A Root Cause Analysis of Agile Practices

At Agile2010 I was chatting with George Dinwiddie about general process related stuff (probably with some reference to Kanban!) and I mentioned an idea I had submitted to a couple of conferences which had never got accepted. George suggested we try it as a Open Jam session, so we did!

The idea is to run a root cause analysis of various agile practices to drill down into why they work and what the benefits to be realised are. So rather than using a 5 whys approach to solve problems, it is used to understand solutions. For example, why do unit test? To minimise defects? Why do we want to minimise defects? To create less rework? Why do we want less rework? etc. The session tied in nicely with another Open Jam run by David Hussman on Dude’s Law, which also emphasised focusing on why rather than how.

Here are the outputs from the 3 practices we picked; Unit Testing, Iterations (Time Boxes) and Limiting WIP. Click to view the album with bigger pictures.

As a general exercise, I found it really useful and interesting. Definitely something to try submitting to future conferences again. The discussion and debate we had, and the surprising tangents we went on, was rewarding and enlightening. I was particularly fascinated by the comparison between Time Boxing and Limiting WIP and the way that creativity came out in both of them through different paths. I hope that by understanding why practice work in more detail, we can avoid following them dogmatically, and be in a better position to solve problems based on context. When a particular practice is not suitable we can draw on other practices which can provide the same benefits.

This is definitely something I want to explore further – hopefully with workshops at future conferences. If you try it out as well, blog your outputs and let me know!

Exploring the Kanban Multiverse at Agile2010

I ran a workshop on “Exploring the Kanban Universe” at Agile2010 with Xavier Quesada Allue. The premise was to setup an example case study and lead participants through visualising different aspects of a project – the different multiverses – over a number of iterations.  Below are pictures of the final boards from the different teams. We encouraged people to think ‘outside the box’ and try and move away from traditional rows and columns approaches.

The highlight for me was the circular design that one of the teams came up with. I thought it was a great solution to the challenge of avoiding kanban boards appearing to show a linear process. With a circular design, a work item can loop round the circle multiple times, with the distance from the centre indicating closeness to completion, and different quadrants indicating the primary focus of work such as analysis, dev, test etc. This was the nicest example of a board as a ‘map’ as opposed to a ‘relational’ representation.

One thing that was reinforced for me was that a Kanban Board should not try to visualise absolutely everything. Its should have just enough information to signal where the issues are, and where the team should look to find out more. In other words it should be able to “point to the gemba”. Thanks to Harada Kiro to reminding me of this after the session. In future runs of the workshop we’d like to try starting with a blank canvas each iteration to avoid teams feeling constrained by trying to show too much and having to build upon previous designs.

I also gave a talk on the subject at the Lean & Kanban Conference in Belgium last week after which Mary Poppendieck pointed out that I’d slipped into referring to some visual management designs as kanban boards when they strictly weren’t because they didn’t limit WIP. Visual Management is a large part of a Kanban System, but not every Visual Management board is a Kanban board.

Agile2010 Bag Packing with Kanban

At Agile2010, as at Agile2009, I went along to help the volunteer bag packing, and use it as an exercise in experimenting with Lean and Kanban ideas. Once again it was a huge success. We completed packing all the bags in (anecdotally) record time, and had great fun in the process.

The video above was put together by Luiz Parzianello and really gives a sense of the energy and enjoyment everyone had. You can also see the “Y” shaped line we put in place and how people moved around and self-organised to keep the materials flowing.

Below are the outputs of the team retrospective, but first, here are my highlights and overall impressions.

  • Even though bag packing is not software development, there was still creativity on the way we solved the problem.
  • Being able to design a successful process in context, whatever the nature of the work, is an important skill.
  • Even with relatively repetitive work, people are motivated when they are involved in designing the work.
  • Clear visibility of bottlenecks (by limiting work in progress) enabled people to move around to keep the flow of material.
  • Measuring throughput in bags per minute (but not setting targets) was a motivator and a predictor of when we would finish.
  • Given the right space, it would be perfectly feasible to pack bags on-demand during registration without needing to pack them up front.

Here are the retrospective notes:

What Worked – Do Again

  • Music
  • 1 person floating around all stations (extra capacity)
  • Y Config
  • Everyone really trying to help
  • Continuity of event planning – better every year – Elastic
  • WIP limits – 4 stacks backlog meant stop & wait
  • 18 people in am / 15 people in pm
  • People taking metrics – live, visible metrics – without warning
  • Paper picking – each one goes under prev
  • Largest on bottom
  • Table splits
  • Breaks
  • Everything organised & stacked with one example on table of each item
  • Handing a stack directly to a person instead of putting on table
  • Continuity & ownership between am/pm – (better)
  • People got to be creative & solve problems
  • Stacking by size
  • Arrange table so no one had to walk
  • Video taping!!!

What Did Not Work – Do Better

  • Tables too short, materials too low
  • Bad sizing on poster
  • Folders came flat, needed to fold
  • Sticker falling out of flyer
  • Bags less than ideal – keeping open – cut hands
  • Started later, slow beginning
  • Still found missing items
  • List still was not accurate, items hard to match
  • Process for matching items to list not efficient
  • Did not have all the bags
  • Table with small things moved too fast
  • Slowing down to obey WIP made it hard to speed up

What To Do Differently – Try

  • Packing on the vendor room – closer to end point
  • Planning on process over email before hand
  • Someone owns planning process early
  • Something to hold bags by handle and open like a rod
  • Big visible labels on boxes – even big colour stickies
  • Do prework on Saturday
  • Insert kanban tokens into inventory so when a signal is found in stack, then supplies can be identified.

Rally and other Announcements

I have a few announcements so I thought I’d group them together into a single post.

  1. Firstly, I’m really excited that I’m going to be joining the Rally team in the UK. I’ve had a good couple of years with Conchango / EMC Consulting, but its time to move on, and I believe Rally are doing some great things in the Lean and Kanban space, especially with the recent acquisition of AgileZen. I’m really looking forward to working with Ryan Martens, Jean Tabaka and the rest of the Agile Coaching team.
  2. Next, I’m going to be at Agile2010 and will be running a workshop Exploring the Kanban Multiverse with Xavier Quesada Allue. Its an evolution of Xavier’s Visual Management workshop from last year’s conference, with some updates that have been used at XP Day London and the Orlando Scrum Gathering. I’m also going to helping out with the bag packing on the Sunday before the conference. This is a repeat of something we did last year where we applied Lean and Kanban thinking to bag packing process and learned a load and had lots of fun in the process. If you’re around the hotel, come and find us and join in!
  3. Finally, the LeanSSC 2010 UK conference was a success and the materials and videos can be found on the Limited WIP Society.

LeanSSC Atlanta 2010 and other Conferences

I’ve just updated my Calendar page with where I’m speaking this year so far (or hoping to), and thought it would be worth adding some more details in a post.

Atlanta 2010 SpeakerThe conference I’m most looking forward to is the inaugural LeanSSC Conference in Atlanta in April (21-23) which is the place to find out about “the next wave of process innovation”: Lean, Pull Systems and Kanban.

If you are interested in applying Lean concepts to software and systems development then this is the conference to attend. It will have the best people in Lean and Kanban, and the best and largest quantity of Lean content. A significant number of the speakers are not part of the regular Agile community so this is your chance to see them. Here’s some other reasons why you might want to go:

  • Learn lean development approaches with a focus on scientific, model based solutions.
  • See how to tailor lean methods to your unique work situation.
  • Find proven approaches that let development and management work together on a system design level.
  • Get pragmatic, actionable advice, delivered by people with field experience presenting metrics and data.

I’ll be giving a new talk on “A Kanban Multiverse”. Here’s the abstract:

Wikipedia defines a Multiverse as the hypothetical set of multiple possible universes that together comprise everything that physically exists: the entirety of space and time, all forms of matter, energy and momentum, and the physical laws and constants that govern them. A Kanban Multiverse can be defined as the hypothetical set of multiple possible Kanban Boards that together comprise everything that physically could be visualised: the entirety of scope and time, all forms of work type, status and flow, and the organisational laws and constants that govern them. This talk will explore how a single Kanban Board might visualise these multiple aspects in a limited and constrained space.

The other exciting conference for me is going to be the Scrum Gathering in Orlando next month (match 8-10). I’m really honoured to have invited to run a deep dive workshop on Kanban. Its going to be structured round what I refer to as the Five Primary Practices (see here and here), with exercises and discussion to explore how Kanban Systems are compatible with Scrum.

The other two confirmed conferences are ACCU 2010 and SPA 2010 where I’ll be talking about Five Steps to Kanban and running a Kanban Game respectively.

Finally, its the Agile2010 submission process at the moment. I have two submissions in, and am a panel member on a third. If you have a user account (why wouldn’t you? :)) please give them feedback to help them get accepted!