[ANN] APLN-MD / gbSPIN February meeting PMBoK vs. Agile . . .
Hillel Glazer
Date & Time:
Tuesday, 24 February 2009, 5:45pm
PMBOK vs. Agile: Sifting Reality From Myth
Description:
With the increased success of Agile processes and the growth of PMI,
project teams are faced with a choice: Use PMI's collected body of
project management practices or use Agile. We are told that PMI
practioners mandate a waterfall command-and-control approach, and that
Agile is the process-free alternative. But a quick look side-by-side
comparison reveals a much more complex situation. What do we make of
facilitative servant-minded PMPs? Could Agilists gain value from PMI
materials? What is the official position of PMI towards Agile?
Columbia, MD
Presenter:
Jesse Fewell.
Donald Buresh
Dear Jesse et al.,
I you really want an answer to the question of whether to use an agile-driven or a plan-driven software development methodology, then I suggest that you read my dissertation. In this work, I collected data on 185 software projects, 95 were known agile-driven project, 53 known plan-driven projects and 37 projects where the software development methodology was not known by these participants. For these 37 projects, a series of three questions were asked, and on the basis of majority, it was inferred that 24 projects were probably agile-driven and 13 were probably plan-driven. The dependent variable was customer satisfaction and the independent variables were product quality, project team effectiveness, and project management effectiveness. A dummy variable was used to distinguish between plan-driven and agile-driven projects. The dummy variable equaled zero for plan-driven projects and one for agile-driven projects. The coefficient of the dummy variable represents the level of autonomous customer satisfaction.
The results were quite telling. When the software development methodology was known, there was no statistically significant difference in autonomous customer satisfaction at the 95% confidence level. When the known and probable projects were combined, the estimated coefficient of the dummy variable was barely statistically significant at the 95% confidence level, where the lower bound of the confidence interval was .001. Because averaged data was employed in the regression, it was argued that the 99% confidence level was the appropriate level to employ. See the arguments by Bonferroni and Brace and Brace-Pellillo for the details. Thus, here is the result. There was no statistically significant difference in customer satisfaction when using the two methodologies.
The title of the book is: Customer Satisfaction and Agile Methods, and you can find the book on Amazon.com, just look up my name. So to answer your question, it does not matter what methodology you use, as long as you follow the methodology carefully, and you provide a quality product, and project team works well together. I hope that this helps.
Donald L. Buresh, Ph.D.
3115 Enoch Avenue
Zion, IL 60099
Tele: 847-872-1659
