For the last year I’ve been driving Agile adoption and use at a small but impactful organization. We sell services and data access for the railroad business in North America and, while in IT, 90%+ of our work is product related.
Early this year I initiated three major efforts: a mainframe migration — about 40 applications to be migrated over four years to a Java/Oracle environment, a Master Data Management (MDM) consolidation of our customer data, and a replacement and reengineering of our Single Sign On (SSO) services. This is in addition to still driving Agile, UX, and a brand-new services strategy that is being defined.
For the mainframe migration project we are using pure Scrum as it fits very well with the effort to be undertaken. The existing application is reviewed with the Subject Matter Expert (SME), functionality assessed; some functionality eliminated; rearchitecture decisions reached; and work defined through user stories.
The SSO and MDM effforts though are not pure software development. SSO entails acquiring a new engine, assessing how to replace the existing one, then taking the steps to deliver this functionality. User Stories (Scrum) are a bit cumbersome for this approach.
The MDM effort includes data modeling and data loading through ETL (Extract, Transform, Load) processes, the definition of data governance and quality processes, and the development of a simple application to view the data. Besides the application development (we are using Scrum for it) the other parts are cumbersome through user stories and Scrum.
Instead of Scrum we are using a modified and simplified CBPM processes. Rather than a full day planning process (Map Day) we undertook one to two hour sessions to define what we knew and got the ball rolling. As the teams are small the users for each deliverable were not defined. While modified CBPM is still very successful. We just implemented our new SSO engine and protected the first migrated application this week. For MDM we have a preliminary loading of the data and a first version of the application. The challenge with MDM has been understanding the data and how it has been loaded into the existing source systems, something that would have taken a long time through traditional data analysis processes
So, don’t be afraid to mix and match approaches and even modify them. Keep to the spirit of Agile.