Summary of Toyota Kata by Mike Rother with excerpts from Puppet Lab 2016 State of DevOps report and questions about SaaS vs. Manufacturing and using kata for DevOps.
2. 2
Vision
“What do we need to do with our
◉ processes
◉ products
◉ services
to meet customer needs.”
3. 333
The Situation:
Basic Principles of Toyota’s Success
The Situation Know Yourself
Improvement
Kata
Coaching Kata Replication Appendix Extensions
4. 4
It’s Hard to See Exactly How Toyota Succeeds
Practices Tools Principles
Management thinking and routines (kata)
(Visible)
(Invisible)
5. 5
Companies that Thrive Long-Term
◉ “The objective is not [just] to win, but to develop the capability of the
organization to keep improving, adapting, and satisfying dynamic customer
requirements.”
◉ Take small steps forward to find the right path (rather than extensive up-
front planning)
◉ Technical innovation is often the cumulative result of many incremental
adaptations
● Think about the difference between the first iPhone and the iPhone 7
◉ Make cost and quality improvements in small steps
◉ Respond to unpredictable and uncertain situations with confidence and
effective action
6. 6
The Constant of Changing Conditions
◉ Toyota improves every process every day at every level of the company
● No process improvement workshops, initiatives, special efforts, campaigns
◉ “Relying on periodic improvements and innovations alone – only improving
when we make a special effort or campaign – conceals a system that is static
and vulnerable.”
◉ A process is either being actively improved or it is degrading
8. 8
Kata Defined
◉ A way of doing something; a method or routine
◉ A pattern
◉ A standard form of movement
◉ A predefined, or choreographed, sequence of movements
◉ The customary procedure
◉ A training method or drill
◉ A way of keeping two things in alignment or synchronization
9. 999
Know Yourself:
Failing at Continuous Improvement
The Situation Know Yourself
Improvement
Kata
Coaching Kata Replication Appendix Extensions
10. 10
Limits of typical approaches
Workshops
● Small number of participants
● Hard to change culture through workshops
● Limited impact on processes
● Hard to keep the momentum
11. 11
Limits of typical approaches
Value-Stream Mapping
● A highly useful tool from Lean Manufacturing
● Process changes happen at the next level deeper
● Can identify a very long list of issues across the value stream – where to start?
● Doesn’t develop our capability
12. 12
Limits of typical approaches
The Action-Item List
● Very widespread approach – easy to do – can fit in our schedule
● Unscientific, ineffective, scattershot
● Can create more variability and instability in processes
● Asks “What can we do to improve?”
● BETTER: “What do we need to do to improve this specific process?”
● Jumping too soon to countermeasures
● Unspoken goal: “Just shut off the problem!” (i.e. don’t worry about causes/improvements)
● Focus: do my action items (i.e. don’t worry about lasting improvements)
● Doesn’t develop people’s capabilities
● Performing the first action item could change the situation
● Priorities may shift
● Previously defined action items may now be counterproductive
14. 14
Philosophy
Nature of
people’s
actions at the
process
Process
outcomes
Consequences
The MEANS The RESULTS
Most of Toyota’s
management
focus is here.
Most of our
management
focus is here.
• Production quantity
• Quality
• Cost
• Productivity
• etc.
• Rewards (or lack of)
• Feedback
17. 17
The vision gives direction
• Vague
• Very far away
Path is unclear and cannot be
planned in advance.
18. 18
The vision gives direction
Where do we want to be next?
Detailed and specific
Target
Condition
19. 19
Limitations of cost/benefit case-by-case
◉ Traditional CBA:
“This proposal is too costly? Then we must do something else.”
◉ Toyota CBA:
“This proposal is too costly? Then we must develop a way to do it more
cheaply.”
20. 20
Limitations of cost/benefit case-by-case
“…without a direction we tend to evaluate proposals individually on their own
merits, rather than as part of striving toward something. This creates that back-
and-forth, hunting-for-a-solution, whoever-is-currently-most-persuasive effect
in the organization.
“Specifically, without a sense of direction we tend to use a short-term
cost/benefit analysis to decide and choose on a case-by-case basis whether or
not something should be done – in which direction to head and what to do –
rather than working through challenging obstacles on the way to a new level of
performance.”
21. 21
Lessons from History
“Toyota’s way of moving forward … is very
much one of
● adaptation and continuous improvement;
● of nurturing processes, products, and
businesses into profitability
● by doing what is necessary
● to achieve target conditions.”
22. 22
Lessons from History
“If our business philosophy and management approach
do not include
● constant adaptiveness and improvement,
then [we] can get
● stuck in patterns that grow
● less and less applicable in
● changing circumstances.”
23. 232323
The Improvement Kata:
How Toyota Continuously
Improves
The Situation Know Yourself
Improvement
Kata
Coaching Kata Replication Appendix Extensions
24. 24
The Improvement Kata steps
Vision
Next
Target
Condition
Current
Condition Obstacles Challenge
Drawing by Mike Rother
(1) “In consideration of a vision, direction, or target,
(2) and with a firsthand grasp of the current condition,
(3) a next target condition on the way to the vision is defined. When we then
strive to move step by step toward that target condition, we encounter
(4) obstacles that define what we need to work on, and from which we learn.”
25. 25
Another way to look at the Improvement Kata steps
Conduct
Experiments
toward the
Target Condition
Grasp the
Current
Condition
Establish
the Next
Target
Condition
Get the
Direction or
Challenge
1
2
3
4
Drawing by Mike Rother
• How should this process operate?
• What is the intended normal pattern?
• What situation do we want to have in
place at a specific future point?
• Where do we want to be next?
26. 26
Too much flexibility can destroy improvement
“[The] process can work around problems and still make the target
output…”
“At Toyota, this sort of flexibility is considered negative, since
problems go unresolved and the process gets into a non-
improving, firefighting cycle.”
“…self-compensating flexibility in processes would strike fear in the
hearts of Toyota managers because of all the problems that go
unnoticed and unaddressed.”
“Flexible systems that autonomously bypass problems are by
their nature non-improving.”
27. 27
Target Condition ≠ Target
Not just a Target
“A target condition is a description
of how the process should operate
in order to achieve the target.”
Target
Condition
28. 28
“Implementation Mode” vs. Improvement Mode
It broke!
We fixed
it!
Now Better? ? ??
??? ???? ?? ??
The Grey Zone
30. 30
Target Conditions provide focus
What could
we improve?
Target
Condition
Without a target condition:
“We could…
…reduce setup time
…clean and organize
…hunt for waste
…apply Kanban
…make a U-shaped line
…etc., etc., etc.”
With a target condition:
“What is preventing us from
having a part here every 16
seconds?”
“The operator has to leave
periodically to get another tray
of components.”
31. 31
Target Conditions alter thinking
Target
Condition
Traditional thinking
• We’re slipping back.
• We need to maintain.
• The operators are responsible.
• We need more discipline!
Target Condition thinking
• We aren’t there yet.
• What is preventing us from
reaching the target condition?
• Management is responsible.
• What is the next step?
32. 32
PREDICTION
Can be tested
ACTION
Conduct the
experiment
EVALUATE
Adjust based on
what you learn
EVIDENCE
Collect facts and data
PLAN
DO
1
23
4
CHECK
(Study)
ACT
THE SCIENTIFIC LEARNING CYCLE
"Plan-Do-Check-Act" or "Plan-Do-Study-Act"
By Mike Rother
33. 33
Example PDCA cycle: “Be in the car an hour after waking up.”
Cycle One
◉ Plan: Be in the car 60 minutes after waking up. (Target Condition)
◉ Do: Wake up and go through the morning routine, get into car.
◉ Check: How long did that take?
◉ Act: Nothing can be done at this point for this cycle:
● Cannot tell where the problem lies
● Cannot make an adjustment to still reach the target condition
34. 34
Example PDCA cycle: “Be in the car an hour after waking up.”
Cycle Two – with a better Target Condition definition
Steps,sequence,times
Plan
minutes
Actual
minutes
Alarm rings / Snooze-button cycles 5 5
Start coffeemaker 3 3
Bathroom routine 15 15
Get dressed 10 10
Make breakfast 7 11
Eat breakfast and read newspaper 10
Clean up breakfast 5
Check calendar and briefcase
contents
3
Leave house and get into car 2
Outcome metric: 60
“What is preventing
us from making
breakfast in seven
minutes?”
35. 35
Problematic attitudes about problems
◉ Traditional Way of Thinking:
● Ensure people are working hard
● Maintain tautness
● Control the person
◉ Toyota Way of Thinking:
● Finding problems means finding ways to improve
● Recognize obstacles early and understand them
● Problem-solving, rapid cycles
● Improve the process
● Work together on a common objective
● “Adopt the right heart”
“It’s always ‘no problem’
until the end, and then
we have a big problem.” sweeping little
problems under
the rug
placing
blame
36. 36
What does “problem solving” mean?
Focus Behavior
Toyota
• Learn about the work system
• Understand the situation
• Improvement toward the vision
• Observe and study the situation.
• Apply only one countermeasure at a
time in order to see cause and effect.
Us
• Stop the problem!
• What are my action items?
• Hide the problem.
• Quickly move into countermeasures.
• Apply several countermeasures at
once.
37. 37
Practicing the Improvement Kata Teaches Scientific Thinking
Practice the Improvement Kata routines to make the basic
skills of scientific thinking more automatic.
That’s the Kata part.
The automatic fundamentals are then a foundation upon which
all sorts of creativity and initiative can proliferate in your team
and organization, to achieve what seemed impossible.
That’s improvisation & creativity!
By Mike Rother
38. 38
The Improvement Kata can be Recursive
Conduct
Experiments (PDCA)
toward the
Target Condition
Grasp the
Current
Condition
Define the
Next
Target
Condition
Get the
Direction or
Challenge
1
2
3
4
Drawing by Mike Rother
As you work to define the Next Target
Condition, you will often need more
information about the Current
Condition
As you PDCA toward the Target
Condition, you may gain insights that
clarify the Target Condition
39. 39
The Improvement Kata: Results
“The original condition was x.
We set a target condition of y.
We achieved z and learned the
following in the process.”
40. 40
When you don’t know how to solve a challenging problem:
“If we already knew the answer, it would just be
an implementation question, and anyone –
including any of our competitors – could do that.
I don’t know the solution to the problem, but I
know how we can go about developing a
solution.”
41. 41
Management Approaches
Traditional thinking
Toyota thinking
• Outcome targets
• Metrics
• Incentive schemes
• ROI-based decision making
• MBO
• Small problems at the source
• Real process details
• Real situations
• In real time
42. 424242
The Coaching Kata: How
Toyota Teaches the
Improvement Kata
The Situation Know Yourself
Improvement
Kata
Coaching Kata Replication Appendix Extensions
43. 43
Management Approaches
Traditional Leadership
Who is
doing what
by when?
ReportingofResults
Goals, targets,
outcome metrics
Toyota Leadership
“Show me.”
The
difference
is here
Goals, targets, outcome
metrics
Mentor/Mentee dialogue
with overlap of
responsibility
Target conditions and
PDCA
Go and See, “Show me”
1 obstacle at a time
1 step at a time
Rapid Cycles
Leading with questions
1-page document / “A3”
44. 44
PLANNING EXECUTING
Understand
the Direction
or Challenge
(from level above)
Grasp the
Current
Condition
Establish the
Next Target
Condition
Experiment
Toward the
Target
Condition
Current State
Value Stream
Mapping
Future State
Value Stream
Mapping
Value Stream
Level
Process
Level
Organization
Level
Longer-Cycle
Experiments
Short-Cycle
Experiments
By Mike Rother
The Improvement Kata Pattern Connects
A Target Condition at one level can be the Direction for the next level
45. By Mike Rother 45
WHAT DEPLOYMENT OFTEN LOOKS LIKE
Don’t try to expand Improvment Kata practice faster
than you can develop internal Coaching Kata proficiency!
Phase I Phase II Phase III
Scouts study
the subject
Form
AG
AG and first coaches
practice the IK
AdvanceGroupmakesamedium-rangeplan
Advance Group conducts bi-weekly reflections
Slice 1 (a process, area, department, VS Loop, etc.)
Slice 2
Slice 3
Slice 4
AdvanceGroupreflectionandnextplan
Increasing number of managers
in the organization who are
proficient as IK coaches
Form an "Advance Group,"
i.e., which practices first
AG works toward a series
of 3 target conditions
(does ~ 25 PDCA cycles)
on real processes
46. 46
Coaching Kata Highlights
Role of Leaders: to
increase the improvement
capability of people.
Management spends well
over 50% of their time on
process improvement.
The mentor guides the
mentee in the use of the
improvement kata.
Training while working on
the real thing
Everyone at Toyota has a
mentor
If the learner hasn’t
learned, the teacher hasn’t
taught.
We can’t see people’s skill-
development needs when
we tell them what to do.
Needs of the mentee and
the situation determine
the next training.
Strategic decisions in sync
with the actual situation at
the process level.
47. 47
Simplified Steps of Toyota’s Practical Problem Solving
1. Pick Up the Problem (Problem Consciousness)
2. Grasp the Situation (Go and See)
3. Investigate Causes
4. Develop and Test Countermeasures
5. Follow Up
48. 48
Detailed Steps of Toyota’s Practical Problem Solving
1. Pick Up the Problem (Problem Consciousness)
a. Identify the problem that is the current priority
2. Grasp the Situation (Go and See)
a. Clarify the problem
a. What should be happening?
b. What is actually happening?
c. Break the problem into individual problems if necessary.
b. If necessary, use temporary measures to contain the abnormal occurrence until the root cause can be addressed.
c. Locate the point of cause of the problem.
Do not go into cause investigation until you find the point of cause.
d. Grasp the tendency of the abnormal occurrence at the point of cause.
3. Investigate Causes
a. Identify and confirm the direct cause of the abnormal occurrence.
b. Conduct a 5-Why investigation to build a chain of cause/effect relationships to root cause.
c. Stop at the cause that must be addressed to prevent recurrence.
4. Develop and Test Countermeasures
a. Take one specific action to address the root cause.
b. Try to change only one factor at a time, so you can see correlation.
5. Follow Up
a. Monitor and confirm results.
b. Standardize successful countermeasure.
c. Reflect. What did we learn during this problem-solving process?
49. 49
The Five Questions - Simplified
1. What is the target condition (the challenge)
2. What is the actual situation now?
3. What problems or obstacles are now preventing you from
reaching the target condition? Which one are you
addressing now?
4. What is your next step? (Start of next PDCA cycle)
5. When can we go and see what we have learned from
taking that step?
50. 50
The Five Questions – Detailed
1. What is the target condition (the challenge)?
a. What do we expect to be happening?
2. What is the actual situation now?
a. Is the description of the current condition measurable?
b. What did we learn from the last step?
c. Go and see for yourself. Do not rely on reports.
3. What problems or obstacles are now preventing you from reaching the target condition? Which one are you addressing
now?
a. Observe the process or situation carefully.
b. Focus on one problem or obstacle at a time.
c. Avoid Pareto paralysis: Do not worry too much about finding the biggest problem right away. If you are moving ahead in fast cycles, you will find it
soon.
4. What is your next step? (Start of next PDCA cycle)
a. Take only one step at a time, but do so in rapid cycles.
b. The next step does not have to be the most beneficial, biggest, or most important. Most important is that you take a step.
c. Many next steps are further analysis, not countermeasures.
d. If next step is more analysis, what do we expect to learn?
e. If next step is countermeasure, what do we expect to happen?
5. When can we go and see what we have learned from taking that step?
a. As soon as possible. Today is not too soon.
How about we go and take that step now?
(Strive for rapid cycles!)
51. Card is downloadable at:
http://www-personal.umich.edu/~mrother/KATA_Files/5Q_Card.pdf
By Mike Rother
52. 52
The “A3”: a written document to support mentor/mentee dialog
◉ Theme and Business Case
◉ Current (Initial) Condition
◉ Target Condition
◉ Moving from Current to Target Condition
◉ Metrics
◉ Proposals
◉ Plans
◉ Key points from reflections
◉ Signatures
54. PDCA CYCLES RECORD
Obstacle:
Date, step & metric What do we expect? What happened What we learned
Learner:
Process:
Coach:
(Each row = one experiment)
DoaCoachingCycle
ConducttheExperiment
56. 56
Developing Improvement Kata Behavior in Your Organization
◉ Be clear about what you are undertaking:
● The goal is to develop consistent behavior patterns across the organization
◉ Organization culture needs to shift, but it’s a “30-degree shift”
◉ Become an experimenter – the kata can suit most organizations through
experimentation
57. 57
Developing Improvement Kata Behavior in Your Organization
◉ What will not work (significant limitations):
● Classroom training – example: sports
● Workshops
● Having consultants do it for you
● Looking to metrics, incentives, and motivators to bring the desired change
● Reorganizing
58. 58
How to Experiment
◉ Use actual work processes
◉ Focus on three main factors:
● The Improvement Kata – how we want people to act
● The Coaching Kata – how we teach people to act that way
● Urgency – the priority for acting this way
◉ Use the improvement kata to start implementing the improvement kata
◉ If the improvement kata is not working properly, the coaching needs
adjustment
59. 59
Tactics Mike Rother Has Been Using
◉ Learn to do before learning to coach
◉ Get the senior managers to practice the improvement kata ahead of others
◉ Establish an Advance Group
● Assess the stability of a production process (measure it)
● Ask “what is preventing this process and the operators from being able to work with a
stable cycle”?
◉ Training through frequent coaching cycles and the five questions
60. 60
Building the Sense of Achievement
Sense of belonging Challenging work Working together
Seeing potential for
improvement
Clearly defined end state
Empowered to make
change
Sense of responsibility
Recognition for process
and results
Frequent challenges
61. 61
Able to TEACH it
AWARE of it
Able to DO it
There is a LEARNING PROGRESSION
By Mike Rother
Target Condition:
Student is aware of the approach to
continuous improvement and adaptation.
Activity:
Introductory training course
Target Condition:
Student can do improvement kata.
Activity:
Repeatedly practice improvement kata
Target Condition:
Student can do coaching kata.
Activity:
Repeatedly practice coaching kata
Is student far enough along to learn to coach?Gate
62. 62
Improvement and Coaching Kata: Common Obstacles
It’s hard for people to resist making lists of action items.
The five key questions are often difficult for senior leaders to internalize.
We like doing but not checking and adjusting.
We jump into solutions and skip over careful observation and analysis.
People assume the mentee must figure out what solution the mentor has in mind.
The unclear path to a target condition is uncomfortable for many people. Most like a
clear plan in advance even though that is only a prediction.
Recursive iterations (redoing steps) are uncomfortable. People feel they did something
wrong when asked to look again or repeat a step, yet this is very important for learning
and seeing deeply.
63. 63
Improvement and Coaching Kata: Common Obstacles
Many people will view this effort as just another project, rather than as developing a
new way of managing.
At the start, it seems like this effort is adding more work on top of daily management
duties, as opposed to it being a different way of conducting daily management.
At the start, coaching cycles often take too much time, becoming burdensome. After
establishing a target condition, a coaching cycle can often be completed in 15 minutes.
Less is more.
Rather than making a list of steps, take one next step and then see where that takes you.
Conduct your coaching cycles standing up at the process and do not let them turn into endless talk
sessions.
Go through the five questions, find the next step, and that is the end of the coaching cycle. Take
the next step as soon as possible.
66. 66
Start with the Value Stream
◉ Which value stream (production deliverable)?
◉ What are the processing steps?
◉ Dedicated or shared?
◉ Where do items wait, ready (inventory)?
◉ How does each process know what to produce (information flow)?
◉ What are the “loops” in this value stream?
67. 67
Then focus on one process
Summarize the current condition:
◉ Assess customer demand and determine line pace
● Customer takt (demand rate)
● Planned cycle time
◉ First impressions of the process
● Is there 1 x 1 flow?
● Are each operator’s work steps the same from cycle to cycle?
● Is output consistent?
◉ Is machine capacity sufficient?
◉ Is the process stable or is there a “hidden factory”?
◉ What is the necessary number of operators (if process were stable)?
70. 70
Comparison of Manufacturing vs. SaaS
◉ Similarities
● “Systems” thinking
● Standardization
● Scalability
● Repeatable Processes
● Value Stream – moving “code” from creator to customer
● Pull model – make some “inventory” ready
● downstream teams pull content as they have capacity
◉ Differences
● SaaS does not actually “deliver products”, per se
● B2B SaaS does not actually deliver to consumers (with some exceptions)
● “Code” ≠ “Parts”
● SaaS has many different “product lines” requiring different “factories”
72. 72
DevOps – faster with better quality
◉ Throughput
● Deployment frequency
● Lead time required for changes
◉ Stability
● Mean time to recover (MTTR)
● Change failure rate
74. 74
DevOps + Improvement Kata = ?
◉ Without the Kata:
● 6% less time on unplanned work/rework
● 5% less time on other work
● 11% more time on new work
◉ With the Improvement Kata?
Editor's Notes
Poll – how far did you get?
Ack my status as a READER.
Overview of the book’s structure.
These slides are available to you as a reference/refresher.
1
2 and 3
“Well, it could be that our production true north is theoretical and not achievable, but that does not matter. For us it serves as a direction giver, and we do not spend any time discussing whether or not it is achievable. We do spend a lot of effort trying to move closer to it.”
“The improvement kata should be depersonalized and have a positive, challenging, no-blame feeling.”
“What is preventing the operators from working according to the standard?”
1000 Andon pulls
The kata doesn’t stultify.
“There is no combination of outcome metrics and incentive systems that by themselves will generate continuous improvement and adaptation.”
“There is no combination of outcome metrics and incentive systems that by themselves will generate continuous improvement and adaptation.”