Those are the take aways of a techie after #LKNL13:

Main take aways

• When a measure becomes a target, it ceases to be a good measure (variation on Goodhart’s law)
• The best metric to prioritise is the cost of delay
• Keep your options open as long as possible, push back commitment as late as possible to gather the most insight first
• Success hinges on being wrong early
• If human are involved, you have a complex environment
• Bottlenecks:
  • They are not an actual issue, but their impact can be
  • Use them to keep resource available (not use the team all the time at 100%) and to do some non mission critical work
• Moving a person (eg. go to another team standup) is a lot easier and more efficient than moving knowledge by eg. documentation.

Kanban in a nutshell

• Principles of KB
  1. Start with what you do know
  2. Agree to pursue evolutionary changes 
  3. Initially, respect roles responsibilities and job titles
  4. Encourage acts of leadership at all levels
  5. (1-3 are to go around resistance to change, or more accurately, resistance to be changed)
• Concepts of KB
  1. Service-orientation (DBA service, architecture service)
  2. Service delivery involves workflow
  3. Work flows through a series of information discovery activities
• Practices of KB
  1. Visualise
  2. Limit wip
  3. Manage flow
  4. Make policies explicit
  5. Implement feedback loops
  6. Improve collaboratively, evolve experimentally

One book came again and again as recommended: The Principles of Product Development Flow, by Don Reinertsen.

The slides can be found on lknl13 page.

Talk by talk take aways

David J. Anderson, Modern Management Method

Modern management means that innovation changed the way of doing something – this is not just about a new tool.

Dave Snowden, Complexity: the new paradigm in decision making

• Successful companies do not follow the recipes of other. What lead to success in one company does not mean it will lead to success for your company.
• When a measure becomes a target, it ceases to be a good measure (variation on Goodhart’s law)
• Intrinsic motivation is destroyed by extrinsic targets.
• Look into cynefin framework. Systems divided in 5:
  • Complicated (known unknown, good practices)
  • Simple (known knowns, best practices)
  • Complex (knowable (by experimentation) unknowns, emergent)
  • Chaotic (unknowable unknowns, novel)
  • Disorder: transition phase or when a preference is imposed
  • Collapse from complacency: sudden and unanticipated collapse into chaos from apparent security
• Complexity cannot be eliminated, only absorbed.
• Complexity is about flow. Need to manage the system as a whole, not in part.
• You can intervene on complexity (experimental action), while keeping in mind:
  • Needs to be coherent (not necessary right)
  • Safe to fail
  • Small and tangible
  • Managed as a portfolio (many experiments in parallel)
  • Sometimes oblique approach
  • Includes naive approach
  • A few high risk/high return

Steve Tendon, Unity of purpose and community of trust

• Different incentives and metrics create internal conflicts => different units will have different agendas, because driven by different incentives.
• There should be 1 system wide metric: throughput metric
• Need a system view that avoids local optimisations

Gaetano Mazzanti, People as bottlenecks

• The issue is not actually the bottlenecks, but their economical impact.
• Trust != hero culture. Trust can scale, heroes cannot

Donald Reinertsen, Shooting crooked arrows at moving target in the fog

• Close gaps on basis of economics, not ideology (is it cost effective to complete the last 5% of this feature?)
• Seek and respond to disconfirming information (confirmation bias)
• Think about the economics of creating exploitable options
• Buy information in small batches,act upon any new information
• Improve response time by decentralising the information, the authority and the resources necessary to quickly redirect programs
• Encourage and reward initiative
• Inform your decisions with economics and statistics

Troy Magennis, Cycle time analytics

• Success hinges on being wrong early
• Always compare model vs. actual

Joshua Bloom, What is the value of social capital?

• Emergent slack, resilience with benefit: use bottleneck to create ‘free’ time of people to non mission critical work.
• Moving a person (eg. go to another team standup) is a lot easier and more efficient than moving knowledge by eg. documentation.
• Your software will only have qualities already present in the organisation. The organisation itself needs to have these qualities
• Teaming (verb, a dynamic activity) is more important than teams.

Olav Maassen – Risks and decisions – the ‘When’ rather than the ‘How’

• Option theory: postpone commitment to the latest possible. Keep all options open
  • Options have value
  • Options expire
  • Never commit early unless you know why
• Always ask yourself: how does this create more options to me? How does this allow me to push decision to a later date to get more information?

David J. Anderson, Kanban and evolutionary management: Lessons we can learn from Bruce Lee’s journey in martial arts.

• See principles, concepts and practices above.
• Bruce Lee’s JKD is based on teh same principles: use what works, test in real life

Astrid Claessen, Gamestorming your retrospectives

• Danger in retro: boredom/familiar feeling, hence gamestorming
• Gamestorming is actual work

Ajay Reddy & Dimitar Bakaradzhiev, not everything that counts can be counted and not everything that can be counted counts

• feasibility of measuring
  • Your problem is not as unique as you think
  • You have more data than you think
  • You do not need as much data as you think
  • It is easier than you think to get the data you need
• Good measures
  • Lead time
  • Throughput
  • WIP
  • Cumulative flow diagram (representation of wip at each stage in the system)
  • Flow efficiency (work time against lead time, indication of waste)