De-escalating_IT_Project.pdf

Loading ...
Global Do...
News & Politics
contributed articles
october 2009  |   vol. 52  |   no. 10  |   communications of the acm     131
doi: 10.1145/1562764.1562797
by donal flynn, Gary Pan, mark keil, and  
maGnus mährinG
Taming runaway informaTion Technology (iT) projecTs 
is a challenge that most organizations have faced and 
that managers continue to wrestle with. These are 
projects that grossly exceed their planned budgets and 
schedules, often by a factor of 2-3 fold or greater. Many 
end in failure; failure not only in the sense of budget or 
schedule, but in terms of delivered functionality as well.2
Runaway projects are frequently the result of 
escalating commitment to a failing course of action,11 a 
phenomenon that occurs when investments fail to work 
out as envisioned and decision-makers compound the 
problem by persisting irrationally.1 Keil, Mann, and 
Rai7 reported that 30-40% of IT projects exhibit some 
degree of escalation. To break the escalation cycle, 
de-escalation of commitment to the failing course of 
action must occur so that valuable resources can be 
channeled into more productive use.5 But, making  
de-escalation happen is neither easy nor intuitive. 
This article briefly examines three approaches that 
have been suggested for managing de-escalation. By 
combining elements from the three approaches, we 
introduce a de-escalation management maturity 
(DMM) model that provides a useful 
framework for improving practice.
approach 1—the crisis  
management approach
Iacovou and Dexter4 follow a crisis man-
agement approach in formulating their 
de-escalation management guidelines. 
Their work is based on the results of a 
3-round Delphi survey of 38 IT consul-
tants, who ranked 10 actions to pursue 
in response to tackling a runaway proj-
ect. They emphasize three objectives 
that underlie a successful turnaround 
strategy: operational containment, cred-
ibility restoration, and organizational 
learning. The de-escalation actions they 
advocate are shown in Table 1. 
approach 2—the change  
management approach
Pan et al9,10 follow a change manage-
ment approach in formulating their 
de-escalation guidelines. Their work 
is based on a case study of UKCouncil, 
a U.K. local government organization 
that proposed an electronic procure-
ment system for efficiency reasons and 
a desire to be the first such organization 
to purchase goods and services elec-
tronically. During project testing, the 
project stalled due to a disagreement 
between the users and the vendor, who 
demanded an additional £150,000 for 
“redesigning the software again”. Pay-
ment was made and the project con-
tinued until the problems resurfaced 
and project development was halted. 
By examining how project individuals 
surrendered their commitment to the 
previous failing course of action and 
accepted a joint agreement to a turn-
around strategy, they formulated five 
important lessons for managing de-
escalation as shown in Table 2. 
approach 3—the Problem  
solving approach
Montealegre and Keil8 follow a problem 
solving approach in formulating their 
de-escalation management guidelines. 
Their work is based on the results of a 
case study involving the computerized 
baggage handling system at Denver 
de-escalating 
it Projects: 
the dmm 
model
132    communications of the acm    |   october 2009  |   vol. 52  |   no. 10
contributed articles
lation discipline and maturity needed 
to fully adopt the model. To provide a 
roadmap for managers who want to 
implement the model in their orga-
nizations, we provide a de-escalation 
management maturity (DMM) model 
(see Table 4) that is based on the famil-
iar CMM model, and which captures 
the progressive levels of preparedness 
International Airport. Based on their 
analysis of this case, they described 
a four-phase model that captures the 
process of de-escalation and includes 
key triggering activities for each phase. 
This model was later applied to the 
well-known Taurus case involving the 
London Stock Exchange and was found 
to fit that case as well.6 The four phases 
of the de-escalation process, as seen 
from the problem solving perspective, 
are described in Table 3. 
comparing the three approaches 
Both the problem solving and the 
change management approaches fo-
cus only on de-escalating the project 
at hand. Neither of these approaches 
is likely to result in the organizational 
learning that is necessary to avoid fu-
ture mishaps. In contrast, the crisis 
management approach, while it ad-
dresses organizational learning, does 
not cover detection, and tends to focus 
on the technical, rather than people, 
issues associated with implementing 
a new project plan. So what approach 
should a manager use? Figure 1 pres-
ents an integrated model that incorpo-
rates the best features of the three ap-
proaches and can guide managers.
The two first steps of the model 
(prevent and detect) are about working 
proactively to minimize the risk for es-
calation and exposing any escalation 
that does occur as early as possible. For 
prevention, advanced project manage-
ment practices (reporting and control 
measures, project ownership, a proj-
ect office) are essential. For detection, 
norms and rewards for reporting bad 
news about a project are key. The fol-
lowing three steps (disrupt, re-evaluate 
and implement) concern wholesale re-
thinking of the project plan and devel-
opment of a new plan, and constitute 
a continuous effort to manage change 
that helps move stakeholder commit-
ment from a failing course of action 
to a new and viable course. Finally, 
the step learn and the feedback loop, 
implement improved practices, move the 
focus of de-escalation management 
from the runaway project at hand to 
organizational learning and improve-
ment of project management and de-
escalation management practices.
Most organizations will not be in a 
position today to follow the integrated 
model because they lack the de-esca-
needed to manage de-escalation.a
Level 1 of the DMM model requires 
that organizations have people who 
know how to apply basic project man-
agement tools and techniques, es-
table 2:  de-escalation lessons based on change management approach
de-escalation lesson
description based on ukcouncil case study
1. Unfreezing commitment to 
a failing course of action
People move away from a failing course of action to a new course 
of action when they believe the present course of action is failing. 
Such movement is viewed as an instance of organizational change 
management. one’s mindset has to change before “unfreezing” of 
commitment to the failing course of action can occur.  
2. Committing to a new 
course of action
In the UKcouncil case, the project champion, who was not the project 
manager, played a crucial role in the commitment transformation process. 
being external to the project, he was less committed to the failing course 
of action and therefore more willing to rescope project objectives.
3. Providing psychological 
safety nets
Managers must reduce the severity of penalties for failure and provide 
assurance of little or no recrimination. In the UKcouncil case, project 
members discussed openly their shortcomings after learning no individual 
would be punished for earlier problems. 
4. Developing new attitudes 
and behaviors
being open and receptive towards the new course of action may help 
project members surrender their faith in the failing course of action. role 
models are useful in initiating changes and developing new attitudes.  
In the UKcouncil case, the project champion became identified as a 
role model and his stance reflected the new attitudes and behaviors to 
be learned by project members. For instance, he influenced the project 
members to work coherently by rescoping project objectives before 
turning the troubled project around.
5. Refreezing commitment to 
a new course of action 
besides changing attitudes, project members should alter their behaviors 
to make the new course of action successful. In the UKcouncil case, the 
project champion assessed the level of stakeholder support, and built a 
favorable consensus toward the turnaround strategies before carrying out 
the action plans.
table 1:  de-escalation actions based on crisis management approach
de-escalation action
description
1. Develop a recovery plan
Freeze project if necessary and develop a recovery plan that may 
restore some of the lost credibility, re-build trust in the team’s ability to 
complete the project, obtain renewed legitimacy from owners and senior 
management, and enable the involved parties to learn from the experience.
2. Manage the scope
Prioritize project by focusing on critical features and removing bells and 
whistles. Project managers should focus on critical path deliverables and 
use the 80/20 rule.
3.  Reevaluate project 
leadership
review project leadership and replace project manager if necessary 
to strengthen PM skills, restore project legitimacy, and enhance the 
likelihood of success.
4.  Re-estimate the business 
case and consider 
cancellation
revise the business case and re-examine the project feasibility, ignoring 
sunk costs.  cancel the project if there is no longer a solid business case.
5. Re-plan the project 
re-plan the project budget and schedule using proven estimating methods.
6.  Manage users’ 
expectations
establish realistic user expectations as this will influence system 
acceptance and is crucial during project turnaround.
7.  Formulate an open 
communications plan
clearly communicate the project delay and proposed recovery plan to all 
stakeholders. Avoid finger pointing and finding scapegoats. 
8.  Break the remainder of the 
project into small chunks
establish clear milestones and break the remainder of the project into 
small chunks.
9.  Deal with the people 
issues of the project team
replace team members who lack necessary skills. reassure the 
remaining ones who may become the target of user frustrations and 
whose morale is negatively impacted by the dismissal of their colleagues.
10.  Incorporate corrective 
practices in the 
development process
Improve the firm’s development practices to reduce the future likelihood 
of similar incidents by conducting staff training, and modifying estimating 
techniques and monitoring policies.
a  The DMM is not designed to compete with or substitute 
for the more broad-based capability maturity model 
(CMM). The DMM is designed with a much narrower 
focus on de-escalation management.
contributed articles
october 2009  |   vol. 52  |   no. 10  |   communications of the acm     133
pecially those that focus on project 
planning. Moreover, the culture of the 
organization must support the applica-
tion of these techniques. The project 
plan must be regarded as changeable 
and it must be updated frequently so 
that it provides a true reflection of proj-
ect status.
Level 2 of the DMM model requires 
that organizations maintain a project 
baseline that allows for detection of 
deviations from the project plan. Task 
status must be tracked regularly and 
compared against the project plan, as 
this is the best way to prevent project 
escalation. Prevention also involves the 
pre-appointment of external advisors 
capable of reviewing the entire project 
or specific vendors. Note that achiev-
ing levels 1 and 2 involve nothing more 
than building and exercising basic 
project management skills. Yet, these 
skills provide the foundation for reach-
ing higher levels on the DMM model. 
As we move up through Levels 3, 4, and 
5, the organization must focus more on 
soft skills that improve change man-
agement and project status reporting 
capabilities.
Level 3 of the DMM model requires 
that organizations have experienced 
project managers who not only pos-
sess the technical skills needed to be 
a good project manager, but also the 
communication and change manage-
ment skills needed to execute project 
plans in environments where they lack 
direct authority over those who must 
be on board in order for the project to 
be successful. 
Level 4 of the DMM model requires 
that organizations develop a culture 
where it is not only permissible to com-
municate bad news but where bad 
news reporting is actually encouraged. 
Some organizations have made prog-
ress in this regard by reminding partic-
ipants that project review meetings are 
for “bad news only” reporting. While 
there are a number of both simple and 
sophisticated approaches to project 
monitoring that can be brought to bear 
at Level 2, the ultimate effectiveness 
of these approaches hinges on reach-
ing Level 4. The organizational culture 
must be supportive of bad news report-
ing so that task and project level status 
updates are honest. 
Level 5 of the DMM model requires 
that organizations pay more than lip 
service to the value of organizational 
learning. Cross-project learning needs 
to be facilitated, for example through 
work sessions where experiences are 
shared and new ideas for de-escalation 
management are generated. In addi-
tion, people in roles outside project 
teams (such as methodology experts, 
project owners, and project manage-
ment office personnel) need to learn 
about experiences and insights that 
can help improve project management 
practices in a systematic way.
How should an organization decide 
on the DMM level that is appropriate 
for them? In our view, this depends 
on the nature of the organization, the 
frequency and criticality of complex IT 
projects, and the skill base of its man-
agers. While attaining high levels on 
the DMM may be desirable, doing so 
table 3:  de-escalation Phases based on Problem solving approach
de-escalation phase
description based on denver international airport case study
1. Problem recognition
De-escalation requires an understanding that something is wrong with 
the present course of action. At Denver, the presence of unambiguously 
negative feedback marked a turning point, leading the city to rectify the 
problems in the project. external events, for example government and 
vendor investigations, triggered a general reassessment of the project..  
2. Re-examination of prior 
course of action
Generally, what holds decision makers to the course of action is a lack of 
understanding of the true magnitude of the problem. Making costs more 
visible can help force a re-examination of the prior course of action.  During 
this phase, managers redefine the problem in order to change course.  
In the Denver case, a key decision which facilitated progress to the next 
phase was to establish a task force to find a short-term alternative.
3. Search for alternative 
course of action
A decision-maker has to convince other stakeholders of the need to 
change course before the new course of action can be implemented.   
At Denver, an external consultant was hired to identify and legitimize 
an alternative; this made it easier for city officials to reduce their 
commitment and blame the contractor bAe, an easy scapegoat. 
4. Implementing an exit 
strategy
this phase begins with operational plans that emerge as a way to resolve a 
perceived crisis, but subsequently provide a way to retreat from the failing 
course of action.  At Denver, a manual baggage handling system was put 
into place and United Airlines was left on its own to work with bAe on a 
semi-automated system for use on its concourse.  In this way, city officials 
were able to disassociate themselves and the city from the project.
figure 1:  integrated de-escalation management model
134    communications of the acm    |   october 2009  |   vol. 52  |   no. 10
contributed articles
requires a significant investment of ef-
fort. Level 4 requires an open culture 
for discussing problems (which may 
have unintended consequences for or-
ganizational life) and level 5 requires a 
learning culture and potentially expen-
sive procedures such as post-mortems. 
Soft skills are required on both these 
levels to persuade internal and exter-
nal stakeholders to change set ways of 
thinking and embedded procedures. 
Ultimately there is no free lunch here 
and organizations must weigh the 
costs of instituting de-escalation man-
agement procedures against the costs 
associated with runaway projects. 
References
 1. Drummond, H. Deescalation in decision making: a case 
of a disastrous partnership. J. of Management Studies 
32, 3, (1995), 265-281.
 2. Ewusi-Mensah, K. Critical issues in abandoned 
information systems development projects. Comm. 
ACM 40, 9, (1997), 74-80.
 3. Iacovou, C.L. and Dexter, A.S. Managing runaway 
projects: a Delphi survey. Proceedings of the 6th 
European Conference on Information Systems, W. R. 
J. Baets (ed.), Aix-en-Provence, France, (1998), 1140-
1153.
 4. Iacovou, C.L. and Dexter, A.S. Turning around 
runaway information technology projects. California 
Management Review 46, 4, (2004), 68-88.
 5. Keil, M., and Robey, D. Turning around troubled 
software projects: an exploratory study of the 
deescalation of commitment to failing courses of 
action. J of Management Information Systems 15, 4, 
(1999), 63-87.
 6. Keil, M., and Montealegre, R. Cutting your losses: 
extricating your organization when a big project goes 
awry. Sloan Management Review 41, 3, 2000, 55-68.
 7. Keil, M., Mann, J., and Rai, A. Why software projects 
escalate: an empirical analysis and test of four 
theoretical models. MIS Quarterly 24, 4, (Dec. 2000), 
631-664.
 8. Montealegre, R., and Keil, M. De-escalating 
information technology projects: Lessons from the 
Denver international airport. MIS Quarterly 24, 3, 
(2000), 417-447.
 9. Pan, G., Pan, S., Newman, M., and Flynn, D.J. 
Escalation and de-escalation of commitment to 
information systems projects: Insights from a project 
evaluation model. European Journal of Operational 
Research 173, 3, (2006a), 1139-1160.
 10. Pan G, Pan S, Newman M, and Flynn D.J. Escalation 
and de-escalation of commitment: A transformational 
analysis of an e-government project. Information 
Systems Journal 16, (2006b), 3-21.
 11. Staw, B., and Ross, J. Knowing when to pull the plug, 
Harvard Business Review. March-April, (1987), 68-74.
Donal Flynn (donal.flynn@manchester.ac.uk) is a senior 
lecturer in the Manchester Business School, University of 
Manchester, UK
Gary Pan (garypan@smu.edu.sg) is practice assistant 
professor in the School of Accountancy, Singapore 
Management University, Singapore.
Mark Keil (mkeil@gsu.edu) is the Board of Advisors 
Professor of Computer Information Systems, J. Mack 
Robinson College of Business, Georgia State University, 
Atlanta, GA.
Magnus Mähring (magnus.mahring@hhs.se) in an 
assistant professor in the Department of Management 
and Organization, Stockholm School of Economics, 
Stockholm, Sweden and affiliated professor at the Ecole 
de Management Stratsbourg, Stratsbourg, France.
© 2009 ACM 0001-0782/09/1000 $10.00
table 4:  de-escalation management maturity (dmm) model
leVel
model maPPinG descriPtion
Level 1 –  
Discipline to 
change project 
plan 
Maps to step D
organizations must embrace basic project management tools 
and techniques, especially those that focus on project planning.  
organizations that construct a project plan at the outset of a 
project but never make changes to it have not yet reached level 
1 in terms of the DMM model.   
Level 2 –  
Discipline to detect 
deviations from 
project plan and 
prevent escalation 
Maps to steps A 
and b
organizations must embrace project management tools and 
techniques that allow for detection of deviations from project plan 
and proactive prevention of escalation.  organizations that do not 
maintain a project baseline or do not employ techniques such 
as traffic light reporting and earned value analysis have not yet 
reached level 2 in terms of the DMM model.  rigorous monitoring 
and control of projects is the hallmark of level 2 organizations. 
Level 3 –  
Discipline to 
execute project 
plan
Maps to step e
organizations must develop a cadre of excellent project 
managers and become skilled in the art of change management. 
Without these project managers and associated change 
management skills, organizations will be unable to execute 
project plans successfully.  organizations that are not project-
centered and do not have well-functioning project management 
organizations (PMos) to promote best practices have not yet 
reached level 3 in terms of the DMM model.
Level 4 –  
Discipline to 
encourage bad 
news reporting 
and to change 
attitudes and 
behaviors
Maps to steps b 
and c
organizations must develop an environment that encourages 
frank and open communication concerning project status.  once 
problems have surfaced, managers must be able to disrupt any 
groupthink tendencies that have resulted and get people to see 
problems for what they are.   tendencies toward rationalized 
continuation must be challenged by offering alternative courses of 
action that highlight opportunity costs associated with blind pursuit 
of a failing course of action.  organizations that lack free and open 
communication and do not tolerate those who question the status 
quo have not yet reached level 4 in terms of the DMM model.
Level 5 –  
Discipline 
to engage in 
organizational 
learning 
Maps to steps F 
and the feedback 
loop depicted in 
the model
organizations must place a value on organizational learning and 
must develop mechanisms for inter-project learning.  only in this 
way will organizations be able to learn from past mistakes and 
prevent future cases of escalation from occurring.  organizations 
that have not embraced open and honest post-mortem 
evaluations have not reached level 5 in terms of the DMM model