Professional Documents
Culture Documents
Participants: Henrik Kniberg, Mattias Skarin, Reza Farhang, Tomas Bjrkholm, Mats Plahn,
Joakim Sunden, Fredrik Sj, Anders Wide, Ulrik Gade, Jesper Thaning, Sune Gynthersen,
Gabrielle Benefield, Thomas Blomseth, Bent Jensen, Jacob Jensen, Martin Josephat, Mary
Poppendieck, Tom Poppendieck.
Mary Poppendieck:
http://tech.groups.yahoo.com/group/leandevelopment/message/4022
1. Fujitsu Measure at very fine level of detail. Perhaps due to need to bill and estimate
accurately
2. Beware of metrics overuse might scare away good people
3. Cultural dimensions aversion to risk at all levels. In rest of world, CIOs face same
constraints
4. CEO totally involved in Kaizen at Fujitsu.
5. Hard ones last. They reform team each project so improved productivity may be due
to team jelling.
6. Be careful of their data. They could easily be gaming the metrics.
7. Interesting that they use function points/scale.
8. Manitoba govt SAP strategy of replacing best people before starting SAP project impl.
9. 7x productivity improvement is probably real.
10. Customer satisfaction was not addressed.
11. Doing things right vs. doing the right things.
12. Would not want to work there. Just because they call it TPS does not mean it is good.
13. Importance of stable teams is lost on them.
14. Solution to other companys problems may not be solution to your problem.
15. To be competitive they need to be more efficient
16. They did not seem to know much of agile software development. Lean <> Agile. Agile
is not the only way to apply Lean to development.
17. They were doing what they needed to do for their problems. Success breeds
defensiveness. Need to step back.
18. The fuzzy front end lasts long, implementation short
19. Is stable team necessary for stable process.
20. Change of HR policies at Toyota in 80s
21. Increase of Temp workers at Toyota. Scandinavia has more fluid labor market.
22. Concept of bulls-eye. What are the primary parts of game. Bulls eye helps prioritize
work during implementation.
23. Data driven decision making for better or worse. Fujitsu may go too far but need more
of this in most places
24. Would not work in environment with high employee turnover.
25. Standardization can support improvements. When all do same process, easier to
improve. With no standard, have no baseline. Not all of their work was standardized but
they were very able to leverage standard and special frameworks.
26. Seem to be templatizing web development as most similar providers do.
27. Wait till they have a design before they make a fast, accurate estimate.
28. No discussion of creating the right design. Apparently they do Govt contracts which
contain the design.
29. They did a pretty direct translation of TPS to software probably not with enough
translation. They seemed to lack enough system thinking.
30. Experimentation with estimation techniques. (Systematic in Ahuus DK is doing similar
experiments.)
Software as a factory
- http://www.bestbrains.dk/Blog/2009/04/21/SoftwareAsAFactory.aspx
Lean Study Tour 2009 - Day 2 (Sune Gynthersen)
- http://www.bestbrains.dk/Blog/2009/04/22/LeanStudyTour2009Day2FeelingPrivileg
ed.aspx
Mary Poppendieck:
http://tech.groups.yahoo.com/group/leandevelopment/message/4031
Day 2 Group reflection:
Toyota Museum
1. Evolution of the loom.
2. TPS is a machine just like the looms
3. TPS Cartoon movie was great.
4. Tolerance for innovation during growth even in absence of immediate return.
5. Attention to details was impressive. Relentless improvement.
6. Looms are almost as hard as cars.
Mary Poppendieck:
http://tech.groups.yahoo.com/group/leandevelopment/message/4032
Katayama San
1. Thought that things were so obvious that it was mystery why other cos dont to it
2. Kaizen = DNA
3. Hours not headcount
4. responsibility throughto whole process
5. Passion mentioned many times
6. in a mature industry, have to act as mature decvelopers
7. Worked chaos first, discipline in execution
8. =Concept at rather high level. Concept is hard part
9. Handshake = Guarantee
10. Ideas of Value engineering and Value analysis was new to group.. but standard in
hard goods design. Not in SW. Applies to any kind of system. Cost per part
committed. Like point estimates. Cost allocated to part, have to come up with design
that will meet cost.
11. lead time is pretty short. Game takes longer.
12. decision making process they are really good at it.
13. If we dont get better we dont survive.
14. Iteration is a bad idea in hardware.
15. Avoid information overload. Summarizing is vital.
16. Cycles of discovery, not iterations.
17. Role of planning division. Clue about how CE knows what is needed. KS had passion
for motorsports. He sold concept to company. Gut feel is important but proposal
needs to be very detailed with data to make point.
18. Family car CE needs to love his family., sports car CE needs to love to drive fast.
19. Attitude toward young people
20. BAD NEWS FIRST, hard items first. Report system, not just policy.
21. DNA many loose entrepreneurial DNA. How to change DNA