(Mis)understanding flow with metrics - Johan Karlsson
Download as PPTX, PDF2 likes1,084 views
This document discusses the importance of understanding product development flow using metrics. It argues that organizations should not be dogmatic about which metrics they use and should trust people over data unless data is being used for a good reason. The document suggests we may be experiencing a paradigm shift in how metrics are used, from a focus on performance metrics to using scientific methods to discover what actually affects outcomes. It maintains that organizations that can discover and understand their actual product development flow will have a competitive advantage.
(Mis)understanding flow with metrics - Johan Karlsson
- 1. (Mis)Understanding ‘Flow’ with Metrics Johan Karlsson Brewing Agile 2014 johan.karlsson@hansoft.com
- 2. Structure Fascist Deep Experience from the Trenches Done some coding Creative Genius
- 4. Why do we need to understand flow with metrics? 94 % Percentage of problems caused by system vs people
- 6. A B Desired state
- 8. Time Performance “There is nothing new to be discovered in physics now. All that remains is more and more precise measurement.” Lord Kelvin 1900 “Hmm….” Einstein, publisher of a paper on relativity theory a few years later. Are we experiencing a paradigm shift in how metrics are being used?
- 9. Current paradigm: Working software is the primary measure of progress* Next paradigm: Do not stay dogmatic about which metrics that are being used * Principles behind the Agile Manifesto: http://agilemanifesto.org/principles.html
- 10. Current paradigm: “In God we trust, all other bring data”* Next paradigm: In people we trust, only bring data you care about for a very good reason *Famous quote, often attributed to Deming
- 11. Current paradigm: Performance is reviewed with metrics Next paradigm: The Death of Reporting: Use scientific methods to discover and teach to stakeholders what actually affects outcome Tip: Arne Roock - Learning from Fake Charts: http://vimeo.com/80365303
- 12. Automatic Creation and Maintenance Powerful data automation and visualization solutions Easier and Cheaper Technology High Cost Low Cost Whiteboards, stickers etc Manual Creation and Maintenance Data warehousing, traditional reporting solutions Improved Agile/Lean Practices
- 13. The usual suspects in current paradigm when asking about metrics (outside of code stuff) Time Time Work remaining # New Bugs
- 14. Work-in-process Development-in- process Lead time 0 10 20 30 40 Time in days Number of items or e.g. story points
- 15. A B Desired state Flow in theory
- 16. Flow in reality a A B d+c d d f Desired state ?
- 17. A good method to understand and uncover complex flows is a competitive advantage
- 18. Actors ● Stakeholders (customers, investors) ● Product Owner ● Team Structures Processes ● Backlogs ● Sprints ● Hierarchy ● Budgets ● Ceremonies: o Stand-ups, planning, reviews etc ● HR, IT, budgeting processes
- 19. Take an Economic Perspective Revenue / Delivered story points Time Cost of delay _____________ _____________ _____________ _____________ _____________ _____________ _____________ _____________ _____________ $$$ $ Time Size S XXL Grooming process + WSJF Yesterday’s Weather Outcome
- 20. Measure validity of advice (example DEEP) Average Size No Prio Low Prio Medium Prio High Prio
- 21. Do not stay dogmatic Truly informed decision making The death of Reporting Organizations that discover their actual flow will be kings and queens in the next paradigm.
- 22. Inspiration in addition to our Customers: Donald G. Reinertsen The Principles of Product Development Flow 2012 Douglas W. Hubbard How to Measure Anything: Finding the Value of Intangibles in Business 2014 Thomas S. Kuhn The Structure Of Scientific Revolutions 1970 Michael Lewis Moneyball 2011
Editor's Notes
- #8: Difficult to talk about metrics without mentioning the lean startup Erick Ries - Lean Startup Actionable metrics Possibly first covered in 1620 by Francis Bacon?
- #9: Knowledge also moves in paradigms Famous example from Physics There is something going on, on a paradigm level, on how we look at these numbers.
- #10: Thought leaders: Douglas Hubbard (How to Measure Anything) - Science was never about having data, it was about getting data. Model your measures around decisions you must make. Examples from the real world: Story: “the only thing we know, is that the things we have been looking at does not help us any more.” Own experience: Many teams struggle with things like burn-downs etc (sometimes for the wrong reasons) but many find it much more useful with Cumulative Flow Diagrams and similar.
- #11: Thought leaders: Datensparsamkeit, Martin Fowler has a good blog post on subject: http://martinfowler.com/bliki/Datensparsamkeit.html Examples from the real world: Story: Moneyball Own experience: The first thing many agile teams do is to throw out old reports, KPI’s and stuff out the door.
- #12: Thought leaders: Deming: Wrong practice: Management by results. (Action based on outcome is not action on the causes of the outcome. Emphasis on cost reduction is an example. Costs are not causes.) Better practice: Use the concepts and tools related to variation to understand and improve the system. Deming estimates that around 94% of the possible improvements belong to the system - the responsibility of management. Examples from the real world: 80+ data scientists working at King Happiness indexes
- #16: A CFD assumes linearity
- #19: Actors = employees, managers, associations etc Structures = authority, responsibility, dependency Processes = methods, codes etc
- #20: Happiness index an example of another outcome