Monday, March 29, I had the opportunity to have dinner with Jeff Sutherland and other RTP leaders. The dinner was a fundraiser for CITCON, which came to RTP in April. This was an opportunity for Agile practitioners and experts to have an informal chat about the challenges and opportunities of using Agile in the world of work.
I was interested in learning more about Systematic, a CMMI level 5 company that implemented Scrum across its entire business. One thing that sets Systematic apart from other companies is that it has really good data to prove that Scrum works, and it is a software company that can execute perfect waterfalls every time. Systematic created hyper-productive teams, and by Jeff’s definition, they are at least 4 times more productive than industry average. They cut TTM in half and the development costs by the same as well.
I asked Jeff how these teams at Systematic were able to do this. What was the secret sauce? Most of the answers are in a published paper here. Jeff highlighted two key metrics and one Agile practice (A-TDD) in addition to Scrum that contributed to their success. On top of this there was intense focus on getting the backlog to a ready-ready state before stories were allowed to enter a Sprint. Systematic introduced a Ready-Ready checklist to evaluate a Story’s readiness to enter a Sprint. The checklist’s three categories; 1) Prepare feature for commitment, 2) Clarify Feature for Development, 3) Prepare feature for Implementation. Everything on one page. A copy of it can be viewed in the paper referenced above.
The two metrics (from LEAN) were process efficiency, and fix time after failed builds.
Simply put, if a User Story takes 1 ideal day to complete, but the elapsed time is 5 days, then you have a process efficiency of 20%. A VSM exposes this rapidly. The goal is to have relentless focus on process efficiency to force it upwards. This means that multi-tasking is driven to as close to zero as possible, and teams have servant leaders who are aggressive about impediment removal.
Fix time after failed builds
From a detected broken build, the time it takes to get the build working. (Akin to LEAN’s “stopping the line”.) Everything stops until the build is fixed. This, complemented with TDD, CI, and A-TDD is a powerful cocktail that drives early and disciplined focus on quality and throughput.
Systematic applies A-TDD to ensure that acceptance is defined in customer terms. This guarantees that POs understand the feature to be delivered from a customer perspective, and development teams understand how success will be attributed in customer terms. This also leads development teams to code just enough to pass the acceptance tests, avoiding unneeded development.
Hyper-productivity then, comes from well prepared Stories, and a LEAN mindset about “stopping the line”. I’ll add that without solid XP practices in place in addition to the above, you would be hard pressed to achieve even moderate gains. Hyper-productivity requires a level of discipline that is rare.