Keys to Successful IT Projects: An Executive Guide
www.construx.com
Copyright Notice
These materials are © 2005-2011 Construx Software Builders, Inc.
All Rights Reserved. No part of the contents of this seminar may be reproduced or transmitted in any form or by any means without the written permission of Construx Software Builders, Inc.
My Perspective
4
My Perspective
Software Engineering ExpertContact with C-Level ExecutivesBusiness Owner/CEOPreoccupation with estimation...
Six Insights that C-Level Executives Already Know
6
Six Insights That C-level Executives Already Know
Software projects are late Software projects are out of control Answers from software staff are often
incomprehensible C-Level executives are rational and intelligent
enough to understand these concepts Answers from software staff often seem
evasive Software is the most unpredictable part of the
business
The Seven Critical Insights
Insight #1
Typical Budgeting Processes Undermine Effective Development
9
Typical Budgeting Processes Undermine Effective Development
Many organizations use “One Phase” budgeting processes
This is a poor match for the nature of software development
10
The Cone of Uncertainty
0.67x0.8x1.0x
0.5x
0.25x
2x
4x
1.5x1.25x
Projectschedule
0.85x0.9x1.0x
0.8x
0.6x
1.25x
1.6x
1.15x1.1x
Time
Project scope(effort, size, features)
11
A Too-Common Response to the Cone…
“You encourage range-based reporting. At three places I have managed, it simply has not been an option. In fact, at one board meeting our Director of Program Management was asked to leave the board meeting because of an attempt to use ranges. The quotation was approximately, ‘You don’t seem to understand that your job is to give me a date. If you can’t give me a date, you can’t do your job.’”
12
Implications of the Cone
Accurate budgets can’t be set before significant preliminary work is done
Accurate budgets can be set after about 10-20% of the total project budget has been expended*
* You won’t know what the total project budget is until you expend the 10-20%, and so you can’t accurately calculate the 10-20% ahead of time
Insight #2
Stage-Gate Processes Are Prevalent In Leading Companies
14
Stage-Gate Processes Are Prevalent In Established Companies
Example companiesSome amount of preliminary work is
necessary before an accurate budget or schedule can be set, because of the Cone
Insight #3
Low Quality Is The Single Largest Cost Driver
16
Low Quality Is The Single Largest Cost Driver
50-80% of the typical project’s expense goes into unplanned rework
The most common reasons for this are Poorly defined requirements, which lead to
changing requirements Inadequate project planning Inadequate project management All of which are exacerbated by committing
too early in the Cone
17
Lifecycle Cost Profile
Requirements Architecture Detailed Design
Construction Testing and Debugging
Rel
ativ
e Ef
fort
Typical Project (Pathological Project)
Advanced Project (Healthy Project)
18
Lifecycle Cost Profile (cont.)
Requirements Architecture Detailed Design
Construction Testing and Debugging
Rel
ativ
e Ef
fort
Typical Project (Pathological Project)
Advanced Project (Healthy Project)
Unplanned rework
Process overhead
19
IBM StudyProjects that focused primarily on schedule tended to overrun their schedule goals and were delivered with low quality; projects that focused primarily on quality tended to meet their schedule goals and were delivered with good quality”
Typical reaction when this is done well: “This project must not have been a good test case”
How to Respond to Low Quality
Insight #4
People (Staff) Exert The Largest Impact On Project Outcomes
21
Significance of Individual Variability
According to Cocomo II calibrations, worst personnel will require 14x as much effort to produce software as best personnel
This basic finding has been replicated many times
Source: Software Cost Estimation with Cocomo II, Barry W. Boehm, et al, Prentice Hall, 2000
1.29x
1.40x
1.43x
1.51x
3.53x
Team Cohesion
Staff’s Platform Experience
Staff’s Language and Tools Experience
Staff’s Applications Experience
Personnel/Team Capability
Insight #5
Software Improvement Works Best When Supported at the Organizational Level
23
Where Can Organizations Improve?
Execution Time Constraint
1.26
1.29
1.31
1.33
1.39
1.40
1.42
1.43
1.43
1.46
1.49
1.50
1.51
1.51
1.52
1.53
1.54
1.63
2.38
3.53
Flexibility of Requirements
Team Cohesion
Development for Reuse
Precedentedness
Architecture and Risk Resolution
Staff’s Platform Experience
Database Size
Staff’s Language and Tools Experience
Process Maturity
Storage Constraint
Platform Volatility
Use of Software Tools
Staff’s Applications Experience
Personnel Continuity
Documentation Match to Lifecycle Needs
Multi-site Development
Required Software Reliability
Product Complexity
Personnel/Team Capability
Within the organization’s control
Within the project’s control
Largely uncontrollable
24
Organizational Level
Many factors must be addressed at the organizational level
Such as, the ConeSome work requires trading off short-
term project performance to benefit long-term organizational effectiveness
Insight #6
The Tradeoffs Between Cost, Schedule, And Quality Might Not Be What You Think They Are
26
Tradeoffs Between Quality and Cost or Schedule
Low quality is already the largest cost driver for many organizations
Attempting to tradeoff quality makes cost and schedule worse
In addition, short schedules increase defects
Focus on improving quality, and cost and schedule will follow
27
Interrelationships Between Cost and Schedule
Shorter-than-nominal schedules increase cost disproportionately
Longer-than-nominal schedules reducecost disproportionately
Larger teams tend to take more time, introduce more defects, and cost more
28
Example: Team Sizes on Similarly-Sized Business Systems Projects
05
101520253035404550
1.5-3 3-5 5-7 9-11 15-20
Team Size
Sch
edul
e (M
onth
s)
0
50
100
150
200
250
300
Effo
rt (S
taff
Mon
ths)
Schedule Effort
Insight #7
Improved Software Practices Offer Exceptionally High ROIs
30
ROI for Selected Practices
Practice 12-month ROI
36-month ROI
Formal code inspections 250% 1200%Formal design inspections 350% 1000%Cost and quality estimation tools 250% 1200%
Long-range technology planning 100% 1000%Productivity measurements 150% 600%Process assessments 150% 600%Management training 115% 550%Technical staff training 90% 500%Source: Capers Jones, Assessment and Control of Software Risks, Prentice Hall, 1994.
31
Overall ROI
Improved software practices pay an average ROI of 500% (including false starts), and continued improvement is sustainable for many years
The best organizations have sustained ROIs of 900% from software improvement initiatives for many years
32
Overall ROI
Much of the ROI comes from avoiding rework (reducing costs)
Some of the ROI comes from improving predictability (increasing business opportunities)
Bonus Insight
Some Organizations Are Doing Very Well Indeed
34
Improved Predictability in Air Force Projects
From a set of U.S. Air Force projects
Project Performance Compared to Estimate
0%
100%
200%
300%
400%
500%
600%
0 1 2 3 4
SW-CMM Level
Actu
al R
esul
ts a
s a
Perc
enta
ge o
f Es
timat
ed R
esul
ts
35
Improved Predictability at Boeing
Estimation Error
Estim
atio
n Er
ror
0%
-150%
+150%
1 2 3 4 5
CMM Level
Historical data used for all project estimates
36
Improved Predictability at Schlumberger
Week
-5Start date
(# of projects)90-1(3)
89-2(0)
89-1 (4)
90-2(3)
91-1(4)
91-2(2)
92-1(3)
0
5
10
15
20
25
30
35
37
Improved Quality at Telcordia
Software Systems Aggregate
47.3
40.2
28.5
9.27.2
3.3 2.6
0
10
20
30
40
50
60
4Q93 4Q94 4Q95 4Q96 4Q97 4Q98 3Q99
Faul
ts/K
FP
38
Reduced Cost at Telcordia
$0.50$0.51$0.51
$0.62
$0.00
$0.20
$0.40
$0.60
$0.80
$1.00
$1.20
$1.40
1993 1994 1995 1996 1997 1998 3Q99
Cos
t/LO
C T
este
d &
Shi
pped
Business Model Target
$0.67
$0.44 $0.46$0.42
$0.53
$0.67
$0.84
$1.19
To Help Achieve These Results in Your Organization:[email protected]@construx.comwww.construx.com(866) 296-6300