Why We Need a New Model for Project Governance
We all know that organizations establish a project management office (PMO) as their main model for project governance. However, this does not provide a competitive advantage that allows an organization to respond to ever-increasing customer needs.
In his video presentation How Kanban Helps Us Move Beyond Traditional PMO during the 2012 Lean Software and Systems Conference, agile coach David Joyce challenged this existing thinking behind projects pioneered by management consultant Henry Gantt in the early 20th century. Joyce says that we need a new portfolio management paradigm that is suited for the 21st century.
According to Joyce, the span of a project is from inception until the customer acknowledges the realization of business value. Agilizing for only the project execution segment means that a team is only doing local optimization, and, at best, will gain unnoticeable improvement in the overall project performance.
Regarding Joyce’s comments, it has been my experience that a project’s inception tends to be notoriously slow in many organizations. This is because the PMO does not typically approve the project until it has estimates and, thus, a rather detailed set of requirements. In his blog post “It Requires Discipline to Keep Inception Short," Scott Ambler recommended a minimalist approach for project inception that requires fewer detailed plans and requirements.
Keep in mind that you can set business goals based on those invalidated requirements. But you can’t expect to have a perfect agile delivery with an excellent plan based on invalidated requirements.
To illustrate the inadequacy of plans, Steve Blank, the father of the lean startup movement, said in this video interview, “You should be testing your [product] hypothesis iteratively and incrementally." Doesn't a project exist in the first place to create something new based on hypothesis? Blank's statement can be equally applicable to organizations, as Eric Ries mentioned in this video. Ries added that his biggest failure was when he worked for a company that had a perfect business plan.
To overcome the inadequacy of traditional PMO, Joyce has proposed a new model based on stable delivery teams that act as long-lived work cells. In this model, the congestion created by "starting" many projects is replaced by managing the flow in order to satisfy the needs of a limited number of high-business-value deliverables.
As a result of limiting the number of business deliveries, you can end up writing less code with fewer unused features. However, limiting the deliveries requires you to make hard decisions about what your team should be working on.
Finally, to enable this new model, management needs to regard a project as an experimental endeavor in which there is no standard process by which a team follows predetermined requirements.
I invite you to comment on what you think the obvious barriers are toward shifting to this new paradigm.