On 22 November, I attended the Agile-Boston Give Thanks for Scrum day. There were great talks by many people using Scrum to make their work and life better. Microsoft hosted the event. My presentation was on the Shu, Ha, Ri of Scaling. Shu is the beginners state (20% improvement, Ha is the intermediate state (>200% improvement) and Ri is the state of mastery (>400% improvement in production and time to market). To get to the Ri state you need to understand Moore's Law.
Moore's Law is fundamental to understanding why Scrum was created and why it scales. Simply put, In 1965 the co-founder of Intel Gordon Moore observed that the number of transistors on a chip was doubling about every two years. That has continued for more than fifty years. The problem was that software development didn't have that kind of exponential growth, instead it was growing literally. In 1993, when Scrum was created, I became chair of the Object Management Group's committee on the problem. We needed to get software on an exponentially improving production curve. Hardware and software experts on the committee collaborated to figure this out and I wrote several papers on this work, for example "The Emergence of a Business Object Architecture" and "Why I Love the OMG."
What they Learned at Intel - NOTHING SCALES!
At the 2016 Construx Software Executive Summit in Seattle, I met the Agile leader from Intel who booted up 25,000 engineers doing Scrum. Scaling transitors on a chip has been in the DNA of Intel since one of the founders wrote Moore's Law. The Agile leader pointed out that Intel learned that "nothing scales" and that you need a "SCALE FREE" architecture to solve the problem. The internet is one example and Scrum is another.
How You Can Tell When Your Scrum is SCALE FREE
When we published the first paper in the IEEE Digital Library on a globally distributed team across North America and Russia, the academic reviewers pointed out that many people had shown improved productivity with process enhancement but no-one had ever published a paper showing linear scalability. In Distributed Scrum: Agile Project Management with Outsourced Development Teams we showed that doubling the number of teams by adding more teams in Russia to a large U.S./Canadian product development effort led to more than a doubling of global production.
During the entire history of software development, this phenomenon of linear scalability has only been demonstrated on Scrum projects. For another example see Scrum and CMMI Level 5: The Magic Potion for Code Warriors. That is because Scrum, when done properly is SCALE FREE by design.
Many Scaling Frameworks are Not SCALE FREE
You can tell if your scaling framework is crippling performance if productivity has not reached a 400% improvement and when you add more teams, productivity does not go up at linearly. You will never get your teams SCALE FREE unless you understand Moore's Law and implement it in software. This phenomenon is also seen in teams outside of software as it is domain independent.
For more detailed information on this issue you can download my presentation - scrum-at-scale-gtfs-22-nov-2016. You can learn more about scaling scrum properly in our Scrum@Scale training in New York and Boston next month.
Gordon Earle Moore (born January 3, 1929) is an American businessman, co-founder and Chairman Emeritus of Intel Corporation, and the author of Moore's law. As of January 2015, his net worth is $6.7 billion. Wikipedia