Friday, July 7, 2017

Facilitating Release Train Design with the ART Canvas Part 3 - Launch Planning

Following a hectic couple of  months launching four new ARTS, its time to conclude my three-part series on facilitating Release Train design with the ART Canvas.  Covering the facilitation of a 1-day workshop, the previous posts dealt with creation of a shared ART Vision accompanied by a supporting design.  In this conclusion, I tackle the closing session of the day: the preparation of a launch plan.

Before commencing the launch planning, we generally shrink the audience.  We release the stakeholders who were critical to establishing a shared vision and design, and bring our focus in to the leadership team who will be executing the launch preparation activities.

Agile is a team sport, and the ART is intended to be a self-organizing team of agile teams.  So, your leadership team needs to work as an agile team too!   The vision workshop is essentially a team formation activity for the leaders, who then have the chance to operate as an agile team as they prepare for the launch.

Since our final objective for the workshop is to generate a launch plan, the metaphor that makes most sense is that of a short “enablement PI” which is executed by the leadership team with the support of the product owners.  Thus, the final segment of the workshop is a mini PI Planning.  I open with the following challenge objective for their enablement PI:

“Your product is an Agile Release Train that can fulfil the vision we’ve defined here today.  8 weeks from now (or whatever launch date we’ve set), 150 people are going to walk into a room to spend 2 days in their first PI Planning.  What will make this a success?”
To assist them in their planning, I then distribute the following Features before sending them into breakouts to create a set of stories and objectives that will realize the objective:







The depth to which the plan is developed depends on our time-boxing.  I’ve generally run this workshop over the course of a single day, which means their set of stories and backlog will be very rudimentary given the constrained time-box they will have for breakouts.  I have increasingly become tempted to extend it to a second day, allowing time to create a more effective plan, establish PI objectives, identify dependencies and create a Program Board.  This would also allow some time on the 2nd day to work through some team formation activities with the Leadership Team such as:

  • Development of a social contract
  • Design of the team Kanban
  • Planning of team demo approach (the team should be demonstrating back to stakeholders on launch preparation progress)
  • Agreement on standup timing
  • Iteration planning for the first iteration

Closing the session

In closing, we should have the following outcomes:

  • A name and launch date for the ART
  • A completed ART canvas, ready to be hung on a wall (and eventually digitised)
  • An agreement for the newly formed ART leadership team to function as an agile team during the launch preparation
  • An appropriately scary launch preparation backlog that motivates movement by the leadership team and is accompanied by an agreed “Agile monitoring” process using Team Demos to enable progress transparency and feedback. 

I like to close out with a final check-in.  I love the definition of consensus Jean Tabaka taught me: “I can live with that, and support it”.  It’s a great way to bring a day of very hard work to a close, perhaps using a roman vote.  “We’ve formed a vision, an ART design and a plan today.  There’s a lot of hard work ahead of us but I’d like to check that we have achieved consensus before we leave this room.”



3 comments:

  1. Nice article Mark - I especially like the features you've defined as these are easy to understand and communicate to stakeholders.

    BTW, your "System Team" feature is duplicated above.

    ReplyDelete
  2. Thanks Stuart,

    Replaced the duplicate with the relevant Feature now :)

    Regards,
    Mark

    ReplyDelete
    Replies
    1. Thanks - I was wondering whether it was a duplicate or another feature was missing!

      Delete