Globo.com Weekly Talks - Sprint Risk Forecasting

583 views

Published on

Palestra sobre Agile Risk Management e Sprint risk forecasting no Globo.com Weekly Talks

0 Comments
2 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
583
On SlideShare
0
From Embeds
0
Number of Embeds
4
Actions
Shares
0
Downloads
6
Comments
0
Likes
2
Embeds 0
No embeds

No notes for slide

Globo.com Weekly Talks - Sprint Risk Forecasting

  1. 1. globo Sprint risk .com forecastingThursday, November 1, 12 2
  2. 2. e se fosse possível prever o resultado do seu sprint...?Thursday, November 1, 12 3
  3. 3. Thursday, November 1, 12 4
  4. 4. Thursday, November 1, 12 5
  5. 5. Risk is the potential that a chosen action or activity (including the choice of inaction) will lead to a loss (an undesirable outcome). [...] Potential losses themselves may also be called "risks". Almost any human endeavor carries some risk. http://en.wikipedia.org/wiki/Riskgot risk?Thursday, November 1, 12 6
  6. 6. Thursday, November 1, 12 7
  7. 7. “Risk Management is part of Prince2, part of PMBOK, and part of the CMMI, but you dont often see it addressed explicitly in books on agile methods. I think thats strange.” Jurgen AppeloThursday, November 1, 12 8
  8. 8. Thursday, November 1, 12 9
  9. 9. Risk Matrix Agile Risk Matrix Risk Register Agile Risk Register Risk Rating Agile Risk Rating Risk Mitigation Agile Risk MitigationThursday, November 1, 12 10
  10. 10. Risk Matrix Agile Risk Matrix Risk Register Agile Risk Register Risk Rating Agile Risk Rating Risk Mitigation Agile Risk MitigationThursday, November 1, 12 10
  11. 11. BULLSHIT! BULLSHIT!BULLSHITThursday, November 1, 12 11
  12. 12. Thursday, November 1, 12 12
  13. 13. Thursday, November 1, 12 13
  14. 14. ... agile software development is a risk management strategy, because all principles of risk management are nicely covered when youve implemented an agile process. Boris GlogerThursday, November 1, 12 14
  15. 15. Thursday, November 1, 12 15
  16. 16. Thursday, November 1, 12 16
  17. 17. Project risk management ‣ Risk management do planejamento do projeto (Product Backlog) ‣ Forecast que monitora quantos sprints temos em backlog ‣ Numero ótimo: maior que 2,5 menor que 4 ‣ Inventado pelo Danilo Bardusco Backlog Total Velocidade média * 3 Ref: http://macaubas.com/agile/backlog-readyThursday, November 1, 12 17
  18. 18. Backlog ready - exemplo ‣ Velocity: ‣ 25 story points ‣ Product backlog no início do sprint: ‣ 120 story points ‣ Backlog ready: ‣ 120 / (25 * 3) = 160% ‣ Significa que no início deste sprint, havia backlog para 4.8 sprints Ref: http://macaubas.com/agile/backlog-readyThursday, November 1, 12 18
  19. 19. Backlog ready - na prática Backlog ready 200% 175% 187% 170% 150% 125% 118% 100% 75% 88% 71% 50% 25% 40% 0% 1 2 3 4 5 6 Backlog ready Minimo desejável Ref: http://macaubas.com/agile/backlog-readyThursday, November 1, 12 19
  20. 20. e se fosse possível prever o resultado do seu sprint...?Thursday, November 1, 12 20
  21. 21. Sprint risk forecastingThursday, November 1, 12 21
  22. 22. Sprint risk forecasting Complexidade média Taxa de sucessoThursday, November 1, 12 21
  23. 23. Sprint risk forecasting ‣ Risk management do planejamento do sprint (Sprint Backlog) ‣ Métrica de forecast do sprint ‣ Numero ótimo: depende do time ‣ Inventado por mim Insight - há uma relação inversa entre o tamanho médio das histórias planejadas no sprint backlog e a taxa de entrega do sprint! Ref: TBD - http://macaubas.com/Thursday, November 1, 12 22
  24. 24. Sprint risk forecasting ‣ Complexidade média: ‣ Média aritmética do sprint backlog (início do sprint) ‣ Complexidade planejada: 26 pontos, 7 histórias ‣ Complexidade média: 26/7 = 3.25 Ref: TBD - http://macaubas.com/Thursday, November 1, 12 23
  25. 25. Sprint risk forecasting ‣ Taxa de sucesso ‣ % de pontos entregues de fato, em relação ao planejado ‣ Complexidade planejada: 26 pontos ‣ Complexidade entregue: 27 pontos (time entregou um ponto a mais do que o planejado) ‣ Taxa de sucesso: (26/27)*100 = 104% Ref: TBD - http://macaubas.com/Thursday, November 1, 12 24
  26. 26. Tamanho médio - na prática Tamanho média das histórias 6 5.6 250% 4.9 5 200% 4.0 4 3.3 3.2 150% 104% 107% 100% 2 100% 77% 69% 1 50% 0 0% 1 2 3 4 5 Complexidade Media Taxa de sucesso Ref: http://macaubas.com/agile/backlog-readyThursday, November 1, 12 25
  27. 27. E daí?Thursday, November 1, 12 26
  28. 28. Sprint risk forecasting ‣ É possível prever o resultado do sprint (taxa de sucesso) baseado em dados históricos! ‣ Regressão linear simples (least square estimates), ‣ Entrada = complexidade média das histórias do sprint atual. ‣ Entrada = histórico de compl. média & taxa de sucesso ‣ Saída = forecast da taxa de sucesso do sprint que está começando Ref: TBD - http://macaubas.com/Thursday, November 1, 12 27
  29. 29. Sprint risk forecasting Sprints 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Compl. planejada 20 32 16 17 37 30 36 40 47 50 34 51 19 21 45 24 40 41 Compl. entregue 50 43 9 32 29 35 31 40 53 62 46 46 20 24 45 34 42 47 Qtd planejado 4 9 6 5 12 8 9 8 14 17 10 16 9 10 20 10 14 14 Qtd entregue 20 17 6 13 13 13 8 9 17 22 14 17 10 13 21 15 15 18 Complexidade 1.00 1.88 2.67 1.31 2.85 2.31 4.50 4.44 2.76 2.27 2.43 3.00 1.90 1.62 2.14 1.60 2.67 2.28 Media Taxa de sucesso 250% 134% 56% 188% 78% 117% 86% 100% 113% 124% 135% 90% 105% 114% 100% 142% 105% 115% 19o Sprint: Taxa de sucesso prevista: Complexidade planejada: 48 102% Quantidade de histórias: 16 Complexidade média: 3 Ref: TBD - http://macaubas.com/Thursday, November 1, 12 28
  30. 30. Thursday, November 1, 12 29
  31. 31. Thursday, November 1, 12 30
  32. 32. Sprint risk forecasting in action Taxa de sucesso prevista Nível de risco maior ou igual a 95% Baixo entre 95% e 80% Médio entre 80% e 70% Alto menor que 70% Fujam para as colinas! Ref: TBD - http://macaubas.com/Thursday, November 1, 12 31
  33. 33. Sprint risk forecasting ‣ Massa de dados: ‣ Coleta em 2 times por 18 sprints (G1 e Ego) ‣ Coleta em 3 times por 9 sprints, e contando (Webmedia) ‣ Slideshare: http://slidesha.re/Q7KjVy Ref: TBD - http://macaubas.com/Thursday, November 1, 12 32
  34. 34. globo.comThursday, November 1, 12 33

×