Diese Präsentation wurde erfolgreich gemeldet.
Wir verwenden Ihre LinkedIn Profilangaben und Informationen zu Ihren Aktivitäten, um Anzeigen zu personalisieren und Ihnen relevantere Inhalte anzuzeigen. Sie können Ihre Anzeigeneinstellungen jederzeit ändern.
Daniel S. Vacanti
ActionableAgileTM
daniel@actionableagile.com
@danvacanti
Actionable Agile Metrics
for Predictability
First, let me set the record
straight…
“When will it be done?”
“When will it be done?”
Date (number of days)
Elapsed Time
“When will it be done?”
Date (number of days)
Story Points / Velocity
Stop me if you’ve heard this
one before…
“Relative Complexity is the
best predictor of how long it
takes an item to complete”
Why do we even bother with
Story Points?*
I’m going to suggest something
radical…
As an example:
How long does it take you to get
to work in the morning?
“It depends…”
Here’s
a spot!!
4 4 2
Active Done
Testing Done
DoneActive
DevelopmentReady Analysis
3
Here’s
a spot!!
Try an experiment for
me…
Calendar Time
Elapsed Time
Try the same thing
for your process
Backlog Analysis Develop Test Deployed
Start Timer Stop Timer
This chart is called a
(Cycle Time) Scatterplot
If you track nothing else, track the
date that an item starts and the date
that an item completes
(for all work items)
That will give you a
measure of the flow
metric of
Cycle Time
Cycle Time is the
amount of
elapsed time
it takes for a given work item
to complete
“When will it be done?”
for a single item is best answered
by
Cycle Time
“Huh?”
How does generating a
chart like this help us
answer
“When will it be done”?
How do we make
sense of this
randomness?
Your process is “random”.
Therefore, you can’t think
deterministically.
You need to think
probabilistically.
What does it mean to
think probabilistically?
Let’s try another experiment…
Thinking probabilistically means
acknowledging that there is
more than one
possible future outcome
How many people can we “expect”
to be standing after 3 flips?
12.5%
Does that mean 12.5% was the
only possible outcome?
Backlog Analysis Develop Test Deployed
2
Start Timer Stop Timer
Backlog Design Develop Test Deployed
2
Exactly
this dot?
Exactly
this dot?
Or Exactly
this dot?
Will Item #3
end up as…
Backlog Design Develop Test Deployed
2
There is more than one possible outcome
for Item #3 while it is sitting in the back...
What are the possible
outcomes?
40
How do we make
sense of this
“randomness”?
41
50th Percentile = 6 days or less
42
85th Percentile = 15 days or less
43
95th Percentile = 22 days or less
44
50th Percentile = 6 days or less
95th Percentile = 22 days or less
85th Percentile = 15 days or less
Scatterplot Percen...
85th Percentile = 33 days or less
85th Percentile = 13 days or less
What factors
affect
Cycle Time?
WIP
Avg Cycle Time =
Avg WIP
Avg Throughput
What else?
Poor Pull Policies,
Blockers,
External Dependencies,
…
“Actionable Agile
Metrics for
Predictability”
https://leanpub.com/actionableagilemetrics
What does a 19th century
Yorkshire cotton
industrialist have to do
with the Manhattan
Project?
For next time…
QUESTIONS?
Daniel S. Vacanti
http://www.actionableagile.com
daniel@actionableagile.com
@danvacanti
All charts created by:
Thank-you!
Actionable Agile Metrics for Predictability - Daniel Vacanti
Actionable Agile Metrics for Predictability - Daniel Vacanti
Actionable Agile Metrics for Predictability - Daniel Vacanti
Actionable Agile Metrics for Predictability - Daniel Vacanti
Actionable Agile Metrics for Predictability - Daniel Vacanti
Actionable Agile Metrics for Predictability - Daniel Vacanti
Actionable Agile Metrics for Predictability - Daniel Vacanti
Actionable Agile Metrics for Predictability - Daniel Vacanti
Actionable Agile Metrics for Predictability - Daniel Vacanti
Nächste SlideShare
Wird geladen in …5
×

Actionable Agile Metrics for Predictability - Daniel Vacanti

1.957 Aufrufe

Veröffentlicht am

Actionable Agile Metrics for Predictability
“When will it be done?” That's the first question customers ask once work is started. Your predictability is judged by the accuracy of your answer. Think about how many times you’ve been asked that question and how many times you’ve been wrong. That you’ve been wrong more times than right is not necessarily your fault. You have been taught to collect and analyze the wrong metrics. Until now.

About Daniel Vacanti
Daniel Vacanti is a 20+ year software industry veteran who has spent most of his career focusing on Lean and Agile practices. In 2007, he helped to develop the Kanban Method for knowledge work and managed the world’s first project implementation of Kanban that year. He has been conducting Lean-Agile training, coaching, and consulting ever since. In 2011 he founded ActionableAgile (previously Corporate Kanban) which provides industry-leading predictive analytics tools and services organizations that utilize Lean-Agile practices. In 2015 he published his book, “Actionable Agile Metrics for Predictability”, which is the definitive guide to flow-based metrics and analytics. Daniel holds an M.B.A. and regularly teaches a class on lean principles for software management at the University of California Berkeley.

Veröffentlicht in: Technologie
  • Login to see the comments

Actionable Agile Metrics for Predictability - Daniel Vacanti

  1. 1. Daniel S. Vacanti ActionableAgileTM daniel@actionableagile.com @danvacanti Actionable Agile Metrics for Predictability
  2. 2. First, let me set the record straight…
  3. 3. “When will it be done?”
  4. 4. “When will it be done?” Date (number of days) Elapsed Time
  5. 5. “When will it be done?” Date (number of days) Story Points / Velocity
  6. 6. Stop me if you’ve heard this one before…
  7. 7. “Relative Complexity is the best predictor of how long it takes an item to complete”
  8. 8. Why do we even bother with Story Points?*
  9. 9. I’m going to suggest something radical…
  10. 10. As an example: How long does it take you to get to work in the morning?
  11. 11. “It depends…”
  12. 12. Here’s a spot!!
  13. 13. 4 4 2 Active Done Testing Done DoneActive DevelopmentReady Analysis 3 Here’s a spot!!
  14. 14. Try an experiment for me…
  15. 15. Calendar Time Elapsed Time
  16. 16. Try the same thing for your process
  17. 17. Backlog Analysis Develop Test Deployed Start Timer Stop Timer
  18. 18. This chart is called a (Cycle Time) Scatterplot
  19. 19. If you track nothing else, track the date that an item starts and the date that an item completes (for all work items)
  20. 20. That will give you a measure of the flow metric of Cycle Time
  21. 21. Cycle Time is the amount of elapsed time it takes for a given work item to complete
  22. 22. “When will it be done?” for a single item is best answered by Cycle Time
  23. 23. “Huh?”
  24. 24. How does generating a chart like this help us answer “When will it be done”?
  25. 25. How do we make sense of this randomness?
  26. 26. Your process is “random”. Therefore, you can’t think deterministically. You need to think probabilistically.
  27. 27. What does it mean to think probabilistically?
  28. 28. Let’s try another experiment…
  29. 29. Thinking probabilistically means acknowledging that there is more than one possible future outcome
  30. 30. How many people can we “expect” to be standing after 3 flips? 12.5% Does that mean 12.5% was the only possible outcome?
  31. 31. Backlog Analysis Develop Test Deployed 2 Start Timer Stop Timer
  32. 32. Backlog Design Develop Test Deployed 2 Exactly this dot? Exactly this dot? Or Exactly this dot? Will Item #3 end up as…
  33. 33. Backlog Design Develop Test Deployed 2 There is more than one possible outcome for Item #3 while it is sitting in the backlog
  34. 34. What are the possible outcomes?
  35. 35. 40 How do we make sense of this “randomness”?
  36. 36. 41 50th Percentile = 6 days or less
  37. 37. 42 85th Percentile = 15 days or less
  38. 38. 43 95th Percentile = 22 days or less
  39. 39. 44 50th Percentile = 6 days or less 95th Percentile = 22 days or less 85th Percentile = 15 days or less Scatterplot Percentiles (making a forecast)
  40. 40. 85th Percentile = 33 days or less
  41. 41. 85th Percentile = 13 days or less
  42. 42. What factors affect Cycle Time?
  43. 43. WIP
  44. 44. Avg Cycle Time = Avg WIP Avg Throughput
  45. 45. What else?
  46. 46. Poor Pull Policies, Blockers, External Dependencies, …
  47. 47. “Actionable Agile Metrics for Predictability” https://leanpub.com/actionableagilemetrics
  48. 48. What does a 19th century Yorkshire cotton industrialist have to do with the Manhattan Project? For next time…
  49. 49. QUESTIONS? Daniel S. Vacanti http://www.actionableagile.com daniel@actionableagile.com @danvacanti
  50. 50. All charts created by: Thank-you!

×