Project Management and Development

Project Management and Deployment

Database implementation specialist Richard Winter, in his article titled "Think Systematically," provides a compelling argument for why proof-of-concepts lower risk and reduce the chance of poor decisions. He does so through the a compelling analysis using decision trees combined with probabilities of outcomes and expected value calculations to demonstrate the economics of two distinct decisions, a full-scale proof-of-concept or work that proceeds immediately to implementation. Winter demonstrates the delays associated with proof-of-concepts are well worth the increase in expected value of a project.

In his example of a data warehouse implementation project, Winter makes several critical assumptions. First, he assumes that there's a 50/50 chance that the project will be a success if proceeding directly to building the data warehouse. Secondly, the proof-of-concept will produce results that also have a 50/50 chance of showing that the project will meet all critical business requirements. Finally, if the proof-of-concept produces results that show that the project will not meet all business requirements, the project will either be modified or postponed/cancelled.

Although a simplification of a complex decision tree analysis, the following points are worth yof note in Winter's calculations. The loss associated with a system failure without a proof-of-concept is much higher than the cost of a proof-of-concept. In Winter's example, the proof-of-concept is $2 million vs. The cost of the failure which is $10 million. For a project that moves beyond the proof-of-concept, the probability of success increases because necessary adjustments are identified. These modifications may either lower the business benefit of the solution or increase the technical costs to deliver the project. Winter uses what he refers to as "pessimistic assumptions" of an eighty percent reduction in the business benefits and a three percent increase in technical costs as guidelines for developing these estimates.

Winter ends by admitting that a proof-of-concept may delay initial release of the project by as much as three to six months. But, he believes that the expected value of the project will increase by almost fifty percent. He believes that proof-of-concepts provide a more solid technical approach and make more economic sense. Thus, Winter concludes that the use of time and resources to reduce risks for large-scale projects is a wise decision.