90 CHAPTER 2 / STRATEG IC MANAGEMENT AND PROJECT SE LECTION

D IRE C TED ·R E A DIN G

FROM EXI'EIUENCE: LINKING
PROJECTS TO STRAfhGY*
R. L. Englund and R. J. Graham

Grow th in organi zations typically results from successful manage ment cooperatio n will sure ly be reflected in the
projects that generate new products, services, or procedures. behavior of project lea rn s, and there is little chance that
Managers are increasingly concerned about getting bener project managers alone can resolve the problems that ari se.
resuhs from the projects under way in their organi zations and A council concept is one mechanism used at HP to
in gelling beHer cross-organizati onal cooperation. One o f the establi sh a strategic direction for projects spanning organi -
most vocal complaints of project managers is that projects zatio nal boundaries. A counci l may be permanent or tem -
appear almost randomly. The projects seem unlinked to a porary. asse mbled to so lve strategic issues. As a resu lt, a
coherent strategy, and people are unaware of the tota l counci l typ ica lly will in volve upper managers. Usually its
number and scope of projects. As a result, people feci they ro le is to set direc ti ons, manage mU ltiple projects or a set
arc working at cross- purposes, on [00 many unneeded proj - of projects, and aid in cross-organ izationa l issue resolu-
ects, and on too many projects generally. Selecting projects ti on. Several of these council -like acti viti es become evident
for their strategic emphasis helps reso lve such feelings and through the examples in thi s article.
• is a comer anchor in putting together the pieces of a puzzle Employing a co mprehensive and systemati c approach
that create an environment for successful projects [6]. illustrates the vast and important influence of upper man-
This artic le covers a series of ste ps for linking projects to agement tea mwork on project success. Increas ing ly evident
strategy. These steps constitute a process that can be applied are co mpan ies who initiate portfo li o se lection committees.
to any endeavor. Incl uded throughout are suggestions for We suggest that organi zati ons begin by developing councils
action as well as guidelines to navigate many pitfall s along to work wi th project managers and to implement strategy.
the path. Process too ls he lp illustrate ways to prioritize proj- These councils exercise leadership by articulating a vision,
ects. The lessons learned arc from consulting with many fi rms discu ss ing it with the project managers. asking them their
over a long time period and from personal experiences in conce rn s about and needs for implementing the strategy, li s-
applying the lessons within Hewlett-Packard Co mpany (HP), tening carefully to them. and show ing them respec t so they
a $40 billion plus company where two thirds of its revenue become engaged in the process. In thi s way, upper manag-
derives from products introduced within the past 2 years. ers and project managers deve lop the joint vision that is so
necessary for implementation of strategy.

fhl' I mporr.Uh.T of { ppcr j\lanagc.·llll·nt

L'Cl1ll\\ ork
O nce the upper manage ment tea m is established, they
Deve loping cooperatio n across an o rgani zati on requires ca n follow a process to se lect sets of projects that achieve
that upper managers take a systems approach to projects. organi zati onal goa ls. They are then ideally positioned to
That means they look at projects as a system o f interre- impl ement co nsistent priorities across all departments.
lated activities that combine to achieve a co mmon goa l. The Figure I represents a mental model of a way to structure
co mmon goal is to fulfill the ove rall strategy of the orga- thi s process. O utputs frorn the four steps interrelate in a
ni zati on. Usuall y all projects draw from one resource pool. true systems approach. This model comes from experience
so they intcrre late as they share the sa me resources. Thus. in researching and applying a thorough approach to all the
the system of projects is itse lf a proj ect, with the s maller issues encountered in a complex organ ization. It is both
projects being the activities that lead to the larger project simple in concept and complex in ri chness. The authors use
(o rgani zatio nal ) goal. the mode l both as an educational tool and to facilitate man -
Any lack of upper managemcnt tea mwo rk reverbe r- agement teams through the process.
ates throughout the organizati on. If upper managers do not
, , - . / )' . I
model desired behaviors. there is little hope th at the rest H'ih •• 'H..
of the organ ization can do it for them. Any lack of upper 1\7:,' t /)"" lIT I' First, identify who is leadi ng the

*Reprinted from Journal of ProdUCI/IIII0YlIIiOll Managemelll with permission. Co pyright Elsevier Science Publishers.
 DIRECTE D READING )1

/
• Use
• Fully fund
, l._t
should do
" •
•
People
Goals
Categor,es
• Communicate '- ./ •
• Update • Criteria

/ ""\ / ""\
2._t
4 Do it! can do
'- ./ ./
Rejects
• Prioritized list
• DeSi"red mix
/
3. Il<cide " • Ust projeclS
• Requirements•
• Decision • Capac ity
• In-pIan '- ./
• Critical few Fi~tlrt' I A systematic approach to selec ting
I Outillan projects.

process and who should be on the management team. More of the team leads to fru stration. wheel spinning. and eve ntual
li me spent here putting together a " mission impossible" team disintegration of the whole process. This pattern is so preva-
pays di vidends later by gelling up-front involvement of the lent that clarity of the goal or strategy is applied as a filter
people who will be affected by the decisions that will be before agreeing to facilitate teams through the process.
made. Take care no t to overlook any key-but-nol-so-visible Organize the projects into categories that will later
players who later may speak up and jeopardize the plan. make it easier to faci litate a decision-making process.
This team may consist solely of upper managers or may Wheelwright and C lark [14J suggest using grids where the
include project managers, a general manager, and possibly axes are the extent of product change and the ex tent of pro-
a custo mer. Include representation of those who can best cess change. Some organizati ons use market segments. The
address the key opportunities and ri sks facing the organiza- benefit to this effort is that seeing all projects and possible
tion. Idea lly they control the resources and are empowered projects on a continuum allows checking for completeness,
to make decisions on all projects. The leader needs to get gaps, opportunities. and compliance with strategy. This
explicit commitment from all these people to participate might also be a good time to encourage "out-of-the-box"
actively in the process and to use the resulting plan when thinking about new ways to organize the work. Use cre-
making related decisions. Be aware that behavioral issues ative di scussion sessions to capture ideas about core com-
beco me super urgent. This process hits close to home and petences, competitive advantage, and the like to determine
may have a se vere impact on projects that people care per- a set of categories most effective for the organization. For
sonally about. Uncertainty and doubt are created if manage- example, the categories might be:
ment does not tread carefull y and pay attention to people
concerns. Evolutionary or deri vative-sustai ning, incre mental,
The team beg ins by listing a ll projects proposed and enhancing.
•
under way in the organi zati on. Many times thi s step is Platform- nex t generation, highl y leveraged; and
a revelation in itself. A usual reac tion is, "I didn ' t realize Revolutionary or breakthrou gh- new core product,
we had so many projects going on." The intent is to survey process, or business.
the fi eld of work and begin the organi zing effort, so avoid
goi ng inlO detailed di scussion about specific projects at this The actual products in Figure 2 were introduced to the
point. market over lime in alphabetical order and positioning
The team clarifies or develops the goals expected from shown. Although the figure represents a retrospective view,
projects. Be careful not to get constrained through consid- it illustrates a successful strategy of sequencing projec ts and
ering only current capabilities. Many teams get sidetracked products. There is a balanced mix of breakthrough products,
by statements such as "We don't know how to do that," such as A, followed by enhancements, B through E, before
effectively curtailing di scussion on whether the organiza- moving on to new platforms, F through H, and eventually
tion ought to pursue the goal and develop or acquire the developing a new architecture and product family with L. At
ca pability. Rather, the discussions at this stage center around the time, thi s strategy was improvisational [1] ; it now rep-
organizational purpose. vision, and mission. This is a crucial resents a learning opportunity for planning new portfolios.
step that detennines if the rest of the project selection process No one area of the grid is overpopulated, and where large
can be successful. In the authors' experience, those organi- projects exist there are not too many of them.
zations with clear, convincing, and compelling visions about Another reason to organize projects inlO these "strategic
what they should be doing move ahead rapidly. Any lack of buckets" is to better realize what business(es) the organi -
understanding or commitment to the vision by a member zation is in. Almost every group the authors work with get

•
 •

<;2 C HAPTER 2 / STRATEG IC MANAG EME NT AND PROJ ECT SELECTION

Extent of Product Chance

New Enhancement
New Breakthroughs
,r ,
I

L'<. ,p o Next generation

or platform
K
A ./
Extent of
Proce..
Ch.n,e GO
/"",
D 10rE
O
Enhancements,
hybrids , and
./
J
cO derivatives

DO Bubble diagram of a produc t grid for one HP

Incremental di vision. Size of bu bble = size of project.

caught in the "tyranny of the OR" instead of e mbraci ng model from 1 to 10. small impact scores a 2. strong a 6.
lhe "geniu s of Ihe AND" [2]. In trying lO do lOO man y proj- critical to the success of one busi ne ss an 8. and critical to
ects and facing the need to make tradeoffs among them. the success of multiple businesses a 10. Most likely all pro-
• the deci sion becomes this OR tha t. In reality, most orga- posed projects mee t meaningful spccillcati ons and provide
ni zatio ns need a balanced portfo li o that creates co mplete val ue to the orga ni zatio n. The task is to develop toug h cri te-
solutions for their customers. They need to do this AND ria to select the best of the best.
that. The way 10 achieve thi s goal is to set limits on the size Some orga ni zation s use narratives to describe how eac h
of eac h category and then foc us efforts o n selecting the best project contributes to th e vision; others use numeri cal scores
set of projects within each category. The collecti ve set of o n whether o ne project is equal. modera te. or strongly bet-
catego ri es becomes the desired mix. a way of framing the ter than another. It is also helpful {Q se t thrl!s holds or limits
work of the orga ni zation. The ideal percentage that co nsti- for projects that will be considered for the plan. These help
tutes the size of each category can be determined from the to sc ree n out projects so that later prioriti zati on effort s can
collec ti ve wisdom of the tea m or perhaps through experi - focus o n fewer projects.
mentation . The organi zation can learn the right mix over Writin g a thorough descripti on of each criteri on helps
time but o nl y if it makes a concerted effo rt to do so. ensure understanding of the intent and expec tations of data
Within eac h category, determine criteri a that can assess th at mu st be suppli ed to fulfill it. One tea m of three or four
lhe "goodness"---<l ualily o r beSI fil-of choices for the plan. people at HP spent 5 days work ing only o n thl! criteria they
A criteri on is a standard o n which a co mparati ve judgment were to use for deci sion-mak ing. And this was onl y the
or decision may be based. Because the types of projects beginni ng; they next in volved custo mers in the sa me di sc us-
and the objecti ves within categori es may be quite different. sion before reaching consensus and bt!gin ning to evaluate
deve lo p uniqu e criteria for each category or have a core choices. An "Aha" occ urred when people found they werl!
set of criteri a that can be modified. Many teams neve r get wrong to assume that everyo ne meant the sa me thing by
to the poi nt of develo ping or clarifyi ng crite ria. and they terms suc h as packaging; so me used wider de finiti ons than
usuall y want to discuss projects before ag reei ng o n criteria; others did. and the mi sunderstanding onl y surfaced th rough
reversing the order is much more effective . group discussion. Asked if the selectio n process eve r fail ed
Several works o n research and deve lo pment proj- the learn, its leader replied, " If the results didn ' t make se nse.
ect selectio n [8.9. 12] provide a robust set of criteria fo r it was usua lly because the criteria were n' t we ll defined."
considerati on. Examples inc lude strateg ic positio ning. Unfortunately, most team s do not exhibit the sa me patience
probability of success, market size, and availability of staff. and discipline that allowed thi s team to be successful.
Most im portant is to identify the criteria that are of greatest Before movin g to the nex t st.ep. the team should estab-
sign ifi cance to the organi zation; fewer are better. However. lish relative importance among criteria. Assign a weight -
teams usually need to brainstorm man y criteri a before ing fac tor for eac h criteri on. All criteria are important but
focusing on the few. some more so th an others. The exa mple in Figure 3 is the
The role of each criterio n is to help compare projects. result of one team's brainstormin g sessio n th at ultimate ly
not speci fy them. Selec t criteria that can measurabl y co m- led to se lecting fo ur criteria. Breakout grou ps subsequentl y
pare how projects support the organizational strategy. Fo r defined eac h cri terion with subcriteria. T hey also devised
example. o ne criteri on may be degree of impact on HP scoring methods to apply the criteria. Co llectivel y they then
business as interpreted by a general manager. On a scaling determined the respec ti ve weighti ng o r importance of eac h
 DI RECTE D REA DI NG J 4

Customer Satisfaction (28%) Employee Satisfaction (7 %)

• Improves service levels • Improves employee knowledge
• Results in more consisll! nt and acc urate • In creases e mployee effici ency or
i nformati0 nltransactions etTecti ve ness
• He lps ensure services are deli vered as • Impro ves work/life balance promi sed
ex pected • Positi ve impact to employee survey
• Helps balance workload

Business Value (46 %) Process Effectiveness (19 %)

• Achieves results th at are critical for a • Enables e mployees to do things ri ght
specific window of opportunity the first time
• Minimi zes ri sk for imple mentation and • Increases the use of technology for
ongoing sustainability se rvice deli very
• Improves integ rati on and relationships • Reduces manual work and non-value
with partners added acti vities
• Provides a posi ti ve ROI in < 2 yrs • Increases empl oyee self-suffici ency
• Ali gns with business goal s
ill { Sample criteri a and wei ghting, plus subcriteri a, developed by one HP team.

criteri on (see the Process Tools sec ti on ror how they did and costs instead of acce pti ng assumptions that may ha ve
thi s). Unlike threshold criteria that "gate" whether a projec t bee n put together casuall y. It is important for each me mber
is go or no-go, all projec ts ha ve to sati sfy selecti on criteria of the team to assess the quality of the data. looking closely
to some extent. Weighting of criteria is the technique th at at sources and the tec hniques for gathering the data. When
can optimize and determine the best of the best. Anothe r pUlling cost fi gures together. consider using acti vity-based
"Aha" lhal helped learn s gel lhrou gh lhe hu rdle 10 develo p costing models instead of traditi onal models based on parts.
effecti ve criteria is when they reali zed the task at thi s point direct labor, and ove rhead. Activity-based costing includes
is "weighting. not gating." the communi cati ons. relationship building. and indirect
It is the authors' experience that criteria. while uni - labor costs that usuall y are required to make a project
versally desired, are usuall y lacking or not formali zed. successful.
One bene fit of effecti ve criteri a is the shaping effect it has The team needs to constantl y appl y sc ree nin g criteria
on behav ior in the organi zat ion. When people know how to reduce the num ber of projects that will be analyzed in
projects will be scored, they tend to shape proposal s in detail. Identify existing projects that can be canceled, down - .
positi ve ways to meet the criteria bener. A pitfall is when scaled, or reconceived becau se their resource consumption
people play games to establi sh criteria that support personal exceeds initial expectati ons. costs of material s are hi gher
agendas. Then it is up to the leader to identify and questi on than ex pec ted, or a co mpetiti ve entry to the market changed
these tactics. Remind people to support the greater good the rules of the gamc. The sc reening process helps eliminate
of the organi zati on. Signifi cant effort could be devoted to projec ts that require ex tensive resources but are not justi -
the be havioral aspec ts that become relevant when deciding fied by current busi ness strategies; maybe the projects were
upon criteria; suffice to say. be warned th at thi s is a touchy co nceived based on old paradi gms about the business. The
area to approach with se nsiti vity and persuasiveness. team can save di sc ussion time by identifyin g must-do proj-
ec ts or ones that require simple go/no-go decisions, such
\ The nex t step for the as legal. personnel. or environmental projects. These fall
team is to gather data on all projec ts. Use similar factors ri ght th rough the sc ree ns and into the allocati on process.
when desc ribing each project in order to ease the evaluation Determine if some projec ts ca n be poStponed until others
process. Engage people in ex tensive anal ys is and debate to are complete or until new resources or funding beco me
get agree ment on the major characteri stics for each proj- av ailahle. Can projec t deli ve rables be obtained from a
ect. This is a time to ask basic questi ons about product and suppli er or subcontractor rather than internall y? ln vo lve
project types and how they contribute to a diversifi ed set of customers in di sc ussions. The team constantl y tests project
projects. Recxamine customer needs, future trends, co m- proposal s for ali gnment with organi zati onal goals.
mercial opportun ities. and new markets. The person consol- It is not necessary to constrain the process by using
idating the data should challenge asse rti ons about benefits the same criteria across all categori es of projects. In fact.

•
 •

'J-! C HAPTER 2 / STRATEGI C MANAGEMENT AND PROJ ECT SELECTION

so me learns found that different cri teria for eac h category the rej ects looked good, then the projects that were acce pted
o f projects was more effective. Al so, consider adjusting must be excellent.
the weighting of criteria as proj ects move throug h their All the actions in this step of the process are intended to
life cycles. Kumar el al. [71 documented research show- screen many possible projects to find the critical few. The
ing that .the mOSI significanr variable for initial sc ree ning team may take a path through muhiple screens or take mul-
of projec ts is the ex tent to which "project objectives fit the tiple passes through sc reens with different criteria to come
o rgani zation's g loba l corporate philosophy and strategy." up with a short list of viable projec ts. Figure 4 represents
Ot her fa crors, such as availabie science and technology, one scenario where Screen I is a coarse sc reen that chec ks
become significant later during the commercial evaluation for impact on the strategic goa l. Subsequent screens apply
stage. A big " Aha" experienced by some tca ms when con- other cri teria when more data are available. Any number of
fronted with thi s data is that they usually did it the other screens may be applied. up to the num ber II. until the team
way around . That explains why they got into trouble- by is satisfied that the remainin g projects relate to co mpelling
focusing on tec hnology or financial factors before determin - business needs. These steps actually save time because the
ing the link lO stralegic goals. next section on analysis can gel quite extensive if all pos-
Cooper (and others before him) report that top-perfoml- sible projects go through it.
ing companies do nOl use financial methods for portfolio It usually is necessary to go through seve ral validation
planning. Rather, they use strategic ponfolio management cyc les before finishing the next step: the upJX!r management
methods where strategy decides project selection [3]. This les- team proposes project objecti ves, project teams provide pre-
son is still a hotly debated one, eSJX!cially for those who cling liminary esti mates based on scope. schedule, and resources
to net prese nt value as the single most important criterion. The back to management. management is not happy with this
difficu lty lies in relying upon forecast numbers that are inher- response and makes adjustments, and so on. This exercise in
ently fi ctitious. The authors' eXJX!rience is that teams get much due dili gence is a healthy negotiation process that results
better resu lts tapping their collective wisdom about the merits in more realisti c projects getting through the funnel.
of each project based upon tangible assessments against stra-
tegic goals. Using computed financial numbers more often 'lJilt, it "i 1h. zd( I The next step is to
J',..(1/1 r:tJ.
leads to argu ments about computation methods and reliability compare estimated resource requirements with avai lable
of the data, resu lting in unproductive team dynamics. resources. A spreadsheet is useful to depict allocation of
The nex t part of gathering data is to estimate the time resources according to project priority.
and resources required for eac h potenti al and existing proj- Part of the analysis is qualitative: Consider the oppor-
ect. Get the data from past projects, stati stical projecti ons, tunity costs of committing to short-term, opportunistic, or
or simulations. The HP Project Management Initiative par- poorly conceived projects that take resources away from
ti cularly stresses in its organizati onal initiatives to get accu-
rate bottom-up project data from work breakdow n structures
and sc hedules. Reconcile this data with top-down projec t
goals. Document assumptions so that resource require-
ments can be revisited if there are changes to the basis for
an assumption. For new or unknown projects, make a best
Many choices (projects)
o 0
csti malc. foc using fi rst on the investigati on phase with the
intent to fund only enough work to determine feasibility. Screen 1
Screen 1: fit to goals (criteria)
The tea m can revisit the estimates when more information
becomes availab le. Constantly improve estimation acc uracy o 0 0
ove r time by tracking actuals with estimated task durations. Screen 2: market too
small, no competence,
00 0 Screen 2
Nex t, the team identifies the resource capaci ty both within partner available?
and outside the organization that will be available to do proj-
ects. Balance project with nonprojcct work by using realistic Screen n : technology
numbers for resource availability, taking into account other fit, breakthrough, Screen n
marketing effort.
projects. vacations, meetings, JX!rsonal appointments, and
other interruptions. Tip: a wise planner consumes no more
than about 50% of a JX!rson's available time.
One assess ment about the quality of projects in a port-
folio is to look at the rejects. In a Story attributed to HP
o 0 The Critical Few!

founder Bill Hewlett, he once established a single metric I· i~Hn' ~ Application of criteria scree ns during a funneling
for how he wou ld evaluate a portfolio manager's JX!rfor- process eliminates the trivial mall)' projects from the critical
mance. He asked to see onl y the rejects. He reasoned that if few that the organization can realistically complete.
 D IRECTE D READI NG '1-"

future prospects that may be a better fit strategically. Al so, Wheelwright and C lark [14] cite an organi zatio n that
avoid selecting "glamorous" new ideas over addressing reduced the number of its development projects from 30
the tough issues from ongoing projects. Some people lack the to 11 : "The changes led to some impressive gains ... as
stamina to deal with the details of implementati on and so are co mmercial development productivity improved by a factor
ready to jump to a new solution at the sli ghtest glimmer o f of three. Fewer products meant more actual work got done,
hope from the latest techno logy. This is a recipe for disaster. and more work meant mo re products." Addressing an inter-
Also, be careful to balance the important projects rather than nal project management conference, an HP Executive Vice
giving in to urge nt. but not so important, demands. President emphasized the need to focus on doing fewer proj-
Documenting all the findin gs and supporti ve data using ects, especially those that are large and complex: "We have
a common set o f descripti ve factors makes it easier to com- to be very selecti ve. You can manage cross-organizational
pare similar factors ac ross projects. Use a "project charter" complex programs if you don' t have very many. If you have
form or a template where all infonnation about each proj- a lot of them with our culture, it just won' t worle First of all,
ect, its sponsors. and key characteri stics is recorded. we need to pick those opJX>rtunities very, very selectively. We
The team can now prioriti ze the remaining projects. need to then manage them aggressively across the company.
Foc us on project benefits befo re costs; that way the merits That means have joint teams work together, strong projec t
of each project get full considerati on. Later include costs management and leadership , constant reviews, a frame-
to determine the greatest va lue for the money. Co mpute work, a vi sion, a strong owner- all those things that make a
overall return from the set of projects, not from indi vidual prog ram and project success ful. " Subsequentl y, a number of
projects, becau se some projects may ha ve greater strategic organi zati ons sought help from the HP Project Management
than monetary value. Requiring each and every proj ect to Initi ative to systematicall y reduce 120 projects down to 30.
promi se a hi gh financial return actually dimini shes coopera- Another organi zatio n went from 50 projects down to 17. It
tion across an organi zatio n. Al so, optimize return over time appears counter-intuiti ve, but by pri oriti zing and mo re care-
and continuity or uniformity of revenue from the projects. fully selec ting projects, organi zati o ns actuall y get more
Some future projects mu st be fund ed early to ensure a rev- projects compl eted.
enue stream when current projects taper off. Figure 5 illustrates a document that captures the output
Using previously agreed-u pon criteria and weighting of this process. Record projects that are full y funded in an
factors, the team compares each project with every other aggregate project pl an (ill -plall). In a separate sectio n or
one within a category. Repeat the process for each criterio n. another document, list projects fo r future considerati on
See the di scussion and example later in th is article about (out-plan); also capture and communicate reaso ns fo r
using an anal ytical hi erarchy process (AHP) to facilitate delayin g or not fundin g projects. The plan of record
thi s step. Consider usin g so ftware to compute res ults-an (POR) is both a process and a tool used by so me organi -
ordered list of projects within each category. A pitfall to zatio ns at HP to keep track of the total list of projects. It
avoid that engenders fear among the team is showing one lists all projects under way or under consideratio n by the
list th at prioriti zes all projects from top to botto m. People entity. If a proj ect is funded and has resources assigned, it
get concerned when their project is on the line. 11 is not fair has achieved in-plan status. Proj ects below the cuto ff line
to compare internal development projects with high gross- of available resources or that have not yet achi eved priority
ing products; keep them separated and within their respec- status are o n the out-plan. The fi gure al so categori zes the
ti ve categories. projects and speci fi es the des ired mix .
Finally, the team is read y to decide whi ch projects to Project managers at HP desc ri be one benefit of the POR
pursue. Be prepared to do fewer projects and to commit process as identifying gaps between required and actual
complete resources required by proj ec ts that are se lec ted. resources. For fl ex ible changes, the process gets all people
Decide o n a mix of projects co ns istent with busin ess into the co mmuni cati ons loop. If people want to add some-
strategy, such as 50% pl atfo rm projects, 20% deri va- thing, the management team has to decide what should be
ti ve projec ts, 10% breakthroug h projects, and 10% part- deleted. The process he lps two di visions that work together
nerships. Note that th ese tota l o nl y 90%; takin g so me agree on one prioriti zed list instead o f two . They utili ze
lesso ns fro m fin ancia l portfo li o man age ment, di ve rsify direct e lectroni c connec tions for bollo m-up entry o f proj-
the set of projec ts by investing in so me speculati ve proj- ects and resources by all project managers into a centrali zed
ects. The team may not be sure whi ch markets o r tec h- admini strati on poin t
nolog ies will grow, so buy an "optio n" and mak e a s mall
!;np!l:mo/t r/i( PIon. No job is complete until it is acted
investment to in vesti gate the poss ibilities. Include exper-
imen tal projec ts. It is al so impo rtant to leave a s mall upon. The team needs to "evangelize" all others in the orga-
pe rce nt o f deve lo pment capaci ty unco mmitted to tak e ni zation to use the aggregate project plan or POR to guide
advantage o f unex pected o ppo rtuniti es and to deal with people who plan work, make decisions, and execute projects.
Although it may be countercultural to do so, do not starve
cri ses whe n they ari se.

•
 ,

ID Strategic Category Priority Project Head Ct aa

fob

22 ~
...

~ lZ 19 26 1
,. ..., M

u~~
~

~l 2
..
, .'" ~

:::-
~

,•
IS IS 22 1017 2!1L 1 16 21 ~ l3 20 21
,,,
5
I Plattonn (Mix = 40%)
,
, ,
, I
, ,
, ,, , ,
2 In Plan I Proj F 2 ,,, ,,, ,I ,,, ,,
, ,,, ,,
,
,,
, ('J
,,, ,,, ,,, ,,, ,, ,, ,, ,, l:
3 2 Proj G 2
,, ,, ,, ,,
,
,,
,
,, ,,, ,
,
,, :;;
4 3 Proj H 4 , , , , , , , -l
5 4 Proj J ,
,,
,
,,
,
,,
,
,,
,
,, ,,
, ,,, ,,, '"'"
5 ,, ,
,, , , ,,, ,, ,, ,,, ,, N
6 5 Proj K 3 , , , , , ,
,, ,, ,, ,, ,, ,, ,, ,, "-
7 Out Plan Next Step , , , , , , , , V>
,, ,, -l
,,, ,,, ,,, ,,, ,, ,,
8
9 Enha""e IMix = 20%}
,,,
,
,,, ~ : ,,
, ,,, ,,,
, ,
:, ...
,
,,
, S
10 In Plan I Proj B 2 ,,, ,,
,
,,
, ,,
,
,,
, ,
,, ,,
, ,,, '"-
Cl
,,, ('J
II Proj C ,,, ,,, ,,
,
,,
,
,,
,
,,
,
,,
,
2 I
,, l:
,, ,, ,, ,, ,, ,, ,, ~
12 3 Proj 0 I , , , , , , , , z~
,,, ,,, ,, ,, ,, ,, ,, ,,
13 4 Proj E 2 , , , , , , Cl
,, ,,, ,, ,, ,, ,, ,,, ,,
14 5 Proj I I
,,
,
,, ,
,
,
,
,, ,,
, ,
,, ,,
,
,, '"
l:
15 Out Plan Fat City
,,
,
,,
, ,
,, ,,
,
,,,
, ,
,,
,
,,
,
,, '"-lZ
16 , , , , , ,
17 R .. D IMix =30%} ,,,
,,
,,,
,,
:,
, ,
,,,
,,,
,,,
,,
,,,
,,
,
,,
,,
t- ,,
~
Z
v
18 In Plan 1 Proj A 7 , , , , , , ,
,,, ,,, ,,, ,,, ,,, ,,, ,, ,,, '"b
19
20
2 Proj L 5 ,,
,
,,
,
,,
,
,,
,
,,
,
,,
,
,
,, ,,
,
-'"
('J
21 ,,, ,,, ,,
,
,,
,
,,
,
,,
,
,,
,
,,
, -l
,, ,, ,, ,, ,, ,, ,, ,, V>

22
,,
,
,,
, , , ,
,,
,
,,
,
,,
,
,, '"r
23 Out Plan Blue Sky , , ,, ,, , , , , '"-l
('J

24 ,
,,
,,
,,
,
,, ,,
,
,
,,
,
,,
,,,
,, ,,
,,
, ,
,,
,, ,,
,
,,
-
0
25 Infrastructure (Mix = 10%) , , ,, , , , , , Z
,, ,, , , ,, ,, ,, ,, ,,
26 In Plan I Bus. Plan I , , , , , , , ,
,, ,, ,, ,, ,, ,, ,,, ,,
27 2 Portfolio 1 , , , , , , ,
,, ,, ,, ,, ,, ,, ,, ,,
28 3 Update plan I , , , , , , , ,
,, ,, ,, ,, ,,, ,, ,, ,,
29 , , , , , , ,
,,, ,,, ,,, ,,, ,, ,, ,, ,,
30 , , , ,
31 Out Plan Corner Office
,,, ,,, ,,, ,,
, ,,
,
,,
,
,,
,
,,
,
'I An exa mple plan of record showing the mix of projects in priority order and th e time line for each project.
 D IRECTED READI NG

committed projects of the resources they need. The team or the goals. the first task was to identi fy which criteri a to enter
resJX>nsible upper managers need to enforce the plan by full y into the decision· mak ing process. Afte r give·and· take
staffing commiued projects; that now becomes possible beca· di sc ussion, they decided that the criteria were customer
use fewer projects are happening simultaneously. Also, use the satisfacti on. business va lue. process effecti veness. and
plan to identify opJX>rtunities for leverage across projects or employees ati sfaction.
for process rt.--enginccring. Mmch people skills to project cate· Next. the criteria we re ranked according to priorit y by
gories to tap their strengths and areas for contribution. making pairwise compari sons between them. Which is
The team or a program manage ment offi ce needs to the more desirable criteri on and by how much. custo mer
maintain th e plan in a central place, such as a projec t offi ce satisfac ti on or business value? Process effecti veness or
or online. Make it known to. and accessible by, all people employee sati sfacti on? Business va lue or process effec ti ve·
in the orga ni zati on doing projec ts. subj ec t to confidentialit y ness? These questions were asked about all possible pairs.
requirements. All th e work to this point may go for naught Eac h potenti al projec t or service then was scored under·
if the process. the steps, and the results are not widely neat h each criteri on. and decisions were made about wh ich
co mmunicated. projects to include in the portfolio, based upon existing
The same people who develop the plan are also the resources. This team went on to create a POR similar to
ones who can best update it periodically. perhaps quarterl y Fig ure 5.
or as changes occur. Usc tools such as an online shared A detai led ex planation for computin g the priority scores
database to gather data directl y from project managers and the final rank ordering list can be quite com plex.
about resources needed for eac h project. This system can in volvi ng eige nva lues and eige nvec tors. so it is much
be used both to gather data when developing the plan and easier to get a software package (Expe rt Choice 14]) Ihat
to update it. View the plan as a " li ving document" that does the co mputations. As an alternati ve, a spreadsheet
accurate ly reflects current realities. could be constructed to normalize the num bers.
The challenge for HP and many companies is to '·mas· This process appears co mplex and analytical but is
ter both adapti ve innovati on and co nsistent executi on ... easy whe n the soft wa re handles the computati ons. and
again and again and again ... in the context of rel entless the manage ment team concentrates on the co mparisons.
change . ... Staying on top means remaining poised on It is thorough in guidin g the tea m to consider all criteri a.
the edges of chaos and ti me . .. These edges are places of both emoti onal and logical. and to apply them to all proj-
adap ti ve behavior. They arc also unstabl e. This instabi lity ec ts. One team rejected the process as too analytical. so be
mea ns that managers have to wo rk at staying on the edge" awa re that it docs not work for everyo ne.
[ IJ. The advice is clear: the plan is indispensable as astra· The key benefit in doing thi s process is the improved
tegic guideline. bu t don' t fall in love with it! Be prepared quali ty of dialogue that occ urs among the manage ment team
to adapt it and to co mmun icate the changes. members. In fac ilitating a number of teams at HP throu gh
this process. each one ac hieved far more progress than they
thought possible. People admit that they become addicted to
One tool that can assist in the decision· making process is the the AHP process. They immediately buy the software. The
AHP [10]. Because of the interac ti ons among man y factors systematic approac h is feasible whether selec ti ng products
affecti ng a complex decision. it is essential to identify the for a prod uct line. projects that comprise a portfolio. or the
important fac tors and the degree that they affect eac h other best supplier or candidate for a job. In reality. the disc ussions
before a clear decision can be made. The AHP helps struc- arc more va luable than the analysis. The process in this case
ture a complex situation. identify its criteria and Olher inlan· provides the discipline that makes the dialogue happen.
gible or concrete factors, measure the interactions among Frame [5) offe rs an alternati ve "poor man 's hi erarchy."
them in a simple way. and synthesize all the infonnation to He puts se lec ti on criteria along th e side as well as across
obtain prioriti es. The priorities then can be used in a bene- the top of a grid. If the criteri on on the side is preferred
fit -to-cost determination to decide whi ch projects to selec t. 10 th e one on the top, put a I in the cel l. If the criteri on on

The AHP orga ni zes feelings and intuition alongside logic in top is preferred. put a 0 in the cell. Diagonals arc blanked
a structu red approach to decision· making- helpful in com· out where criteri a would be compared to themselves.
plex situations where it is difficult to co mprehend multiple Below the diagonal. put the opposite va lue from corre·
variabl es together. An indi vidual or tea m foc uses on one cri· spo nding ce lls above the diagonal. Then add up the num-
terion at a time and applies it step by step across alternati ves. bers across the rows to get total scores. which provide a
A number of sites across HP find value in using AHP. rank order. One tea m at HP modified thi s process to replace
In another exa mpl e. a tea m got together to choose the I s and Os with an actual count of how ) 8, pcople voted
among a set of services they will offer to customers. More in each pairwise co mparison of alternati ves. Aga in. they
choices were avai lab le th an the organi zati on had ca paci ty added up the rows and normali zed the resu lts fo r a priority
to suppon. After defini ng organizational strategy or product order and we ighted ranking (Figure 6).

•
 •

98 C HAPTER 2 / STRAT EG IC MANAGEMENT AND PROJ ECT SELECT ION

Total
Business Customer Technology Employee Votes %

Business
•
••• 16 16 18 - 50 46

Customer 2 ••• 13 15 - 30 28
--t
Technology 2 5 14 - 21 19
***
Employee 0 3 4 *** -- 7 7

"igun' () A si mplifi ed hi erarchy used by o ne HP tea m to weight criteria.

This simplified hi erarchy is espec ially he lpful for And if people sense that the leader does not authenti -
weighting criteria. h can be used for prio ritizing projects cally believe in the elements, such as the goals. the process,
when applied to one criterion at a time. It becomes bulky or the tool s, they are hesitant to follow with any enthusi-
and less useful when appli ed to multiple projects over asm. When the leader lacks integrity and exhibits incon-
multiplec riteria. gruity between words and actions, people may go through
the moti ons hut do not exert an effort thai achieves mean-
] ~arrj(:n. t (1 II11I'!t.' 111t.'Jlti.lli(H} ingful results.

Now for a reality check. The model depicted in this arti- En'lblcr~ f()t" l:,rfn,~I:i\'(" ImpkmellLati()1l
cle is thorough, and it integrates objective and subj ec-
ti ve data. When all is said and done, however, people It is possible (0 lead people Ihrough (his change process
may throw o ut the results and make a different dec ision. if the leader asks many questi ons, listens to the concerns
Sometimes the reason is a hunch, an instinct, or simply a o f all people involved , and seeks to build support so that
desire to try something different. Sometimes people have people feel they have an active role in deve loping the
a pet project and use the process to justify its existence, process [9]. A Hex ible process works better than a ri gid
or a hidden agenda may be at play-perhaps the need to one. Cultivate "champions" who have the credibility and
maneuver among colleagues, trading projects for fa vors. fo rtitude to carry the process across the o rgani zatio n.
Politics at this stage cannot be igno red . nor are they likely Believe that change is possible.
to di sappear. It is imperati ve for leaders to become skilled When the e ffort appears too massi ve, o ne approach
in the political process. Any attempt at leading change in is to go after the low-hanging fruit. Start with one of the
how an organization links projects to strategy is bound to more pressing issues and use the general concepts of thi s
meet res istance. The concept receives almost unanimous model to address it. Still have a vision for what the organi-
intellectual support. Implementing it into the heart and soul zation ultimately can achieve but understand that patience
of all people in the organization is another story. It goes and pacing are necessary to get there. Consider also that
against the cultural norms in many organizations and con- thi s process is hierarchical- it can be applied singularly or
jures up all kinds of resistance if the values it espouses are co llecti ve ly, up or down the organization.
not the no rm in that organization. The path is full o f pit- For people who get frustrated when all linkages are not
falls, especially if information is prese nted carelessly or present, the authors urge teams and indi viduals to "just
perceived as final when it is work in process. do it." Small change s in initial conditions have enonnou s
Some people resist because the process is too anal yti ca l. consequences. Eventually successes or small wins are
Some want dec ision-making to be purely interactive, intui- noti ced. The practices start to permeate an organization.
tive, or the purvi ew of a few people. A complete process This can happen in the middle, mo ve up, and then over to
cannot be forc ed upon people if the organization has more other organizations. Incidentall y, a corporate group like
immediate concerns or unresolved issue s. Res istance HP's Project Management Initiative helps facilitate thi s
occ urs when there is no strategy, the strategy is unclear, or transformation. We do thi s by acting as a conduit for suc-
people are unco mfortable with the strategy. Work on the cess stories and best practices.
process may come to a standstill when people reali ze how Over the long run , we believe that organizations that
much work is involved to full y link proj ects to strategy. If foll ow a process similar to the one desc ribed increase their
the pain is not great enough with the statu s quo, people are odds for greater success. This happens because team s of
not goin g to be ready to change. people following a systematic process and using convinci ng
 DIRECfED READ ING I)l)

data to suppon their arguments more often produce better 10. Saaty, T. L. Decisioll Makillg for Leaders. Pittsburgh,
results than individu als. Their projects ha ve more visibility, PA: RWS, 1990.
and the quality of dialogue and decision-making improve. 11. Stacey. R. D. Managing the Unknowable: Strategic
The power of using cri teria that are tightly linked with Boundaries Between Order and Chaos ill Organil,iltions.
strategy and known by everyone in the organi zation is San Francisco: l ossey-Bass Publishers, 1992, p. 62.
the mitigating effect it has to guide behavior in construc-
tive ways. Having a process means it can be replicated 12. Tunle, Q . C. Implememing Concurrelll Project Man -
and improved over time until it is optimi zed. It also means agement. Englewood C liffs, Nl: Prentice Hall , 1994.
other people can learn the process and coac h others. 13. Westney. R. E. Computerized Management of Multiple
thereby c reating a learni ng organizati on. Small Projects. New York : Dekker, 1992.
14, Wheelwright, S., a nd K. Clark. "C reating Project Plans
to Focus Product Development." Harvard Business
I. Brown. S. L. , and K. M. Eisenhardt. Competillg Oil the Review, March- April ( 1992).
Edge: Strategy as Structured Chaos. Boston: Harvard
Business School Press, 1998.
Questions
2. Co llins. J. C. and J. I. Porras. Built to Last: Success-
ful Habits of Visionary Companies. New York: Harpcr- 1. Why a re successful projects so imponant to Hewlett-
Co llins, 1994. Packard?
3. Cooper, R. G., S. l . Edgett, and E. l. Klei nsc hmidt.
2. How far should an eval uati on tea m go in trying to quan-
Portfolio Managemelll for New Products. Reading,
tify project contributi ons to the finn's mission or goals?
MA : Addison-Wesley, 1998.
What is the role of financial se lecti on criteria in HP's
4. "Expert Choice," Pittsburgh, PA: Expe rt Choice Inc. project se lection process?
(see IVww.expertchoice.com).
3. Conside rable attention is paid to the measures HP uses
5. Fra me, J. D. The New Project Mallagemem: Tools for
to evaluate its projects. Is the aim of carefull y defining
an Age of Rapid Challge. Corporate Reengineering. these measures to si mplify the project se lection process
(111{1 Other Business Realities. San Francisco: Jossey-
or so me thing else?
Bass Publishers, 1994.
4. What do the aggregate project plan and the plan o f
6. Graham, Robert l . and Randall L. Englund. Crea tillg
record illustrate to upper management?
all Environment for Successful Projects: The Quest to
Manage Project Managemelll. San Francisco: Jossey- 5. When sho uld out-p lan projects be reconsidered for
Bass Publishers, 1997. inclusio n?
7. Kumar. v., e t al. "To Tenninate or Not an Ongoing 6. What was you r impression of the impact that HP's proj -
R&D Project: A Managerial Dilemma." IEEE Trallsac- ec t selectio n process had on the num ber of projects
tiolls all Engineering Mllllagemelll 279 ( 1996). underway? How do you expect HP would score on proj - •

8. Martino. l . R&D Project Selectioll . New York: Wiley, ect ma nage ment maturity?
1995 . 7. How did the new project selection process ha ndle non-
9. O'Toole, J. Leading Change: Overcoming the Ideology nume ric type projects? Ri sk? How did thi s new process
of Comfort and the Tyranny of Custom. San Francisco: alter new project proposals at HP?
lossey- Bass Publishers, 1995.