Postmortem

1
Postmortem

• Project Greenberry Industrial Purchase Order
  SystemTeam Thomas Pham, Aaron Veden, Tammy
  Fuchigami

2
5 Steps to the postmortem analysis

• 1. Conduct Survey
• 2. Collect Objective Data
• 3. Conduct A Debriefing Meeting
• 4. Conduct A Project History Day
• 5. Publish Results

3
1. Conduct survey

• Used for large teams
• Starting point Defines the scope and captures
  project team members interests
• 3 guiding principles
• 1. Dont ask for more then you need.
• 2. Dont ask leading questions
• 3. Preserve anonymity
• Sample Questions (Collier, DeMarco Fearey
  1996)
• Did project meetings make effective use of your
  time?
• ( ) always ( ) sometimes ( ) rarely ( ) never
• Did you feel empowered to participate in
  discussions regarding issues that affected your
  work?
• ( ) always ( ) sometimes ( ) rarely ( ) never

4
2. Collect objective data

• According to Collier DeMarco and Fearey (1996)
  three kinds of measurements were cost, schedule,
  and quality.
• Cost Example
• -Total lines of code usually by function
• Schedule Example
• -History of events that cause the schedule to
  slip.
• Quality Example
• -Faults found during each development activity.

5
Collect Objective Data(CONT.)
6
Collect objective data (cont.)
7
Collect Objective Data(cont.)

• Meeting Deliverables
• Introduction (Aaron)
• User Stories (Aaron)
• Tasks (Thomas Tammy)
• Submitting hardcopy (Thomas)
• Submitting electronically (Thomas)
• Validation Requirements left (N/A)

8
3. Conduct a debriefing meeting

• A facilitator to keep meeting on track
• Loosely structured
• What went well and what didnt.
• Items that werent covered in the survey.
• Find the root cause of negative and positive
  effects.
• Members get things off their chest.

9
ConDUCT A Debriefing meeting(cont.)

• Well
• Perfect attendance
• Good compatibility flow in meetings
• Meeting deliverables
• Not Well
• Leaving out parts in assignment
• Submitted electronically late
• Improvement
• Read assignment instructions more carefully

10
4. Conduct a project history day

• Limited number of participant.
• Identify the root causes of the key problem.
• Prioritized list
• Participant are people who know something about
  the problem.
• Outside participant may attend.
• Schedule predictability.

Pfleeger Atlee 2006
11
5. Publish results

• Share insights with team.
• Open letter to managers, peers, developers that
  share Postmortem results.
• Project Description
• All Post mortem positive findings
• Three worst factors that kept team from meeting
  goal
• Improvement activities