agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Empirical Process Control
An important but often overlooked aspect of agile product
development
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Agenda
• Introduction
• Defined Process Control
• Hands-on
• material needed 5 pens/pencils + one A4 paper
• Empirical Process Control
• Cynefin Framework
• Applied to Scrum
• NUMMI a concept GM failed to copy
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Niels Verdonk
Agile Coach
niels.verdonk@agile42.com
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
agile42 - The Agile Coaching Company
We make your agile transition succeed
agile42 has a proven approach to successful agile implementations. We support you with a
full range of coaching services and specialized trainings.
We work locally & worldwide
Connected coaching for international
teams
•Coaches across the globe
•Deliver consistently across internationally
distributed organizations
•Support different time zones
•Recognize cultural diversity.
Agile Transition
Agile change for the whole organization
•Transformation across entire organizations
•At individual, departmental and management
level
•Assessment
•Set goals and define strategy for change
•Guide organizations through the transformation
•While building knowledge for sustainability
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Defined Process Control
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Input -> Activity ->
Output
Production-based process
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Defined Process Control
• Sequence of activity chains
• Input Activity Output➟ ➟
• Each activity requires specific skills
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Defined Process Control
• Sequence of activity chains
• Input Activity Output➟ ➟
• Each activity requires specific skills
• Flow is pre-defined and predictable
• Time passed / time estimated = efficiency.
• Handoff after each activity, responsibility transfers to the next role
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Empirical Game
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
G
am
e
Pencil Game
• Put the A4 paper in front of you and draw a circle like so:
• Take pens in one hand and close your eyes and let go
• With your eyes closed try to predict where they fell
• How many are completely outside the circle?
Needed: 5 Pens/Pencils and a sheet of A4 paper
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
G
am
e
Pencil Game
• Put the A4 paper in front of you and draw a circle like so:
• Take pens in one hand and close your eyes and let go
• With your eyes closed try to predict where they fell
• How many are completely outside the circle?
• Now repeat with your eyes open
• How many are completely outside the circle?
Needed: 5 Pens/Pencils and a sheet of A4 paper
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
G
am
e
Pencil Game
• Put the A4 paper in front of you and draw a circle like so:
• Take pens in one hand and close your eyes and let go
• With your eyes closed try to predict where they fell
• How many are completely outside the circle?
• Now repeat with your eyes open
• How many are completely outside the circle?
• Try to change 1 thing to improve
• Now how many are completely outside the circle?
Needed: 5 Pens/Pencils and a sheet of A4 paper
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Empirical Process Control
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Experiment and
Analyze
R&D based approach
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Empirical Process Control
• Based based on empirical measurement
• Measure results in a defined interval of time
• Conduct Safe-to-Fail experiments
• Allows for stabilization and optimization
• While improving the outcome iteratively
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Empirical Process Control
• Based based on empirical measurement
• Measure results in a defined interval of time
• Conduct Safe-to-Fail experiments
• Allows for stabilization and optimization
• While improving the outcome iteratively
• Effective in unordered, complex systems
• With behavior which can’t be predicted upfront
• But retrospectively we can identify patterns in behavior
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Cynefin Framework
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Cynefin Framework
• Developed by Dave Snowden
as a sense making model
• Cynefin is a Welsh word, which
means 'habitat' or 'place’
• The true meaning is we all have
different pasts and
backgrounds
• The name describes the
evolutionary nature of complex
systems
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Simple
Disorde
r
Cause and Effect relationship
exists, are predictable and
repeatable
Domains in a Nutshell
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Simple
Complicated
Disorde
r
Cause and Effect relationship
exists, are predictable and
repeatable
Cause and Effect relationship
exists, but are not self-evident,
analysis is required
Domains in a Nutshell
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Simple
ComplicatedComplex
Disorde
r
Cause and Effect relationship
exists, are predictable and
repeatable
Cause and Effect relationship
exists, but are not self-evident,
analysis is required
Cause and Effect are only
obvious in hindsight, with
unpredictable emergent
outcomes.
Domains in a Nutshell
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Simple
ComplicatedComplex
Chaotic
Disorde
r
Cause and Effect relationship
exists, are predictable and
repeatable
Cause and Effect relationship
exists, but are not self-evident,
analysis is required
Cause and Effect are only
obvious in hindsight, with
unpredictable emergent
outcomes.
No Cause and Effect can be
determined. We need to act
very quickly in order to stabilize
the system again.
Domains in a Nutshell
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Simple
Sense
Categorize
Respond
Cynefin Framework
Disorde
r
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Simple
Complicated
Sense
Categorize
Respond
Sense
Analyze
Respond
Cynefin Framework
Disorde
r
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Complex
Simple
Complicated
Sense
Categorize
Respond
Sense
Analyze
Respond
Probe
Sense
Respond
Cynefin Framework
Disorde
r
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Complex
Simple
Complicated
Chaotic
Sense
Categorize
Respond
Sense
Analyze
Respond
Probe
Sense
Respond
Act
Sense
Respond
Cynefin Framework
Disorde
r
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Cynefin Framework
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Empirical Process Control in Scrum
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Software Development is Complex
• Software Development today is almost always in the Cynefin
Complex domain
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Software Development is Complex
• Software Development today is almost always in the Cynefin
Complex domain
• On a technical level we are dealing with:
• Various modules, gems, libraries, etc.
• Interaction with 3rd
party systems and services
• We are not repeating an existing known process
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Software Development is Complex
• Software Development today is almost always in the Cynefin
Complex domain
• On a technical level we are dealing with:
• Various modules, gems, libraries, etc.
• Interaction with 3rd
party systems and services
• We are not repeating an existing known process
• On a functional level we are dealing with:
• Changing requirements and priorities due to market changes
• We cannot predict the behavior of global internet users
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Developer Insights
• In waterfall solutions are created upfront including detailed technical
design.
• In practice developers encounter imperfections in the technical design
• Also on a functional level things are not always accurate
• The results were that we needed to charge our clients extra (remember
CRs?)
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Developer Insights
• In waterfall solutions are created upfront including detailed technical
design.
• In practice developers encounter imperfections in the technical design
• Also on a functional level things are not always accurate
• The results were that we needed to charge our clients extra (remember
CRs?)
• In Scrum cross functional experts work together to solve small
problems
• The problems defined in User Stories are solved collaboratively
• They can apply what they learned already in the next user story
• The technical design is emerging
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Incremental Product Development
• An incremental process is not by definition an empirical approach
• You can still use Scrum to deliver a pre-defined software solution
• This has benefits, you will still deliver working software each Sprint
• But you cannot deal with changes in your market
• Or know how your pre-defined solution will be received by global users
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Incremental Product Development
• An incremental process is not by definition an empirical approach
• You can still use Scrum to deliver a pre-defined software solution
• This has benefits, you will still deliver working software each Sprint
• But you cannot deal with changes in your market
• Or know how your pre-defined solution will be received by global users
• If you apply an empirical approach to Product Development
• Product Owners evaluate the delivered User Stories together with the team
• The Backlog can be re-prioritized accordingly and stories deleted or added
• We can work towards a Minimal Viable Release and validate it with our users
• We release sooner and incorporate the feedback to create a better products
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Continuous Improvement of the Process
• After each Sprint we gather feedback on how team can work more
effective
• This allows us to continually improve the way we work
• We should inspire teams to try things, conduct safe-to-fail experiments
• Some experiments will be successful, they can incorporate this in the
process
• Some experiments will fail, and they can stop doing this
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Continuous Improvement of the Process
• After each Sprint we gather feedback on how team can work more
effective
• This allows us to continually improve the way we work
• We should inspire teams to try things, conduct safe-to-fail experiments
• Some experiments will be successful, they can incorporate this in the
process
• Some experiments will fail, and they can stop doing this
• It’s vital for organizations to show their commitment to support teams
• By showing they are acting on organizational impediments outlined by
the teams.
• The progress on these items should be visible
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Short story: NUMMI Factory
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Toyota and GM partnered in the US
• The ideas behind Agile and Scrum come from Lean Manufacturing based on the
NUMMI plant in the US using the Toyota Production System (TPS).
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Toyota and GM partnered in the US
• The ideas behind Agile and Scrum come from Lean Manufacturing based on the
NUMMI plant in the US using the Toyota Production System (TPS)
• Toyota was building high quality cars in Japan. In the US however GM had a terrible quality
track record rate and quality was poor. In the 80s GM closed down a factory in Fremont
with one of the poorest quality, and an unmotivated workforce
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Toyota and GM partnered in the US
• The ideas behind Agile and Scrum come from Lean Manufacturing based on the
NUMMI plant in the US using the Toyota Production System (TPS)
• Toyota was building high quality cars in Japan. In the US however GM had a terrible quality
track record rate and quality was poor. In the 80s GM closed down a factory in Fremont
with one of the poorest quality, and an unmotivated workforce
• When the US were going to increase import taxes on Japanese cars, they partnered with
GM to reopen the Fremont plant. They introduced TPS and trained the workforce. It was a
huge success
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Toyota and GM partnered in the US
• The ideas behind Agile and Scrum come from Lean Manufacturing based on the
NUMMI plant in the US using the Toyota Production System (TPS)
• Toyota was building high quality cars in Japan. In the US however GM had a terrible quality
track record rate and quality was poor. In the 80s GM closed down a factory in Fremont
with one of the poorest quality, and an unmotivated workforce
• When the US were going to increase import taxes on Japanese cars, they partnered with
GM to reopen the Fremont plant. They introduced TPS and trained the workforce. It was a
huge success
• What everyone considered strange was that Toyota would partner with GM, a competitor,
and allowed them to see the reasons for their success
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Toyota and GM partnered in the US
• The ideas behind Agile and Scrum come from Lean Manufacturing based on the
NUMMI plant in the US using the Toyota Production System (TPS)
• Toyota was building high quality cars in Japan. In the US however GM had a terrible quality
track record rate and quality was poor. In the 80s GM closed down a factory in Fremont
with one of the poorest quality, and an unmotivated workforce
• When the US were going to increase import taxes on Japanese cars, they partnered with
GM to reopen the Fremont plant. They introduced TPS and trained the workforce. It was a
huge success
• What everyone considered strange was that Toyota would partner with GM, a competitor,
and allowed them to see the reasons for their success
• GM staff who had seen the plant in Fremont, tried to copy the process in their other plants.
They failed miserably! They only copied the outline of the plant, but failed to realize it was
the teamwork and the focus on continuous improvement which was the key!
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Toyota and GM partnered in the US
• The ideas behind Agile and Scrum come from Lean Manufacturing based on the
NUMMI plant in the US using the Toyota Production System (TPS)
• Toyota was building high quality cars in Japan. In the US however GM had a terrible quality
track record rate and quality was poor. In the 80s GM closed down a factory in Fremont
with one of the poorest quality, and an unmotivated workforce
• When the US were going to increase import taxes on Japanese cars, they partnered with
GM to reopen the Fremont plant. They introduced TPS and trained the workforce. It was a
huge success
• What everyone considered strange was that Toyota would partner with GM, a competitor,
and allowed them to see the reasons for their success
• GM staff who had seen the plant in Fremont, tried to copy the process in their other plants.
They failed miserably! They only copied the outline of the plant, but failed to realize it was
the teamwork and the focus on continuous improvement which was the key!
• Moral of the story, implementing an empirical approach to continuous improvement is
hard, but key to the success of an agile implementation such as Scrum
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Thank you :-)
agile42 | We advise, train and coach companies building
software www.agile42.com | All rights reserved. Copyright © 2007 - 2009.
Questions? & Answers!
For any further comment and or question, feel free to contact
us: netherlands@agile42.com
Further References:
agile42 Website http://www.agile42.com/
Cynefin Framework: http://www.cognitive-edge.com

Empirical proces control

  • 1.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Empirical Process Control An important but often overlooked aspect of agile product development
  • 2.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Agenda • Introduction • Defined Process Control • Hands-on • material needed 5 pens/pencils + one A4 paper • Empirical Process Control • Cynefin Framework • Applied to Scrum • NUMMI a concept GM failed to copy
  • 3.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Niels Verdonk Agile Coach niels.verdonk@agile42.com
  • 4.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. agile42 - The Agile Coaching Company We make your agile transition succeed agile42 has a proven approach to successful agile implementations. We support you with a full range of coaching services and specialized trainings. We work locally & worldwide Connected coaching for international teams •Coaches across the globe •Deliver consistently across internationally distributed organizations •Support different time zones •Recognize cultural diversity. Agile Transition Agile change for the whole organization •Transformation across entire organizations •At individual, departmental and management level •Assessment •Set goals and define strategy for change •Guide organizations through the transformation •While building knowledge for sustainability
  • 5.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Defined Process Control
  • 6.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Input -> Activity -> Output Production-based process
  • 7.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Defined Process Control • Sequence of activity chains • Input Activity Output➟ ➟ • Each activity requires specific skills
  • 8.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Defined Process Control • Sequence of activity chains • Input Activity Output➟ ➟ • Each activity requires specific skills • Flow is pre-defined and predictable • Time passed / time estimated = efficiency. • Handoff after each activity, responsibility transfers to the next role
  • 9.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Empirical Game
  • 10.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. G am e Pencil Game • Put the A4 paper in front of you and draw a circle like so: • Take pens in one hand and close your eyes and let go • With your eyes closed try to predict where they fell • How many are completely outside the circle? Needed: 5 Pens/Pencils and a sheet of A4 paper
  • 11.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. G am e Pencil Game • Put the A4 paper in front of you and draw a circle like so: • Take pens in one hand and close your eyes and let go • With your eyes closed try to predict where they fell • How many are completely outside the circle? • Now repeat with your eyes open • How many are completely outside the circle? Needed: 5 Pens/Pencils and a sheet of A4 paper
  • 12.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. G am e Pencil Game • Put the A4 paper in front of you and draw a circle like so: • Take pens in one hand and close your eyes and let go • With your eyes closed try to predict where they fell • How many are completely outside the circle? • Now repeat with your eyes open • How many are completely outside the circle? • Try to change 1 thing to improve • Now how many are completely outside the circle? Needed: 5 Pens/Pencils and a sheet of A4 paper
  • 13.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Empirical Process Control
  • 14.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Experiment and Analyze R&D based approach
  • 15.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Empirical Process Control • Based based on empirical measurement • Measure results in a defined interval of time • Conduct Safe-to-Fail experiments • Allows for stabilization and optimization • While improving the outcome iteratively
  • 16.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Empirical Process Control • Based based on empirical measurement • Measure results in a defined interval of time • Conduct Safe-to-Fail experiments • Allows for stabilization and optimization • While improving the outcome iteratively • Effective in unordered, complex systems • With behavior which can’t be predicted upfront • But retrospectively we can identify patterns in behavior
  • 17.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Cynefin Framework
  • 18.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Cynefin Framework • Developed by Dave Snowden as a sense making model • Cynefin is a Welsh word, which means 'habitat' or 'place’ • The true meaning is we all have different pasts and backgrounds • The name describes the evolutionary nature of complex systems
  • 19.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Simple Disorde r Cause and Effect relationship exists, are predictable and repeatable Domains in a Nutshell
  • 20.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Simple Complicated Disorde r Cause and Effect relationship exists, are predictable and repeatable Cause and Effect relationship exists, but are not self-evident, analysis is required Domains in a Nutshell
  • 21.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Simple ComplicatedComplex Disorde r Cause and Effect relationship exists, are predictable and repeatable Cause and Effect relationship exists, but are not self-evident, analysis is required Cause and Effect are only obvious in hindsight, with unpredictable emergent outcomes. Domains in a Nutshell
  • 22.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Simple ComplicatedComplex Chaotic Disorde r Cause and Effect relationship exists, are predictable and repeatable Cause and Effect relationship exists, but are not self-evident, analysis is required Cause and Effect are only obvious in hindsight, with unpredictable emergent outcomes. No Cause and Effect can be determined. We need to act very quickly in order to stabilize the system again. Domains in a Nutshell
  • 23.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Simple Sense Categorize Respond Cynefin Framework Disorde r
  • 24.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Simple Complicated Sense Categorize Respond Sense Analyze Respond Cynefin Framework Disorde r
  • 25.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Complex Simple Complicated Sense Categorize Respond Sense Analyze Respond Probe Sense Respond Cynefin Framework Disorde r
  • 26.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Complex Simple Complicated Chaotic Sense Categorize Respond Sense Analyze Respond Probe Sense Respond Act Sense Respond Cynefin Framework Disorde r
  • 27.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Cynefin Framework
  • 28.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Empirical Process Control in Scrum
  • 29.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Software Development is Complex • Software Development today is almost always in the Cynefin Complex domain
  • 30.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Software Development is Complex • Software Development today is almost always in the Cynefin Complex domain • On a technical level we are dealing with: • Various modules, gems, libraries, etc. • Interaction with 3rd party systems and services • We are not repeating an existing known process
  • 31.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Software Development is Complex • Software Development today is almost always in the Cynefin Complex domain • On a technical level we are dealing with: • Various modules, gems, libraries, etc. • Interaction with 3rd party systems and services • We are not repeating an existing known process • On a functional level we are dealing with: • Changing requirements and priorities due to market changes • We cannot predict the behavior of global internet users
  • 32.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Developer Insights • In waterfall solutions are created upfront including detailed technical design. • In practice developers encounter imperfections in the technical design • Also on a functional level things are not always accurate • The results were that we needed to charge our clients extra (remember CRs?)
  • 33.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Developer Insights • In waterfall solutions are created upfront including detailed technical design. • In practice developers encounter imperfections in the technical design • Also on a functional level things are not always accurate • The results were that we needed to charge our clients extra (remember CRs?) • In Scrum cross functional experts work together to solve small problems • The problems defined in User Stories are solved collaboratively • They can apply what they learned already in the next user story • The technical design is emerging
  • 34.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Incremental Product Development • An incremental process is not by definition an empirical approach • You can still use Scrum to deliver a pre-defined software solution • This has benefits, you will still deliver working software each Sprint • But you cannot deal with changes in your market • Or know how your pre-defined solution will be received by global users
  • 35.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Incremental Product Development • An incremental process is not by definition an empirical approach • You can still use Scrum to deliver a pre-defined software solution • This has benefits, you will still deliver working software each Sprint • But you cannot deal with changes in your market • Or know how your pre-defined solution will be received by global users • If you apply an empirical approach to Product Development • Product Owners evaluate the delivered User Stories together with the team • The Backlog can be re-prioritized accordingly and stories deleted or added • We can work towards a Minimal Viable Release and validate it with our users • We release sooner and incorporate the feedback to create a better products
  • 36.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Continuous Improvement of the Process • After each Sprint we gather feedback on how team can work more effective • This allows us to continually improve the way we work • We should inspire teams to try things, conduct safe-to-fail experiments • Some experiments will be successful, they can incorporate this in the process • Some experiments will fail, and they can stop doing this
  • 37.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Continuous Improvement of the Process • After each Sprint we gather feedback on how team can work more effective • This allows us to continually improve the way we work • We should inspire teams to try things, conduct safe-to-fail experiments • Some experiments will be successful, they can incorporate this in the process • Some experiments will fail, and they can stop doing this • It’s vital for organizations to show their commitment to support teams • By showing they are acting on organizational impediments outlined by the teams. • The progress on these items should be visible
  • 38.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Short story: NUMMI Factory
  • 39.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Toyota and GM partnered in the US • The ideas behind Agile and Scrum come from Lean Manufacturing based on the NUMMI plant in the US using the Toyota Production System (TPS).
  • 40.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Toyota and GM partnered in the US • The ideas behind Agile and Scrum come from Lean Manufacturing based on the NUMMI plant in the US using the Toyota Production System (TPS) • Toyota was building high quality cars in Japan. In the US however GM had a terrible quality track record rate and quality was poor. In the 80s GM closed down a factory in Fremont with one of the poorest quality, and an unmotivated workforce
  • 41.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Toyota and GM partnered in the US • The ideas behind Agile and Scrum come from Lean Manufacturing based on the NUMMI plant in the US using the Toyota Production System (TPS) • Toyota was building high quality cars in Japan. In the US however GM had a terrible quality track record rate and quality was poor. In the 80s GM closed down a factory in Fremont with one of the poorest quality, and an unmotivated workforce • When the US were going to increase import taxes on Japanese cars, they partnered with GM to reopen the Fremont plant. They introduced TPS and trained the workforce. It was a huge success
  • 42.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Toyota and GM partnered in the US • The ideas behind Agile and Scrum come from Lean Manufacturing based on the NUMMI plant in the US using the Toyota Production System (TPS) • Toyota was building high quality cars in Japan. In the US however GM had a terrible quality track record rate and quality was poor. In the 80s GM closed down a factory in Fremont with one of the poorest quality, and an unmotivated workforce • When the US were going to increase import taxes on Japanese cars, they partnered with GM to reopen the Fremont plant. They introduced TPS and trained the workforce. It was a huge success • What everyone considered strange was that Toyota would partner with GM, a competitor, and allowed them to see the reasons for their success
  • 43.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Toyota and GM partnered in the US • The ideas behind Agile and Scrum come from Lean Manufacturing based on the NUMMI plant in the US using the Toyota Production System (TPS) • Toyota was building high quality cars in Japan. In the US however GM had a terrible quality track record rate and quality was poor. In the 80s GM closed down a factory in Fremont with one of the poorest quality, and an unmotivated workforce • When the US were going to increase import taxes on Japanese cars, they partnered with GM to reopen the Fremont plant. They introduced TPS and trained the workforce. It was a huge success • What everyone considered strange was that Toyota would partner with GM, a competitor, and allowed them to see the reasons for their success • GM staff who had seen the plant in Fremont, tried to copy the process in their other plants. They failed miserably! They only copied the outline of the plant, but failed to realize it was the teamwork and the focus on continuous improvement which was the key!
  • 44.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Toyota and GM partnered in the US • The ideas behind Agile and Scrum come from Lean Manufacturing based on the NUMMI plant in the US using the Toyota Production System (TPS) • Toyota was building high quality cars in Japan. In the US however GM had a terrible quality track record rate and quality was poor. In the 80s GM closed down a factory in Fremont with one of the poorest quality, and an unmotivated workforce • When the US were going to increase import taxes on Japanese cars, they partnered with GM to reopen the Fremont plant. They introduced TPS and trained the workforce. It was a huge success • What everyone considered strange was that Toyota would partner with GM, a competitor, and allowed them to see the reasons for their success • GM staff who had seen the plant in Fremont, tried to copy the process in their other plants. They failed miserably! They only copied the outline of the plant, but failed to realize it was the teamwork and the focus on continuous improvement which was the key! • Moral of the story, implementing an empirical approach to continuous improvement is hard, but key to the success of an agile implementation such as Scrum
  • 45.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Thank you :-)
  • 46.
    agile42 | Weadvise, train and coach companies building software www.agile42.com | All rights reserved. Copyright © 2007 - 2009. Questions? & Answers! For any further comment and or question, feel free to contact us: netherlands@agile42.com Further References: agile42 Website http://www.agile42.com/ Cynefin Framework: http://www.cognitive-edge.com

Editor's Notes

  • #3 We do an experiment I tried to work with things you might have around Please see if you can find 5 pens and a piece of paper (A4 size) If you don ’ t have 5 pens, 2 or 3 will also work.
  • #4 Introduction of me: Agile coach Worked in the industry over 15 years, have a development brand Etc.etc.
  • #7 If you have to produce the same thing day in day out - an assembly line might work. Then a defined process might work for you. Based on separated activities with a defined input and output.
  • #8 The outcome must be known and well defined, it already has been achieved more than once. Since the order of the activities has a clear order, delay of one step has a direct impact on the total chain of activities and throughput of production. This is why each activity is monitored and measured for efficiency. This type of process control works well with ordered systems in simple or complicated domains
  • #9 The outcome must be known and well defined, it already has been achieved more than once. Since the order of the activities has a clear order, delay of one step has a direct impact on the total chain of activities and throughput of production. This is why each activity is monitored and measured for efficiency. This type of process control works well with ordered systems in simple or complicated domains Handover, with the responsibility now resting on the recipient
  • #13 Off course you should only change 1 thing at a time.
  • #15 Creating software is a creative exercise. It is about solving the problem ONCE. There is no production issue in IT. The Team is ONLY responsible for Software Quality (not hours worked, not customer satisfaction, etc. etc.)
  • #16 Use the example how many cars drive on the A2.
  • #17 Use the example how many cars drive on the A2. Feedback: Keywords and bullets - enthousiaster
  • #19 Although the Cynefin Framework is too big a topic for this webinar (Dave Snowden offers a 4 day course), But I wanted to show you the need for an Empirical approach. In SW dev we are mostly in the complex domain in todays software development projects, Sense making, opposed to categorization models. Sensemaking iswhat we do when individuals are attempting to make sense of observed data.
  • #20 This slide should probably be taken out.
  • #21 This slide should probably be taken out.
  • #22 This slide should probably be taken out.
  • #23 This slide should probably be taken out.
  • #24 Simple - Ordered system, we are experienced, we sense and decide what to do Complicated - Ordered system, we are not experienced, we analyze (with expert) and decide what to do Complex - Hard to predict, no causality,
  • #25 Simple - Ordered system, we are experienced, we sense and decide what to do Complicated - Ordered system, we are not experienced, we analyze (with expert) and decide what to do Complex - Hard to predict, no causality,
  • #26 Simple - Ordered system, we are experienced, we sense and decide what to do Complicated - Ordered system, we are not experienced, we analyze (with expert) and decide what to do Complex - Hard to predict, no causality,
  • #27 Simple - Ordered system, we are experienced, we sense and decide what to do Complicated - Ordered system, we are not experienced, we analyze (with expert) and decide what to do Complex - Hard to predict, no causality,
  • #28 Although the Cynefin Framework is too big a topic for this webinar (Dave Snowden offers a 4 day course), But I wanted to show you the need for an Empirical approach. In SW dev we are mostly in the complex domain in todays software development projects, Sense making, opposed to categorization models. Sensemaking iswhat we do when individuals are attempting to make sense of observed data.
  • #33 If you have worked in on a waterfall project as a developer, you probably encountered that the technical design was often less than perfect in practice. Also on a functional level things sometimes didn’t add up, and required changes to the design. Changes were hard, documentation and technical design needed to be updated, and we needed to charge our clients extra (remember dealing with CRs?)
  • #34 If you have worked in on a waterfall project as a developer, you probably encountered that the technical design was often less than perfect in practice. Also on a functional level things sometimes didn’t add up, and required changes to the design. Changes were hard, documentation and technical design needed to be updated, and we needed to charge our clients extra (remember dealing with CRs?)
  • #35 A MVR is a trimmed down version of the end product, by focussing on what is really needed.
  • #36 A MVR is a trimmed down version of the end product, by focusing on what is really needed. Feedback: Minder text in bullets
  • #37 When a team gives you as a manager some suggestions with what they can work more efficiently, it would be stupid to ignore them, right? Even failed experiments have value, they are still learning's and they may provide an inspiration for new experiments.
  • #38 When a team gives you as a manager some suggestions with what they can work more efficiently, it would be stupid to ignore them, right? Even failed experiments have value, they are still learning's and they may provide an inspiration for new experiments.
  • #40 You probably know
  • #41 Toyota had little defects and low call back. GM had lots of defects and high callback rate. Workforce in Fremont was also considered the worst of the GM plants
  • #42 The workers were very happy no more strikes the quality rates were almost as good as the factories in Japan
  • #43 They also thought it was strange they hired a lot of the unmotivated staff back.
  • #44 Some of the GM staff had worked in NUMMI or were trained by Toyota Photograph every inch.
  • #45 Mention Cargo Cult.