Evaluating Project Success

PROJECT RETROSPECTIVES: EVALUATING PROJECT SUCCESS, FAILURE, AND EVERYTHING IN BETWEEN
© 2005 University of Minnesota MIS Quarterly Executive Vol. 4 No. 3 / September 2005 361
FAILURE TO LEARN1
“Life can only be understood backwards, but it
must be lived forwards.” (Kierkegaard)2
Project managers are caught in a paradox. On the one
hand, they are working too hard to pause and reflect
on effective project management practices. On the
other hand, such reflection and corresponding learning
are key to better project management. According to
the widely-quoted Chaos Chronicles,3
only 34% of IT
1
Jeanne Ross was the accepting Senior Editor for this article.
2
The author gratefully acknowledges Barb Wixom, Mike Morris, Peter
Todd, Steve McConnell, Jerry Deville, Jeanne Ross, Barbara McNurlin,
and the reviewers for their helpful suggestions on improving this
article. My sincere thanks to the graduate students that conducted, and
the organizations that participated in, the retrospectives that made this
article possible.
3
Chaos Chronicles, Version 3.0 The Standish Group International,
West Yarmouth, MA, 2003.
Standish Group definitions:
projects undertaken by Fortune 500 companies are
completed successfully. In other words, almost two-
thirds of the 13,522 IT projects surveyed suffered
from one or more of the following:
• total failure
• cost overruns
• time overruns, or
• a rollout with fewer features or functions than
promised.
The good news is that this 2003 result is a 100+%
improvement over the 16% rate reported in 1994, and
outright failures have declined from 31% to 15%
• Successful = The project is completed on time and on
budget, with all the features and functions as originally
specified.
• Challenged = The project is completed and operational, but
over budget, over the time estimate, and/or with fewer
features and functions than initially specified.
• Failed = The project is cancelled before completion.
Executive SummaryR. Ryan Nelson
University of Virginia The fact that the majority of IT projects fail on at least one measure of success, and
that billions of dollars in project waste is reported each year, suggests that there is a
critical need for improving the way we manage these projects. The sobering truth is
that the secret to more successful project management has been right in front of us
the whole time – learning from the past.
A retrospective (a.k.a. a postmortem) is a formal method for evaluating project
performance, extracting lessons learned, and making recommendations for the
future. A comprehensive retrospective considers three process-based measures of
project success: whether it came in on schedule (time), whether it came in on budget
(cost), and whether the requirements were met (product). It also considers three
outcome-based measures of success: whether the resulting product or service was
actually used (use), whether the project helped prepare the organization for the
future (learning), and whether the project improved efficiency or effectiveness of the
client organization (value).
This article presents several retrospectives of IT projects to illustrate the importance
of evaluating project success from these multiple dimensions, as well as from
different stakeholder perspectives. Doing so can lead to some valuable lessons in the
form of “failed successes” (process success + outcome failure), and “successful
failures” (process failure + outcome success).2
PROJECT RETROSPECTIVES:
EVALUATING PROJECT SUCCESS,
FAILURE, AND EVERYTHING
IN BETWEEN
1

(challenged projects account for the remaining 51%).
However, the fact that two out of three projects fail on
at least one criterion suggests that there is a critical
need to improve what is arguably the most important
IT competency area: project management.4
Failure to learn from mistakes has consistently been a
major obstacle to improving IT project management.
As Boddie wrote in 1987:
“We talk about software engineering but reject
one of the most basic engineering practices:
identifying and learning from our mistakes.
Errors made while building one system appear
in the next one. What we need to remember is
the attention given to failures in the more
established branches of engineering. In
software projects, as in bridge building, a
successful effort can do little more than affirm
that the tools and methods used were
appropriate for the task. By the same token,
failed projects need more than explanation or
rationalization if they are to teach any
lessons.”5
RETROSPECTIVES
A retrospective is a formal method for evaluating
project performance, extracting lessons learned, and
making recommendations for the future. The word
“retrospective” means looking back on,
contemplating, or directed to the past. In the IT
industry, retrospectives go by many names. One
popular term is postmortem, from the Latin for “after
death.” In addition to the obvious negative
connotations attached to this label, IT projects don’t,
or at least aren’t supposed to, end with death. Rather,
they should bring something to life. As an alternative,
the Latin term postpartum, meaning “after birth,” is
sometimes used, again with its own set of negative
associations. Various branches of the military use their
own terms: After Action Review or Post Engagement
Redress (Army), Navy Lessons Learned or Hot Wash
Up, and C-GULL (Coast Guard Uniform Lessons
Learned). While each term has its following, the terms
4
As defined by the Project Management Institute (PMI), a project is a
temporary endeavor undertaken to create a unique product, service, or
result. As such, projects represent the central unit of work in the field of
information technology.
5
J. Boddie, “The Project Postmortem,” Computerworld, 21:49,
December 7, 1987, pp. 77-82. Engineer Henry Petroski states in To
Engineer is Human: The Role of Failure in Successful Design, (New
York: St. Martin’s Press, 1985): “I believe that the concept of failure is
central to understanding engineering, for engineering design has as its
first and foremost objective the obviation of failure. Thus the colossal
disasters that do occur are ultimately failures of design, but the lessons
learned from those disasters can do more to advance engineering
knowledge than all the successful machines and structures in the
world.”
retrospective and post-implementation review seem to
be the most descriptive, without implying success or
failure.
Another argument for using the term retrospective is
that it is not limited to the post-implementation phase
of a project. In fact, retrospectives conducted
following critical milestones in a project’s life cycle
can either confirm that the project is on track or
suggest mid-course adjustments before it’s too late. In
some cases, an interim retrospective may conclude
that a project should be terminated, avoiding the
dreaded “runaway” label.6
Why Retrospectives Are Important
Retrospectives offer a variety of potential benefits,
including the following:
• Organizational learning – Get the collective story
out (synergistic learning) and ensure that
individual stakeholders hear the whole story, not
just their personal experience.
• Continuous improvement – Facilitate
improvements in processes, procedures, and
culture.
• Better estimating and scheduling – Capture actual
data on size, effort, and time to use in calibrating
future estimation models and practices.
• Team building – Acknowledge and repair
relationship issues as appropriate.
• Improved recognition and reflection – Pause and
reflect on accomplishments before proceeding to
“solve the next problem.”
Why Retrospectives Aren’t Done
Regardless of what you call them, and despite their
obvious potential benefits, most organizations rarely
conduct formal retrospectives, outside of the military.7
The most obvious reason is the natural human desire
to put the past to rest and go on to something new.
Indeed, the IT profession seems to subscribe to the
philosophy of the famous writer Oscar Wilde, who
once stated, “The past is of no importance. The
present is of no importance. It is with the future that
we have to deal.”
6
The following reference provides an excellent discussion on the
consequences of runaway projects and recommendations for recovering
before it’s too late: C.L. Iacovou and A.S. Dexter, “Turning Around
Runaway Information Technology Projects,” California Management
Review, 46:4, Summer 2004, pp. 68-88.
7
As reported by J. Boddie (cited earlier); also reported by T. K. Abdel-
Hamid and S. E. Madnick in “The Elusive Silver Lining: How We Fail
to Learn from Software Development Failures,” Sloan Management
Review, 32:1, Fall 1990, pp. 39-48; and based on years of personal
experience.

Enterprises are also reluctant to allocate additional
time and money to a project after the system is
completed. This reluctance is particularly profound if
the project is seen as a failure: management will only
approve the retrospective if its benefits are quantified
beforehand.
Finally, most retrospectives are poorly done, which
doesn’t help overcome either the social or the
financial obstacles. As a project manager in one large
consulting firm lamented, “Our post-implementation
reviews tend to be witch hunts, where the innocent get
punished and the guilty get promoted!” In other cases,
retrospectives are seen as merely “checklist items.”
Enterprises conduct them but do not apply the lessons
learned. In these cases, it is certainly difficult to see
the value.
To address these perceived shortcomings, many large
IT services firms have developed proprietary
methodologies for conducting retrospectives. The
retrospective process has also recently received
attention in the Project Management Institute’s (PMI)
Body of Knowledge8
and the Software Engineering
Institute’s (SEI) Capability Maturity Model.9
THE RESEARCH STUDY: A META-
RETROSPECTIVE
Since the summer of 1999, the University of Virginia
has delivered a Master of Science in the Management
of Information Technology (MS MIT) degree program
in an executive format to working professionals. Over
this time, 357 working professionals have
participated; each has had an average of twelve years
of experience and direct involvement with at least one
major IT project. All 357 participants received
instruction in how to conduct effective retrospectives
(see Appendix A) and were given a framework for
assessing each of the following:
• Project context and description
• Project timeline
• Lessons learned – an evaluation of what went
right and what went wrong during the course of
the project, including recommendations for the
future.
• Evaluation of success/failure
In partial fulfillment of program requirements, the
participants have worked in teams and conducted
retrospectives of recently completed IT projects. Thus
far, 72 retrospectives have been conducted in 57
8
Project Management Institute – Information Systems Specific Interest
Group: http://www.pmi-issig.org.
9
Software Engineering Institute: http://www.sei.cmu.edu/cmmi.
different organizations. These projects have ranged
from relatively small (several hundred thousand
dollars) internally built application development
projects to very large (over $100 million) mission-
critical applications involving multiple external
providers.
When viewed individually, each retrospective tells a
unique story and provides a rich understanding of the
project management practices used within a specific
context during a specific timeframe. However, when
viewed as a whole, these 72 projects provide an
incredible opportunity to understand project
management practices at a more macro level and
generate findings that can be generalized across a
wide spectrum of applications and organizations.
For example, this “meta-retrospective” yielded some
very interesting findings on success criteria and
stakeholder perspectives for the 72 projects studied.
The next section describes both, along with selected
case studies that support key findings and
recommendations.
EVALUATING PROJECT
SUCCESS
While identifying mistakes and recommendations for
improvement seem to be straightforward when
conducting a retrospective, in most cases, evaluating
project success proves to be quite challenging.
Contrary to the methodology employed by the
Standish Group, which considers any project that is
one day late, or one dollar over budget, or one
requirement short of specifications to be unsuccessful,
our experience suggests that determining success
tends to be much more subjective and is wrought with
much ambiguity and political overtones.
To begin with, success is in the eye of the beholder.
The typical IT project may be subject to review by a
host of stakeholder groups, including the project
sponsor, system users, project team, maintenance and
support personnel, internal and external auditors, and
top management. At any point in time, a project may
receive an entirely different opinion on success, and
it’s very unlikely to be a binary one.

Based on both our experience as well as a detailed
review of the literature,10
evaluating project success
should include both process and outcome criteria, as
illustrated in Figure 1.
The three process-related criteria include:
1. Time – The project came in on schedule.
2. Cost – The project came in on budget.
3. Product – The project produced a product of
acceptable quality and met other product-related
specifications, including requirements, usability,
ease of use, modifiability, and maintainability.
The three outcome-related criteria include:
10
The following is a sampling of the most interesting work on
evaluating IT project success-failure: Lyytinen, K., and Hirschheim, R.,
“Information Systems Failures: A Survey and Classification of the
Empirical Literature,” Oxford Surveys in Information Technology, 4,
1987, pp. 257-309; Pinto, J.K., and Slevin, D.P., “Project Success:
Definitions and Measurement Techniques,” Project Management
Journal, 19:1, February 1988, p. 67; Sauer, C. Why Information
Systems Fail: A Case Study Approach, Henley-On-Thames,
Oxfordshire, 1993; Shenhar, A.J., Dvir, D., Levy, O., and Maltz, A.C.,
“Project Success: A Multidimensional Strategic Concept,” Long Range
Planning, 34:6, December 2001, p. 699.
4. Use – The project’s resulting product/service is
being used by its target constituencies.
5. Learning – The project increased stakeholder
knowledge and helped prepare the organization for
future challenges.
6. Value – The project will directly result in
improved efficiency and/or effectiveness for the
client organization(s). Common metrics include
NPV, IRR, EVA, and the balanced scorecard.
Taken together, the six criteria yield a more
comprehensive view of project success. The ultimate
goal of the project manager should be to maximize
stakeholder satisfaction on as many success criteria as
possible. To this end, the relative importance of each
criterion needs to be clearly defined and documented
at the beginning of a project, and revisited as
necessary throughout its life. For example, project
managers need to clearly communicate the inherent
tradeoffs among the three process-related criteria:
time, cost and product (the project trade-off triangle).
Increases in the product corner of the triangle require
corresponding increases in schedule and/or budget.
Likewise, completing a project in less time and/or for
less money than originally estimated usually requires
reducing product functionality.
Figure 1: Project Success Criteria
COST
TIME
PRODUCT USE
LEARNING
PROCESS OUTCOME VALUE
STAKEHOLDER
SATISFACTION

Managing Stakeholder Perspectives
The challenge of managing different stakeholder
perspectives is illustrated in 15 recently completed
retrospectives. In each, five stakeholder groups were
asked to rate the six success criteria on a 10-point
scale (where 1 = not important at all and 10 =
extremely important). Figure 2 presents the interesting
results of this survey.
First, it is obvious that the different groups were
interested in different things. Specifically, project
managers and team members seemed to be more
process-centric, focusing primarily on bringing a
project in on time and meeting specifications. In
contrast, project sponsors and top management were
most concerned about a project’s value to the
organization. Not surprisingly, users cared most about
whether or not the project’s outcome (e.g., system)
would be used by its intended clients. None of the
groups ranked learning (preparing for the future) in
their top three criteria; although all of them suggested
that learning was of at least moderate importance
(ranking 6 or higher).
The top three success criteria for all five groups were
product, use and value (in descending order of
importance). Project cost was ranked lowest overall.
In fact, based on these findings, it can be inferred that
two of the groups – users and team members – didn’t
seem to care how much a project cost, as long as it
met specifications and was used by its intended
clients.
Based on these findings, project managers need to pay
careful attention to the absolute and relative
importance of each success criterion for each
stakeholder group. Furthermore, project plans need to
clearly define measurable metrics for each success
criterion, as well as the optimal time to measure.
While process-oriented criteria can be monitored
throughout the course of a project and evaluated at
project close-out, the outcome-oriented criteria
(benefits realization) probably cannot be evaluated
until well after implementation.
In addition, project managers can use these findings to
educate top managers who are too narrowly focused
on cost and schedule. If maximizing overall
stakeholder satisfaction is important, top managers
should be focusing more on the product, use, and
value criteria.
Failed Successes and Successful
Failures
If the six success criteria were to be evaluated in a
strict, absolute method (similar to that employed by
the Standish group for time, cost, and product), the
result would be a collage of all possible combinations.
In fact, most projects would be judged to have failed
on at least one of the process criteria (approximately
60% of the projects that we studied failed on time,
cost and/or product), and significantly fewer would be
viewed as being either a total success (40%) or a total
failurev(5%).11
11
Note: 11 of the 72 retrospectives were inconclusive with respect to
the success criteria; therefore, percentages are based on a total of 61
projects. Three of the 36 projects that failed on at least one process
criterion were canceled before completion and were counted in both
categories (thus the 5% overlap).
Figure 2: Stakeholder Rankings (& Ratings) of Success Criteria for 15 projects
Stakeholder Groups
Success Project Team Top
Criteria Manager Members Users Sponsor Mgmt Overall
Time 1 (8.67) 2 (7.73) 5 (5.54) 2 (8.13) 4 (7.14) 4 (7.44)
Product 2 (8.47) 1 (8.93) 2 (8.77) 3 (7.93) 5 (7.00) 1 (8.22)
Cost 3 (8.27) 6 (4.48) 6 (3.64) 4 (7.87) 3 (7.33) 6 (6.32)
Value 4 (7.47) 5 (5.59) 3 (7.42) 1 (8.60) 1 (8.73) 3 (7.56)
Use 5 (7.39) 3 (7.08) 1 (9.58) 5 (7.38) 2 (7.43) 2 (7.77)
Learning 6 (7.17) 4 (6.52) 4 (7.11) 6 (7.18) 6 (6.00) 5 (6.80)
N = 15 for all stakeholder groups

Our experience suggests that a more meaningful
determination of project success/failure comes from
analyzing overall stakeholder satisfaction on all six
criteria. When we conducted this analysis on the 72
retrospectives, some interesting patterns arose,
including “failed successes” and “successful failures”
(see Figure 3 for examples of these two patterns).
Failed Successes. Seven projects were seen as
successful from the process perspective (i.e., they met
specifications and came in on schedule and on
budget), but were seen as failures from the outcome
perspective (i.e., they didn’t add enough net value to
the organization). Figure 4 presents a brief description
of four of these case studies. An underlying theme in
these failed successes is the importance of strategic
and business process alignment. In each case, the
successful efforts were derailed by changes in the
macro environment.
The retrospective of Financial Services A. As an
example, Project A in the financial services
organization successfully met the defined business
requirements. But when the business strategy changed,
the key assumptions underlying the application
became invalid, so the application solved a
diminishing business problem.
The retrospective of this project provides a rich
example of the types of benefits retrospectives can
provide. The primary objective of Project A was to
reengineer the financial services firm’s recruiting of
non-exempt or hourly associates to increase hiring
capacity, reduce cycle time, improve the quality of
new hires, and reduce the average cost per hire. At the
time, the firm was undergoing immense credit-card
account and customer growth. Accounts increased
from approximately 5 million in 1994 to more than 33
million in 2000. The firm needed to proportionally
increase its call center headcount to deal with this
huge increase in customer calls. Prior to adopting
Project A, management evaluated the existing
recruiting process and estimated that it would fall 40%
short of meeting the hiring demand.
Project A was a very large project for the firm,
initially carrying the label “BCP” or Big Complex
Project. The project budget was approximately $10
million and at its peak, over 25 business people and 60
IT associates and contractors were working on it.
In addition to generating a myriad of project
management-related lessons, the retrospective process
revealed fresh knowledge about developing
applications on the J2EE platform and working with
an offshore development team. Both sets of lessons
benefited the IT department moving forward.
Specifically, the project team experienced the joys and
pains of working with Navion in Shanghai, China, a
development team 12 hours ahead of them. Based on
the results of the retrospective, the firm wrote a white
paper on the quality of documentation needed to guide
offshore work and the need to verify the quality of
offshore code. This organizational learning was
invaluable because the firm initiated additional
projects using offshore resources and J2EE.
The retrospective also allowed the review team to step
back and evaluate the six success criteria. Their
analysis:
1. Schedule – A long, fuzzy front-end existed before
the commitment to do the project. But once the
schedule was established, it was met. While the
team used no formal scheduling methodology, a
modified waterfall lifecycle eventually emerged. It
was modified to include parallel track development
via overlapping “sashimi” phases, providing time
savings and a significantly compressed schedule.
2. Cost – The project slightly exceeded its original
budget, largely in increased personnel costs. The
excess was deemed to be within an “acceptable
range.”
3. Product – The project team was judged to have
done a good job of negotiating the initial project
scope, freezing requirements early, and managing
a tight change-control process throughout the
project. As a result, the application met
requirements and performed well.
4. Use – Upon implementation, the tool was used as
originally intended and did fundamentally
reengineer the recruiting process.
5. Value –Initially, Project A was deemed a success
along a number of dimensions:
• Hiring capacity increased immediately and
the target growth of approximately 74% was
nearly achieved by a corresponding 71%
increase.
• The target of increasing the hiring rate by at
least 20% was far exceeded when results
demonstrated an astounding 100% increase.
• The target goal of reducing the 90-day
attrition rate by 75% was met.
• A reduction in cost-per-hire exceeded the
target goal by 12%.
• Hiring cycle time was reduced by 37%, just
short of the desired 45% goal.
In terms of these objective value metrics
established at the onset of the project, Project A

Federal
Government
F
F
F
F
F
?
S
Consulting
F
F
F
?
S
S
S
Midwest
Bank
F
F
F
S
S
S
S
ProcessFailure-OutcomeSuccess
Financial
ServicesB
F
F
F
S
S
S
S
Financial
ServicesA
S
S
F
F
F
S
F
RealEstate
Management
S
S
S
F
F
S
F
Shipbuilder
S
S
S
F
F
S
F
CaseStudies
ProcessSuccess-OutcomeFailure
Manufacturer
S
S
S
S
S
S
F
Figure3:CaseStudiesofProjectSuccess-Failure
Success
Criteria
Time
Product
Cost
Value
Use
Learning
Overall
Stakeholder
Satisfaction
Legend:S=Success;F=Failure

was a success in the months following
implementation. However, roughly one year later,
in response to the downturn in the economy in
2001, the firm changed its hiring strategy,
dramatically. As a result, the advanced features of
the system were no longer needed. The system
was decommissioned in favor of a BPO solution.
The system had been built, and cost justified, on
the assumption that continued business growth
meant ongoing rapid headcount growth. The
possibility of slow growth had been overlooked.
Therefore, while Project A met nearly all of its
stated goals, it failed to create a flexible hiring
and recruitment platform and fell short of
achieving an acceptable return on investment.
6. Learn – At the time of the retrospective, it was
unclear whether or not the firm would make the
institutional changes necessary to reduce the
likelihood of a similar situation in the future. To
this end, the review team recommended better
strategy-project alignment, as well as re-baselining
projects when fundamental assumptions get
altered. When the hiring plans were being
adjusted, Project A could have been re-scoped,
carving out pieces of functionality, thereby
Figure 4: Case Studies of Process Success/Outcome Failure
Manufacturer
• Organization: A joint venture between two large multi-national corporations focused on
delivering factory automation solutions.
• Project: Phase II of an ERP installation and conversion; manufacturing module of SAP’s R/3
product; (1 year, $12.4M).
• Evaluation: Project was a success on all six process and outcome criteria, yet it was viewed
as a failure by the organization’s corporate holding entity that wanted to standardize
corporate-wide on the Oracle ERP platform.
Shipbuilder
• Organization: Largest private US shipyard with 17,300 employees
• Project: TRIPS: Travel Expense InPut System; automated end-to-end travel management
system; (9 months, $100,000).
• Evaluation: A successful process produced a successful product that was never used because
ERP eliminated the need for the TRIPS application.
Real Estate Management
• Organization: Owns, develops, and operates some 300 office properties in 12 markets in the
US; 780 employees; $1.53 billion in market capitalization
• Project: Strategic Lease Processing (SLP); Lotus Notes-based (COTS) customer relationship
management application; implemented to provide strategic advantage in leasing vacant space;
(2 years)
• Evaluation: A successful process produced an application that met specifications but didn’t
successfully integrate with business processes; no ROI due to lack of use. As stated by the
firm’s CTO, the application is “100% effective at 100% participation and 0% effective at 95%
participation.”
Financial Services A
• Organization: A publicly held, Fortune 200, bank holding company offering consumer-lending
products; among the largest providers of MasterCard and Visa credit cards in the world.
• Project: Redesign and automate the global call center associate recruitment processes to
increase hiring capacity, reduce cycle time, improve quality of hires, and reduce cost per hire;
(20 months).
• Evaluation: The project delivered a scalable solution that met original business requirements
and was delivered on-time without massive cost overruns; but the application was
decommissioned due to changes in corporate staffing strategies.

creating downward scalability, and long-term
business value.
From a typical project manager’s or team member’s
process-centric perspective, Project A was a very
successful IT project. Yet, from a top manager’s or
sponsor’s outcome-centric perspective, it was a
failure. Indeed, there is a lot more to project
management than meeting specifications and coming
in on time and under budget.
Successful Failures
On the other hand, a project shouldn’t be written off
as a failure just because it doesn’t navigate the
development process well. Projects that cost more
and take longer to complete, yet deliver solutions that
are used to solve business problems and add net value
to the organization, have become a classic scenario in
the world of IT project management. This point was
proven by twenty-six of the projects we studied, and is
exemplified in the four case studies described in
Figure 5.
The retrospective of Financial Services B. Project B at
the financial services organization provided a rich
example of successful failures. It involved the
mission-critical credit card collections system whose
primary objective was to prevent losses and cure
accounts that either go over limit or become
delinquent by restoring the accounts to under limit
and/or to their current status.
Figure 5: Case Studies of Process Failure/Outcome Success
Financial Services B
• Organization: (same as Financial Services A)
• Project: Internal development of a system to enable rapid development, testing, deployment,
and measurement of collections strategies, to continually improve collections performance; (20
months, $5.7M).
• Evaluation: A change in project manager when the project reached “red status” resulted in a
product with a 13-month ROI and a more adaptive organization, despite being six months late
and costing more than twice the original estimate.
Midwest Bank
• Organization: One of the 20 largest banks in the U.S.; FY99 assets ~ $82B
• Project: Develop an automated system for efficient retrieval and storage of images, charge
card vouchers, and documentation and processing of the associated workflow; (Fixed-price
contract with large service provider; 2 years).
• Evaluation: Stakeholders were satisfied with a working product that was used by its intended
clients despite experiencing 75% of the common mistakes outlined in Figure 2.
Consulting
• Organization: A 200-person IT consulting firm with 95% of its revenue coming from contracts
with the federal government.
• Project: Redesign the Federal Transit Administration’s Web site; (17 months, $270,000).
• Evaluation: Poor estimation and ineffective scope management resulted in a $100,000 loss on
a fixed price contract; yet the project was considered a “loss leader” important for follow-on
business.
Federal Government
• Organization: Army Transportation Agency; manages surface transportation of military
equipment and supplies worldwide.
• Project: COTS-based global distribution system with predictive event management, using
large systems integrator; (estimated: 9 years, $100M).
• Evaluation: This runaway project was killed after the first year of this nine-year project;
successful de-escalation was achieved after changing the project manager and a key project
sponsor.

In January 2002, the firm embarked on an effort to
restructure its business processes and underlying IT
systems for collections and pre-collections strategies.
The goal of Project B was to “Enable rapid
development, testing, deployment, and measurement
of collections strategies to continually improve
collections performance.”
The project was justified by regulatory compliance
and financial return. The pre-existing collections
business processes and systems had grown
increasingly complex to manage; they could not
support concurrent implementation of tests nor
accurately track returns of new collections strategies.
Furthermore, to adhere to corporate controls and
governance, the firm needed to better trace its test
results and returns. The business case for rapid
strategy testing and implementation was an after-tax
3-year NPV of $6.7 million. The present value of that
expected cost savings was $12.4 million; the required
investment was $5.7 million.
The retrospective team found that the project
encountered a number of issues early on. For one,
although the Sr. Vice President of Financial Solutions
(FS) originally requested and sponsored the project,
no business executive was initially responsible for
project delivery. The IT staff formulated the project
plan, scope, and deliverables without significant
business involvement. As would be expected, the
result was several mismatches between the business
leaders and the IT staff. After two months of project
“wandering,” the senior vice president of FS assigned
a vice president in the sponsoring business unit to be
accountable for delivering the expected results.
This executive set project delivery for late 2002 with a
budget of $2.6 million – before requirements
definition and project estimation. It soon became
clear that these were unrealistic because:
• The scope was larger than originally conceived
• Little business process documentation existed
• Little systems documentation existed
• The requirements kept changing
Thus, the project team had to reverse engineer and
document the existing business processes and IT
systems, which proved to be a large archeological
effort.
During the first nine months of the project, the IT
project management staff also made some key
mistakes. First, they did not communicate or publish
to executives the formal project tracking of key
milestones, risks, requirements and scope changes.
The result: unmanaged business executive
expectations, scope creep, and cost escalation.
Second, the project management team micro-managed
the project and blamed the technical team for missed
milestones, rather than setting clear, realistic goals for
the team. The result: slower progress and team
apathy, low moral, and mistrust.
In October 2002, it became clear that the project
would not be delivered on schedule, so it was given a
“red” status.
A new program director took control in December, re-
baselined the effort and project plan, and established a
formal Program Management Office (PMO) team and
a comprehensive project management program for this
project. This program established formal work tracks,
each with a designated lead, milestone tracking,
communications plan, change management plan, and
risk management plan with mitigation strategies. In
addition, project delivery was revised to deliver in
third quarter of 2003, with a 50% increase in budget.
The business executives immediately challenged these
revisions because they wanted delivery in March
2003. By drawing on formal estimation results and
project tracking metrics, the IT project manager was
able to convince the business leaders that a June
delivery might be possible, but the costs would be
significantly higher. The IT project manager then
tracked the project metrics and frequently
communicated progress to the executives, which
improved executive visibility and understanding. At
the May 2003 executive meeting, the IT project
manager used risk metrics and a formal risk profile to
tell the executives that delivery would slip to
September.
Under this new project management, team morale
improved dramatically, collaboration increased, and
senior management gained confidence, all because of
the project management discipline, the clear goals,
and the commitments being analytically driven.
Project B went live in August and was formally closed
in September 2003. It cost $5.7 million, more than
twice the original estimate.
The key lessons from the retrospective included:
• Define clear scope and requirements before
establishing a delivery date.
• Conduct thorough and authoritative systems
impact analyses.
• Assign a senior-level business executive to the
project at the outset.
• Use a formal project management methodology
with clearly defined metrics, risk management,
and communication plans.

At the time of the retrospective, key business metrics
were still being tracked to determine the expected
return. But the initial results showed:
• $33 million reduction in charge-off accounts.
• Reduced time-to-value and increased capacity
resulted in a 50% increase in number of
concurrent collection strategy tests in production.
Project B was a classic case of an IT project that failed
from a process perspective (and seen as a “runaway
project”), yet it was turned around and yielded a
successful outcome from an overall stakeholder
perspective. As one retrospective participant stated,
“The single biggest factor in [this project’s] success
was bringing in senior, experienced managers who
had the dedication and focus to get the job done!”
Indeed, the lessons learned led to managing projects in
new ways, including policies requiring business
ownership and procedures for handling “red status”
projects.
As these case studies demonstrate, evaluating project
success can be challenging. While all relevant
stakeholders may occasionally consider a project to be
a complete success (or failure), most IT projects will
receive a mixed report card. These case studies also
demonstrate the valuable learning from a
retrospective, regardless of project success.
CONDUCTING A RETROSPECTIVE
Based on our six years of experience, as well as the
opinions of professional facilitators,12
retrospectives
should be conducted after completing each major
milestone in large projects, and one to two weeks after
implementation. Benefits realization should be
conducted later, when business value metrics can
legitimately be assessed.
Retrospectives should be scaled and located based on
the size and nature of the project. Thorough discussion
and thoughtful analysis can usually be completed in
one to three days. Location options include on-site,
off-site local, and off-site residential; each has its
advantages and disadvantages.
The two most important retrospective decisions are
the choice of participants and facilitator. All major
stakeholder groups need to be represented, but all the
participants do not need to be in all the sessions. For
example, it is often useful to hold anonymous
feedback sessions with only the project team members
present. The ideal facilitator is someone who was not
a member of the project team, has subject matter
12
See the following reference for a detailed review of how to conduct a
project retrospective: N. L. Kerth, Project Retrospectives: A Handbook
for Team Reviews, (New York: Dorset House Publishing, 2001).
expertise (e.g., systems development experience), and
is a superb group facilitator (e.g., with the ability to
create an open and constructive atmosphere).
As detailed in Appendix A, retrospectives should
begin with a complete description of the project,
including its context, the methodology and technical
approaches taken, and an analysis of the various
project stakeholders – that is, their power (i.e.,
influence over others and direct control of resources),
level of interest, and degree of support/resistance. The
project timeline provides a temporal explanation of
project events including significant milestones,
deliverables, and momentum fluctuations.
A lessons-learned section carefully evaluates what is
going wrong (or went wrong) with the project and
makes recommendations for improvement. A
thorough retrospective also includes a root-cause
analysis that uncovers the observed failure symptoms
and when they were observed. Examples of failure
symptoms in the two financial services cases included
lack of strategic alignment (Project A) and lack of
business ownership (Project B). Take-a-ways such as
these can quickly improve upcoming projects.
As previously discussed, evaluating a project’s
success criteria typically proves to be the most
difficult, yet potentially valuable, component of a
retrospective. A thoughtful analysis evaluates all the
criteria (time, cost, product, use, learning, and value),
determines which combination of process-outcome
criteria best describes the project, and suggests the
lessons learned for each stakeholder group.
CONCLUSION
Given the critical role that project management plays
in the field of information technology, we need to
accelerate our progress on this typically slow and
painful learning process. To this end, project
retrospectives need to evolve beyond simple checklists
of what went right and wrong to become more
analytic, as exemplified by the case examples in this
article. Managers need to recognize that virtually
every project experiences some successes and some
failures. Yet, regardless of the level of success or
failure, every project should contribute to
organizational learning and continuous improvement.
Based on the findings of this research, stakeholder
perspectives on the several dimensions of success
need to be considered, carefully documented at the
beginning of a project, and revisited periodically
throughout the project, including at least one post-
implementation retrospective. Project managers need
to become adept at negotiating the corners of “the

tradeoff triangle” – time, cost and product – and the
impact of those tradeoffs on the three outcome criteria
– use, learning, and business value. The project
charter should include the negotiated success metrics,
the project dashboard should enable real-time
monitoring of the metrics, and the project
retrospective should document the actual results,
concluding with overall stakeholder satisfaction.
There is a lot to be gained by looking in the rear view
mirror from time to time. As George Santayana,
philosopher, poet, and cultural critic once said, “Those
who cannot remember the past are condemned to
repeat it.” We do this all too often in IT, as evidenced
by the Chaos Report of the Standish Group, which
estimates annual project waste at over $55 billion. If
the retrospective process could reduce this waste by
only ten percent, perhaps project managers could
afford to pause and reflect more often!
ABOUT THE AUTHOR
R. Ryan Nelson (rnelson@virginia.edu) is a
Professor, Director of the Center for the Management
of Information Technology (CMIT), and Director of
the MS in the Management of Information
Technology program at the McIntire School of
Commerce of the University of Virginia. He received
his Ph.D. in business administration from the
University of Georgia in 1985 and spent five years on
the faculty of the University of Houston before joining
UVa. His research has been published in such journals
as the MIS Quarterly, Communications of the ACM,
Journal of Management Information Systems,
Information & Management, International Information
Systems, and Data Base. He currently serves on the
editorial board of the MIS Quarterly Executive.
APPENDIX A. IT PROJECT
RETROSPECTIVE TEMPLATE
• Organization Name and Description
• Project Name and Description
• Why was the project undertaken in the first
place? Initial vision? Business case?
• What external competitive or environmental
factors were significant and why?
• What business metrics were supposed to go up
or down as a result?
• What were the institutional dynamics regarding
governance and funding?
• Organization map and stakeholder analysis
• Type of project (e.g., system, business
application, shrink-wrapped product)
• Business criticality (e.g., strategic,
infrastructure, maintenance, etc.)
• Basic approach (e.g., internal development,
COTS)
• Lifecycle approach (e.g., waterfall, prototyping,
code and fix, spiral)
• Methodology(ies) employed (e.g., object-
oriented, agile)
• Technology (e.g., architecture, hardware,
software, development tools)
• Size (e.g., lines of code/function points,
effort/team size, budget)
• Project Timeline
• How do the project’s phases, main facts, critical
events and inflection points unfold when laid
out chronologically? What happened and in
what sequence? What is the factual storyboard?
• What would a trend line of the project’s
momentum look like over time (+/-)?
• Lessons Learned
• Common mistakes checklist
• Symptoms of failure (lack of strategic
alignment, lack of stakeholder involvement,
poor planning, poor execution) and root-cause
analysis
• Recommendations for the future
• Evaluation of Success/Failure (time, cost, product,
use, learning, value)
• Retrospective Participants and Authors
• Keywords; Search Tags
• Appendix – Project Artifacts
• Project charter, business case, and/or white
paper
• Organizational map (including program office,
project team, stakeholders)
• Project history, log documents or issues
management summaries
• Project status meeting summaries
• Size, effort, schedule, budget estimates (and
relevant assumptions)
• Size, effort, schedule, budget actuals
• Architectural schematics
• System design documents
• Release plans (what functionality was scheduled
in R1, R2, etc)
• Conversion and implementation plans
• Project completion documents
• Audited business case results
• Interim retrospective(s), lessons learned, or
other documentation

Evaluating Project Success

More Related Content

What's hot (20)

Similar to Evaluating Project Success (20)

More from Baker Khader Abdallah, PMP (20)

Recently uploaded (20)

Evaluating Project Success