Group and team leads functioned as the Product Owners for all seven teams, while I worked as the ScrumMaster. I felt strongly that Scrum was an important framework to implement within these teams and I was willing to take the risk to champion it. Although the result of taking this risk was positive, I nearly didn’t survive a full quarter of ScrumMastering seven teams!
Working with Danube consultants Michael James and Dan Rawsthorne, we determined that having volunteer ScrumMasters would be important to each team’s success and to their own sanity. First, we worked with Intel management to make sure that the role of ScrumMaster was valued in the performance appraisal system as “real engineering work,” rather than administrative overhead. Secondly, those who stepped up to take ScrumMaster roles did so on teams in which they did not have a technical stake. This helped prevent any conflicts of interest between their own technical projects and their facilitation responsibilities. Budget did not exist for ScrumMasters to give up their own engineering work in favor of ScrumMaster work. However, support was lent for a lighter change in product role for those who stepped up to shepherd the process in ScrumMaster positions.
At the end of three months, there were three additional ScrumMasters to manage seven teams. Additionally, an eighth team had volunteered to start using Scrum.
After approximately five months, scaling work across the Scrum teams became one of the largest challenges. Prior to adding the additional five teams that were formed throughout the remainder of the year, the organization needed more knowledge on how to manage the dependencies between multiple teams and facilitate better inter¬team communication. Danube was again retained to develop a customized Scrum scaling course for some of the original participants of the Certified ScrumMaster course along with the senior managers who missed the first class. This day-long training reviewed major principles of release planning, sprint planning, and, in particular, scaling across multiple teams. We again took the “learn, try, inspect, and adapt” approach to this scaling.
After learning a few “best practices” for scaling, we took the issue back to the teams to try one of the scaling models from class and then tailored the approach to the teams’ real world environments. After adding some roles to handle technical