Roadmap planning

As you might know, different levels of planning apply to an Agile SW development (actually, beyond what you can see in the picture, you have also the Sprint Planning and the Daily Planning).

It’s continuous planning with the different levels influencing each other by feedback both between levels and across levels in order to respond to changes. 

It’s a Rolling Wave Planning approach, where upfront planning is reduced to the minimum indispensable, to what is “just enough, just in time”.

As one can see, despite the myth about Agile meaning “coding with no planning”, there’s probably much more planning in Agile than in a waterfall approach.

In preparing for battle, I have always found that plans are useless but planning is indispensable.         

-Dwight D. Eisenhower

Last week we experimented a new way of Roadmap planning, aiming at understanding what will probably be contained in the coming releases.

So far this kind of planning had always been done mainly by the business guys, based on their level of understanding of customer requirements, their perception of the possible solutions and partly on wishful thinking.

This time we decided to do it differently, trying to take decisions or at least draw possible scenarios based on facts as much as possible. Once again the best recipe turned out to be a collaborative approach.

The decision was to call for a 4-hours workshop (it was actually 2 slots of 2 hours each) attended by the Product Managers, the Product Owner Team and a bunch of system architects.

The Product Managers brought to the workshop a list of prioritized requirements to analyze.

We decided to dedicate no more than a defined time-box to each requirement in priority order, so that we did not spend more time than strictly necessary at this very early stage of understanding.

The time-box was divided in 3 parts:

1.      1/3^rd to explain the requirement and clarify background and business value

2.      1/3^rd to sketch possible MMFs or epics

3.      1/3^rd to estimate the MMF or epics, a relative estimation using as a baseline the actual story points spent by our teams to deliver the MMFs in the last 2 releases

Everything was done collectively. After the workshop, given the number of sprints contained in the coming releases, the PO team did a date first planning, by considering the actual velocity statistics of our organization.

Based on that, we determined scope scenarios for the coming releases, considering the worst and the best cases in our recent history.

The feedback from participants and stakeholders both on the approach and the outcome was pretty good and I must say it looks very promising to me as an Agile coach.

At least it meant providing “just enough, just in time” information (even if rough) based on the current knowledge and understanding of requirements and above all on real data coming from our actual historical performances.