Meeting Objective: Mike Kennon will present a talk on managing agile projects. It will explore a deterministic project management paradigm (traditional project management) and one that recognizes that software development is a discovery process. The talk will draw largely from Mike's experience directing the project office for a $30 million eBusiness. He will be able to talk about specific techniques for managing the relationships with external stakeholders including executive oversight.
Comments on presentation material: What made this presentation interesting for me was the focus on managing external relationships. Mike did NOT discuss general day to day project management activities nor agile methodologies. His main points were regarding the communication and control perspectives of the project sponsors. He compared the deterministic project model where we feel we can predict the future and the non-deterministic project model where we acknowledge that software development is a discovery process. He makes a point that even if we think we are proceeding in a waterfall fashion that we eventually revert to a non- deterministic mode as we decide what to do when we make a new realization.
We still need the development team to work within a deterministic model where we work on the part we can predict and stay within our given constraints. Where we need to change is the approach of the larger project group that includes the IT sponsor and the business sponsor. Here we need to acknowledge that we don't know everything up front. We especially need to focus on where the project will deliver value to the business and how it will improve the process. This needs to be redetermined as new discoveries are made. The development team pays attention to the value of the project and operating costs and technical solutions and monitors & reports changes discovered. The development team can then make recommendations on how to improve the value of the project. The decisions between alternatives and changes in priorities or assignment of resources then become the control points that the business sponsor has. This provides the relief to the frustration that everyone has about how long it takes to develop software by allowing the managers/sponsors to reassess the new information and to focus on the trade offs between value and cost.
We wish to legalize the non-deterministic model and make good business decisions in the early inception and elaboration phases of the project. Mike recommends making obvious allowances for the undiscovered. This means to budget for unknown use cases, budget for unknown requirements, and budget for technical solutions to become more complex than originally anticipated. You can build better trust when you accurately estimate what you do know and leave these allowances for what you do not know. There is a tie here to iterative development where each iteration is focused on where you will gain the most business value and where there are technical risks. Your baseline is the simplest possible solution and you must consciously choose when a more robust solution provides justifiable value.
To summarize with key phrases:
"Software development is a discovery process"
"not the best solution, the affordable solution"
"invent successful outcomes"