Linear scalability of a large global project was noted for the first time in 2007 in a peer-reviewed paper by Victorov et al. (1) and subsequently by Systematic (2) and Xebia (3). This is in direct contrast to the findings of Fred Brooks in the “The Mythical Man-Month” (4) where developer productivity decreased as the number of people on a project increased leading to Brook’s Law – “Adding people to a late project makes it later.” Brooks asserts that this is due to exploding numbers of communications paths with more people. This phenomenon has been observed on 100% of non-Scrum software projects and for the majority of Scrum projects.
This talk will describe the history of how Scrum was used by Sutherland and Schwaber (5) across many companies leading up to the linearly scalable global project in 2007. Since then recent experience with Saab (6), Toyota, 3M, GE, transportation and oil companies, and other domains will be described. Codification of these experiences into a systematic way of deploying Scrum across domains, across organizations, and across the globe will be shared. The essentials for developing a “scale-free” organization architecture will be emphasized based on input from the Agile leaders of the largest known Scrum implementations at Intel and SAP.
Scale-free architectures are commonly seen in biological systems and Scrum is based on my experience with evolutionary biology as a principal investigator on NCI cancer research grants. For example, the internet is a scale free architecture as every node is like every other node with no extra bureaucracy introduced at higher levels that could break the internet. The experts at Intel have pointed out that putting billions of transistors on a chip requires the same mode of thinking and have reminded us that Scrum properly implemented must work the same way to be scalable.
Agile leaders interested in taking the performance of their collocated and distributed teams to the next level will find great value in this presentation.