Agile Business Change Blog Thoughts on Agile Strategic Business Change and Agile Delivery

14

JUL

Implementing Basic Governance

14 JUL 2011 | Posted in ADAPT, agile governance, agile project management, lean, planning, project velocity | Author Rob Smith | 2 Comments

Tweet

Mail

Last week I wrote a small blog post again pointing out the need for governance of Agile projects. In response Joseph Motha commented that my post would be more interesting if I explained how governance could be achieved. So here goes...

Governance is clearly a subject too broad to address in a single post so I’ll start here and build upon this in the future. As an initial starting point it’s important to establish an approach to planning and tracking. Unfortunately the standard approach within Agile (e.g. Scrum and XP) focuses primarily on the output from sprints and loses sight of the wider picture, thereby providing insufficient commitment or reporting. However, the basic estimation and tracking techniques are valuable and provide the basis for something more appropriate.

The approach is as follows:

1. Create a full backlog for the project. This does not have to be detailed and will include epic stories, however it should broadly cover the planned scope. This is then estimated using relative measures (points) and key delivery milestones are agreed. The resultant plan is then used as the baseline for the project. The expectation is that it will change, but variance and change can be measured against it.
2. Change reporting is an important part of any governance approach, it provides visibility and control to external stakeholders. However, it is essential that reporting does not impact the responsiveness of the development team and therefore change to the baseline is only reported if it falls outside of agreed tolerances. This provides space for the development team to manoeuvre. On the Guardian project the tolerances were set at 25% for both stories and iterations i.e. stories that became either 25% bigger or smaller than originally planned were reported (you may choose to vary this for smaller stories) and iterations that delivered 25% more or 25% less were regarded as exceptional.
3. As the sprints are completed, progress is mapped on a burn-up chart and shown relative to the original plan, as are the changes to the milestones (i.e. change in the number of points required to deliver a milestone) and projected delivery dates (intersection between extrapolated delivery and milestones).

From a governance perspective this gives a number of key indicators: performance against plan implies an ability to achieve the commitment; variability of milestones implies the level of (un)certainty in the plan. Additionally, the change process provides a mechanism for those external to the team to understand key decision points, “we’ve improved the overall solution because of X and it had impact of Y”

Within the plan it is also important to recognise variability in team size and output (holidays, sickness, ramp-up etc.) in order to avoid unrealistic performance being extrapolated (particularly the negative perception after a slow start). There is also value in reporting actual velocity and optimistic velocity (as described in Applying Optimistic Velocity) and in setting project KPIs against which stories can be prioritised and change can be evaluated.