Vous êtes sur la page 1sur 13


From Experience: Linking Projects to Strategy

Randall L. Englund and Robert J. Graham

There is a dramatic rise in the use of project management as organizations shift

to provide customer-driven results and systems solutions. Some implementations
of project management have been successful, whereas others are spectacular
failures. A common occurrence in many organizations is too many projects being
attempted by too few people with no apparent link to strategy or organizational
goals. Research and experience indicate that the support of upper management is
critical to project success. This article reviews actions that upper managers can
take to create an environment for more successful projects in their organizations.
Specifically, the authors discuss practices for upper manager teamwork and offer
a complete model for selecting projects that support a strategic emphasis. The
article includes experiences from within Hewlett-Packard Company. © 1999
Elsevier Science Inc.

Introduction out are suggestions for action as well as guidelines to

navigate many pitfalls along the path. Process tools

rowth in organizations typically results from help illustrate ways to prioritize projects. The lessons
successful projects that generate new prod- learned are from consulting with many firms over a
ucts, services, or procedures. Managers are long time period and from personal experiences in
increasingly concerned about getting better results applying the lessons within Hewlett-Packard Com-
from the projects under way in their organizations and pany (HP), a $40 billion plus company where two
in getting better cross-organizational cooperation. One thirds of its revenue derives from products introduced
of the most vocal complaints of project managers is within the past 2 years.
that projects appear almost randomly. The projects
seem unlinked to a coherent strategy, and people are
unaware of the total number and scope of projects. As
a result, people feel they are working at cross-pur- The Importance of Upper Management
poses, on too many unneeded projects, and on too Teamwork
many projects generally. Selecting projects for their
strategic emphasis helps resolve such feelings and is a Developing cooperation across an organization re-
corner anchor in putting together the pieces of a puzzle quires that upper managers take a systems approach to
that create an environment for successful projects [6]. projects. That means they look at projects as a system
This article covers a series of steps for linking of interrelated activities that combine to achieve a
projects to strategy. These steps constitute a process common goal. The common goal is to fulfill the over-
that can be applied to any endeavor. Included through- all strategy of the organization. Usually all projects
draw from one resource pool, so they interrelate as
Address correspondence to Randy Englund, Hewlett-Packard Com-
they share the same resources. Thus, the system of
pany, 1501 Page Mill Road MS 5MW, Palo Alto, CA 94304. projects is itself a project, with the smaller projects
J PROD INNOV MANAG 1999;16:52– 64
© 1999 Elsevier Science Inc. All rights reserved. 0737-6782/99/$–see front matter
655 Avenue of the Americas, New York, NY 10010 PII S0737-6782(98)00046-0
1999;16:52– 64

A council concept is one mechanism used at HP to

BIOGRAPHICAL SKETCHES establish a strategic direction for projects spanning
Randy Englund is a project manager at Hewlett-Packard Company organizational boundaries. A council may be perma-
(HP) in Palo Alto, California, USA. He is a member of the Project
Management Initiative team that provides corporate-wide leader- nent or temporary, assembled to solve strategic issues.
ship for the continuous improvement of project management across As a result, a council typically will involve upper
the company. Randy joined HP in 1978 and has been a member of managers. Usually its role is to set directions, manage
the Project Management Initiative since 1991. He develops and
facilitates workshops, trains intact teams, and provides consulting
multiple projects or a set of projects, and aid in cross-
on project management practices to product and process developers organizational issue resolution. Several of these coun-
in HP businesses. Randy was a program manager in computer cil-like activities become evident through the exam-
systems and personal computer product development, in systems
marketing, and in manufacturing. He led teams to bring complex
ples in this article.
development systems to market; develop a hardware system prod- One example at HP was a cross-organizational
uct life cycle; resolve computer system architectural issues; and council pulled together to resolve input/output (I/O)
identify, document, and apply best practices. He was a session architectural issues for a new line of computer sys-
keynote speaker at the World Congress on Project Management, a
speaker at PDMA International Conferences, and invited to speak at tems. A computer architecture is the underlying struc-
many other professional conferences. Prior to HP, Randy spent 10 ture that directs hardware to implement software com-
years with General Electric Company in field service engineering. mands. The processor architecture was solid, but
Randy has a B.S. in Electrical Engineering from the University of
California at Santa Barbara and an M.B.A. in Management from
portions of the I/O were vague, broken, or undefined.
San Francisco State University. He is a member of the Project Hundreds of technical issues were logged against the
Management Institute (PMI) and formerly on the Board of Direc- architecture. Individual project teams were optimizing
tors for the Product Development and Management Association
solutions to fit their objectives (organizational subop-
timization). Meanwhile, the overall architecture was at
Bob Graham is currently an independent management consultant in risk of getting out of control. Because this architecture
the areas of project management and organizational change. Previ- provided the structure for a new platform of minicom-
ously he was a senior staff member of The Management and
Behavioral Sciences Center at The Wharton School, University of
puter products, impact on the product family would be
Pennsylvania. While at Wharton he taught in the M.B.A. and Ph.D. enormous.
programs and was a part of the Wharton Effective Executive One project manager (a champion) took the initia-
program teaching Project Management to practicing executives. tive to convene an upper-manager council. The coun-
Bob has held Visiting Professor positions at both the University of
Bath, in England, and the University of the German Armed Forces cil accepted ownership to resolve the set of interrelated
in Munich, Germany. Bob continues as Adjunct Professor at both issues. People accepted membership on the council
the University of Pennsylvania and as a part of the Project Man- because they came to understand the strategic impor-
agement Unit at Henley Management College in England. His
previous book is entitled Project Management as if People Mat-
tance of the mission. The council authorized groups of
tered. His latest book, co-authored with Randy Englund, is entitled engineers to study, propose, review, and accept solu-
Creating an Environment for Successful Projects: The Quest to tions. It first established a set of priorities and con-
Manage Project Management. Bob has B.S. in Systems Analysis
from Miami University, as well as an M.B.A. and Ph.D. in Oper-
straints to guide the study groups. The council met at
ations Research from the University of Cincinnati. He was also a least once a month to review progress and make
Post-Doctoral Fellow at The Wharton School. In addition, he has an changes. When several issues bogged down, it autho-
M.S. in Cultural Anthropology from the University of Pennsylva- rized an escalation path to two managers who would
nia. He earned Project Management Professional (PMP) certifica-
tion from the Project Management Institute. listen to the arguments and make decisions. Because
of the tremendous impact on time to market of projects
dependent on the outcome, the council kept appropri-
ate pressure on making progress. At the end of the
being the activities that lead to the larger project resolution phase, it enthusiastically supported a cele-
(organizational) goal. bration party for the hard work contributed by hun-
Any lack of upper management teamwork reverber- dreds of engineers. They listened to recommendations
ates throughout the organization. If upper managers do from a retrospective analysis and took action on sug-
not model desired behaviors, there is little hope that gested improvements, applying them to subsequent
the rest of the organization can do it for them. Any projects that were initiated to resolve additional issues.
lack of upper management cooperation will surely be Over time the process improved dramatically and led
reflected in the behavior of project teams, and there is to reduced anxiety about the chaotic state of the ar-
little chance that project managers alone can resolve chitecture.
the problems that arise. This systematic approach illustrates the vast and
1999;16:52– 64

important influence of upper management teamwork pays dividends later by getting up-front involvement
on project success. Increasingly evident are companies of the people who will be affected by the decisions that
who convene portfolio selection committees. We sug- will be made. Take care not to overlook any key-but-
gest that organizations begin by developing councils to not-so-visible players who later may speak up and
work with project managers and to implement strat- jeopardize the plan. This team may consist solely of
egy. These councils exercise leadership by articulating upper managers or may include project managers, a
a vision, discussing it with the project managers, ask- general manager, and possibly a customer. Include
ing them their concerns about and needs for imple- representation of those who can best address the key
menting the strategy, listening carefully to them, and opportunities and risks facing the organization. Ideally
showing them respect so they become engaged in the they control the resources and are empowered to make
process. In this way, upper managers and project man- decisions on all projects. The leader needs to get
agers develop the joint vision that is so necessary for explicit commitment from all these people to partici-
implementation of strategy. pate actively in the process and to use the resulting
plan when making related decisions. Be aware that
Process for Project Selection and Prioritization behavioral issues become super urgent. This process
hits close to home and may have a severe impact on
Once the upper management team is established, they
projects that people care personally about. Uncertainty
can follow a process to select sets of projects that
and doubt are created if management does not tread
achieve organizational goals. They are then ideally
carefully and pay attention to people concerns.
positioned to implement consistent priorities across all
The team begins by listing all projects proposed and
departments. Figure 1 represents a mental model of a
under way in the organization. Many times this step is
way to structure this process. Outputs from the four
a revelation in itself. A usual reaction is, “I didn’t
steps interrelate in a true systems approach. This
model comes from experience in researching and ap- realize we had so many projects going on.” The intent
plying a thorough approach to all the issues encoun- is to survey the field of work and begin the organizing
tered in a complex organization. It is both simple in effort, so avoid going into detailed discussion about
concept and complex in richness. The authors use the specific projects at this point.
model both as an educational tool and to facilitate The team clarifies or develops the goals expected
management teams through the process. from projects. Be careful not to get constrained
through considering only current capabilities. Many
What the Organization Should Do and How to teams get sidetracked by statements such as “We don’t
Know When You Are Doing It know how to do that,” effectively curtailing discussion
on whether the organization ought to pursue the goal
First, identify who is leading the process and who and develop or acquire the capability. Rather, the
should be on the management team. More time spent discussions at this stage center around organizational
here putting together a “mission impossible” team purpose, vision, and mission. This is a crucial step that

Figure 1. A systematic approach to selecting projects.

1999;16:52– 64

determines if the rest of the project selection process tioning shown. Although the figure represents a retro-
can be successful. In the authors’ experience, those spective view, it illustrates a successful strategy of
organizations with clear, convincing, and compelling sequencing projects and products. There is a balanced
visions about what they should be doing move ahead mix of breakthrough products, such as A, followed by
rapidly. Any lack of understanding or commitment to enhancements, B through E, before moving on to new
the vision by any member of the team leads to frus- platforms, F through H, and eventually developing a
tration, wheel spinning, and eventual disintegration of new architecture and product family with L. At the
the whole process. This pattern is so prevalent that time, this strategy was improvisational [1]; it now
clarity of the goal or strategy is applied as a filter represents a learning opportunity for planning new
before agreeing to facilitate teams through the process. portfolios. No one area of the grid is overpopulated,
Organize the projects into categories that will later and where large projects exist there are not too many
make it easier to facilitate a decision-making process. of them.
Wheelwright and Clark [14] suggest using grids where Another reason to organize projects into these “stra-
the axes are the extent of product change and the tegic buckets” is to better realize what business(es) the
extent of process change. Some organizations use mar- organization is in. Almost every group the authors
ket segments. The benefit to this effort is that seeing
work with get caught in the “tyranny of the OR”
all projects and possible projects on a continuum al-
instead of embracing the “genius of the AND” [2]. In
lows checking for completeness, gaps, opportunities,
trying to do too many projects and facing the need to
and compliance with strategy. This might also be a
make tradeoffs among them, the decision becomes this
good time to encourage “out-of-the-box” thinking
OR that. In reality, most organizations need a balanced
about new ways to organize the work. Use creative
discussion sessions to capture ideas about core com- portfolio that creates complete solutions for their cus-
petences, competitive advantage, and the like to deter- tomers. They need to do this AND that. The way to
mine a set of categories most effective for the organi- achieve this goal is to set limits on the size of each
zation. For example, the categories might be: category and then focus efforts on selecting the best
set of projects within each category. The collective set
• Evolutionary or derivative—sustaining, incremen- of categories becomes the desired mix, a way of fram-
tal, enhancing. ing the work of the organization. The ideal percentage
• Platform—next generation, highly leveraged, and that constitutes the size of each category can be deter-
• Revolutionary or breakthrough—new core product, mined from the collective wisdom of the team or
process, or business.
perhaps through experimentation. The organization
The actual products in Figure 2 were introduced to can learn the right mix over time but only if it makes
the market over time in alphabetical order and posi- a concerted effort to do so.

Figure 2. Bubble diagram of a product grid for one HP division. Size of bubble 5 size of project.
1999;16:52– 64

Within each category, determine criteria that can only surfaced through group discussion. Asked if the
assess the “goodness”— quality or best fit— of choices selection process ever failed the team, its leader re-
for the plan. A criterion is a standard on which a plied, “If the results didn’t make sense, it was usually
comparative judgment or decision may be based. Be- because the criteria weren’t well defined.” Unfortu-
cause the types of projects and the objectives within nately, most teams do not exhibit the same patience
categories may be quite different, develop unique cri- and discipline that allowed this team to be successful.
teria for each category or have a core set of criteria that One team lost energy at this point; in recognizing
can be modified. Many teams never get to the point of the power that the criteria would exert on project
developing or clarifying criteria, and they usually want selection, team members realized they were still not
to discuss projects before agreeing on criteria; revers- comfortable with the goal. A time-out was taken to
ing the order is much more effective. reassess the team’s approach. It would have been futile
Several works on research and development project to push ahead on details until the big picture was clear.
selection [8,9,12] provide a robust set of criteria for The manager was frustrated that the team did not
consideration. Examples include strategic positioning, achieve consensus on criteria during this session, but
probability of success, market size, and availability of the team truly was not ready. Going through the pro-
staff. Most important is to identify the criteria that are cess with an outside facilitator at an offsite meeting
of greatest significance to the organization; fewer are helps through these rough spots.
better. However, teams usually need to brainstorm Before moving to the next step, the team should
many criteria before focusing on few. establish relative importance among criteria. Assign a
The role of each criterion is to help compare weighting factor for each criterion. All criteria are
projects, not specify them. Select criteria that can important but some more so than others. The example
measurably compare how projects support the organi-
in Figure 3 is the result of one team’s brainstorming
zational strategy. For example, one criterion may be
session that ultimately led to selecting four criteria.
degree of impact on HP business as interpreted by a
Breakout groups subsequently defined each criterion
general manager. On a scaling model from 1 to10,
with subcriteria. They also devised scoring methods to
small impact scores a 2, strong a 6, critical to the
apply the criteria. Collectively they then determined
success of one business an 8, and critical to the success
the respective weighting or importance of each crite-
of multiple businesses a 10. Most likely all proposed
rion (see the Process Tools section for how they did
projects meet meaningful specifications and provide
value to the organization. The task is to develop tough this). Unlike threshold criteria that “gate” whether a
criteria to select the best of the best. project is go or no-go, all projects have to satisfy
Some organizations use narratives to describe how selection criteria to some extent. Weighting of criteria
each project contributes to the vision; others use nu- is the technique that can optimize and determine the
merical scores on whether one project is equal, mod- best of the best. Another “Aha” that helped teams get
erate, or strongly better than another. It is also helpful through the hurdle to develop effective criteria is when
to set thresholds or limits for projects that will be they realized the task at this point is “weighting, not
considered for the plan. These help to screen out gating.”
projects so that later prioritization efforts can focus on It is the authors’ experience that criteria, while
fewer projects. universally desired, are usually lacking or not formal-
Writing a thorough description of each criterion ized. One benefit of effective criteria is the shaping
helps ensure understanding of the intent and expecta- effect it has on behavior in the organization. When
tions of data that must be supplied to fulfill it. One people know how projects will be scored, they tend to
team of three or four people at HP spent 5 days shape proposals in positive ways to meet the criteria
working only on the criteria they were to use for better. A pitfall is when people play games to establish
decision-making. And this was only the beginning; criteria that support personal agendas. Then it is up to
they next involved customers in the same discussion the leader to name and question these tactics. Remind
before reaching consensus and beginning to evaluate people to support the greater good of the organization.
choices. An “Aha” occurred when people found they Significant effort could be devoted to the behavioral
were wrong to assume that everyone meant the same aspects that become relevant when deciding upon cri-
thing by terms such as packaging; some used wider teria; suffice to say, be warned that this is a touchy
definitions than others did, and the misunderstanding area to approach with sensitivity and persuasiveness.
1999;16:52– 64

Figure 3. Sample criteria and weighting, plus subcriteria, developed by one HP team.

What the Organization Can Do current business strategies; maybe the projects were
conceived based on old paradigms about the business.
The next step for the team is to gather data on all The team can save discussion time by identifying
projects. Use similar factors when describing each must-do projects or ones that require simple go/no-go
project in order to ease the evaluation process. Engage decisions, such as legal, personnel, or environmental
people in extensive analysis and debate to get agree- projects. These fall right through the screens and into
ment on the major characteristics for each project. the allocation process. Determine if some projects can
This is a time to ask basic questions about product and be postponed until others are complete or until new
project types and how they contribute to a diversified
resources or funding become available. Can project
set of projects. Reexamine customer needs, future
deliverables be obtained from a supplier or subcon-
trends, commercial opportunities, and new markets.
tractor rather than internally? Involve customers in
The person consolidating the data should challenge
discussions. The team constantly tests project propos-
assertions about benefits and costs instead of accepting
als for alignment with organizational goals.
assumptions that may have been put together casually.
It is not necessary to constrain the process by using
It is important for each member of the team to assess
the quality of the data, looking closely at sources and the same criteria across all categories of projects. In
the techniques for gathering the data. When putting fact, some teams found that different criteria for each
cost figures together, consider using activity-based category of projects was more effective. Also, con-
costing models instead of traditional models based on sider adjusting the weighting of criteria as projects
parts, direct labor, and overhead. Activity-based cost- move through their life cycles. Kumar et al [7] docu-
ing includes the communications, relationship build- mented research showing that the most significant
ing, and indirect labor costs that usually are required to variable for initial screening of projects is the extent to
make a project successful. which “project objectives fit the organization’s global
The team needs to constantly apply screening crite- corporate philosophy and strategy.” Other factors,
ria to reduce the number of projects that will be such as available science and technology, become
analyzed in detail. Identify existing projects that can significant later during the commercial evaluation
be canceled, downscaled, or reconceived because their stage. A big “Aha” experienced by some teams when
resource consumption exceeds initial expectations, confronted with this data is that they usually did it the
costs of materials are higher than expected, or a com- other way around. That explains why they got into
petitive entry to the market changed the rules of the trouble— by focusing on technology or financial fac-
game. The screening process helps eliminate projects tors before determining the link to strategic goals.
that require extensive resources but are not justified by Cooper (and others before him) report that top-
1999;16:52– 64

performing companies do not use financial methods Next, the team identifies the resource capacity both
for portfolio planning. Rather, they use strategic port- within and outside the organization that will be avail-
folio management methods where strategy decides able to do projects. Balance project with nonproject
project selection [3]. This lesson is still a hotly debated work by using realistic numbers for resource availabil-
one, especially for those who cling to net present value ity, taking into account other projects, vacations, meet-
as the single most important criterion. The difficulty ings, personal appointments, and other interruptions.
lies in relying upon forecast numbers that are inher- Tip: a wise planner consumes no more than about 50%
ently fictitious. The authors’ experience is that teams of a person’s available time.
get much better results tapping their collective wisdom One assessment about the quality of projects in a
about the merits of each project based upon tangible portfolio is to look at the rejects. In a story attributed
assessments against strategic goals. Using computed to HP founder Bill Hewlett, he once established a
financial numbers more often lead to arguments about single metric for how he would evaluate a portfolio
computation methods and reliability of the data, lead- manager’s performance. He asked to see only the
ing to unproductive team dynamics. rejects. He reasoned that if the rejects looked good,
The next part of gathering data is to estimate the then the projects that were accepted must be excellent.
time and resources required for each potential and All the actions in this step of the process are in-
existing project. Get the data from past projects, sta- tended to screen many possible projects to find the
tistical projections, or simulations. The HP Project critical few. The team may take a path through mul-
Management Initiative particularly stresses in its or- tiple screens or take multiple passes through screens
ganizational initiatives to get accurate bottom-up with different criteria to come up with a short list of
project data from work breakdown structures and viable projects. Figure 4 represents one scenario where
schedules. Reconcile this data with top-down project Screen 1 is a coarse screen that checks for impact on
goals. Document assumptions so that resource require- the strategic goal. Subsequent screens apply other cri-
ments can be revisited if there are changes to the basis teria when more data is available. Any number of
for an assumption. For new or unknown projects, screens may be applied, up to the number n, until the
make a best estimate, focusing first on the investiga- team is satisfied that the remaining projects relate to
tion phase with the intent to fund only enough work to compelling business needs. These steps actually save
determine feasibility. The team can revisit the esti- time because the next section on analysis can get quite
mates when more information becomes available. extensive if all possible projects go through it.
Constantly improve estimation accuracy over time by It usually is necessary to go through several valida-
tracking actuals with estimated task durations. tion cycles before finishing the next step: the upper

Figure 4. Application of criteria screens during a funneling process eliminates the trivial many projects from the critical few that
the organization can realistically complete.
1999;16:52– 64

management team proposes project objectives, project Some future projects must be funded early to ensure a
teams provide preliminary estimates based on scope, revenue stream when current projects taper off.
schedule, and resources back to management, manage- Using previously agreed-upon criteria and weight-
ment is not happy with this response and makes ad- ing factors, the team compares each project with every
justments, and so on. This exercise in due diligence is other one within a category. Repeat the process for
a healthy negotiation process that results in more re- each criterion. See the discussion and example later in
alistic projects getting through the funnel. this article about using an analytical hierarchy process
(AHP) to facilitate this step. Consider using software
to compute results—an ordered list of projects within
Analyze and Decide on Projects each category. A pitfall to avoid that engenders fear
among the team is showing one list that prioritizes all
The next step is to compare estimated resource re- projects from top to bottom. People get concerned
quirements with available resources. A spreadsheet is when their project is on the line. It is not fair to
useful to depict allocation of resources according to compare internal development projects with high
project priority. grossing products; keep them separated and within
Part of the analysis is qualitative: Consider the their respective categories.
opportunity costs of committing to short-term, oppor- Finally, the team is ready to decide which projects
tunistic, or poorly conceived projects that take re- to pursue. Ask what you should do, not what you can
sources away from future prospects that may be a do. Especially in high-tech industries, people are often
better fit strategically. Also, avoid selecting “glamor- tempted to include a new technology without being
ous” new ideas over addressing the tough issues from sure that customers are interested or will get value
ongoing projects. Some people lack the stamina to from the investment. Be prepared to do fewer projects
deal with the details of implementation and so are and to commit complete resources required by projects
ready to jump to a new solution at the slightest glim- that are selected. Decide on a mix of projects consis-
mer of hope from the latest technology. This is a tent with business strategy, such as 50% platform
recipe for disaster. Also, be careful to balance the projects, 20% derivative projects, 10% breakthrough
important projects rather than giving in to urgent, but projects, and 10% partnerships. Note that these total
not so important, demands. only 90%; taking some lessons from financial portfo-
Documenting all the findings and supportive data lio management, diversify the set of projects by in-
using a common set of descriptive factors makes it vesting in some speculative projects. The team may
easier to compare similar factors across projects. Use not be sure which markets or technologies will grow,
a “project charter” form or a template where all infor- so buy an “option” and make a small investment to
mation about each project, its sponsors, and key char- investigate the possibilities. Include experimental
acteristics is recorded. projects. It is also important to leave a small percent of
The team can now prioritize remaining projects. development capacity uncommitted to take advantage
Focus on project benefits before costs; that way the of unexpected opportunities and to deal with crises
merits of each project get full consideration. Later, when they arise.
include costs to determine the greatest value for the Wheelwright and Clark [14] cite an organization
money. Compute overall return from the set of that reduced the number of its development projects
projects, not from individual projects, because some from 30 to 11: “The changes led to some impressive
projects may have greater strategic than monetary gains...as commercial development productivity im-
value. Requiring each and every project to promise a proved by a factor of three. Fewer products meant
high financial return actually diminishes cooperation more actual work got done, and more work meant
across an organization. For example, a computer sys- more products.” Addressing an internal project man-
tem division depends on an interface card from the agement conference, an HP Executive Vice President
networks division to produce a whole product. But if emphasized the need to focus on doing fewer projects,
the networks division has other priorities, it may not especially those that are large and complex: “We have
commit to developing the card. Such situations require to be very selective. You can manage cross-organiza-
the prioritization process to happen higher in the or- tional complex programs if you don’t have very many.
ganization. Also, optimize return over time and con- If you have a lot of them with our culture, it just won’t
tinuity or uniformity of revenue from the projects. work. First of all, we need to pick those opportunities
1999;16:52– 64

very, very selectively. We need to then manage them under consideration by the entity. If a project is funded
aggressively across the company. That means have and has resources assigned, it has achieved in-plan
joint teams work together, strong project management status. Projects below the cutoff line of available re-
and leadership, constant reviews, a framework, a vi- sources or that have not yet achieved priority status are
sion, a strong owner—all those things that make a on the out-plan. The figure also categorizes the
program and project successful.” Subsequently, a projects and specifies the desired mix.
number of organizations sought help from the HP Project managers at HP describe one benefit of the
Project Management Initiative to systematically re- POR process as identifying gaps between required and
duce 120 projects down to 30. Another organization actual resources. For flexible changes, the process gets
went from 50 projects down to 17. It appears counter- all people into the communications loop. If people
intuitive, but by prioritizing and more carefully select- want to add something, the management team has to
ing projects, organizations actually get more projects decide what should be deleted. The process helps two
completed. divisions that work together agree on one prioritized
Figure 5 illustrates a document that captures the list instead of two. They utilize direct electronic con-
output of this process. Record projects that are fully nections for bottom-up entry of projects and resources
funded in an aggregate project plan (in-plan). In a by all project managers into a centralized administra-
separate section or another document, list projects for tion point.
future consideration (out-plan); also capture and com-
municate reasons for delaying or not funding projects. Implement the Plan
The plan of record (POR) is both a process and a tool
used by some organizations at HP to keep track of the No job is complete until it is acted upon. The team
total list of projects. It lists all projects under way or needs to “evangelize” all others in the organization to

Figure 5. An example plan of record showing the mix of projects in priority order and the time line for each project.
1999;16:52– 64

use the aggregate project plan or POR to guide people ecution. . .again and again and again. . .in the context
who plan work, make decisions, and execute projects. of relentless change. . ..Staying on top means remain-
Although it may be countercultural to do so, do not ing poised on the edges of chaos and time. . .These
starve committed projects of the resources they need. edges are places of adaptive behavior. They are also
The team or the responsible upper managers need to unstable. This instability means that managers have to
enforce the plan by fully staffing committed projects; work at staying on the edge” [1]. The advice is clear:
that now becomes possible because fewer projects are the plan is indispensable as a strategic guideline, but
happening simultaneously. Also, use the plan to iden- don’t fall in love with it! Be prepared to adapt it and
tify opportunities for leverage across projects or for to communicate the changes.
process reengineering. Match people skills to project
categories to tap their strengths and areas for contri- Process Tools
The team or a program management office needs to One tool that can assist in the decision-making process
maintain the plan in a central place, such as a project is the AHP [10]. Because of the interactions among
office or online. Make it known to, and accessible by, many factors affecting a complex decision, it is essen-
all people in the organization doing projects, subject to tial to identify the important factors and the degree that
confidentiality requirements. All the work to this point they affect each other before a clear decision can be
may go for naught if the process, the steps, and the made. The AHP helps structure a complex situation,
results are not widely communicated. identify its criteria and other intangible or concrete
The same people who develop the plan are also the factors, measure the interactions among them in a
ones who can best update it periodically, perhaps simple way, and synthesize all the information to
quarterly or as changes occur. Use tools such as an obtain priorities. The priorities then can be used in a
online shared database to gather data directly from benefit-to-cost determination to decide which projects
project managers about resources needed for each to select. The AHP organizes feelings and intuition
project. This system can be used both to gather data alongside logic in a structured approach to decision-
when developing the plan and to update it. View the making— helpful in complex situations where it is
plan as a “living document” that accurately reflects difficult to comprehend multiple variables together.
current realities. An individual or team focuses on one criterion at a
The challenge for HP and many companies is to time and applies it step by step across alternatives. A
“master both adaptive innovation and consistent ex- number of sites across HP find value in using AHP.

Figure 6. An analytical hierarchy process, with definitions on the left and examples on the right.
1999;16:52– 64

Figure 6 depicts the structure of the hierarchy. The The key benefit in doing this process is the im-
examples along the right-hand side represent choices proved quality of dialogue that occurs among the
available to the HP Project Management Initiative. management team members. In facilitating a number
In another example, a team got together to choose of teams at HP through this process, each one achieved
among a set of services they will offer to customers. far more progress than they thought possible. People
More choices were available than the organization had admit that they become addicted to the AHP process.
capacity to support. After defining organizational They immediately buy the software. The systematic
strategy or product goals, the first task was to identify approach is feasible whether selecting products for a
which criteria to enter into the decision-making pro- product line, projects that comprise a portfolio, or the
cess. After give-and-take discussion, they decided that best supplier or candidate for a job. In reality, the
the criteria are customer satisfaction, business value, discussions are more valuable than the analysis. The
process effectiveness, and employee satisfaction. process in this case provides the discipline that makes
Next, the criteria were ranked according to priority the dialogue happen.
by making pairwise comparisons between them. Frame [5] offers an alternative “poor man’s hierar-
Which is the more desirable criterion and by how chy.” He puts selection criteria along the side as well
much, customer satisfaction or business value? Pro- as across the top of a grid. If the criterion on the side
cess effectiveness or employee satisfaction? Business is preferred to the one on the top, put a 1 in the cell. If
value or process effectiveness? These questions were the criterion on top is preferred, put a 0 in the cell.
asked about all possible pairs. Diagonals are blanked out where criteria would be
Each potential project or service then was scored compared to themselves. Below the diagonal, put the
underneath each criterion, and decisions were made opposite value from corresponding cells above the
about which projects to include in the portfolio, based diagonal. Then add up the numbers across the rows to
upon existing resources. This team went on to create a get total scores, which provide a rank order. One team
POR similar to Figure 5. at HP modified this process to replace the 1s and 0s
A detailed explanation for computing the priority with an actual count of how 18 people voted in each
scores and the final rank ordering list can be quite pairwise comparison of alternatives. Again, they
complex, involving eigenvalues and eigenvectors, so it added up the rows and normalized the results for a
is much easier to get a software package (Expert priority order and weighted ranking (Figure 7).
Choice [4]) that does the computations. As an alter- This simplified hierarchy is especially helpful for
native, a spreadsheet could be constructed to normal- weighting criteria. It can be used for prioritizing
ize the numbers. projects when applied to one criterion at a time. It
This process appears complex and analytical but is becomes bulky and less useful when applied to mul-
easy when the software handles the computations, and tiple projects over multiple criteria.
the management team concentrates on the compari-
sons. It is thorough in guiding the team to consider all
criteria, both emotional and logical, and to apply them Barriers to Implementation
to all projects. One team rejected the process as too
analytical, so be aware that it does not work for Now for a reality check. The model depicted in this
everyone. article is thorough, and it integrates objective and

Figure 7. A simplified hierarchy used by one HP team to weight criteria.

1999;16:52– 64

subjective data. When all is said and done, however, across the organization. Believe that change is possi-
people may throw out the results and make a different ble.
decision. Sometimes the reason is a hunch, an instinct, When the effort appears too massive, one approach
or simply a desire to try something different. Some- is to go after the low-hanging fruit. Start with one of
times people have a pet project and use the process to the more pressing issues and use the general concepts
justify its existence, or a hidden agenda may be at of this model to address it. Still have a vision for what
play—perhaps the need to maneuver among col- the organization ultimately can achieve but understand
leagues, trading projects for favors. Politics at this that patience and pacing are necessary to get there.
stage cannot be ignored, nor are they likely to disap- Consider also that this process is hierarchical—it
pear. It is imperative for leaders to become skilled in can be applied singularly or collectively, up or down
the political process. Any attempt at leading change in the organization. A mental model of linking projects to
how an organization links projects to strategy is bound strategy is like fractals and chaos theory. As a viewer
to meet resistance. The concept receives almost unan- moves through layers, each is a reduced-size copy of
imous intellectual support. Implementing it into the the whole, exhibiting all its similar but chaotic traits—
heart and soul of all people in the organization is unpredictable and sensitive to small changes. The
another story. It goes against the cultural norms in leader invoking this process in action experiences both
many organizations and conjures up all kinds of resis- order and disorder. Behavioral patterns appear in ir-
tance if the values it espouses are not the norm in that regular but similar forms. Amidst unpredictable ac-
organization. The path is full of pitfalls, especially if tions, however, we find patterns of similar behavior.
information is presented carelessly or perceived as The process or the behaviors do not vary across the
final when it is work in process. layers as much as the type of projects on which they
Some people resist because the process is too ana- are utilized. It is not necessary for every level in an
lytical. Some want decision-making to be purely in- organization to apply the process, but it is much more
teractive, intuitive, or the purview of a few people. A effective if they do. Be accountable to take action
complete process cannot be forced upon people if the where you are. Expect to work within a realm of
organization has more immediate concerns or unre- “bounded instability” [11]. Each and every team, in-
solved issues. Resistance occurs when there is no dividual, or organization benefits by using the process.
strategy, the strategy is unclear, or people are uncom- For people who get frustrated when all linkages are
fortable with the strategy. Work on the process may not present, the authors urge teams and individuals to
come to a standstill when people realize how much invoke the power of one and “just do it.” Small
work is involved to fully link projects to strategy. If changes in initial conditions have enormous conse-
the pain is not great enough with the status quo, people quences. Eventually successes or small wins are no-
are not ready to change. ticed. The practices start to permeate an organization.
When people sense that the leader does not authen- This can happen in the middle, move up, and then over
tically believe in any of the elements, such as the to other organizations. Incidentally, a corporate group
goals, the process, or the tools, they are hesitant to like HP’s Project Management Initiative helps facili-
follow with any enthusiasm. When the leader lacks tate this transformation. We do this by acting as a
integrity and exhibits incongruity between words and conduit for success stories and best practices.
actions, people may go through the motions but do not Over the long run, we believe that organizations that
exert an effort that achieves meaningful results. follow a process similar to the one described increase
their odds for greater success. This happens because
teams of people following a systematic process and
Enablers for Effective Implementation using convincing data to support their arguments more
often produce better results than individuals. Their
It is possible to lead people through this change pro- projects have more visibility, and the quality of dia-
cess if the leader asks many questions, listens to the logue and decision-making improve. The power of
concerns of all people involved, and seeks to build using criteria that are tightly linked with strategy and
support so that people feel they have an active role in known by everyone in the organization is the mitigat-
developing the process [9]. A flexible process works ing effect it has to guide behavior in constructive
better than a rigid one. Cultivate “champions” who ways. Having a process means it can be replicated and
have the credibility and fortitude to carry the process improved over time until it is optimized. It also means
1999;16:52– 64

other people can learn the process and coach others, • Creates a model for linking projects to strategy and
thereby creating a learning organization. supports it with authenticity and integrity.
In summary, the successful complete upper man-
ager: References
• Knows that projects without strategic emphasis of- 1. Brown, S. L. and Eisenhardt, K. M. Competing on the Edge: Strategy
as Structured Chaos. Boston: Harvard Business School Press, 1998.
ten end in failure. 2. Collins, J. C. and Porras, J. I. Built to Last: Successful Habits of
• Develops an upper management team to oversee Visionary Companies. New York: HarperCollins, 1994.
project selection. 3. Cooper, R. G., Edgett, S. J. and Kleinschmidt, E. J. Portfolio Man-
agement for New Products. Reading, MA: Addison-Wesley, 1998.
• Focuses on the goals of what an organization 4. “Expert Choice.” Pittsburgh, PA: Expert Choice Inc. (see www.exp-
should do before limiting choices by considering ertchoice.com).
only what the organization is capable of doing. 5. Frame, J. D. The New Project Management: Tools for an Age of Rapid
• Works to develop a system of projects and links Change, Corporate Reengineering, and Other Business Realities. San
Francisco: Jossey-Bass Publishers, 1994.
them to organizational strategy. 6. Graham, Robert J. and Englund, Randall L. Creating an Environment
• Guides the development of consistent criteria that for Successful Projects: The Quest to Manage Project Management.
San Francisco: Jossey-Bass Publishers, 1997.
are used to prioritize projects.
7. Kumar, V., et al. To terminate or not an ongoing R&D project: A
• Selects projects based on comparative priority rank- managerial dilemma. IEEE Transactions on Engineering Management
ing of contribution to strategy. 279 (1996).
• Reduces the total number of projects to minimize 8. Martino, J. R&D Project Selection. New York: Wiley, 1995.
possible disruption. 9. O’Toole, J. Leading Change: Overcoming the Ideology of Comfort
and the Tyranny of Custom. San Francisco: Jossey-Bass Publishers,
• Knows that a system of projects utilizes a common 1995.
resource pool and that the pool may be abused 10. Saaty, T. L. Decision Making for Leaders. Pittsburgh, PA: RWS, 1990.
without cooperation across the organization. 11. Stacey, R. D. Managing the Unknowable: Strategic Boundaries Be-
tween Order and Chaos in Organizations. San Francisco: Jossey-Bass
• Develops a system to manage the resource pool and Publishers, 1992, p. 62.
reward interdepartmental cooperation. 12. Turtle, Q. C. Implementing Concurrent Project Management. Engle-
• Allows unallocated capacity in a resource pool for wood Cliffs, NJ: Prentice Hall, 1994.
emergencies and for creativity. 13. Westney, R. E. Computerized Management of Multiple Small
Projects. New York: Dekker, 1992.
• Believes in the power generated by a learning or- 14. Wheelwright, Stephen and Clark, Kim. Creating project plans to focus
ganization. product development. Harvard Business Review March–April (1992).