1. Agile in the Bathtub Gaetano Mazzanti @mgaewsj Agile42 Gama-Tech
2. “too many things to do” “we are too slow” “poor quality”sounds familiar?
3. 18 months ago “I would like to coach Agile & Lean in an non- software context”
4. on May 2011 I got a call“we need help”
5. May 2011 “we are too slow” “poor quality” “too many things to do”“engineers! those lazy ineffective slackers”
6. products: luxury hydromassage bathtubsand showers, italian market leaderorg & process: traditional top-down org,phase-gate process, rigidly deﬁnedproceduresfocus: cutting costs, downsizing, outsourcing
7. the challengeAgile & Lean coaching in an non-software context “let’s try it!”
8. step Iwho, what, how... ... and why :)
9. assessment interviews +Draw Your Process +pre-mortem retro
10. interviews outcome
11. draw your process
12. dot voting critical steps
13. draw your process most critical stepsfeasibility (especially determining costs and schedule)late product modiﬁcationsproduct concept phase (too long, vague inputs)
14. pre-mortem retro
15. pre-mortem outcomeoptimistic planningdelays in approvalscontinuous changes in prioritiespeople moved to different projectsprojects ‘forgotten’ in a dormant statespecs change, even late in the projectproduction costs higher than expected
16. summing it upmost operational issues somewhat knownsome dismissed as irrelevant, some simply notaddressed for the time beingdispirited employeesmanagement caught completely off-guard byreported lack of transparency and communicationand by employee’s demotivation
17. step IIbeginning the transition
18. timeline !
19. active learning “we want prescriptions not principles” Agile & Lean Values interactions, communication, self-organization queues, WIP limits initial misunderstandings and annoyance by people “multitasking is a used to being told justnatural and effective what to doapproach to address overload” (!)
20. forced managers to read this book
21. mixed feelingsenthusiasm for the possible upcoming (r)evolutionskepticism about the chances to really changeanything
22. step IIIAgile & Lean for Product Development
23. single project vs project portfolio
24. initial goalsmap the real process visualize all ongoing activitiesexpose dysfunctions measure and analyze data
25. a perfect recipe for failure Development/ Concept Production Prototyping CKP CKP CKP CKP CKP I II III IV V DRW DRW DRW I II IIIfeatures, cost and shipment date set here Involved Depts: Engineering, Operations, Quality Control, Test Lab
26. work-item type analysis project typical average size duration delay (months) (months) small 6-9 2-3 medium 9-12 3-6 large 12-18 6-9 small modiﬁcations / /each project leader is in charge of multiple projects
28. demand analysisnew product conceptsrecurring events (e.g. exhibitions)sales opportunities difﬁcult to predict, typically take priority over any other project
29. time for a board!
30. our ﬁrst try color = size S M L
31. the magic of visualizing capacity demand waiting forCKPIII approval initial design, planning, outsourcing, mold design, etc.
32. immediate action/winsprojects reviewed in terms of business priority,cost of delay, risk and complexitysome projects stopped; others accelerated andmoved quickly to CKP III
33. one week later
34. we used this board for three weeks tounderstand which kind of information we couldget from itwe started tracking time information to collect(partial) cycle time data
35. some issues with our boardoverloading and queues exposed but only partiallypreventedpriorities still ‘suggested’ by managementwith no objective selection criteriaproject ownership changes during the project life,this was not visible on the board
36. time for daily standups andweekly retrospectives
37. standups & retrosinitially, engineers found these activities useless andboringstandups were too long and repetitious as they wereused just to report status to a managerso we changed approach and started focusing only onhighlighting problems and collaboratively decide whatto do
38. the path to a new boardassign value and priorities to projects based on businessvalue, cost of delay, due dates, etc.deﬁne and set WIP limits (“wow!”)pull approach to select and move forward projectspolicies to rule transitions from one state to anotherdeﬁne relevant data to represent on cards to help intaking decisions and deﬁning further policies
39. a new board color = project leader
42. reached WIP limit!
43. WIP limitseach project leader is in charge of multiple projects at a given timeinitially assigned WIP = 1 to each project but this was not very usefulproject characteristics including size, cardinality and parameters like machining orforming complexity were then used to weigh WIP (deﬁned by engineers...)this resulted in each project inducing 0,2 to 2 in terms of WIPspeciﬁc WIP limits have been set for resources (mostly hardware and used also forstandard production) like molding machines, test lab and ﬁnal prototypingsome of these limits are not constant but may vary in time as they have to lookahead to take into account standard production needs and to anticipate the loadinduced by projects in progresswhenever WIP limits are reached, a project is tagged with a blue square symbol andput on hold until resources are available (waiting time is measured)
44. WIP limits
45. WIP limits ][ ] ][
46. too many columns? we tried multiple simpler boards it did not work (people rejected it)reality is complex, better to see it clearly
47. metrics used(sub)cycle timeactive vs. idle timecumulated WIP by project leader per projectphase (to check perceived vs measuredoverburden)
48. decreased (sub)cycle time Development/ Concept Production PrototypingCKP CKP CKP CKP CKP I II III IV V DRW DRW DRW I II III -50% -30% -20% -25%
49. sm M 0.50$ 1.00$ 1.50$ 2.00$ 2.50$ 3.00$ ay 0.50$ 1.00$ 1.50$ 2.00$ 2.50$ 3.00$ +1 Jul+10$Aug+10$ Ju $ n+ Ju $ 10 l+1 0 0 al lpSep+10$ Au $ g+Oct+10$ 10 Se $ ro p+Nov+10$ j 10 Oc $Dec+10$ t+1 No 0$ ecJan+11$ v+ 1 De 0$Feb+11$ c+ 10 tsMar+11$ Ja $ n+ 11 DRW$I$Apr+11$ Fe $ DRW$III$ b+ 1May+11$ M 1$ ar +1 1Jun+11$ Ap $ r+1 Jul+11$ M 1$ ay +1Aug+11$ 1$ LCL$ avg$ UCL$ LCL$ avg$ UCL$ DRW$III$ DRW$I$ 0.40$ 0.60$ 0.80$ 1.00$ 1.20$ 1.40$ 1.60$ 1.80$ 1.00$ 1.20$ 1.40$ 1.60$ 1.80$ 2.00$ Sep,10$ Jun,10$ Oct,10$ Jul,10$ Nov,10$ Aug,10$ Dec,10$ Sep,10$ before... Jan,11$ Feb,11$ Oct,10$ Mar,11$ Nov,10$ Apr,11$ Dec,10$ May,11$ Jun,11$ Jan,11$ Jul,11$ Feb,11$ CKP$IV$ DRW$II$ Aug,11$ Mar,11$ Sep,11$ Apr,11$ Oct,11$ May,11$ Nov,11$ Dec,11$ Jun,11$ LCL$ avg$ LCL$ avg$ UCL$ UCL$ CKP$IV$ DRW$II$
50. ec ts rojsm al lp ...after DRW$I$ DRW$II$ 3.00$ 2.00$ 1.80$ 2.50$ avg$ 1.60$ avg$ 2.00$ 1.40$ LCL$ LCL$ 1.50$ 1.20$ UCL$ UCL$ 1.00$ 1.00$ pre$avg$ pre$avg$ 0.80$ DRW$I$ DRW$II$ 0.50$ 0.60$ $ $ 1$ $ $ 2$ $ $ 1$ $ $ $ $ 2$ $ 2$ 2$ 2$ 11 11 11 11 11 11 11 11 12 12 12 t+1 +1 t,1 +1 ,1 r+1 r,1 g+ p+ v+ c+ g, p, v, c, n, b+ b, n ar ar Oc Oc Ap Ap De De No No Au Au Se Ja Se Ja Fe Fe M M DRW$III$ CKP$IV$ 3.00$ 1.60$ 2.50$ 1.40$ avg$ avg$ 1.20$ 2.00$ LCL$ LCL$ 1.00$ UCL$ 1.50$ UCL$ 0.80$ pre$avg$ pre$avg$ 1.00$ 0.60$ CKP$IV$ DRW$III$ 0.50$ 0.40$ Oct,11$ Nov,11$ Dec,11$ Jan,12$ Feb,12$ Mar,12$ Apr,12$ $ 1$ $ $ $ $ 2$ 2$ 11 11 11 12 12 t+1 +1 r+1 p+ v+ c+ n+ b+ ar Oc Ap De No Se Ja Fe M previous range
51. extending Agile/Lean to other departmentsthey asked for it :), but obviously they wantedjust to “copy” the practicesin marketing: iteration-based approach toprepare a product-launch (4 weeks)in sales: kanban board to visualize and optimizeﬂow of sales activitiesretrospectives rocked again
52. achievements ings you do, how m any th eate” “it ’s not value you cr t how much bubetter predictability for Small andMedium projects; reduced cycle timesvalue + cost of delay based selectionreduced overburdenengineers are more engaged and efﬁcien cies ismotivated delays and in g fast” “rem oving han ju st goin impo rtant t more
53. differences with sw teamsno pursuing of technical excellenceengineers feeling and acting as individuals.they did not feel they were part of a team.they had their own independent goals and were focusedjust on themnote: standups and retrospectives helped a lot in providing ateam-forming glueno “works on my machine” syndrome
54. open challengespeople are inclined to rest on their laurels andto stop pursuing continuous improvementlead time (concept to cash) is still too longeconomy in Italy has worsened, the companyhas a new Managing Director, still not clearwhat’s going to happen
55. in closingAgile+Lean in non-software context yes!
56. on December 2011 I got another call “we need help”I am doing it again: currently working with a Power Transmission company “look ma, no software!” to be continued...