, if you build it right the first time, quality costs nothing. Think about it.

Quality is conformance to requirements, not goodness or gold-plating. Put another way, quality is doing what you said you were going to do.

All Blogs -- Main Page