Sunday, August 19, 2012

The importance of Scaled Learning to Scaled Agile

I was facilitating a fishbowl discussion on Leadership at the recent LAST Conference when one of the participants suggested that we should take what we know about the importance of "Slack" to improvement in agile teams and plan in 20% slack for Agile Leaders to give them space to focus on improving their leadership.   I was focussing on facilitating rather than contributing at the time, but my immediate internal response was that most senior leaders I know are managed by their diaries.  Their opportunity to create this slack lies in exercising the choice to deliberately reserve space in their schedule for leadership rather than management.

"Deliberate choice" is the phrase that really sums up what I'm trying to explore in this post.  Regardless of your preferred flavour of Agile, at the team level a number of choices are made for you.  Some mechanism such as the sprint retrospective will exist to dictate investment in shared learning and improvement, and others such as standups, sprint planning and WIP visualisation to promote teamwork and "geting the team on the same page".

In my early Agile endeavours introducing Scrum and XP to small teams, I repeatedly experienced the oft-promoted phenomenon of "radical change".  Whilst I began in the "Shu" of blindly following the practices and ceremonies suggested by thought-leaders, I came to realise that this radical change came about through the provision of a "scaffolding" within which the team could self-organise and the deliberate choice to invest time in working as a team to reflect and improve.

In recent years, I have primarily been applying Agile in large high-formality enterprises pursuing scaled implementations.  My constantly-frustrated desire has been to find the secrets to achieving the same radical pace of change in these environments that I saw in my early days.  Whilst I found many inspirations in the writings of Highsmith, Larman and Vodde it has been Dean Leffingwell who has triggered my biggest breakthroughs with his focus on building the "self-organising program".

I often find myself in discussions which critique Dean's model as overly proscriptive and "process-heavy", but what I have found in recent months as I work with people applying it is that it provides the opportunity for scaled level "Shu Ha Ri".  Whilst in theory I have known the underlying principals, it has been in helping people apply the practices and learning through failure that the real breakthroughs have come in moving from being able to talk about the principals to being able to realise them.

I have come to believe that the purpose to creating a "Self-organising team" is to enable rapid learning across the team and constant tuning of dynamics and collaborative optimisation within it.  The scaffolding which permits self-organisation (such as Dean's Scaled Agile Framework) is not the end in itself, but the enabler for scaled learning.


In my last post, I wrote about a program leadership team sprinting their way towards their first PSI planning session.     Shortly after writing the post, the team concluded for a number of valid reasons that formal application of the PSI construct did not lie in the short-term future.   At the same time, there was universal agreement that alternative practices were needed to pursue the principals underlying the PSI in the interim – in particular the PSI planning session and the Program level shared learnings.

This moment in time provided a pivot point in learning for the program.  Where the pace of learning had been slowly and linearly increasing in the preceding 12 months, the learning curve has gone exponential in the last 3.  Following are 3 of the key contributors to this change.

Book Clubs

Whilst there’s nothing new in this idea, in my experience it’s typically introduced as an optional ‘brown-bag’ concept attracting the passionate to invest their own time.  In this context, the general manager felt so passionate about supporting shared learning that it was born as scheduled mandatory meetings on work-time.

The initial book club involved the senior leadership team and used Dean's Agile Software Requirements Book as the text.  Chapters were assigned to logical owners, and each session would cover 1 chapter with the presenter documenting their key insights on the Wiki and other attendees making sure they’d at least read the chapter to be able to contribute to the discussion.  The underlying theme was “What are the key insights, what are the underlying principals, how do we adapt them to apply to our world, and where are we doing well/poorly with current execution”. 

Despite initial push-back by many with over-crowded diaries, these sessions  have become a weekly highlight now.  Each time I attend, I get a warm glow as a coach as I listen to the participants debate improvements at a depth I could not have conceived 3 months ago.  Improvement stories are generated for a dedicated program improvement backlog with the GM acting as product owner.

This has been so impactful, it has spurred the start-up of two more book-clubs in recent days – one for the Scrummasters studying Lyssa Adkins' Agile Coaching book  and one for the tech leads to study Ken Collier's Agile Analytics.

The Innovation Challenge

On the improvement front, two things were painfully apparent.  One was the fact that the ‘improvement actions’ coming out of team retrospectives were the first things to be sacrificed when the following sprint came under pressure, and the other was the fact that teams weren’t learning from each other – the vast bulk of successful improvement was localised.  Inspired by some ideas I came across in the Rockefeller Habits, I pitched the idea of rewarding innovation to the GM and the innovation challenge was born. 

We created an ‘innovation wall’ opposite the kitchen with a space for every team.  Any innovation achieved during the sprint could be posted on the wall regardless of whether it involved removing waste, improving productivity, improving quality or changing the way the team worked together.  At the end of the sprint, the teams would dot-vote for “innovation of the sprint” and the winning team would receive $100 for team celebrations. 


Unity Day

This one is very much a case of saving the best till last.  The key goal of this was to tap into some of the “whole program in the room on the same page” benefits built into the PSI Planning concept.  It has been a wild success.  The session typically runs for an hour and a quarter first thing  on day 1 of each sprint with 70 odd attendees (and a ban on powerpoint).  The agenda varies a little in time allocation and sequencing, but the key elements are as follows:
  • ·      Appreciations from the leadership
  • ·      Voting on the innovation challenge
  • ·      Agile learning activities
  • ·      Report-ins from the huddles
  • ·      Warnings and Tips from technical leadership
  • ·      Upcoming work in the pipeline

Appreciations from the leadership
Cheerfully stolen from my friends at Rally, this is the opener for the session and involves 5 minutes for senior leadership to share good news stories and appreciations for teams either heard from customers or observed themselves in the past sprint.

Voting on the Innovation Challenge
Teams’ innovation sheets are plastered around the walls, and each team gets 1-2 minutes to talk to their key innovations for the sprint.   Contributions vary from team members teaching each other keyboard shortcuts and cross-skilling sessions to implementation of new design patterns and integrations with other applications.

Each team has a set of 8 dots (different colors/shapes for each team to prevent self-voting), and they’re given 5 minutes to wander around the room and dot-vote for their favourite innovation, following which the vendor sponsoring the award for that sprint hands over the goodies.

Despite some humorous attempts at rorting the system with one team promising to spend the whole reward on lollies for their team area and an open invite to others to share in the goodies, it’s become a great way of sharing learning and is now being followed up by brown-bag sessions for deep-dives into the winning innovations.

Agile Learning Activities
Being the shameless coach I am, I usually steal 20-30 minutes of the session to run some kind of game.  Teams are randomly formed, and we run learning activities such as the ball-point game, the penny game, and the agile airplane game.

Probably the most enjoyable moment so far in this segment was the mayhem as 80 or so paper airplanes were flight tested at the end of each iteration of the airplane game. 

Report-ins from the huddles
5 huddles run 2-3 times per week.  Scrum of scrums for the scrum-masters, and discipline huddles for tech leads, test leads, BI leads and ETL leads.  This is a short segment looking for 1-2 minutes of highlights on learnings from the huddle for the preceding sprint.

Warnings and tips from technical leadership
There are usually 2 sheets of butcher paper for this segment.  In short, “things to watch out for” and “things you might benefit from”.  It’s a consolidation of warnings of potential impacts (such as Olympics deployment embargos) and advice of newly available utilities and enablers.

Upcoming work in the pipeline
This segment is basically a heads up on pieces of work that are progressing through the program kanban and either kicking off that day or 1-2 sprints away.  It’s primarily aimed at keeping teams informed of what the near-term roadmap looks like and which teams are focussing on what areas. 

Conclusion
Teams leave the room, grab a coffee and get into Sprint planning.  

Back to the self-organising program

I started this post talking about "deliberate choice".   When you distill down the rambling, I'm talking about a deliberate choice to build "program as team of teams" and foster learning at the program level.  It's an investment decision.  Putting a team reward and recognition program in place to reward innovations is an investment.  Putting a program management team in a room once a week to study a book is an investment.  Putting 70 people in a room for an hour and a half to play agile games and get on the same page at the start of a sprint is an investment.

What are you investing in?  You're investing in creating a sense of team among teams, and in creating learning between teams rather than just within teams.   If you want radical change at scale, you have to break out of local optimisations into system optimisations.  It will only come about by creating a framework for self-organisation and investing in the time to build team and build a learning culture.