Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Using the Troubled Project Recovery Framework: Problem Recognition and Decision toRecoverAuthor(s): Douglas Havelka and T. M. RajkumarReviewed work(s):Source: e-Service Journal, Vol. 5, No. 1 (Fall 2006), pp. 43-73Published by: Indiana University PressStable URL: http://www.jstor.org/stable/10.2979/ESJ.2006.5.1.43 .Accessed: 18/03/2012 21:45
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].
Indiana University Press is collaborating with JSTOR to digitize, preserve and extend access to e-ServiceJournal.
http://www.jstor.org
43@ 2006 e-Service Journal. All rights reserved. No copies of this work may be distributed in print or electronically without express written permission from Indiana University Press.
Using the Troubled Project Recovery Framework:
Problem Recognition and Decision to Recover�
Douglas Havelka
Miami University
T.M. Rajkumar
Miami University
ABSTRACT
Information system development projects continue to be delivered significantly behind sched-
ule, drastically over budget and without meeting specifications. To improve the likelihood of
success for these projects, we propose a staged framework for recovery and rehabilitation.
This framework is composed of four stages that include twelve steps. The four stages are
recognition, immediate recovery, sustained recovery and maturity. We performed a field
study to validate the first stage of the framework, the Recognition Stage. The results of the
study indicate support for the steps of the Recognition Stage and provide detailed “symptoms”
that could indicate trouble during an information system development project. These
symptoms were logically categorized into eleven areas of interest to project managers: Client/
Stakeholder, Goal, Meeting, Team, Task, Project, Project Management, Communication,
Management, Project, and Process.
Keywords: Projects, Project Management, Troubled Projects, Information Systems Development
INTRODUCTION
More than $255 billion is spent every year in the US on information systems development
projects (Standish Group, 2004). However, most information system (IS) development
�. Earlier versions of this paper were presented at the 2004 Americas Conference on Information Systems
and the First International Workshop on IT Project Management, Milwaukee, Wisconsin, USA, 2006.
e-Service Journal
44
projects are not successful in at least one of the key measures: over budget, over schedule,
or do not meet specifications. Some research has shown that only 34% of software devel-
opment meets the three success criteria and as many as 66% of software development
projects are “troubled.” This leads to an annual waste of nearly 50 billion dollars in IS
development projects, $38 billion in lost value and $�7 billion in cost overruns (CHAOS
Chronicles, 2004). Referring to the Standish Group’s Chaos report of 2005, Balestreto,
the CEO of the Project Management Institute, stated that nearly 56 percent of projects
saw cost overruns of more than 50 percent and 84 percent of projects had schedule over-
runs (Cost overrun? Call in a project manager, 2006).
Nearly half of all projects were reported to be challenged, i.e., they were over budget,
behind schedule, and the end product did not meet user requirements. A definition for
projects that are even more deeply troubled has also been suggested, i.e., a runaway project
is one where the schedule, cost, or functionality is twice as bad as what was sought (Glass,
2002). Additionally, projects that are canceled prior to completion, i.e., impaired projects,
make up about �6% of all projects. For the purpose of this study and as suggested by prior
research, the following measures are used as indicative of trouble: (�) being over schedule
by 30%, (2) being over budget by more than 30%, and (3) having an end product that
does not meet user requirements (Whittaker, �999). Regardless of the label used for trou-
bled projects or the rule for classifying a troubled project, it is clear that schedule, cost, and
functionality goals are not adequately achieved by IS project managers.
Some of the confusion regarding troubled projects is attributable to the varying defi-
nitions of project success (Belassi and Tuckel, �996; Klein and Jiang, 200�; Linberg, �999).
Different stakeholders within the project may view success and failure differently (Pintro
and Slevin, �988). IS professionals have deemed systems as successful even when a system
is 4�7% over cost and �93 % behind schedule (Linberg, �999), because their notion of suc-
cess was based upon the knowledge gained by the innovative nature of the solution and the
relationships built within the team. Such views, while deviant from normal perceptions of
project management success, indicate how different groups view success criteria differ-
ently. Other factors include the fact that IS professionals might view success from a narrow,
information technology perspective and miss the overall business and organizational envi-
ronment within which the software solution is being deployed (Day, 2000).
In addition, it appears that businesses do not seem to learn from past mistakes in
software development. They repeat the same mistakes over and over again and are habit-
ually in trouble (Tarek and Madnick, �990). A reason for this may be the fact that a ma-
jority of software development improvement is done at the project management level
rather than at the organizational level (McConnell, 2002). In a study that measured
project management maturity across different industries, including engineering con-
struction, information management and movement (telecommunication), information
systems development, and high tech products and services, information systems compa-
Using the Troubled Project Recovery Framework
45
nies ranked lowest in maturity, achieving only a mean of 3.07 on a scale of � to 5 (Ibbs
and Kwak, 2000). The Software Engineering Institute has developed a model called the
Capability Maturity Model which measures the software development capabilities of
companies (Jalote, �999). Many U.S. companies rank at Level � (initial level) or 2 (re-
peatable level), which are the lowest levels. This is indicative of the lack of standard,
consistent, and predictable project management processes.
To be considered a “recovery,” a project headed toward disaster or cancellation must
at least be able to overcome the initial problem situation and achieve at least average re-
sults (Caper Jones, �995a, �995b). Recovery may not always be possible. Early discovery
of the problem should lead to improved likelihood of success or recovery. It is proposed
that if trouble in projects is not discovered or is discovered and left unattended, the proj-
ect will have a higher likelihood of failure. We contend that in the software industry
problems are seldom detected early enough. Based on this belief we attempt to address
the following research question:
How can organizations identify trouble in an IS development project as early as
possible?
To address this question, we conducted a field study using a nominal group process
to identify symptoms of trouble in IS development projects. The rationale for the field
study is based on the Troubled Project Recovery Framework (Aiyer, Rajkumar, and
Havelka, 2005).
Different perspectives of troubled projects and recovery processes exist. Several dis-
tinct areas of study can inform our efforts to understand the issues confronted by manag-
ers when dealing with the identification and correction of problems as early as possible.
Research and results from the areas of software project management, de-escalation of
project commitment, and crisis management exist. All of these areas contribute to our
understanding of how to improve troubled projects and are the foundations for the Trou-
bled Project Recovery Framework.
Software Project Management
It has been proposed that both long-term and short-term recovery methods are important
(Connell, 2000) for IS project management success. The Capability Maturity Model and
ISO 900� that help all projects across the organization are examples of long-term methods.
These approaches have been found to be effective, but may be inappropriate or difficult to
implement for an ongoing troubled project. For a specific project in trouble, short-term re-
covery methods are needed. Connell suggests the following steps: an initial assessment is
used to identify the areas of trouble in the project and a plan is then developed; the purpose
and final objectives and goals of the project should be reviewed and agreed upon by all
e-Service Journal
46
major stakeholders; modest, stable, and achievable goals for imminent milestones should
be established; realistic schedules are set; the project is tracked with source control; auto-
matic builds are in place; and efficient processes are specified to track and fix bugs. Con-
nell’s method seems to place an emphasis on the programming as opposed to the overall
life cycle of the project. The Troubled Project Recovery Framework adds detailed steps
that project managers or sponsors could follow to these overall quality assurance and man-
agement approaches. Additionally, the Troubled Project Recovery Framework addresses
management and personnel issues as well as the coding-intensive focus of these models.
Lastly, the results of our framework and this study provide managers using CMM or
ISO900� types of programs additional detailed areas for analysis and review of projects.
De-escalation of Commitment
The decision to add resources to a failing or troubled project is referred to as the escala-
tion of commitment. Through de-escalation, troubled projects may be successfully
turned around or sensibly abandoned (Keil and Robey, �999). Some research has looked
at recovery processes from the de-escalation perspective and this research identifies the
following steps as necessary: problem recognition, reexamination, selecting alternative
courses of action, and implementing an exit strategy (Keil and Montealgre, 2000). By
following these steps, organizations should do a better job of supporting projects that are
more likely to succeed and ending projects that have a strong chance of failing. Alternate
scenarios to kill projects have been presented (Kapur, 200�). Kapur’s approach includes:
check for vital signs of the troubled project; meet with decision makers to obtain key
stakeholder and senior management buy-in before a decision to kill or recover the project;
if the decision is to recover, then the manager must create both a recovery plan outlining
the steps needed to bring the vital signs back to acceptable levels and a mechanism to
track project recovery. If the decision is to kill the project, the aim is to develop a cancel-
lation plan, obtain stakeholder approval to the plan, kill the project while causing the
least amount of harm and salvaging usable project components, and conduct a post mor-
tem analysis including team members and stakeholders to glean lessons from the project
so that the mistakes are not repeated again. In general, the de-escalation literature focuses
on how to cancel a project when appropriate and cut losses to the organization. Recover-
ing and rehabilitating the project is only a secondary focus and is not as well developed.
The Troubled Project Recovery Framework (TPRF) overlaps with some of the first steps
in these approaches, but provides more detailed procedures for identifying trouble and
taking immediate steps to correct, if that is the decision made. The TPRF focuses less on
how the decision to recover is made and more on how to achieve recovery. The results of
this study provide managers with more detailed information to identify and manage
trouble on IS projects earlier than they previously might have.
Using the Troubled Project Recovery Framework
47
Crisis Management
Another area of related research is crisis management, which is the process of assessment,
response, mitigation, and relief of crisis situations (Seeger and Ulmer, 200�). Crisis man-
agement is defined as the immediate response to such events, recovery efforts, and miti-
gation and preparedness efforts to reduce the impact of future crises (Summary of a
Workshop on Information Technology Research for Crisis Management, �998). Immediate
response is dedicated to the immediate protection of life and property. It requires urgent
action and the use of coordinated resources and facilities not available to routine prob-
lems. Recovery efforts encompass both short-term activity intended to return vital sup-
port systems of the project to operation and longer-term activities designed to return
infrastructure systems to pre-disaster conditions.
These activities are directly analogous to the recognition and recovery stages of the
Troubled Project Recovery Framework (Aiyer et al., 2005). In troubled software projects,
immediate response might involve recognizing the problem and communicating to the
stakeholders the nature and magnitude of the problem. Recovery efforts include assess-
ing the problems and taking quick corrective action to restore the health of the project,
such as changing team composition when one member of the team creates conflicts
within the team (Brown, Malveau, McCormick III, and Mowbray, �998). Recovery also
includes long-term plans such as reorganizing and implementing a new base plan.
Mitigation includes steps such as documenting the problems, preparing measures
for corrective actions, and preparing specific actions to address each problem. Prepared-
ness is the process of risk management: identifying, analyzing, and elaborating the symp-
toms and creating contingency plans to mitigate the effects of the risk. For IS projects,
documenting the problems identified and the success or failure of corrective actions can
provide lessons learned for these processes and can be used to proactively prepare risk
management procedures to be used in the current and future projects.
Some research has directly focused on crises in IS projects (Boundy and Diamond,
�998). Again, a set of steps is suggested for dealing with the crisis: recognizing and ac-
knowledging the crisis including discussion of the crisis and consideration of solutions to
the problem; identifying the magnitude of the problem and communicating it to the rel-
evant stakeholders; ensuring that the problem is being handled by the right people within
the team including escalating the problem to higher levels; and applying additional re-
sources (such as a consultant) if needed and appropriate; continuing work on other parts
of the project for which people can be spared while keeping the resolution of the crisis as
priority number one; and ensuring that management and stakeholders are updated daily
with regard to the crisis.
The prior work on crisis management, especially in the IS area, is directly relevant to
the recognition and recovery stages of the TPRF. The TPRF recognizes the fact that there
e-Service Journal
48
are times when trouble, or crises, in IS projects is not readily apparent and that it would be
useful to have early warning signs or indicators that trouble is occurring. Once the trouble
is recognized, prescriptions from the crisis management literature could be used to help
address and mitigate the trouble. Overall, the TPRF spans a greater set of activities and
provides a bridge between the more focused crisis management and project de-escalation
literature to the broader software management literature. As can be seen throughout the
literature, a logical, structured approach with specified activities and tasks is recommended
to help recover (or kill) troubled projects. The Troubled Project Recovery Framework pro-
vides just such an approach for use in IS project management (Aiyer et al., 2005).
THE TROUBLED PROJECT RECOVERY FRAMEWORK (TPRF)
The Troubled Project Recovery Framework is composed of four main stages: (�) problem
recognition, (2) immediate recovery, (3) sustained recovery, and (4) maturity. This
framework was built with prior research as a foundation and is similar to the crisis man-
agement approach which has four main stages: immediate response, recovery, mitigation
and preparedness; and the four stages of de-escalation: recognition, reexamination, se-
lecting course of action, and exit strategy (Keil and Montealegre, 2000). The conceptu-
alization of the four-stage �2-step recovery framework is shown in Figure �. The four
main stages are each composed of two to four steps. A short discussion of the four stages
and the twelve steps follows.
Stage 1: Problem Recognition and Decision to Recover
Before any project can be recovered, the fact that the project is in trouble must be recog-
nized. A decision to recover is made and the recovery process begins. The steps of the first
Figure 1. Troubled Project Recovery Framework
Using the Troubled Project Recovery Framework
49
stage include: (�) problem awareness, (2) admission of trouble and seeking help, (3) quick
assessment and review, and (4) decision to recover.
Step � is problem awareness. When projects are in trouble, it is expected they would
display symptoms of the trouble, similar to a patient displaying symptoms of a disease.
Many times some type of trigger or other significant event occurs that demands a re-
sponse or action. This may be in the form of an event, (e.g., a subcontractor quitting the
project) or in the form of a state (red flag) that needs attention (e.g., a missed milestone or
deliverable) that indicates things are not operating as they should. In addition, many or-
ganizations do not use good estimation tools or techniques and lack any kind of histori-
cal measures or metrics to make comparisons during project management (Caper Jones,
�995b). Therefore, projects may get into serious trouble without knowledge of the proj-
ect manager or other stakeholders until the projects are very late in development. These
symptoms may be identifiable and monitored to help project managers recognize and
acknowledge trouble earlier in a project’s life and thereby increase the likelihood of suc-
cessful recovery.
Step 2 is admitting that trouble exists and getting help. According to a survey by
the Center For Project Management (Kapur, 200�), many project managers are afraid of
being labeled as quitters or failures and only 20% have a process for identifying troubled
projects. This suggests that project managers may resist or ignore symptoms until it is too
late for recovery. Denial is an unconscious coping mechanism to block out and avoid
major changes that may have some pain associated with them (Kiechel, �993). Project
managers recognize trouble, but deny they are in trouble (Brady and DeMarco, �994)
and keep silent without reporting the bad news (Smith, Keil, and Depledge, 200�). Proj-
ect managers tend to hold on to a problem too long and do not ask for help, thinking it
makes them look bad (Boundy and Diamond, �998). Keil and Robey (200�) also iden-
tify a deaf effect, i.e., even when the trouble is reported, the manager may fail to hear it.
By ignoring the trouble report, managers may insulate themselves from dealing with the
problem.
Step 3 is to conduct a quick and honest assessment and review of the project status.
This assessment must be a quick review of the project execution and its true status. The
aim is to create as complete and accurate an assessment of the project as possible in a
short time frame. A check for the vital signs of the project must be conducted (Kapur,
200�). This includes a check of planned versus actual achievement on the schedule, re-
sources used, milestones met, and deliverables met. Also a comparison of estimated ver-
sus actual cost of the project to date must be made.
Step 4 is the decision to recover the project or to kill it. Once an assessment is
made, a decision as to whether the project should be recovered or canceled must then be
made. Once the decision to recover is made in consultation with upper management and
stakeholders, an internal commitment to recover must be made. Not only must a deci-
e-Service Journal
50
sion to recover be made, but one must truly believe that project recovery is possible. Buy-
in to the commitment must come from both stakeholders and managers.
Stage 2: Immediate Recovery Stage
In this stage, steps are taken to nurture the troubled project back to stability. The aim in
immediate recovery is to identify and isolate the critical problems and take immediate
corrective or mitigating action. The aim is not to revamp the complete plan. The two
steps in this stage are triage and treatment. These steps may occur in sequence or concur-
rently and iteratively within the recovery process. In the triage step, similar to triage in
major accidents or natural disasters, the most critical problems are identified, isolated,
and prioritized. The major problem may be identified in Stage � and additional problems
or issues may have been identified in the assessment step of Stage �. For example, a trigger
may have been the loss of the project manager and the assessment revealed several missed
milestone deliverables and cost overruns. The issues should be prioritized so that the
treatments in the next step occur appropriately. Closely linked to the triage step is the
treatment step. As the critical problems are identified and prioritized, a treatment action
is determined for each critical problem. For example, the sponsor may decide to forgo an
imminent milestone until further analysis can be performed. The aim is to stabilize the
project and bring the project to a stage from which a full recovery can be made.
Stage 3: Sustained Recovery
The long-term or sustained recovery process can be put in motion once the project is in a
stable stage. The steps include analyzing the project status and creating an issue list with
resolutions for addressing the issues, creating a revised baseline plan and obtaining man-
agement approval, followed by executing and monitoring the revised plan. Building on
the prior steps, there should be an initial inventory of issues that need to be analyzed and
appropriate action planned. By incorporating resolutions into the original project plan
and making further adjustments as necessary, a new baseline plan (revised master plan
for the project) is now created. This might involve a reduction in scope, reprioritization of
goals and objectives, and clarification of the objectives and expectations. A re-estimation
of tasks will be performed, a new schedule will be developed, and resources allocated to
the project accordingly. In addition, the mix of personnel and skills inventory present
should be compared to the project requirements and appropriate actions planned, such as
training in new technologies. The risks involved in the project should be re-evaluated
and plans for mitigating the risk must be established and evaluated. In essence, a com-
plete new baseline plan is drawn up. The revised plan and milestones become the metric
against which success for the project is now measured. The revised baseline must be put
through the normal business value approval process. Management and customer ap-
Using the Troubled Project Recovery Framework
5�
proval for the new baseline must be obtained before proceeding further, i.e., a go-no go
decision must be made. After the revised baseline is established and approved, the plan
should be implemented and the status of the project continually monitored. Implement-
ing the plan may include adding or redeploying resources. This may be human resources
that may involve re-assigning, refocusing, or removing personnel on the team. Other
project resources, such as software development tools or equipment, may need to be pur-
chased or changed in some manner.
Stage 4: Maturity Stage
The final stage is the maturity stage. This stage focuses on the on-going management of
projects in general and is critical for increasing the return on investment in recovered proj-
ects. Post implementation reviews of projects to learn from successful and fruitless actions
are not standard practice (Brady and DeMarco, �994). Research indicates that failures
continue to recur in organizations for various reasons (Fowler and Walsh, �999). Organi-
zational learning should help the individual project manager accomplish three goals: de-
liver a successful project, deliver a series of successful projects, and build project management
capabilities (Kotnour, �999). The detailed steps (or activities) in this stage are: document-
ing the lessons learned from each project, propagating project management knowledge
throughout the organization, and implementing this knowledge in other projects.
Comparison of Framework with Other Existing Frameworks
Other recent research discusses project management using early warnings to help recover
projects (Nikander and Eloranta, 200�). Early warnings reflect probable problems, lead-
ing the decision-maker to identify the underlying root cause of the problem. A response
to the problem follows using standard risk management techniques. These approaches
focus primarily on our problem recognition and decision to recover stage.
The State of Michigan’s project management resource center and others discuss
steps to assess and recover troubled projects (“Assessment, recovery, transition,” 2007;
Sifri, 2003a, 2003b, 2004). Both frameworks appear similar and contain two major
phases: assessment and recovery. The project management life cycle is applied in each of
the phases. Developing an assessment charter, assessment plan and conducting assess-
ment activities are the steps in the assessment phase. The recovery phase includes devel-
oping the recovery plan and implementing the recovery. The activities outlined in the
assessment phase are similar to the actions needed to be done in step 3 and 4 of the
TPRF, and is compatible with the TPRF. TPRF explicitly recognizes the need to per-
form both a short term recovery and a sustained recovery. In addition, TPRF includes
both a problem recognition stage and a maturity stage. The maturity stage emphasizes
the need for organizational learning which is consistent with the project management
e-Service Journal
52
book of knowledge (PMBOK). Organizational learning is of paramount importance to a
comprehensive recovery strategy, since not learning from mistakes dooms us to repeat
them (Iacovou and Dexter, 2005).
THE FIELD STUDY
To assist project managers in applying the TPRF a field study was performed to identify
common symptoms that might cause or contribute to trouble during information sys-
tems development projects. The purpose of the study is to identify and categorize these
symptoms to allow managers to achieve the first stage of the proposed framework, i.e.
recognize and acknowledge that trouble exists. This, in turn, should lead to the later
stages where the set of symptoms could be used to improve, measure, or take corrective
action on projects that are in trouble or may be heading for trouble. These results should
be of interest to business managers and IS project managers dealing with the project
management process as well as to information systems researchers attempting to better
understand the IS development process.
Research Method
The field study consisted of a series of focus groups undergoing a nominal group process
(Delbecq, Van de Ven, and Gustafson, �982). An underlying assumption of the method
is that individuals who perform a task can provide valuable insight into the important
factors influencing their ability to achieve a high level of productivity when performing
the task. This method has been used successfully in several domains including systems
development (Havelka, 2002; Havelka, Arnold, and Sutton, �998; Havelka, Sutton, and
Arnold, 200�).
In the current study, this method was used to identify symptoms of trouble on IS
development projects. The nominal group technique used in the focus groups is a modi-
fication of procedures developed by prior research in the social sciences (Delbecq et al.,
�982). The technique consisted of five steps. All five steps were conducted during a single
meeting (of each group) that took approximately two hours. The technique is discussed
in the next paragraphs and summarized in Table �. Detailed instructions for conducting
the nominal group technique are available upon request.
The first step in the nominal group technique was a general discussion of the subject
being addressed by the group. In this case, the subject chosen for study was the identifica-
tion of symptoms that might indicate trouble on an IS development project. The objective
of the first step was to provide a common understanding of the subject and to determine
the scope of the topic under study. A general definition was given of the IS development
process, the tasks performed during the process, and the outputs of the process. The par-
ticipants were then given the opportunity to clarify their understanding of the subject.
Using the Troubled Project Recovery Framework
53
After any questions were raised and discussed, group consensus was used to define the
process and the boundaries of the problem to be addressed in the subsequent steps.
The second step in the nominal group technique was a period of silent generation of
ideas. This was achieved by presenting a nominal question to be addressed by the group
and asking them to record their thoughts individually and silently on worksheets. The
objective of this step was to generate as many ideas as possible for later discussion. The
participants were encouraged to be creative and to interpret the question based on their
own experiences and knowledge. The question posed to the groups in this step was:
What symptoms can you identify that might indicate that an information system
development project is in trouble, i.e., it is over-budget, behind schedule, or not
meeting specifications?
The benefits of this silent, independent idea generation include: adequate time for
thinking and reflection, social facilitation (i.e., constructive tension by observing other
participants working), avoidance of interruptions, avoidance of prematurely focusing on
particular ideas, sufficient time for search and recall, avoidance of competition, avoidance
of status pressures, avoidance of conformity pressures, and avoidance of choosing between
ideas prematurely (Delbecq et al., �982). The question being asked should be general
Table 1. Nominal group technique used
Step �: The facilitator made general introductions of the participants, explained the purpose of the study and the meeting, presented definitions to be used by the subjects, explained the nominal group process to be performed, and introduced the question to be answered by the subjects. A definition of troubled projects was given and the research question was presented to the group.
Step 2: Each subject was then asked to silently generate as many of symptoms as possible.
Step 3: After �5 minutes, the facilitator began to write the symptoms on a white board or flip charts for all participants to view. The symptoms were elicited from the participants in a round-robin fashion until all the participants’ symptoms had been listed. Only questions related to clarifying the symptoms being listed were allowed at this point and no discussion of the merits or importance of the symptoms was allowed. Participants were encouraged to add to their lists as this step progressed.
Step 4: After all of the symptoms were listed, discussion of the symptoms for clarification of the items and distinction from one another was allowed. Again, discussion of the relative merits or importance of the indicators was discouraged.
Step 5: Each participant was then asked to identify on a worksheet those symptoms generated by the group that they considered “critical” to IS audit quality.
Step 6: Participants were asked to rank the symptoms that they identified in step 5 from most to least important.
Step 7: Participants were asked to fill out a questionnaire for demographic data. Partici-pants were invited to an informal venue for further debriefing and discussion.
e-Service Journal
54
enough that the participants interpret it based on their personal perspectives. To this end,
the leader of the meeting should take special care to avoid answering the question for the
group, thereby focusing the group on a particular viewpoint toward the question.
After the silent generation of ideas is complete, a round-robin recording of the ideas
generated is performed. The leader of the group prepares an aggregate list of the partici-
pants’ ideas by asking for one idea from each person on each round. The aggregate list
was prepared in written format and was presented so that all the participants could view
it. During this step, participants were encouraged to add new ideas to their lists. The
benefits of the round-robin recording in step 3 include: equal participation in the presen-
tation of ideas, increase in problem-mindedness, depersonalization of ideas from person-
alities, increase in the ability to deal with a large number of ideas, tolerance of conflicting
ideas, encouragement of ”hitchhiking” of ideas, and a written record and guide of the
ideas presented. It is thought that by sharing ideas and equalizing participation, the cre-
ativity of the group is increased (Delbecq et al., �982). The round-robin format of elicit-
ing the ideas also establishes an important group behavior. After two or three rounds,
each member in the group is an achieved participant. This sets the precedent for further
active participation.
After recording of the ideas was complete, each idea was discussed for clarification
and definition so that the group could come to a consensus as to the item’s meaning. Dis-
cussion of each idea was performed to share the logic behind the idea, the relative impor-
tance attached to the idea, and to air agreements and disagreements. The benefits of
discussing each idea include: opportunity for clarification and elimination of misunder-
standing, opportunity to provide logic for arguments and disagreements, and recording
of differences of opinion without undue argumentation (Delbecq et al., �982). To ac-
commodate these benefits, the leader should make it clear that disagreements regarding
the relative importance of each idea are acceptable and expected. With respect to this
step, it should be emphasized that the individual responsible for suggesting an idea is not
directly responsible for clarifying the idea. The clarification is a group task and not the
unilateral responsibility of the author of the idea (Delbecq et al., �982).
The final step in the nominal group technique was to gather demographic data
from the subjects and perform an informal de-briefing. Information regarding each sub-
ject’s project management experience, training, IT development experience, education
level, and other demographics were gathered using a questionnaire. The informal de-
briefing usually consists of a comparison of the subjects’ inputs and evaluations of the
group results and the process overall.
There were four focus groups; the first was used as a “pre-test” of the data gathering
method and to establish some baseline constructs. Each group met for approximately
two hours. Group I consisted of five subjects. At the time of the focus group, all the sub-
jects were employed as instructors for a large mid-western university and had significant
Using the Troubled Project Recovery Framework
55
(>5 years each) experience as IS consultants for information services organizations. Group
II consisted of five subjects all of whom worked in the IS department of a large financial
services firm and had acted as the project manager for IS projects for a minimum of six
years (average = 8.4 years). Group III consisted of six subjects from the same IS depart-
ment as Group II. These subjects had at least three years of IS project management expe-
rience and averaged �0.� years of IS project management experience. Group IV consisted
of four subjects from two separate organizations, one a large information services firm
and the other a major telecommunications firm. These subjects had an average of four
years of IS project management experience. The details of the focus groups are summa-
rized in Table 2. The output of these focus groups included a list of symptoms from each
participant and a comprehensive, master list of symptoms for each group.
Field Study Results
The results of the field study will be presented as follows: a general discussion of the focus
group outputs is given and then a discussion of the symptoms identified is given within a
logical categorization of the symptoms. The four focus groups yielded a total of �57 sepa-
rate potential symptoms; the groups identified 39, 5�, 43, and 24 symptoms respectively.
After review, 49 of the symptoms were considered duplicates, leaving �08 unique symp-
toms. Of the �08 unique symptoms identified, only 7 were identified by all four of the
groups, 9 were identified by three of the four groups, 22 were identified by two of the
four groups, and 70 were identified by only one of the four groups.
A few earlier studies have specifically looked at early warning or trouble indicators in
projects. One recent study identifies a dominant dozen of early warning signs of IS project
failure (Kappelman, Mckeeman, and Zhang, 2006). That study specifically looked at
warnings that occured in the first 20 percent of the projects initial calendar. The dozen
warnings primarily were people related and process related factors. People related factors
included lack of top management support, weak project manager, no stakeholder involve-
Table 2. Background of focus group participants
Group Industry
Number of Members in Focus group
IS/Project Management Experience
I University 5>5 years as Information Systems
Consultants
IIFinancial
Services Firm5 >6 years as Project Managers
IIIFinancial
Services Firm6 >3 years as Project Managers
IVTelecommunications and Information Services Firm
4 >4 years as Project Managers
e-Service Journal
56
ment, weak commitment of team, team members lacking requisite knowledge, and over-
scheduling of subject matter experts. Process related indicators included lack of
documented requirements, change control, communication breakdown among stake
holders, resources assigned to a higher priority project, and no business case. In contrast,
other researchers looked for warnings through the entire project life cycle and identified
eleven major categories of early warnings in projects (Nikander and Eloranta, 200�).
These categories included personnel or project group, project manager and management,
project planning, project control and reporting, working within the project, communica-
tion, exposed by parties (stakeholders such as clients, CEO, suppliers), and documents.
Using the prior research as a guide, a set of potential logical categories for the symp-
toms was identified. This set of symptom categories was determined by having indepen-
dent judges (the researchers and colleagues) assign each of the symptoms identified to a
category. The researchers then finalized these categories collaboratively and allowed a
specific symptom to be assigned to multiple categories as appropriate. There were eleven
categories identified in the final model. The categories and the number of symptoms as-
signed to each category are presented in Table 3. Each of these categories is discussed
below and the associated symptoms are presented.
Client-Related Symptoms. The first category of symptoms identified contains symp-
toms that are all related to the client or other stakeholders of the project. These symptoms
are presented in Table 4. These symptoms are in one way dependent upon or driven by
the project’s client(s) or other stakeholders (excluding members of the IS department or
project team members). As can be seen by the symptoms identified, the root cause of
many of these symptoms appears to be related to the client’s commitment, buy-in, and
Table 3. Logical categories and number of related symptoms
Logical Category Number of Symptoms
�) Client or stakeholder-related symptoms �2
2) Project’s goal-related symptoms �0
3) Meeting symptoms 6
4) Team symptoms �3
5) Task symptoms 7
6) Project symptoms �2
7) Project Management symptoms 2�
8) Communication symptoms 5
9) Management symptoms 8
�0) Project portfolio symptoms 7
��) Process symptoms 7
Using the Troubled Project Recovery Framework
57
support of the project. Also, management of client’s expectations and a clear specification
of the project’s goal would appear to be related to the client-related symptoms.
Project Goal-Related Symptoms. The focus group subjects identified ten symptoms
that are included as project goal-related symptoms (presented in Table 5). These symp-
toms are all characteristics or attributes of the project’s objectives and desirable outcomes.
It appears that these symptoms become relevant when there is a lack of clearly defined
purpose, scope, and agreed upon metrics for success for the project. These symptoms
may become more important when there are multiple clients with multiple expectations
from the project.
Meeting-Oriented Symptoms. A handful of symptoms centered on meetings (Table
6). These symptoms appear to indicate trouble in two ways: (�) evidence that team mem-
bers have “given-up” by no longer attending or participating in meetings and (2) an in-
crease in meeting activity with clients and management that indicates something out of
the ordinary is happening. Regardless, changes in the quality, quantity, and scheduling
of meeting may be useful metrics to gauge a project’s health.
Team-Related Symptoms. The team-related symptoms represent characteristics of
the team or team behaviors that may indicate trouble on the project. These symptoms
included items such as the skill-set of the team members as well as team morale and stress
level. These symptoms emphasize the importance of creating and managing a well-
Table 4. Client-related symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Loss or lack of project sponsorship
Sponsor demands end date before scope and schedule are set
Efforts to redefine project after specifications are set
Poorly defined requirements—sponsor continually adding requirements (scope creep)
Different user groups communicate different points of view—adding to the complexity of the project
Lack of subject matter expert or user involvement
Lack of commitment by major stakeholders
Client informs project manager that project is in trouble
Client does not understand their needs
Sponsor does not want to meet with team members
Unrealistic stakeholder expectations
Lack of willingness to compromise
e-Service Journal
58
blended, high-performance team. As Table 7 shows, there are four symptoms in this cat-
egory that were identified by all four groups: high team turnover, high stress levels,
depletion of resources, and morale. Clearly, these symptoms could be useful to IS project
managers as indicators of trouble or risks that may need special attention.
Task-Related Symptoms. Several symptoms were identified that could be considered
task-related. These symptoms were associated with specific tasks and activities that need
to be performed during software development or the assignment, documentation, or
completion of these tasks during the project (Table 8). In general, it appears that most of
these symptoms point to a lack of good project management. Issues with scoping and
task assignment are regularly the responsibility of the project manager. This suggests that
perhaps these symptoms should be monitored by the project’s client sponsor.
Project-Related Symptoms. The next category is made up of symptoms that are re-
lated to the project itself, that is, these are symptoms that are particular or unique to the
Table 5. Goal-related symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Lack of clear scope definition
Efforts to redefine project
A large number of change requests or requirements fluctuations within the project
Project does not meet business strategy
Recurring questions about the value of the project
Lack of success criteria
Lack of definition for quality
Client does not understand their needs
Lack of clearly articulated goals
Conflicting objectives of participants
Table 6. Meeting symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Lack of participation at team meetings
Reduced attendance at project meetings
Too many meetings
Increased meetings with clients to clarify problems
Senior management involvement at project meetings
Change in meeting frequency
Using the Troubled Project Recovery Framework
59
specific project (Table 9). These symptoms include characteristics of the system being
developed, e.g., size or value, as well as symptoms related to the project itself, e.g., ade-
quate budget and resources. Some of these symptoms may also be considered project
management symptoms, such as use of a sound methodology.
Project Management Symptoms. A total of 2� symptoms were identified and pre-
sented in Table �0 that were assigned to the category labeled project management. These
Table 7. Team symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Team fighting
High team turnover
Lack of communication on team
Resources dragging their feet on tasks, hoping tasks will go away
Lack of critical skills
Lack of team buy-in
Different levels of expertise within the team leading to communication problems
Stress
Adding new resources
Depletion of resources (e.g., allocation to other projects)
Morale
Team does not understand requirements
Lack of willingness to compromise
Table 8. Task symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Limited, or lack of, documentation (e.g., not docu-menting tasks completed or technical specifics)
Consistent levels of overtime
Design problems discovered during code review
WBS (work breakdown structure) not assigned
Major defects identified during the inspection—changes in the requirements or solution
High amount of rework or defects
Training deferred or postponed
e-Service Journal
60
symptoms were directly associated with how the project is being managed or the project
management practices followed or established by the organization (or lack thereof). Sim-
ilar to the task-related symptoms, this category may include primarily items that client or
project manager’s supervisor should monitor. One of these symptoms, project status un-
known, was identified by all four groups. The fact that the project’s status is not known,
i.e., budget, schedule, or meeting specs, would be a clear indicator that trouble is pres-
ent.
Communication Symptoms. The next category was made up of symptoms that were
all related to communication (Table ��). This included communication among the team
members, between the team and user groups, and among the team, user groups and
management. Communication during systems development has been shown by previous
research to be critical for establishing correct system requirements and maintaining good
relationships between developers and users. From the symptoms listed, it appears that
communication would lead to errors and possibly re-work or incorrect features.
Management Symptoms. The next category includes symptoms that were all related
to the organization’s general management, including support of upper management and
management issues such as clearly defined roles and responsibilities (Table �2). The im-
portance of support and involvement of management in systems development projects
Table 9. Project symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Over allocation of resources (i.e., more work than a person can reasonably perform)
Constantly changing scope
Lack of budget in terms of people, time, and materials for the project
Failing to use any methodology in that the team or sponsor is not following the method
New technology that is unfamiliar to the team
Risk mitigation delayed until end of the project (i.e., all major symptoms appear at the end of the project)
Scope is too large, indicator of size of project
High CPI (cost/performance index) in that costs are higher than benefits
Lack of budget for training
Team feels lack of ownership
Lack of project value
Budget unknown
Using the Troubled Project Recovery Framework
6�
Table 10. Project management symptoms
SymptomGroup I
Group II
Group III
Group IV
Project status not known due to poor status report meetings
Schedule slippages along critical path
Poor change management
Schedule of work not defined or poorly defined
Revisiting issues thought to be closed
Milestones are not met
Difficulty in assessing the impact of change requests
Variance of time estimates (high or low)
Constantly behind schedule
Deliverables are not met on schedule
Symptoms not identified
Poor estimates
No one to take ownership of issues
Lack of leadership initiatives
Late PM engagements
Key risk issues not addressed
Not able to define team resources up front
Poor management of outsourced resources
Wrong product or deliverable defined in requirements
Expenditures are over budget
Multitasking of resources
Table 11. Communication symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Lack of communication on team
Different levels of expertise within the team – communication problems
Miscommunication between user and team
Lack of “say” or “input”
Escalation increases
e-Service Journal
62
has also received attention in prior research. Management can impact IS project success
in many ways, from controlling resources to setting the tone that a project is or is not
important to the organization.
Project Portfolio Symptoms. The next category included was project portfolio symp-
toms (Table �3). These symptoms are related to how the organization manages its port-
folio of projects and where the project under consideration relates to other projects in the
portfolio. These symptoms impact a project in several ways, including projects doing re-
dundant work or competing for resources. In addition, the value each project may add
Table 12. Management symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Weak project sponsorship (i.e., lack of commitment)
Pressure from upper level management causes stress on team
Reduction in staff, yet project kept on original schedule (i.e., original milestones)
Micromanagement by upper management
Lack of confidence in management
Unclear roles and responsibilities
Inability to commit resources
Multicultural offshore teams with issues such as communications, time zone, cultural habits, language differences
Table 13. Project portfolio symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Difficulties in blending across organizational teams (i.e., teams and/or management from different areas do not identify the problems in the same manner)
Management and project not in sync
Project interdependency
Project has a low priority
No portfolio management in that two similar projects underway
Inability to prioritize project
Putting a project on hold while management works it out
Using the Troubled Project Recovery Framework
63
toward organizational goals may change over time and affect the priority of one project
vis-à-vis another.
Process-Related Symptoms. The last category of symptoms identified are process-
related items (Table �4). These symptoms were primarily related to the IS development
or implementation process being used. These symptoms could be considered the project
manager’s responsibility also; however, most of the items do not appear to be directly
within the project manager’s control, e.g., no quality assurance process in place.
In summary, the results of the field study yielded a total of �08 unique symptoms.
The vast majority was identified by only one of the groups. Only seven symptoms were
identified by all four. The symptoms appear to fall into eleven distinct logical categories.
Potential implications and further exploration of these results are discussed below.
DISCUSSION
Overall, the results of the focus groups suggest that there are different types of “symp-
toms” that may present themselves during a project that could cause or indicate trouble.
Some of the symptoms identified by the groups appear to be symptoms or indicators of
trouble; however, it appears that other items suggested by the groups are clearly the cause
of trouble rather than a symptom of a problem, e.g., lack of management commitment.
To further explore this distinction, each logical category of symptoms is reviewed and the
items identified are further classified as either a likely cause of trouble or a symptom of a
problem (or trouble indicator). In addition, for each category, prescriptions are suggested
that should help project managers and team members address the cause and symptom.
Causes, Symptoms, and Prescriptions
Although the goal of this research project was to identify symptoms that would act as
early warnings for trouble on a project, many of the items identified by the subjects were
more akin to causes or descriptions of problems. In fact, some of the items identified
Table 14. Process symptoms
SymptomGroup
IGroup
IIGroup
IIIGroup
IV
Testing is rushed, limited, or eliminated
Automating process without analysis
No lessons learned from a previous failure
No quality assurance or evaluative process
Lack of definition for quality
Lack of documentation (i.e., results, planning, design)
No R&D prior to planning
e-Service Journal
64
could be considered both a cause and a symptom. For each category of items identified,
we discuss causes, symptoms, and prescriptions for corrective action.
Client-Related Causes, Symptoms, and Prescriptions. As presented in Table �5, the
causes of trouble related to clients are primarily associated with the need for information
from the clients and the on-going support in terms of resources and psychological com-
mitment. For a project to succeed, substantive information from the client (users and
sponsor) is necessary to build a system that performs the desired tasks in a way the clients
can understand. The symptoms identified reflect difficulties in communicating this in-
formation and the complexity involved with building systems that serve multiple user
groups. The prescriptions suggested focus on establishing a common understanding
among clients and the IS staff working on the project, which includes an evaluation of
the business value of continuing the project.
Goal-Related Causes, Symptoms, and Prescriptions. The causes of trouble associated
with the goals of the project (see Table �6) appear to fall into two general types: (�) the
lack of documentation or work to clearly define the purpose, goal, and objectives of the
project and (2) misunderstanding or conflict about the goals of the project. The symp-
toms all center on efforts to correct these shortcomings by redefining and changing the
work and persistent questions regarding the project’s value. The prescriptions suggested
focus on gaining consensus about the value and purpose of the project and specifying in as
much detail as possible the deliverables expected and then obtaining agreement for these.
Meeting Causes, Symptoms, and Prescriptions. The meeting related items all appear
to be symptoms of trouble rather than causes of problems (see Table �7). Some of the
causes of meeting symptoms may include any of the causes identified in other categories.
Table 15. Client/stakeholder causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Loss or lack of project sponsorshipPoorly defined require-ments—sponsor continu-ally adding requirements (scope creep)Lack of subject matter expert or user involvementLack of commitment by major stakeholdersClient does not understand their needsUnrealistic stakeholder expectationsLack of willingness to compromise
•
•
•
•
•
•
•
Sponsor demands end date before scope and schedule are setEfforts to redefine project after specifications are setDifferent user groups communicate different points of view – adding to the complexity of the projectClient informs project manager that project is in troubleSponsor does not want to meet with team members
•
•
•
•
•
Determine specific problem, if necessaryEvaluate the value of continuing projectDefine the scope of the project and specify boundariesIdentify the stakeholders and end-users and solicit support, especially a sponsorReview project specifica-tions and requirements with sponsor and users, obtain signoffs
•
•
•
•
•
Using the Troubled Project Recovery Framework
65
The prescriptions suggested focus on investigating the causes and reviewing assignments
and responsibilities.
Team Causes, Symptoms, and Prescriptions. The causes related to the development
team address communication, competence, and psychological issues (Table �8). In gen-
eral, the lack of communication or skill level impacts the ability of the team to perform
adequately. The symptoms reflect the team’s general frustration and low morale. The
prescriptions suggested focus on motivating and energizing the team and making changes
to the team membership if necessary.
Task Causes, Symptoms, and Prescriptions. The causes of trouble associated with
Table 16. Goal-related causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Lack of clear scope definitionProject does not meet business strategyLack of success criteriaLack of definition for qualityClient does not understand their needsLack of clearly articulated goalsConflicting objectives of participants
•
•
••
•
•
•
Efforts to redefine projectA large number of change requests or requirements fluctuations within the projectRecurring questions about the value of the project
••
•
•
Determine specific problem, if necessaryEvaluate the value of continuing project, determine business objectiveDefine the scope of the project and specify boundariesReview project specifica-tions and requirements with sponsor and users, obtain signoffsSpecify clear metrics for project and system success
•
•
•
•
•
Table 17. Meeting causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Lack of participation at team meetingsReduced attendance at project meetingsToo many meetingsIncreased meetings with clients to clarify problemsSenior management involvement at project meetingsChange in meeting frequency
•
•
••
•
•
Determine the specific problem, if necessaryInterview key staff to determine reasons for absence or low participa-tionReview project specifica-tions and requirements with sponsor and users—rework project schedule and deliverables if changed from original project approved
•
•
•
e-Service Journal
66
tasks, presented in Table �9, address the need to document each person’s responsibilities
and track their progress. Each task should have one individual responsible for its perfor-
mance, making sure that it is performed, not necessarily doing it themselves, with a dead-
line for the task. The symptoms indicated for this category focus on “extra” work that is
performed to complete tasks correctly. This is in the form of rework or overtime, but also
includes poor performance when work does not pass inspections or testing later in the
development process.
Project Causes, Symptoms, and Prescriptions. Table 20 presents the project related
causes, symptoms, and prescriptions. The causes hit upon several different themes in-
cluding: lack of resources, lack of structure in the form of methods and procedures,
Table 18. Team causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Lack of communication on teamLack of critical skillsLack of team buy-inDifferent levels of expertise within the team leading to communication problemsDepletion of resources (allocation to other projects)Team does not understand requirementsLack of willingness to compromise (users & IS)
•
•••
•
•
•
Team fightingHigh team turnoverResources dragging their feet on tasks, hoping tasks will go awayStressAdding new resourcesDepletion of resources (e.g., allocation to other projects)Morale
•••
•••
•
Determine specific problem, if necessaryInterview key staff to pinpoint issues (personal-ity conflicts, overworked)Change team membersEvaluated workloads and compensationPerform team building and communication exercisesReview project specifica-tions and requirements with sponsor and users, obtain signoffs
•
•
••
•
•
Table 19. Task causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Limited, or lack of, docu-mentation (i.e., not docu-menting tasks completed or technical specifics)WBS (work breakdown structure) not assigned
Consistent levels of overtimeDesign problems discovered during code reviewMajor defects identified during the inspection—changes in the requirements or solutionHigh amount of rework or defectsTraining deferred or postponed
Review project methods and practices with teamReview and update project management artifacts, e.g. the WBSEvaluate workloads and compensationReview project specifica-tions and requirements with analysis, design, coding, and testing personnel
•
•
•
•
Using the Troubled Project Recovery Framework
67
vagueness about the project, and the use of new or unfamiliar technology. The symptoms
also reflect a variety of issues including overloading team members with tasks and delay-
ing dealing with problems. The prescriptions for project related problems include evalu-
ating the business case to assess risk versus return for the project.
Project Management Causes, Symptoms, and Prescriptions. The items associated
with project management (Table 2�) reflect common project management issues. The
causes focus on lack of control and the quality of management of resources. Symptoms are
related to the standard overarching project management metrics of time, cost, and work
performed. Prescriptions for project management would include establishing and follow-
ing project management best practices and perhaps changing the project manager.
Communication Causes, Symptoms, and Prescriptions. The communication category
causes are summed up as lack of proper communication among team members and the
clients (Table 22). The symptoms include the frequent escalation of issues to get clarifica-
tion or a decision from higher up in the hierarchy. Prescriptions include establishing for-
mal policies and procedures for information exchange among the key groups involved.
Management Causes, Symptoms, and Prescriptions. Causes associated with general
management included “too much” involvement from upper management, e.g., micro-
managing tasks, as well as “too little” involvement in the form of lack of commitment
and resources (Table 23). The symptoms reflect an inappropriate involvement in the day-
to-day tasks by management. The prescriptions include reviewing the project purpose
and reaffirming management’s support of the project.
Project Portfolio Causes, Symptoms, and Prescriptions. The project portfolio focuses
attention on the fact that one project is usually dependent upon other projects and systems
Table 20. Project causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Constantly changing scopeLack of budget in terms of people, time, and materials for the projectFailing to use any method-ology in that the team or sponsor is not following the methodNew technology that is unfamiliar to the teamScope is too large, indicator of size of projectLack of budget for trainingLack of project valueBudget unknown
••
•
•
•
•••
Over allocation of resources (i.e., more work than a person can reasonably perform)Risk mitigation delayed until end of the project (i.e., all major symptoms appear at the end of the project)High CPI (cost/perfor-mance index) in that costs are higher than benefitsTeam feels lack of owner-ship
•
•
•
•
Define the scope of the project and specify boundariesPerform financial analysis, risk assessment, and business case to determine benefits of continuing projectReview project methods and practices with teamDetermine budget and constraints—assess level of skill for technology being used
•
•
•
•
e-Service Journal
68
Table 21. Project management causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Poor change managementSchedule of work not defined (i.e., poorly defined schedule)Poor estimatesLack of leadership initia-tivesLate PM engagementsKey risk issues not addressedPoor management of outsourced resources
••
••
••
•
Project status not known – due to poor status report meetingsSchedule slippages—along critical pathRevisiting issues that you thought were closedMilestones are not metDifficulty in assessing the impact of change requestsVariance of time estimates (high or low)Constantly behind scheduleDeliverables are not met on scheduleSymptoms not identifiedNo one to take ownership of issuesNot able to define team resources up frontWrong product or deliver-able defined in require-mentsExpenditures are over budgetMultitasking of resources
•
•
•
••
•
••
••
•
•
•
•
Perform an audit of work performed versus planReview (and test) docu-mentation of work performedChange project managersPerform financial analysis, risk assessment, and business case to determine benefits of continuing projectReview project methods and practices with teamDetermine budget and constraints—assess level of skill for technology being usedReview project specifica-tions, requirements, budget with sponsor
•
•
••
•
•
•
Table 22. Communication causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Lack of communication on teamMiscommunication between user and team
•
•
Different levels of expertise within the team—commu-nication problemsLack of “say” or “input”Escalation increases
•
••
Establish policy and procedures to exchange information between users and IS staffInterview key staff to pinpoint issues (personal-ity conflicts, overworked)Perform team building & communication exercisesReview project specifica-tions and requirements with sponsor and users, obtain signoffs
•
•
•
•
Using the Troubled Project Recovery Framework
69
(Table 24). Ensuring that these projects and their ensuing systems can be integrated is a
major consideration in today’s corporate environment. Symptoms reveal difficulties when
two or more projects are addressing the same or similar business objectives. Prescriptions for
addressing these issues include establishing a project portfolio or program management that
manages the entire systems development function for an organization and establishing
knowledge management practices that would allow sharing of information across projects.
Process Causes, Symptoms, and Prescriptions. Table 25 presents the causes associ-
ated with the process used for IS projects. The causes are all in some manner related to
having an established methodology or process for IS project management. None of the
items in this category appeared to be actual symptoms of trouble. The prescriptions sug-
gested emphasize the importance of establishing a professional IS project management
function and following a methodology using best practices.
Further study should be performed to follow-up on these causes, symptoms, and
prescriptions to identify potential patterns that could apply across projects and organiza-
tions for IS projects. Clearly, further research is needed to test whether an item is a cause
or symptom of trouble and the nature of the relationship between causes, symptoms,
problems, and corrective prescriptions.
The Common Set of Symptoms. The fact that only seven symptoms were identified by
all four of the groups suggests that there may be a small set of “critical” symptoms that
Table 23. Management causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Weak project sponsorship (i.e., lack of commitment)Pressure from upper level management causes stress on teamReduction in staff, yet project kept on original schedule (i.e., original milestones)Micromanagement by upper managementUnclear roles and responsi-bilitiesInability to commit resourcesMulticultural offshore teams with issues such as communications, time zone, cultural habits, language differences
•
•
•
•
•
•
•
Micromanagement by upper managementLack of confidence in management
•
•
Review project specifica-tions and requirements with sponsor and users, obtain signoffsEvaluate the value of continuing projectDefine the scope of the project and specify boundariesDetermine who the stakeholders and end-users are and solicit support, especially a sponsor
•
•
•
•
e-Service Journal
70
project managers and sponsors could track as early indicators of trouble. The seven com-
mon symptoms are:
�. Lack of clear scope definition
2. High team turnover
3. Stress
4. Depletion of resources (allocation to other projects)
5. Morale
6. Consistent levels of overtime
7. Project status not known – poor status report meetings
Table 24. Project portfolio causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Difficulties in blending across organizational teams (i.e., teams and/or manage-ment from different areas do not identify the problems in the same manner)Management and project not in syncProject interdependencyProject has a low priorityNo portfolio manage-ment—two projects doing similar things
•
•
•••
No portfolio management in that two similar projects underwayInability to prioritize projectPutting a project on hold while management works it out
•
•
•
Review project specifica-tions and requirements with sponsor and users, obtain signoffsEvaluate the value of continuing project—in context of organization project portfolioIdentify the stakeholders and end-users and solicit support, especially a sponsor
•
•
•
Table 25. Process causes—symptoms—prescriptions
Causes Symptoms Prescriptions
Testing is rushed, limited, or eliminatedAutomating process without analysisNo lessons learned from a previous failureNo quality assurance or evaluative processLack of definition for qualityLack of documentation (i.e., results, planning, design)No R&D prior to planning
•
•
•
•
•
•
•
Follow an established methodology for develop-ment – including appropri-ate analysis and business process redesignExecute project manage-ment best practicesImplement project portfolio and knowledge management practices to improve organizational learning
•
•
•
Using the Troubled Project Recovery Framework
7�
Of these seven items, only one appears to be only a cause rather than a symptom,
i.e., lack of clear scope definition. The remaining items do appear to be good indicators
of some type of trouble. High turnover of personnel on the team could be a sign that the
“rats are fleeing the sinking ship” and the loss of competent personnel makes the project
even more likely to have errors or poor quality. In addition, the project may get a bad
reputation among the IS staff, making it difficult to recruit personnel. Stress and low
morale would both be indicators of the workload and frustration associated with poorly
run projects. The depletion of resources could be a cause of trouble, but as an indicator of
trouble this may indicate that de-funding of the project is occurring. Consistent overtime
is an indicator of rework due to low quality, poor specifications, or changing require-
ments; all of which indicate problems. Not knowing the status of a project indicates that
the project manager is not getting the documentation regarding work performed from
the staff or that the project manager is inadequately updating and reporting progress.
Together these seven items may make a good set of metrics to be used on a project man-
agement or project sponsor’s dashboard.
CONCLUSION
Many information system projects run into trouble. Many can be recovered and rehabili-
tated to success. We outlined the TPRF and performed a field study to validate the first
phase of our framework and provided guidance for managers using the TPRF. The four
stages include recognition, immediate recovery, sustained recovery, and maturity. With-
out recognition, no recovery can take place. The field study identified a broad set of
symptoms that could present themselves during a project that project managers can use
to recognize trouble.
The diversity of symptoms identified may indicate that industry and organization
specific symptoms exist, and some symptoms may be more prevalent or more important
in certain industries or application domains. However, the results also suggest that there
may be some symptoms that are general and relevant to all IS projects. These questions
require additional research into the details of the causes and effects of trouble in IS devel-
opment projects. Additionally, more research is needed to identify successful solutions to
common problems.
REFERENCES
Aiyer, J., Rajkumar, T. M., and Havelka, D. “A Staged Framework for the Recovery and Rehabilitation of
Troubled IS Development Projects,” Project Management Journal, Research Quarterly (36:4), 2005,
pp. 32–43.
“Assessment, Recovery, Transition,” Michigan Department of Information Technology, Retrieved Febru-
ary �0, 2007 from http://www.michigan.gov/documents/ART_200504_�22595_7.pdf.
Belassi, W., and Tuckel, O. “A New Framework for Finding Critical Success/Failure Factors in Software
Development,” International Journal of Project Management (�4:3), �996, pp. �4�–�5�.
e-Service Journal
72
Boundy, G., and Diamond, M. “Keeping Crisis out of Project Management,” Computing Canada, 29,
�998, p. 32.
Brady, S., and DeMarco, T. “Management Aided Software Engineering,” IEEE Software, �994, pp. 25–
32.
Brown, W. J., Malveau, R. C., McCormick III, H. W., and Mowbray, T. J. AntiPatterns: Refactoring Soft-ware, Architectures and Projects in Crisis, John Wiley and Sons, New York, �998.
Caper Jones, D. “Patterns of Large Software Systems: Failures and Success,” IEEE Computer, �995a, pp.
86–87.
Caper Jones, D. Patterns of Software System Failure and Success, International Thompson Computer Press,
Boston, �995b.
CHAOS Chronicles, Research Reports from http://www.standishgroup.com, 2004.
Connell, C. “Healing Sick Software,” 2000, retrieved from http://www.chc-3.com.
“Cost overrun? Call in a Project Manager,” Business Times (4), 2006, retrieved September 2006 from
http://www.pmi.org/prod/groups/public/documents/info/ap_articles.asp.
Day, J. “Software Development as Organizational Conversation: Analogy as a Systems Intervention,” Sys-tems Research and Behavioral Science (�7), 2000, pp. 349–358.
Delbecq, A. L., Van de Ven, A. H., and Gustafson, D. H. “Guidelines for Conducting NGT Meetings,” in
D. R. Hampton, C. E. Summer and R. A. Webber (eds.), Organizational Behavior and the Practice of Management, 4th ed., Scott, Foresman, and Company, Glenview, Illinois, �982, pp. 279–298.
Fowler, A., and Walsh, M. “Conflicting Perceptions of Success and Failure in an Information Systems Proj-
ect,” International Journal of Project Management (�7:�), �999, pp. �–�0.
Glass, R. L. “Failure is Looking More like Success These Days,” IEEE Software, 2002, pp. �03–�04.
Havelka, D. “Requirements Determination: An Information Systems Specialist Perspective of Process
Quality,” Requirements Engineering Journal (6:4), 2002, pp. 220–236.
Havelka, D., Arnold, V., and Sutton, S. “A Methodology for Developing Measurement Criteria for Assur-
ance Services: An Application in Information Systems Assurance,” Auditing: A Journal of Practice and Theory (�7), �998.
Havelka, D., Sutton, S. G., and Arnold, V. “Information Systems Quality Assurance: The Effect of Users’ Ex-
periences on Quality Factor Perceptions,” Review of Business Information Systems (5:2), 200�, pp. 49–62.
Iacovou, C. L., and Dexter, A. S. “Surviving IT Project Cancellations,” Communications of the ACM (48:4),
2005, pp. 83–86.
Ibbs, C. W., and Kwak, Y. H. “Assessing Project Management Maturity,” Project Management Journal, 2000, pp. 32–43.
Jalote, P. CMM in Practice, Addison-Wesley, �999.
Kappelman, L. A., Mckeeman, R., and Zhang, L. “Early Warning Signs of IT Project Failure: The Domi-
nant Dozen,” Information Systems Management (3�:36), 2006.
Kapur, G. “How to Kill a Troubled Project,” 200�, retrieved from http://www.cioinsight.com/print_
article/0,3668,a=�932500.asp.
Keil, M., and Montealgre, R. “Cutting your Losses: Extricating your Organization Losses when a Big Proj-
ect Goes Awry,” Sloan Management Review, 2000, pp. 55–68.
Keil, M., and Robey, D. “Turning Around Troubled Software Projects: An Exploratory Study of the De-
escalation of Commitment to Failing Courses of Action,” Journal of Management Information Systems (�5:4), �999, pp. 63–87.
Keil, M., and Robey, D. “Blowing the Whistle on Troubles Software Development Projectsc” Communica-tions of the ACM (44:7), 200�, pp. 87–93.
Kiechel, W. “Facing up to Denial,” Fortune (�23), �993, p. �63.
Klein, G., and Jiang, J. J. “Seeking Consonance in Information Systems,” Journal of Systems and Software (56), 200�, pp. �95–202.
Using the Troubled Project Recovery Framework
73
Kotnour, T. “A Learning Framework for Project Management,” Project Management Journal Research Quar-terly, �999, pp. 32–38.
Linberg, K. R. “Software Developer Perceptions about Software Project Failure; A Case Study,” The Jour-nal of Systems and Software (49), �999, pp. �77–�92.
McConnell, S. “The Business of Software Improvement,” IEEE Software, 2002, pp. 5–8.
Nikander, I. O., and Eloranta, E. “Project Management by Early Warnings,” International Journal of Proj-ect Management (�9), 200�, pp. 385–399.
Pintro, J. K., and Slevin, D. P. “Project Success: Definition and Measurement Techniques,” Project Man-agement Journal, Research Quarterly (�9:�), �988, pp. 68–72.
Seeger, and Ulmer. “Virtuous Responses to Organizational Crisis,” Journal of Business Ethics (3�), 200�,
pp. 369–376.
Sifri, G. “Assessing and Recovering Troubled Projects, Part �,” ESI Horizons (5:5), 2003a, pp. �–7.
Sifri, G. “Assessing and Recovering Troubled Projects, Part 2,” ESI Horizons (5:7), 2003b, pp. �–5.
Sifri, G. “Assessing and Recovering Troubled Projects, Part 3,” ESI Horizons (5:9), 2004, pp. �–5.
Smith, J., Keil, M., and Depledge, G. “Keeping Mum as the Project Goes Under,” Journal of Management Information Systems (�8:2), 200�, pp. �89–227.
Summary of a Workshop on Information Technology Research for Crisis Management, �998, retrieved from
(http://books.nap.edu/html/itr_crisis_mgmt/index.htm).
Tarek, A. M., and Madnick, S. “The Elusive Silver Lining: How we Fail to Learn from Project Manage-
ment Failures,” Sloan Management Review, �990, pp. 39–49.
Whittaker, B. “What Went Wrong? Unsuccessful Information Technology Projects,” Information Manage-ment and Computer Security (7:�), �999, pp. 23–29.