6. Central Themes
A high-trust culture might not
necessarily emerge, especially if anyone
that cares about things like that feels
obliged to express neutrality
7. Central Themes
How to reconcile Kanban with its place
in a larger context of change, or
reconciliation in general
12. Analytic vs Synergistic
Bad old thing Good new thing
Low trust culture High trust culture
Theory X Theory Y
Extrinsic motivation Intrinsic motivation
Focus on cost Focus on value
Management Fellowship
Single loop learning Double loop learning
13. Rightshifting
A modern healthy culture for product development
and knowledge work doesn’t appear by magic
You need internal change agents or culture
hackers, biased towards freeing knowledge
workers to delight customers
Path towards synergy: rightshifting from analytic
to synergistic
19. Managers
Likely to posses an analytic set of thinking tools
Biased to left-drift
May prefer driving change rather then being
changed
Hold purse-strings
Have momentum behind decision making
20. Workers
Unaccustomed to owning process
May have internal bias to left-drift,
neutrality or right-shift.
Bias likely not realized externally
21. Internal Culture Hackers
Biased to right-shift
May be managers or workers
Likely in the minority
May be responsible for advocating or
“coaching” Kanban
23. The Kanban System
Exhibits characteristics that qualify it as a
participant
Imposes pull
Via WIP limits, imposes things like kaizen,
slack and/or swarming
28. Kanban
Coach Traditional
Management
Kanban
Itself
29. Kanban
Coach Traditional
Management
Kanban
Itself
30. Kanban
Coach Traditional
Management
Kanban
Itself
31. Kanban
Coach Traditional
Management
Kanban
Itself
32. It Is Complicated
Kanban (pull and WIP limits) will result in
situations that compel change
These changes could be analytic/low-trust or
they could be synergistic/high-trust
Analytic changes tend to conflict with pull and
WIP limits, (and may threaten the sustainability
of Kanban itself).
33. Kanban
Coach Traditional
Management
Kanban
Itself
34. Kanban
Coach Traditional
Management
Kanban
Itself
35. There is so much momentum
behind the prevailing analytic
mindset that neutrality seems
like a losing strategy.
36. But David Said...
WHAT KANBAN COACHES
DO, AND DON’T DO:
Stop Advocating! Stop
Evangelizing! Observe
Instead
There Is No Judgment in
Kanban
Kanban consultants respect
the current situation and do
not pass judgment upon it.
38. The idea seems to be:
Start where you start
Just add the Kanban magic sauce
If you were advocating change before
Kanban, then after Kanban you should stop
it
Watch Kanban do its magic
39. My issues with that are...
It assumes we are introducing Kanban for its own
sake, and not as a tool to support a wider initiative
Managers accustomed to analytic thinking are not
going read David’s blog, and do what they always did
Progressive culture hackers with high-trust ideals are
going read David’s blog and sit back and observe
This is not speculation, I have seen it occur
40. The Story
Introduction of Kanban at a medium sized
German software company
Also subject of my MSc dissertation
Does NOT describe a company that
actually reached “synergy” with deep
Kanban
41. “well we only have Theory
X people around here”
“does that mean Agile is not suitable?”
48. 2010
My focus on improving Scrum, mainly by
pushing for cross functional teams and
backlog grooming
Started my dissertation with little enthusiasm
from the PMs
Then the boss read an article about Kanban...
49. The Workshop
Held in February 2011
With external consultants
Learn about Kanban and decide if we
want to do it
50. Goals of Most Participants
Better quality
Better maintainability
51. My Goals
Better cultural and psychological
environment
Better motivation
Better communication and collaboration
between project and development
department
57. The Kickoff Meeting
Balancing new development and
maintenance
Swimlanes?
Identified 11 process steps and suggested
WIP limits significantly below current WIP
59. One of the Initial Designs
Column WIP Limit
Understand and Assess 10
Analyze 10
Offer 50
Product backlog
Prepare for Sprint Planing 8
Development including acceptance tests 8
Internal acceptance 8
Review, test, improve and integrate 4
Wait for Freeze
Prepare for client release including documentation 20
Deployment 20
In production
60. First Few Weeks
Split into 2 systems, New Development
and Maintenance
Each “feature” is a project of around
20-30 stories
Total WIP limit is equivalent to 5000
stories, or 100 per person
61. Other Calculations
Excluding buffers like Offer and Product
Backlog: 40 stories per person
Working backwards from sensible 2
stories per person implies total WIP limit
of 100 stories, or 4 features.
65. Observations
Project and Development departments
working together ok, but less “agile”. Project
culture dominates.
The board is a hybrid of portfolio and team
level. Most activities left and right of
development are independent of feature
“size”.
66. The First Retrospective
Batch Size
Still no measurable lead time for either
New Development or Maintenance
Global WIP
67. The First Retrospective
Management reported that
“development” must be a problem
because cards move quicker through the
other columns, and linger in development
71. 3-Day Workshop with David
Initial focus on culture
I somehow got diverted towards the
“metrics”
David’s workshop got me back on culture
72. 2nd Retrospective
Developers lament: “It just feels like a
continuous stream of work. We miss projects”
Managers lament: “We need more detailed
planing documentation.” Suggest a planning
phase where we plan projects upfront
Managers also need: More control. Suggest
more formal sign-off between “phases”
73. I Suggested
More Agility
I called it “reintegrating Scrum”
Cross-functional teams, delivering small
enough stories that they can be delivered
in 30 days or less.
75. What We Committed To
A more formal approach to projects and
planning
A more formal “quality gate” approach
Teams specializing on sub-domains (but
maybe not cross-functional)
76. What Management Decided to
Implement
A more formal approach to projects and
planning
A more formal “quality gate” approach
Teams specializing on function
77. Other Observations
Push rather than pull
Conflict and stress over WIP limits and
deadlines
No tolerance or acceptance of slack or
swarming
Increasing time pressure from management
78. The Realization
Management continue to work and think
they way they always did. Since before,
and during Scrum, and now with Kanban
Culture dominated by fear and mistrust
82. Shift Focus Towards the Hard Bits
Current focus on principles and practices
good for training and initial system design
Description on the tin, less relevant for on-
going change efforts at the coal face
Even trainers should probably spend 5% on
Kanban itself and 95% on “propaganda”
83. One of Our Current Best Hopes
for the Future of the Workplace
We have a chance here to really change the
world
There is a lot of momentum behind 19th century
management, and neutrality wont beat it
Kanban should play the central role in
everything that the rightshifters and Stoosians
are trying to do
Things have teeth and look scary when you go deep.\nGreet Audience - Hello everyone, nice to be here.\nI am new at this, to me you all look as scary as this fish. But as Deming said, drive out fear.\n
Not being a consultant means I am slightly more trustworthy. Less likely to exhibit a bias towards positive experiences.\nThe photo is me as a victim of traditional management thinking. My experience as a software developer taught me that the world of management needs to radically change.\n
For those that like letters of the alphabet, I can probably say how each are significant. Especially the MSc and KCP\nAlso a quick note about what I am doing now, what interests me (Rightshifting, Stoos), and what I intend on getting into in 2013 (Games and Storytelling).\n\n
It is a 50 minute talk.\n0-5 This intro\n5-15 Central themes\n15-25 Definitions\n25-40 The story - Mostly based on my experiences at Bewotec\n40-45 Conclusion\n45-50 Questions\n
Trainers and consultants that help with initial system more concerned with the description of Kanban. A lot of the principles and practices, and a lot of the discussion around kanban are mostly relevant for that initial introduction. Workers, managers and change agents on the ground are concerned with very different things 6 months later.\nMention how in discussions online, I might say something based on my experiences, and a trainer or consultant will reply saying I am wrong and use the description of Kanban as proof.\n
A lot of the principles and practices imply a neutrality that might be more appropriate for trainers and external consultants than workers, managers and change agents on the ground.\n
Kanban doesn’t stand alone. It is just one tool that a company may decide on to support a wider system of change. Depending on the goals of Kanban, and the way it is introduced it might be hard to reconcile the way a Kanban coach is expected to work, with the way a scrum master, manager, developer or culture hacker might otherwise work.\n\nDavid actually said at the 3-day workshop that he tries to talk people out of “kanban initiatives” because the focus should be on the culture not the tool. But I don’t know how to reconcile that with neutrality. \n
\n
Don’t forget to mention the productivity, waste and ROI in change bits.\nAnd why I like the model.\n
\n
The second point is perhaps a reaction to the idea that we can only change culture via changing the management system.\n\n
Low trust to high trust culture was discussed at DAs masterclass. (He implied high trust results from Kanban. My experience suggests that doesn’t automatically happen).\nI personally tend to be biased towards the things on the right to the extent that I think things on the left can be actively bad.\n\n\n
1. Or with Kanban alone.\n2. That second bit was what I learned as a consultant, without that don’t even try.\n
\n
Explain shallow and deep, but mention the intention to have at least wip limiting should be there before calling it Kanban.\n\nNow the actual change happens in step 6, assuming Kanban is actually being used as a tool to support change.\nJudging by the discussion around Kanban, people seem to be using it mostly as a workflow management tool.\n\nAlthough as an aside, I don’t know how it is possible to have a sustainable Kanban system with WIP limits and pull, but without an effort to improve and evolve. I suspect it works in small teams ok, but I don’t think it is sustainable with a wide Kanban system.\n
Thank Bernadette Daria and the limited WIP society for the diagram.\nPeople have stopped talking about shallow or deep and are looking more at these detailed Kiviat charts, but its a little too fine grained for me. I like deep and shallow, because there is a fundamental qualitative difference between kanban systems that focus on managing the flow of work, and those that primarily support change.\n
\n
Managers and Workers separate, because even though this split is unsavory, it exists, and considering that managers have the resources to bring in consultants etc. By managers and workers, I exclude the few who are internal culture hackers\n\nCulture hackers is what I call change agents\n
\n
They have probably learned to delegate their decision making upwards\n
People that know about change models, the true use of Kanban, and may be conflicted about how to reconcile their calling as a culture hacker with the neutrality associated with Kanban coaching.\n\nThey might not be as refined as how they express their bias.\n
\n
Similar to a manager making decisions, it can be overridden, it may depend on acceptance.. It is really not that much different to a manager “making suggestions” or asking people to do things.\n\nI think in a deep, wide kanban system, pull, and WIP limits are much more shocking and radical things to impose on an organization than anything an organization has to do to set up a scrum team.\n
\n
If Analytic and Synergistic are too controversial, or too abstract, lets call it low-trust and high-trust, as David Anderson uses that in his workshops.\n
We can assume the average organization is either moving towards, or sitting stably at 1.\nPerhaps as a smaller company it started off as Ad Hoc or even Synergistic, and as they grew they had scaling issues they needed to solve and they solved it with traditional management.\n
Forces towards right-shifting usually come from culture hackers.\nIf right-shifting doesn’t occur, then traditional management must be pushing back.\nSometimes culture hackers might succeed.\n
Now if we introduce Kanban, the culture hacker is now a Kanban coach, who read that a good Kanban coach should be neutral.\nLets also assume Kanban itself is neutral.\nManagers that have always acted according to the analytic mindset don’t feel compelled or obligated to either neutrality, or to start thinking differently.\nThey may have agreed to change, and Kanban might set things up for change, but the changes they make will result in left-drift.\n
To see rightshifting occur, or a high-trust culture to emerge, we need to see either Kanban coaches exhibiting a force to rightshift....\n
or management suddenly start thinking differently...\nI think some people believe this might happen when a manager looks at the card wall or the CFD, maybe it does in some cases...\n
or Kanban itself acts as a force to rightshift... or as a force causing a high-trust culture to emerge.\nMany people imply that this is absolutely not the case, and are adamant that Kanban merely “shows us things”, and doesn’t force anything, and it is up to people to act on what they see. \nDavid Anderson, I think, implies Kanban itself will help result in a high trust culture to emerge...\n
I wont yet claim that Kanban is a force to right-shift, or bring about a high-trust culture, but lets say it compels change of either kind, but is biased against changes inspired by the analytic mindset, and towards changes inspired by the synergistic mindset.\n
Maybe I don’t understand the point of neutrality, but my ideal would be if all three of these components, plus any other participants are biased towards rightshifting and developing a high-trust culture.\n
People don’t voluntarily change, even if they look at a task board or a CFD. I think Kanban is biased, and I think Kanban coaches should be.\n
\n
\n
This was the subtitle of my dissertation.\n
This seems to imagine that analytic managers, after years of commanding and controlling, are going to suddenly look at a Kanban board or a CFD and start thinking differently, and presumably providing the required leadership with the Kanban coach in an assisting role.\n\nIt also conflicts with what a lot of other people say, that Kanban just shows you things, and it is up to people to actually make the improvements. But with all the Kanban experts sitting back and observing, it leaves a power vacuum that will only be filled with the same old analytic decision making that built the current system.\n
1. Presumably, if other aspects of the wider initiative involve pro-active pursuit of certain goals, then it may be hard to work out what to keep doing and what to stop doing. i.e. which actions are “part of Kanban” and which actions are part of the wider initiative.\n\n2. The type of people who explicitly choose theory x management, because it suits how they see the workforce.\n\n3. The people who probably suggested Kanban in the first place, understand it the best, and end up taking on the coaching role. They are going to observe the analytic thinking managers hijack Kanban and use it to accelerate the left-drift\n
1. Mention the size of the company and the departments\n3. In fact they seemed determined to reject it. \n
I think anyone can become a “theory y employee” once management adopt a theory y management approach, i.e. treat people with respect\n
\n
\n
\n
\n
One interesting point is that we got 2 consultants in. One focused on cultural and management aspects, and the other one on technical. The former didn’t last as long.\n\nI can’t call it Scrum-but, but it was the type of Scrum where they do the task breakdown first and estimate it in hours. Instead of measuring velocity they add up everyones hours and multiply it by 0.85 to represent 85% productive time and 15% meetings.\n
One other interesting experience was seeing a presentation from the project department about how we work, and seeing a classical waterfall style diagram. I was shocked as I thought we were a Scrum organisation. It was pointed out that Scrum happens inside this little “development” box in the middle\n
\n
\n
\n
Everyone looked at me like I was crazy. “How will replacing Scrum with Kanban achieve that?”\n
1. But our customers only want to upgrade once or twice a year anyway...\n
Basically doing more work in the same amount of time, and making more money. I don’t actually think that is the sweet spot of Kanban. Kanban can reduce lead times given a throughput. But increasing throughput is usually best left to things outside Kanban. Better technical practices etc.\n
1. We spent too much time on the easy bits in my opinion. From Scrum we already new about moving cards around etc. there was too little discussion about the actual changes and how they would be handled. I asked but that was apparently outside the scope of Kanban itself.\n2. Kanban would track larger MMFs, and a development column in the middle would be where Scrum happens at the story level.\n3. Confusion...\n
2. As a new release starts.\n
1. No single process\n
\n
\n
Total WIP here is 138, plus unlimited columns.\n
1. Was not a change to underlying system, but to the Kanban system itself\n
2. 4 features.. which is exactly how many new development teams we have, 1 feature per team seems sensible.\nFor me it was clear we should have tracked things at the story level, or set up a very small portfolio system to track features.\n
Note the 2 main changes. We noticed there was a lot of inventory in the Manual Integration test stage in the first month. Then in June we noticed there was a lot of inventory in Preparing Customer release and customer acceptance. No change to system, just a focus on addressing the symptom.\n\nThis is also typical of portfolio systems, not much flow.\n\nOh and at the beginning of june someone took the first board away for some reason.\n
\n
Maintenance looks a lot better. Notice we only started counting the backlog in May.\n
1, There were more managers used to making decisions in the project departments. The Scrum walled garden at least allowed developers to have a voice and make decisions at a certain level.\n
This was a month later\n1. I wanted to address this\n2. In one month no card had yet traveled across the board.\n3. In allocating WIP across maintenance and new development, we forgot to account for the fact that new development features are 20 to 30 times larger than maintenance tasks.\n
\n
Made sense to split the Kanban system into a portfolio system and a normal system. One tracking features and one tracking stories. This was rejected.\n\n
\n
\n
\n
\n
\n
\n
\n
Management had a secret meeting (not all of them were at the retro)\n
\n
2. Management actively favors those that work how they like, and “punish” those who speak up at retrospectives.\n
A developer had a good idea, and was excited to help\nAnother coach suggested he prepare an A3\nDrama over where to book the 2 hours time\nA new project would have to be set up for it in the time tracking system - but that needs to be “escalated” to senior management\nHe backs off of course, and there were comments floating around about people not being committed to critical projects so they might be better suited for less interesting projects.\nYou cant be neutral about things like that\n