With most projects, initial specifications are vague and subject to multiple changes. Project managers encounter delays in receiving customer revisions as well as inconsistent or conflicting requirements. This leads to delay, rework and unhappy customers. Most misunderstandings are generated early in the project but are only detected much later. The result is increased fulfillment time, cost escalation, and ultimate lower quality.
There is a tradeoff between cost and quality. How can you optimize this? Think of it this way: Total Cost is the sum of the costs of two activities: conformance and non-conformance. Conformance costs: evaluations (e.g., reviews, early testing), resource positioning (e.g., requirements analysis, training, evaluation of subcontractors), implementation (e.g., design, construction). Non-conformance costs: re-testing, troubleshooting, rework, handling recalls, customer complaint resolution.
According to J.M. Juran, the right way to optimize the total cost of quality is not to focus on one of these factors but to consider both at the same time. The Scrum framework implements Sprints as an essential mechanism to consider both simultaneously. (Sprint is basic unit of development in the Scrum development methodology and other agile development methodologies. Sprints typically last from one week to a month.) The engineering activity requirements of Scrum are embedded within Sprints (as described below).
To illustrate the ideas of Mr. Juran, let’s take a look at the following two scenarios:
Scenario 1: Minimize the cost of non-conformance. It is nearly impossible to create a perfect product in the real world. It is certainly possible to attempt perfection and thus achieve zero non-conformance, but the cost would be extremely high and thus not economical. See Figure 1.
Scenario 2: Minimize the cost of conformance. Attempting this would reduce cost but result in poor quality. This would raise the cost of non-conformance to extremelyhigh levels, so this scenario would likewise be uneconomical.
Figure 1 shows us the relationships of “cost” to “degree of perfection” but separates the two cost components, i.e. conforming and non-conforming. This shows us that there is some Optimum point where a “Cost of Quality” can be estimated based on the tradeoff between the two cost components.
Figure 1: Cost of Quality
Traditional plan-driven projects using waterfall or V-model approaches frequently try to attain ultimate perfection (Scenario 1). These seldom succeed and ultimately, if attempted, lead to very high conformance cost. In practice, these models are often poorly implemented, resulting in very high cost of non-conformance (Scenario 2).
Compared to Scrum, CMMI is often perceived as the more heavyweight methodology, with emphasis on extensive planning and formalism (as in Scenario 1). This is not so, however. We’ll demonstrate that Scrum and CMMI both match and complement one another according Mr. Juran’s principles.