
Most of the waterfall based project management thinking assumes that planners know enough about what the customer wants and estimate with enough accuracy to create a huge waterfall schedule way in advance.
True or False?
I certainly think it's true. A good example is large, predefined releases which go into low-level details. With predefined releases we are assuming that we know how project execution will play out well ahead of time.
My experience has shown me that we don't know, especially with large complex projects there are too many interfaces involved and decisions during execution that change the reality of what adds value from the customer perspective.
This creates all kinds of hassles regarding which scope goes into which release, which really doesn't matter when it comes to the end customer. Unless it's going to change the timeline for which these features or capabilities are available to the end user, it doesn't really matter which release it gets pocketed into.
What If?
So what if projects treated validated learning, a concept from Eric Ries' book "The Lean Startup", as the primary goal of the project? What if the primary purpose was to learn through iterations and then provide what the end-user wants, as opposed to just having the delivery of certain scope on time and on budget?
Could we set up a project management system in such a way that learning, validated learning, is the primary driver for project planning and execution, and still be confident that we are going to meet the projects scope schedule and budget requirements?
I haven't had an opportunity to do this myself fully from end-to-end on a large project. I've only been able to implement these types of principles within a subset of a larger project. So I can't say for sure. But I have heard from many people who are implementing projects in a Lean organization, using the type of approach that I'm talking about here.
What do you think about the importance of learning as a major part of your project management methodology? Deming's Plan>Do>Check>Act cycle is (sometimes) applied to PM processes, but how often is it applied directly to assessing the customer definition of value? Please discuss in the comments below!



