E6S-158 Project Schedule Planning Part 1 - Near PERTfection Task Duration Estimation

Like / Dislike  **We'd Appreciate Your Opinion**


Email me: aaron@e6s-methods.com

Leave a Review! http://bit.ly/E6S-iTunes;

Donations: http://bit.ly/E6S-Donate


Intro:  Welcome to the E6S-Methods podcast with Jacob and Aaron, your weekly dose of tips and tricks to achieve excellent performance in your business and career.  Join us as we explore deeper into the practical worlds of Lean, Six Sigma, Project Management and Design Thinking.  In this episode number 158, we insert ourselves back into project schedule planning with "Near PERTfection," and compare several methods of task duration estimation.  If you're just tuning in for the first time, find all our back episodes on our podcast table of contents at e6s-methods.com. If you like this episode, be sure to click the "like" link in the show notes.  It's easy.  Just tap our logo, click and you're done. Tap-click-done!  Here we go. http://bit.ly/E6S-158 Leave a Review! http://bit.ly/E6S-iTunes


I            Recall: Project Schedule Planning

a.       Break work down into estimable tasks (i.e. Work Breakdown Structure - WBS)

i.      Episode 039: Work Breakdown Structure (WBS) ***Bite-size Project Planning With the WBS*** http://bit.ly/E6S-039

ii.      Episode 41: WBS in LSS http://bit.ly/E6S-041***How Do You Eat the Lean Six Sigma Elephant...? Wicked Bite-Sized with the WBS.***

b.      Estimate the task time and arrange in Network diagram to find the critical path and float times

i.      E6S-098 Network Diagramming - Know your float before you crash . http://bit.ly/E6S-098

c.       Develop project schedule

II         But how do we estimate task duration?

a.       Historical data, given a detailed history

b.      Anecdotal data. Gathering input from others

c.       Bottom up, educated guesses


III      In this episode we'll concentrate on different methods of duration estimation and how they can impact the overall timeline.

a.       3-point estimates

b.      Monte Carlo Simulations

c.       Rule of Pi

IV      Consider this example where the critical path determine by 3 tasks un series that each take 5 days, most likely

a.       Intuition- add up the "Most Likely" - 5+5+5=15

b.      3-point estimates (simple) - add up the estimated averages

i.      (optimistic + most likely + pessimistic)/3 at each step

1.      (3 + 5 + 9)/3 = 5.667

2.      All summed up --> 5.667+5.667+5.667=17

ii.      SD = (optimistic - pessimistic)/6

1.      Convert 15 Days to Z-score with SD and mean of 17 Days to determine probability ("NORMSDIST" function in excel or Statistics Tables)

2.      Estimated Mean +/- 2 × SD is approximately 95%

c.       3-point PERT analysis (Project Evaluation and Review Technique) add up the estimated averages

i.      (optimistic + 4*most likely + pessimistic)/6 at each step

1.      (3 + 4*5 + 9)/6 = 5.333

2.      All summed up --> 5.333+5.333+5.333=16

ii.      SD = (optimistic - pessimistic)/6

1.      Convert 15 Days to Z-score with SD and mean of 16 Days to determine probability ("NORMSDIST" function in excel or Statistics Tables)

2.      Estimated Mean +/- 2 × SD is approximately 95%


d.      Monte Carlo = Identify each process step as a distribution.  Simulate 20,000 trials and measure probability based on aggregated distributions (converges on normal)

i.      Triangular Distribution Estimates.

ii.      Beta PERT Distribution Estimates

e.       Unofficial rule of Pi?  - Sum it up, and multiply by Pi (3.14)

V         So what does this all mean?

a.       If you do not incorporate SD, you can almost guarantee you'll underestimate the total duration

b.      When reporting likely durations, consider giving the 95% answer for total duration, or some range.

c.       Hand 3-point calculations are not very difficult for aggregating step durations if you have excel,

d.      Most methods tend to agree somewhat around the 95% range, especially the PERT 3-point and Beta Monte Carlo methods

e.       Monte Carlo offers greater precision if there is significant historical data on task duration

Outro: Thanks for listening to episode 158 of the E6S-Methods podcast.  Stay tuned for episode 159 where we continue our discussion on project schedule planning, this time the "agile" way.  Don't forget to click "like" or "dislike" for this episode in the show notes. Tap-click-done!  If you have a question, comment or advice, leave a note in the comments section or contact us directly. Feel free to email me "Aaron", aaron@e6s-methods.com, or on our website, we reply to all messages.  If you heard something you like, then Clammr and share it.  Don't forget you can find notes and graphics for all shows and more at www.E6S-Methods.com. "Journey Through Success. If you're not climbing up, you're falling down."    Leave a Review! http://bit.ly/E6S-iTunes


Like / Dislike  **We'd Appreciate Your Opinion**