23. It lives! It moves!
Kanban the Hard Way
March 2013
Balance Flow
24. It lives! It moves!
Kanban the Hard Way
March 2013
Balance Flow
25. It lives! It moves!
Kanban the Hard Way
March 2013
Respect Balance Flow
26. Simple rules combine to give us the
behaviour we desire
Kanban the Hard Way
March 2013
Transparency Flow
27. Simple rules combine to give us the
behaviour we desire
Kanban the Hard Way
March 2013
Transparency
28. Your process can evolve faster and more
safely than you may realise
Kanban the Hard Way
March 2013
Transparency
29. Your process can evolve faster and more
safely than you may realise
Kanban the Hard Way
March 2013
Transparency
30. Every limit, every policy
is a leverage point
Kanban the Hard Way
March 2013
Transparency
31. A Little Queuing Theory
Work in Progress
Average Cycle Time =
Throughput
Where: John Little
how long it takes one item to
Cycle Time (CT) =
go through the system
Work in Progress (WIP) how many items are in
Kanban the Hard Way
= the system at any time
how many items are
Throughput (TH) =
March 2013
produced per unit of time
Understanding Flow
32. Theory and practice meet in a CFD
120
100
On hold
Proposed
80
Prioritised
Ready for
Dev
Dev
60
Testing
Ready for
Kanban the Hard Way
Release
40 Released
Implemented
20
March 2013
0
Transparency Understanding Flow
33. Kanban works with your process
…and on it
Kanban the Hard Way
March 2013
34. “The work will teach you how to do it”
(Estonian proverb)
Kanban the Hard Way
March 2013
Understanding
35. How we wish change worked
There
Kanban the Hard Way
Here
March 2013
Understanding
36. What change really feels like:
the J Curve
Kanban the Hard Way
March 2013
Understanding
37. What change really feels like:
the J Curve
Kanban the Hard Way
March 2013
Understanding
38. Evolutionary change with
many small J’s
Increasing “fitness”
Kanban the Hard Way
Increasing capability
for change
March 2013
Understanding
39. Decades of experience in these small J’s of
“safe to fail” experiments
SPC: Shewhart, Deming
TPS: Ohno, Shook, Rother
TOC: Goldratt
Lean: Womack & Jones, Liker, Reinertsen
Six Sigma: Motorola, GE
Systems Thinking:
Kanban the Hard Way
Deming, Ackoff, Gall, Meadows, Senge
Complexity Science: Snowden
Beyond Budgeting: Bogsnes, Hope & Fraser
March 2013
Agile, XP, etc: Beck, Cockburn, Jeffries et al
Lean Startup: Ries
Collaboration Understanding
40. Decades of experience in these small J’s of
“safe to fail” experiments
Plan - Do - Check – Act
Check - Plan - Do
Look - Ask - Model - Discuss - Act
Define - Measure - Analyse -
Kanban the Hard Way
Improve - Control
Build - Measure – Learn
TOC’s 5 Focusing Steps, POOGI
March 2013
Collaboration Understanding
41. Foundational principles
that make evolutionary change possible
1. Start with what you do now Understanding
2. Agree to pursue incremental,
evolutionary change Agreement
3. Initially, respect current roles,
Kanban the Hard Way
responsibilities & job titles Respect
4. Encourage acts of leadership at
March 2013
all levels in your organization Leadership
– from individual contributor to senior management
42. Core practices
that stimulate and sustain change
Transparency
1. Visualise
Balance
2. Limit Work-in-Progress
Flow
3. Manage Flow
Customer Focus
4. Make Policies Explicit
Kanban the Hard Way
Transparency
5. Implement Feedback Loops
6. Improve Collaboratively,
March 2013
Collaboration
evolve experimentally
43. Core practices
that stimulate and sustain change
1. Visualise Transparency
4. Make Policies Explicit
5. Implement Feedback Loops
2. Limit Work-in-Progress Balance
Kanban the Hard Way
3. Manage Flow Flow
Customer Focus
6. Improve Collaboratively,
March 2013
evolve experimentally Collaboration
44. Kanban works within your system
…and beyond it
Kanban the Hard Way
March 2013
45. Kanban works…
with your organisation
and on it
with your process
and on it
Kanban the Hard Way
within your system
March 2013
and beyond it
46. Keywords
Organisation
Visualisation, visual management
Knowledge discovery process
Sense-making
Self-organisation
Process
Pull system, kanban system, flow
Kanban the Hard Way
Bottleneck, variation, variety
Emergent behaviour, leverage points
Evolutionary change; fitness
Cumulative Flow Diagram (CFD)
March 2013
System
Kanban Method
Models for collaborative improvement (several, growing)
47. References
Kanban the Hard Way
March 2013
Kanban, David J. Anderson
http://agilemanagement.net/index.php/kanbanbook/
48. Resources
kanbandev
http://finance.groups.yahoo.com/group/kanbandev/
kanbanops
http://finance.groups.yahoo.com/group/kanbanops/
#kanban
https://twitter.com/#!/search/realtime/%23kanban
Kanban the Hard Way
djaa.com (agilemanagement.net)
leankanbanuniversity.com
limitedwipsociety.org
March 2013
49. Thank you.
Questions?
Kanban the Hard Way
March 2013
Mike Burrows (@asplake)
mike@djaa.com
http://positiveincline.com
Notes de l'éditeur
"kanban" refers to the tokens This is "knowledge work" made visible, even physical
Not all work is alike!
The primary organisation is the vertical one, organising work items according to what they need most
Easy! This one needs development
Appropriate developers with the capacity to work on them
This one needs rework so that its testing can finish
So we keep asking “what does this work item need” until we’ve make sense of it all. Then we establish an order, so that work moves rightwards as it approaches completion.It’s a way to visualize what in product development circles is known as the “knowledge discovery process”. It’s about finding a realistic and appropriately-detailed abstraction of the process, one that is based on how things really operate in the real world, not based on imposing some idealised workflow.That’s the vertical swimlanes…
Related work grouped togetherPerhaps a project, or simply a large item broken in to smaller itemsThese swimlanes can come and go as needed
We might maintain a special area for tracking work that is out with suppliers or other teams.For some teams, managing these external dependencies is a huge part of what they do, and their kanban systems must reflect that if it’s going to be much use to them
So now what do we see?Work happening, rework needed, work that is stuck,work that is waiting. It’s all there!How much work is in play, where it is, where the issues areThis is visual management; the visualisation providing high quality information for the fast, pattern-matching part of our brains. We get attuned to our boards, and we are very quick to recognize that something is amiss.
It is really important to understand that it’s work, not people, that is organised in our visualisationSometimes, what we don’t see is as important as what we do seeBy not visualising it we avoid constraining it, let it find its natural shape.By making it seem less important than the things we do visualise, we allow it to change, perhaps even encourage it to changeSelf organisation is HUGE! From the perspective of the wider organisation it’s key to finding fitness & creating resilience. From the team’s perspective it greatly enhances the opportunity for people to find meaning in their work.
Kanban as a sense-making tool, leading to visualization of the knowledge discovery processWe're making space for collaborative, self-organised problem-solving around what is actually needed right nowSelf-organisation, fitness, resilience, meaning. It’s all good If we’re in the business of building better workplaces, this seems a pretty good place to start.What we need to see now is some dynamics. Process in action, process evolution in action too
When a work item reaches the right hand side it is complete, delivered. Feels good!
New work can move up to occupy the gaps left behind, up to the work-in-progress limits (the numbers in brackets here on each column). What we’re seeing in operation is a kanban system (a type of pull system), one specially adapted for knowledge work. Although many of the aims are the same, this is not Toyota’s kanban!
We see how pull gives us a nice unforced flow. Work flows rightwards, while the gaps (availability signals) flow back upstream. It seems that the whole board is connected. We see work itemsproceed at a good pace across the board.
We see how pull gives us a nice unforced flow. Work flows rightwards, while the gaps (availability signals) flow back upstream. It seems that the whole board is connected. We see work itemsproceed at a good pace across the board.
We see how pull gives us a nice unforced flow. Work flows rightwards, while the gaps (availability signals) flow back upstream. It seems that the whole board is connected. We see work itemsproceed at a good pace across the board.
We see how pull gives us a nice unforced flow. Work flows rightwards, while the gaps (availability signals) flow back upstream. It seems that the whole board is connected. We see work itemsproceed at a good pace across the board.
Looking at the overall workload, we never have more in the system than we can deal with effectively. It's not just that welimit painful multi-tasking and over-commitment, we avoid wasteful over-production. Once again, the benefits are shared between the team and the wider organisation.
Looking at the overall workload, we never have more in the system than we can deal with effectively. It's not just that welimit painful multi-tasking and over-commitment, we avoid wasteful over-production. Once again, the benefits are shared between the team and the wider organisation.
The overall design of the kanban system, then the detail of the WIP limits Selection policies Quality & completeness criteriaOverall performance is the product of these, the team, and of course the work itself
Down in the bottom left corner we have a very brief description of the rules that determine which stickies get pulled in which order. The system is explicitly designed to deal with variety. Obviously this is important for the customer, but it matters to the team and to the wider organisation too! Systems that can deal comfortablt with variety are much better places to live in. You avoid the disruption and waste that goes with re-organising for each new project, and you get benefits in flow too. Work is delivered faster for the same cost.Let’s see just how easy change can be!
Team demo before testThat was easy!
WIP limit from 4 to 3That was easy!
What Kanban is exposing here are “Leverage points”. These are places in the structure of a system where change can be implemented; what we have here are leverage points whose impact can far outweigh the cost of change. Just from our little animation you can see that WIP limits play an vital role in determining the speed at which work moves. And when we reduce it we also heighten our awareness of the bumps and bottlenecks in the road. This is good, but sometimes you can have too much of a good thing! Sometimes we carefully increase WIP to protect bottlenecks. We have to keep in mind too that the mix of people and work will change, so we keep limits under review all the time.The green stickies are our criteria for allowing work to move from one column to the next. These can be a great catalyst for creativity. I have direct experience of teams adopting simple, quality-enhancing practices from their own initiative.Down in the bottom left corner we have a very brief description of the rules that determine which stickies get pulled in which order. The system is explicitly designed to deal with variety. Obviously this is important for the customer, but it matters to the team and to the wider organisation too! Systems that can deal comfortablt with variety are much better places to live in. You avoid the disruption and waste that goes with re-organising for each new project, and you get benefits in flow too. Work is delivered faster for the same cost.Let’s see just how easy change can be!
We have to be a little careful in how we apply this mathematical law to our work, but in general:More WIP does mean longer lead timesReducing WIP tends to reduce lead times.That’s good!Reduce WIP too far though and throughput tends to suffer. That cusp is where improvement happens!
So if constant change is so necessary, let’s do itdeliberately, like we mean it!Cycles of improvement: hypothesis followed by validationIf the key question to organising work was “What does this one need?”, the key question for process improvement is “How will we know?”We see in this chart the story of one real team’s journey. A team that at the beginning didn’t really have things under control. A team that soon found that it didn’t know clearly enough what the “Testing” column actually meant. A team that was unsure about making releases. A team that worked on stuff that the business had already lost interest in.Also, a team that improved, sometimes radically, sometimes in barely noticeable increments. A team that took responsibility. Here were see a 7 month lead time reduce by more than half, and throughput actually increase in spite of the overhead of supporting a new system. You can’t see here (but I know) that the team soon got to a point where they could make releases at will, and where lead times were measured in weeks and then days, not months.That team was my team, and I’m proud of them!
And it’s a good job that we can make it that easy. Evolutionary change is very necessary.These sources of unpredictabilityadd frustration for customers, pain for workers, and expense for the organisationRevealing and addressing these sources of unpredictability in economically sensible ways is win all round!End result: speed with smoothness, predictability.For a while anyway! It’s a constant battle in an ever-changing world.
Don’t waste the bad stuff, learn from it“How is this even possible?” is a good question to ask. We look for solutions that make good outcomes more likely without adding too great an economic overhead. Not deluding ourselves that we can always fix outcomes at the outset, but making sure that we’re maximising feedback and learning at every opportunity.
Making your policies and parameters open to challenge and improvement Constantly giving you feedback, provoking the right conversations NOT overlaying yet more process, INSTEAD helping work to flow more smoothly
Let’s talk a little about getting started with Kanban
Not an exhaustive list!
PDCA or PDSA (popularized by Deming, though he called it the Shewhart cycle) is most recognized one. Cf the scientific method ("hypothesis"–"experiment"–"evaluation" or plan, do and check).CPD and LAMBDA start not with planning but with looking at the work as it currently is, perhaps the most appropriate approach to knowledge work.DMAIC (from Six Sigma) is interesting because of the explicit relationship between improvement and control, control here referring to systems for detecting that things aren’t working as well as they should.BML (from Lean Startup) deserves credit for moving the focus away from narrow definitions of “value” or “done” to an emphasis on learning (restoring a key element of Lean that sometimes gets lost).
Helping you find balance & resilience, to the benefit of team, customer & wider organisationMaking expectations clearer and more manageableKeeping risk where it can be managed most effectively