Transcript Document

When Things Go Right
Cornell’s PeopleSoft 8.9 Upgrade
Lisa Stensland
Manager, CIT Project Management Office
May 15, 2008
Agenda
• Challenge
• Change in Methodology and How We
Manage Resources
• Results
Challenge
• Two critical projects competing for resources
– Upgrade PeopleSoft (HR, Payroll, Contributor
Relations) from 8.0 to 8.9
– Implement PeopleSoft Student Administration
modules
• Technical implementations could not overlap
• They both must be done ASAP
Analysis
• Conducted forecasting of schedule and
budget for multiple options
– Upgrade first, then implement Student Admin
…or…
– Implement Student Admin, then upgrade the
whole thing
• Decision
– Upgrade first, do it as quickly as possible
What We Knew
• The forecasting effort indicated that the PS
8.9 Upgrade project would take
approximately 15-18 months to complete
– Late Summer 2005 - Spring 2007
• We spent 3 months doing more detailed
planning, which resulted
– Targeted completion in Winter of 2006
– Estimated $5M budget
What We Knew
• The timing of a Fall 2006 upgrade was not
ideal
– Competing business cycles
– A “tax/fix version” of PS comes out at that time,
and how would we address that?
• How do we bring the schedule in further when
all our planning indicates that we can’t?
I did some digging….
Critical Chain Project Management
• A way to schedule and track a project that
encourages:
– Aggressive scheduling
– Team Focus
– Team ownership of the project commitment
The old way….
Well…about 5 days
Create list
of tasks
+ I am working on
another project
+ I get interrupted alot
+ Something usually
goes wrong
So…10 days!
Project Manager asks,
“How long will
each task take?”
Safety in estimation
• Time needed to protect the work estimate
commitment from:
– Murphy’s Law - what can go wrong, will go wrong
– Distraction
– Multi-tasking
Parkinson’s Law
Work expands to fill the time
allotted
“Student Syndrome”
Many people will start to fully
apply themselves to a task at
the last possible moment before
a deadline.
What do these factors do to a project?
• It is normal to focus on task ‘due dates’
• Student Syndrome -> Late starts
• Parkinson’s Law
– If nothing goes wrong, the task will finish on time
– If something goes wrong, the task will likely finish
late
• Late tasks on the critical path will delay the
project
What should be done differently?
Safety
Safety
Safety
End
Safety
Remove the safety from the individual tasks
Committed
Finish
Earliest Possible Finish
Move it to the end…the Project Buffer
Refocus the team on…
• Starting tasks on time
• Completing tasks as aggressively as possible
• Managing the amount of Project Buffer that is
consumed
– Project Buffer is consumed when a task on the
critical path is late
– Project Buffer is replenished when a task on the
critical path is early
9/27/2005
10/4/2005
10/11/2005
10/18/2005
10/25/2005
11/1/2005
11/8/2005
11/15/2005
11/22/2005
11/29/2005
12/6/2005
12/13/2005
12/20/2005
12/27/2005
1/3/2006
1/10/2006
1/17/2006
1/24/2006
1/31/2006
2/7/2006
2/14/2006
2/21/2006
2/28/2006
3/7/2006
3/14/2006
3/21/2006
3/28/2006
4/4/2006
4/11/2006
4/18/2006
4/25/2006
5/2/2006
5/9/2006
5/16/2006
5/23/2006
5/30/2006
6/6/2006
6/13/2006
6/20/2006
6/27/2006
7/4/2006
7/11/2006
7/18/2006
7/25/2006
8/1/2006
8/8/2006
8/15/2006
8/22/2006
8/29/2006
9/5/2006
9/12/2006
9/19/2006
9/26/2006
10/3/2006
10/10/2006
10/17/2006
10/24/2006
10/31/2006
11/7/2006
11/14/2006
11/21/2006
11/28/2006
12/5/2006
12/12/2006
12/19/2006
% Buffer Utilization
PS 8.9 Upgrade Fever Chart
100.0%
90.0%
December
80.0%
70.0%
60.0%
50.0%
October
40.0%
30.0%
20.0%
10.0%
July
0.0%
Status Date
Status Date
16
15
11
10
6
5
7/28/2006
7/21/2006
7/14/2006
7/7/2006
6/30/2006
6/23/2006
6/16/2006
6/9/2006
6/2/2006
5/26/2006
5/19/2006
5/12/2006
2
5/5/2006
4/28/2006
4/21/2006
4/14/2006
4/7/2006
3/31/2006
3/24/2006
3/17/2006
3/10/2006
3/3/2006
2/24/2006
Buffer Utilization
PS 8.9 Upgrade Fever Chart
20
7/28
19
18
17
7/21
14
13
12
7/14
9
8
7
7/7
4
3
Final cutover decided:
July 28 - 31
1
0
What Had to Change?
• Estimate aggressively by removing safety
from individual tasks
– Trust management to not hold staff accountable to
aggressive estimates without safety
• Put safety where the project can use it
• Minimize multi-tasking
• Track progress using buffer consumption
– Not individual milestones!
Benefits
•
•
•
•
•
Protection against Murphy’s Law
Took advantage of early finishes
Team protected the project buffer
Opportunity for team to focus
Sponsor visibility to what is possible vs
realistic
• Better visibility to when project is in trouble
The Results?
• Team cutover to PeopleSoft 8.9 in July 2006
– Our most aggressive possible go-live!
• Project delivered 40% under-budget
– $3M spent
– $2M saved
• Staff time not needed - $1.3M
• Unused contingency - $500K
Moral of the story…
• Encourage aggressive planning by
developing trust between management and
the team
• Allow staff to focus on one task at a time
rather than multitask across projects
• Implement mechanisms that encourage the
team to protect the project commitment,
rathern than individual commitments
More information…
• Critical Chain Project Management
– http://www.focusedperformance.com/articles/ccpm
.html