2. Managing Project InterfacesKey Points for Project Success Peter W. G. Morris' One of the most impodlnt qualities of a project manager is a mature understanding of the way projects develop. This allows the nature of project activities to be better understood, problems to be seen in perspective, and needs to be assessed ahead of time. To somc cxtcnt this understanding of project development is intuitive, thorigh i t clcarly also depends upon specialist knowledge of the project's tech:~ologpand industry. It can, houlcver, also be acquired in large part from formal study of the development process of projects, since all projects, regardless of size or type, follow a broadly similar pattern of development. The organizational frametvork underlying a project's development is the subject of this chapter. The intent of the chapter is to illustrate the types of issues that are normally encountcred as a project develops and to suggest ways in which these issues should be handled. The extent to which these issues can be related to project success is then discussed. THE SYSTEMS PERSPECTIVE AND PROJECT MANAGEMENT The most pervasive intellectual traditiorl to project management, whether in organization, planning, control, o r other aspects, is without doubt the systems approach. Dr. Peter Moms ic Director of the Major Projects Association, a member of Oxford University's Faculty of Physical Sciences and an Associate Fellow c l Templeton College, he Oxford Centre for Management Studies, Oxford, England. He undertook research into project management at Manchester University, England, in the late I%&, gaining his Ph.D. in 1972. llc has worked both as a manager and as a consult~nton a variety of projects around the world ranging from telecommunications and petrochemicaJ projects in the Mideast, N o r ~ hAfrica, and Europe; steel projects in Latin America; and construction, MIS, and aerospace projects in North America and Europe. Prior to returning to England in 1934, he was responsible for the international business of the Anhur D. Lit~leProgram Systems Management Company, Cambridge, Massachusetts. He is the co-author of The Annfo~riyo Major Projects, published by John Wiley & Sons Ltd.; 15%7. A system is an assen~blageof people, things, information, or other attributes, grouped together according to a particular system "objective." Thus, one has an electrical system, the digestive system, a highpressure \veatlicr system, an air conditioning system, a \ileapons system, a system for winning at cards. A systcrn may be logically broken down into a number of subsystems, that is, asse~nblagesof people, things, information, or organizations required to nchie\*e a defined system srrbobjecti\)e, like the switching, outside plant, building, transniission, and subscriber subsystems in a tclephone system. The subsets of each subsystem may then be identified-cables, poles, microwa\re, and transmission and distribution equipment for the transmission subsystem-thereby creating sub-subsystems. Subsets of these subsets may then be idptified, and so on. Propcrly organized and managed, the overall system acts in a way that is greater than the sum of its parts. The systcrns approach emphasizes treating the system as a \\)hole. The systems approach has its origin in the late 1920s and 1930s. Biologists noticed similarities in the way that living organisms interacted with and controlled their environments. At the same time, similar p3tterns were observed by Gcstalt psychologists in the \\lay the human tnind organized sensory data. Both tlie mind and living organisms have to adapt to changes in their environment. Systcms of his type arc known as "open" systems. Before long it was sccn that all social systems operate as open systems.' During the 1950s, work in economics, psychology, sociology, atithropology, and other disciplines de\teloped these open-system ideas by elaborating such concepts as self-organization, purposive systcrns, the importance of goals and objectives, the hierarchical classification of systems and subsystems, and the importance of systems' boundaries and interfaces (2)*. At the same timc, this "systemic" view of the world was enriched by a parallel (but initially separate) set of disciplines which had their origin in ,the induslrial and military applications of the scientific method during and immediately after World \Var 11. This was the essentially numeric set of disciplines, such as cybernetics, control theory, operOpen systcrns are "open" to the cflects of their environment. On the other hnnd, closed systems. which are the other major system type (including, for example. much of physical chemistry and many types o f machines). operate independently o f their environment. In open systems, events rather than things are strucfured; there is a constant energy and information exchange between the system and its environment; the syslem organizes to minimize entropic decay; equilibrium with the environment is achieved through a process known as homtostasiS; and there is a tendency towards difrerentiation. Closed systems operate in almost exactly the opposite manner (I).' Numbered references are given at the end of this chapter. ations research, systems analysis, and systcms engineering, concerned with modeling real-life situations so that complex behavior could be more accurately described and forecast. Slowly both streams merged, encour-' aged greatly by the enormous growlh in the ability of the computer to apply these systems ideas with powerful effectiveness, s o that the systems approach is now an established and vigorous influence on managemen; and research. The systems perspective has contributed substantially to the development of project management. Firstly, the systems emphasis on viewing a system a s a whole has frequently becn behind the recognition of the nccd for an acrosr -the-board integrating role-that is, for managcrnent itself (3).2 Secondly, systems thinking has shown how projects should \ircrk as successfully regulated organizations, for example, the need for clcarly defined ~ S j e c t i v e sthe , recognition that projects are organizations in constant change, and the need to define and manage major subsystems and their interfaces. A third important contribution is that the dynamic control nceds o f projects a&' now bctter undcrstood-the importance of feedback, Ihc progressive devclopmcnt of information and multilevel projcct control. And a fourth contribution is the widespread use of systems lechrliques-s,ystems analysis, systems engineering, work breakdown structures, and simulation models. Interface Management, as i t is used in project manngcmcnt today,] is an outgr0~~1.h of the first two of these influences of systems thinking on project management. Interface Management identifies the follo~ving: The subsystems 10 be managed on a project, The principal sobsystcm interfaces requiring managcnlent atlcntion The development ofproject management by the U . S . military i s an ill~strittion:the systems ideas developed initially for technical purposes were adapted to grnerate the organizational flexibility and control missing in the existing military bureaucracy. This can be seen in each ofthe steps in the U.S.military's development o f project management-the dcvelopmcnt of !.be U S A F and Navy program management practices (4). particularly for the I C B M programs ( 5 ) and for Polaris (6) around 1952-1955; Peck and Schcrer's study of the U.S. and Soviet \\.capons procurement processes in the I;+te 1950s (7); the dcvclopmcnt o f PERT on Polaris in I Y 5 R ; the i n t r o t l ~ ~ c t i oofn projcct organizations in the Navy, Air Force. and Arm). in the late 1950s/early 1960s; hfcblamara's extensive s ~ u d yand implementation o f program rn:inagement and project control ~echniquesin the early 1960s; and I-sird's and Packard's process-oriented focus on the needs ofthe total projec~life cycle in the late 1960s~carl~~1970s (8). ' Interface hlanagement i s ~ c n c r a l l yused now in a broader sense t h a i i t war ten or tu,cnly years ago. I n the 1960s and early 1970s Interface Management generally refcncd simply l o ensuring that system interfaces matched (i.e.,had the same specifications, w'erc not missing any equipment, data, etc.). Today'it is used i n the sense of defining syslems-orpanizalional, managerial, and ~echnical-and of ~ c t i v e l ymanaging their intcrrcbtion~hip:. The term " i n l c h c c management" i s relatively common in high-technology projects, such as information systems and aerospace; i t i s much.rarer on conrtruction projEcts. T h e ways in which these interactions should be managed successfully. The emphasis on identifying key interfaces and on focusing on interface performance has grown a s it has been irrcreasingly realized that all projects share a common pattern of interfaces derived from a common pattern of subsystem interaction. This is true no matter \\that the type of project, be it'a theater production, an aid program, an clection or a major capital investment program. There a r c three sets of subsystems on any project: those deriving from the project's life cycle, its management levels, and its operational charactcristics. PROJECT LIFE CYCLE4' Projcct manngcmcnt teaches that to achieve the decircd project objective one must go throl~gha specific process. There is no cxccption to this rule. 'I'he process is kno\rfn as thc projcct "life cycle." Projccts (like pcople) have a lifc cycle that involves 3 gradual buildup a s definitions are cstnblishcd id \{,orking characteristics developed, a fullbodicd irnplcmcntation as thc \vork is ;lcconiplishcd, and a phasing out a s thc work is completed and the project winds do\vn. While thcre are various definitions of the project (or program) life cycle. (Figure 2-1). the csscntial scqucncing is invariant, althotrgh (as with pcople) sometimes not f~lllyrccogni7cd or rc<pcctcd. A project <tarts ;is an incipicrit idea which is cxplor-cd for fin;\ncinl and. 1echnic;il fcasibilit y in t I,e Prcfinsibi1ir~~lI;'cn ~ i h i l i r y~ f n i e ~ . A p i t c i tis~ dccidcd, locations chosen, fin~ncingarranged, overall schedule a n d budgct ngrcctl, and prcliminary organizations set up. At t h e end c ' this phase therc shoi~ldbe a formal "gotno-go" decision. I n the next phase, Dcsigtt, the \vork is organizationally and managerially similar to the first phssc only it is n1ol.c comprehensive and detailed. T h e technic;?l definition of thc projcct is expanded (;~lbeilgenerally still at a fairly strategic lcvel); schedule, budget, and financing are ~.cappraiscd;contracting strategy is defined; pcrmits are sought, and infrastruct~rrcand logistics systems arc dcfincd. . Note that an interface i s tcchnicnlly defined as the space httwcen interacting subsystc .. Even though there might be a commoll set of subs).stems on all projects. this does not neccszrrrily mean there will be a common sct of interfaces. The extent that thcre is de?- ids on the commonality of subsystem intcrac~ion.This chapter will show that s~~bsystcm interaction docs in fact follow a common p;trtcrn on most projects. ' . Thc impticalions of the life cycle are Ircared in r,rcalcr detail in Char lcr 9. , . Kettey'r model 191 is essent~altythat o f a program life cycle. He dtfferentiales product control from prolect control. prDduct control being concerned wilh the definitions of the end product parameters and project control being concerned wtth the process of making that product. ~ e e d Economics Concept - Feasibility - Definition - Implementation Procurement - Turnover L L Project Control - * Product Control /--~l~~~i~~-~~s~tion+ operation -1 Wearne'a model (101 is typical of industroal projecls. "The nature and scale of activities change a! each stage but the &!ages usually overlap." Of pnrl~cul'drtnterest are the Demand and the Use. Ma~ntenance.and Records h Experience stages. stnce these a r e often overlooked i n project management lterat~onw ~ t h ~and n between the early stages is not unllkety. Demand \ experience f I Evaluation Research Maintenance A Development - Design Contracts \ Commirrioning / / Manufacture & construction Figure 2-1. Views of !he projecdprogram life cycle. In phase three, Mnnrrfactrtre, Corrstrrrction, arrd Irrstnll(~tion(or Prodrrcriort), equipmept is procured, civil work is undertaken, and equipment and facililies are installed. This phase diners dramatically from the previous two. First, whercas the Design and Fensibilify phases wcrc organic and evolutionary in character,-the Prodrrction phase is highly mechanistic Morris: 111) concentrated o n the lersibility-design-implemenlationphases lo make ~ r apbints l about the nature-of work within and between phases. The design phase i? brsically "organistic." for errmplr, wh~iethe producl~onphase is more "mechanistic." Crossing the design-production inlerfacr, wilh the Ielt~ngof ~mplcmenlationcontracts, is a major transition point in tne project life cycle. F u l l Operations Installation Substantially Percent Complete Companies 100% I Ill Stage I FEASIBILITY Project Formulation Fearability Studies Strategy Design & Appraisal Staye I1 P L A N N I N G & DESIGN Base Design Cost & Schedule Contract Terms & Conditions Detailed Planning Stage Ill PRODUCTION Manufacturing Delivery Civil Works Installation Testing IV * Stage I V TURNOVE A & START-UP Final Testing Maintenance Kenner's d~agram(121 illustrates the f~nanciatIlfe ot a product devclopment program (me) Kerzner comments on the w r y the l ~ f e cycle varies between systtms, projects and producls; apart from d~tferencesin terrntnotogy. 8 major d~lfcrenceis In the overlapping of phases. Pure I basic I rtrearch i E C U a I 1 I Applied research I I I I I I I Definition 1 I -Operational Figure 2- 1. (cottr;t~rrcd) L I Divestment I (13). The aim is not to develop new technical options but .d build as efficiently a s possible the thing which has been defined in the Design phase. Second, there is a large-sometimes vast-expansion in organization (whereas thcre may have bcen only dozens o r hundreds of persons active in the first two phases, there may be thousands or even tens of thousands involved in this third phase.) And third; the characteristic mode of control changes from one of "estimating" costs and durations to one of tight "monitoring" of quality, schedule, and cost to keep actual performance within the target estimates. The fourth and final phase, Tiirn-Over ur~dSfarf-Up, oxferlapsthe third phase and involves planning all the a ~ l i * ~ ; t i qeccssary cs for acceptance and operation of the project. Successfully sy-chronizing phases three and four can prove a major management exercise. The cost of capital locked up in the yet urlcommissioned plant, and the opportunity costs of both underutilized operating systems such a s sales, operating plant, personnel, etc., and a possible diminishing strategic advantage while con~petitors develop rival products, can prove enormous. Between each of these life-cycle phascs there are distinct "change points" (\+!hat shall later be called "dynamic project interfaces"): to D ~ s i g t ~the : "go" decision. From Prefe./pasiDilirylFca.~ibilify From Design to Prod~rction. From Prod~rcfiot~ to Tirrn-Over & Sfrrrf- Up. The project or? either side of thcse change points is dramnticnlly different-in mission, size, technology, scale, and rate of change-and these differences create their own particular different characteristics of work, personal behavior, and direction and control needs. Thus, importantly, the manngcmcnt style of each of the four main life-cycle phases is significantly diffcrct~t. PROJECT MANAGEMENT LEVELS The four phases have a set and important managerial relation to each stage is 'highly "inst ituother. The work of the Prefiasibilif~~lFea~ibiIify tional" (top management) in kind-decisions taken in this phase will later havc an overriding Impact on the health of the invzsting entcl-prises. In Design, the work is of a "strategic" nature, laying the axes upon which the detailed, "tactical" work, of the third, Prodrrcfion phase will rest. ,Interestingly, the fourth phase, Trrrn-Ooer & Sfurr-Up, exhibits a nlixture of all three managerial levels of work: institutional, stratcgic, and tactical. These three levels of management activity have bcen. recognized as distinct levels of management since at least the t i n e of Talcott Parsons, the eminent American sociologist. Parsonsmade the point that each of the three levels has an essential role to play in any successfully regulated enterprise: the technical/tactical level (111) manufactures the product; middle management (11) coordinates the manufacturing effort; at the institutional level-(I) top management connects the enterprise to the wider social system (14). Each of these three management levels has a iundamental role to play in the rnanagcment of every project (although it is true that the levels tend to become more blurred on the smaller projects). Yet surprisingly, most project management literature deals only with Lcvels II and 111. There is little in the literature that treats such Level I issues as: the role of the owner and his financer; relations with the media, local and fedcral government, regulatory agencies, lobbyists, and community groups; the sizing and timing of the project in relation to product demand and the cost of finance-all issues that became crucially important during the 1970s and are particularly so in the 1 9 8 0 ~ . ~ The distinction between Levels I1 and I is quite critical since it is csscntially the distinction between the project and its outside \vorld (Figure 2-2). Levcls 11 and 111 deal almost exclusivcly with such familiar project activities as engineering, procurement, installation, testing, and start-up-Level 111 providing the tcchnical input, Levcl I1 providing both a buffer from the outside world arid guidance in how to avoid external pitfalls. But no.project erists in isolation from outside events. Level I provides the coordination of the project with outside events and institutions. Level I actors typically include the project owner and his finance team, government agencies, commwqity groups, very senior project management, and one or two special project executives specificnliy charged with external affairs, such as Public Relations and Legal Counsel: The involvement of cach of these Levcls is different during each of the major phases of the projcct life cycle. During [he Prefii7sibi!irylFeasibility stage, the owner and his learn (Levels I and 11) have to make crucial decisions about the technical performance and business advantages they are to get for thcir investment-and indeed, whether the project should "go" at all. Once the decision to go ahead is taken the weight of the \(pork moves to the design leam (Levels I1 a'nd 111). During Prodlrctiort engineering reaches a detailed level. Both project management (Lcvel 11) and technical staff (Levcl 111) arc now at full stretch, while top management (1,cvel I) takes a more rcduccd "monitoring" role. Finally, during TurnOver a n d Srnrr-Up, all three Levels are typically highly active as ehgi'Such "institutional" project management issues are preeminently the concern of thk UK hl;~jorProjects Alsoci:~lion(15). They are treirled in this handbook in Chapters R. 13, 18. and 22. Figure 2-2. T h e three lcvcls of projcct management. nee1 ing work gets completed (Level I I I ) , often under intense management pressure (Level II), while high-level coordination is required at the owner level in coordinating the initiation of Start-Up activities. The responsibilities of thcsc 1,cvels thus focus on two main nrcas of activity: Levels 11 and 111 on the tecl~nicaland middle management work within the project, and Level I on the senior nlanqgemcnt work at the project/outsidc world interface. PROJECT OPEPJTIONAL SUBSYSTEMS T h c work of these two essentially dislinct lcvcls of project managcmcnt activity tcnds to follow a pallern which is similar on nlitny projects. A1 the project/outsidc world level, the concern is to ensure that the project is commercially viable and, a s far a s possible, is provided with the conditions and resources neccssnry to sl~ccccd.The principal arcns of work at tliis lcvcl arc 8 Ensuring satisfactory Project DcJirtiliotr (which includes both technical content, cost and sche.dule requirements) Preparing for Oprrations & b h i ~ t r e n a n c e Preparing for Sales & Murkelir~g Ensuring appropriate 0.-ganizrtriort structur'es and systcms, both for the project and for operations Facilitating relations with important Orrsidc Grorrps such a s povernmen! and community groups, financial institutions, and the media Ensuring appropriately skilled A4nnpo\vcr for both the project and future operations Ensuring that thc total enterprise is Cor~intcrc~iall~~ Sorrrid and "adequately financed." Work within the project, on the other hand, focuses morc on nccomplishing the tasks within the strategic parameters develqped and managed by senior management. At the intra-project level, the principal sulisystems are Realizing the desired Projccl DeJ~~irion-i.e., assuring that the project is produced to technical specification, on time, and in' budget Creating the Org~niznrionneeded to execute the project-this includes both the fot-ma1 organization structures, contractual rclationships, systems of information flow and control procedures, and also informal patterns of working relationships and communication Minimizing external disruptions from the E11virort11rcr11-by, for euample, acquiring adequate materials inventory to provide buffer stocks against delivery disruptions, I~andlingupion negotiations, ob-' taining necessary regulatory approvals, or warning top management of future financing needs6 Providing adequate Infrasfrr~cr~rrea n d Logisrics to accomplish the project (facilities, transportation, communication, utilities). The cfTects of environrncnt on a project can be profound, and continue to provoke keen theoretical analysis and discussion. Of padicular interest is the problem of how organizations behave in a constantly and rapidly changing environment. Theorists d e s c n i such environments as turbulent and call the type ofsystems that operate in such environments "multistable" (16). Large or complex projects in parlicular s ~ ~ f r mnny er ofthe conscqucnces predicted for mullistable systems, such as large subsystem interaction, continuous objectives redefinition, rich internal fecdbnck processes. high impact o f exlemrl factors (onen causing [,he subsystems to have to act i n an npparcnrly kss-than-rotionul wry), and rubsrrntin1 o~.gsnizationalchange, onen of r sreplunction size. These chamcten'stics can be found an major projects such as the Concordc and Norlh Sea Oil projects, nuclear power projects, and many dcfcnse projects (17). STATIC AND DYNAMIC INTERFACES The likcly existcnceof these subsystems in a project, no matter how it unfolds, enables us to categorize certain interfaces as on-going or "staticv-they are not a function of the way the project bcvelops but reprcscnt relationships between on-going subsystems (like engineering and procurement, o r Level I and Level 11). There is another group of interfaces, however, which arise only as a function of the pattern of activity interdependencies generated by the way the project develops. Thcsc we may identify as life-cycle or "dynamic" interfaces. Dynamic intcrfaces between life-cycle (or activity) subsystems are of thc utmost importance in project managcment, first becausc of the continuous in~portanceof thc clock in all projects, and second because early subsystems (like Dcsigrt) have a managerially dominant role on subsequent ones (like Aln~irr/acrrrrir~g).Dynamic interrelationships require careful handling if minor mistakes in early systems are not to pass unnoticed and snowball into larger ones latcr in the project. Boundaries should be positioned where there are major discontinuities in technology, territory, time, or organization (1 8). Major breakpoints in 'the project life cycle-as, for instance, between each of the four major phases, and also between activity subsystems within each phase (for example between manufacture, inspection, delivery, warehousing, installation, and testing)-provide important dynamic interfaces. These serve as "natural" check points at which management can monitor performance. Most major dynamic interfaces are in fact used in this way: for example, the Project Feasibility Report, the initial Project Technical Design, thc formulation and negotiation of the "Prsduction" contracts, and Testing and Hand-Over. Review points such as design-freeze points, estimates-to-complete, and monthly progress reports may also bc introduced for purely control purposes without there being any "natural" discontinuity. Each in its own way reprr*.cntsa response by project management to control the project's momentum across its dynamic interfaces. Whereas the important dynamic interfaces are relatively sharply,defined for Level I1 and 111 management, at Level I they are less distinct. Level I management is certainly partly driven by the anatomy of the project's internal development, but it also has its own dynamic intcrfaces for each of its own principal subs'ystems. Operation, Sales, many of the Outside Groups, Manpower, and Finance and Commercial issues each have their own often distinct life cycles. (For example: the process of recruiting and training manpower; preparing annual financial plans.) Thus at Lcvel 1, dynamic interfaces do not become less impartant; rather they become more varied and less clearly dcfined. They are still crucial to the project's success. . Static interfaces too are less clearly defined at Levcl I than at Levels I1 and 111, partly due to the wider scope ofconcera of Level I (which gives rise to much mullifacted subsystem interaction, as, for cxample, between Operations, Salcs, Manpower, and Finance) and partly due to the disruptive effect of the outside environment. Figure 2-3 sketches thc three principal sets of project subsystems which have now been identified: the three levels of management, static subsystems, and dynamic subsystems. PROJECT INTEGRATION Some .interfaces are clearly larger and more important than others. Organization theorists describe the size of an interface not in terms of, for instance, a small change point or a major one, but in terms of the degree of difleretlrinrion between subsystcms. Typical measures of differentiption include differences in "STATIC" S U B S Y S T C M S Project definition Saln snd marketing Operalions a rd mainrenance Organization Manpower proups CommerI~l and financial conditions - P "Project" . , ... . Time LEVEL! ' . . . Fessibilit'y .,. B. PROJECT LEVEL'II . MANAGEMENT ,;: LEVELS 8 " . "0perat;ons and maintenance'' .:; urnover an : : Production . , L E V E L 111 ' . ,I ,.;... .. ': C. " D Y N A M I C SUBSYSTEMS A121. INTRA-PROJECT "STATIC" SUBSYSTEMS .- 1.: . . :, . \a:.. . . .. . . . . . . - . . ~&ironkntlnf,ncructurr . c k~'f?~i,*n ~ t m norganization ; .... ... .,....;: . . .. . . .. . . . . .. . ' 1 Figure 2-3. The three sets of projecl subsysrcms. . Organization structure Interpersonal orientations Time horizons Goals and objectives (19). Thus, a mechanized infantry brigade c a n be differentiated clearly from a local community opposition group on all dimensions. The R&D wing of a conlpany can be similarly differentiated from the marketing wing. The architect on a building project can be differentiated from the building contractor. A mechanized infantry brigade will go to some lengths to avoid having to integrate with a local opposition group but R&D has to integrate with marketing quite frequently, and it is inevitable that architects must integrate with contractors. Why? Because the activities of the groups create ceflain technical, organizational, and environmental intcrdependeticies. These interdependencies may be almost accidental or may be deliberately organized. Ititegr(~fiotrbeco~ncsimportant whcn the tlcgrce of or-ganizritional intcrdcpcndcnce becomes significant. Research has shown that tightcr organizational integration is necessary when The goals and objcctives of an cntcrprisc bring a need for difrcicnt groups to work closcly together The environment is complex or changing rapidly The technology is uncertain or complex The en~el-pi-iseis changing quickly The enterpl-isc is organiza~ionallycomplcx (20). The amount of integration actually required at an interface depends both on the size of differentiation across the interface and on how much "pulling together" the interfacing subsystems need. Certain project subsystems can be differentiated from one another quite markedly. For example, the projectloutside groups interface is marked by very strong differences in time horizons, goals and objcctives, interpersonal orientations, and structure--this is why the conflict over many envIronmenta1 and regulatory issues is so drastic on many large projects: the aims and mores of the cr~vironmentalistsare far, far removed from those who are trying to build the project. The design group often functions quite differently from rhc construction group-the former's intercst might be elegant enginccring, time might not be money, and quality might be paramount; the construction crew might be of less elitist thinking, have strong incentivcs'l~otto waste time, and might often work in a tourher organizational milieu. Similarly, there are majcr differences in perspective between operations and project personnel, bctweep the project fi- nance team and project engineering, between a construction manager and project management, and so on. It is, in short, possible to"establish the degree of differentiation between esch one of the project's interacting subsystcms, and. in so doing thereby establish which are the principal projcct intelfaces (21): Despite the insights of managcmtnt theorists, choosing the degree of integration-the amount of "pulling together1'-required across an interface still calls for considerable judgment. This is inevitable. There is no easy answer to the question, "How much management is enough?" There arc some pointers, howevcr. James D. Thompson, in a classic book (22), observed that there are three klnds of interdeperidency, each requirihg its own type of integration. The simplest,"'poo~ed," only requires that people obey certain rules or standards. The second form, "sequential," rcquires that interdependencies be xchcdrrled. "Reciprocal" interdependence, the moqt complex kind, requires nlrrtrral odjrrsrt~tcntbetween parties (Figure 2-4). I n project terms, subsystems which are in continuous 1. POOI.ED I N T E R D E F E N D E N C E Participants pool rmourccs etc.. c.9.. membership of r club Coordination by standards. rules 2. SEOUE'NTIAL I N T E R D E P E N D E i K E Participrcts follow each other 3. RECIPROCAL INTERDEPENDENCE Prnicipants intcrrd Coordination by committee or otha such liriwn devise. I.*., by mufurl rdjustment Figure 2-4. The three lypes of interdependence. . interaction require liaison in order to achieve the necessary integration, whcrens those that just follow on from one another can follow plans and schedules. There is a range of devices which can be used to achieve liaison (23): Liaison positions Task forces Special teams Coordinators (or per1:lanent integr'ators) Full project management Matrix organizations. Each of these options provides stronger integration than the last. The primary function of iiaisort poiitioris is to facilitate communication bct\veen groups. Other than this, the liaison position carries no real aurhority and little responsibility. Task forrcs are much stronger. Task forccs provide mission-oriented integration: a group is formed specifically for a particular task and upon completi~nof the task the group disbands. Special tentns are like task forces but httend to regularly recurring types ' of problems raiher than specific issues. A coordirrntor, or permanent integrator, provides a similar service a s a liaison position but has some formal authority. Iqe exercises this authority over the decision-making processes, however, not over the actual decision makers thcmselves. This is a subtle point, but an important one, and it often causes difficulties in projects. The coordinator cannot command the persons he is coordinating l o take specific actions. That authority rests with their functional manager. IJe can, however, influence their behavior and decisions, either through formal means such as managing the project's budget and schedule, approving scope changes, ctc., or through informal means such as his persuasive and negotiating skills. The full prqiecf ninnoger role upgrades the authority and responsibility of the integration function to allow crossfunctional coordination. The integrator-the project manager-now has authority to order groups directly to take certain actions o r decisions. Matrix organizations are, by general consenf, considered about the most complex form of organization structb~re.h4atrix structures provide for maximum information exchange, man~gemcntcoordination, and resource sharing (24). Matrixes achieve this by having staff Sccount simultaneously to borh the integrating (project) managers and the functional managers whose work is being integrated. Both project managcrs and functional managcrs have authority and responsibility over the work, albeit there is a divisioil of responsibility: the functional manager is responsible for the "what" and "by whom"; the project manager decidcs the lor program1 manager Figure 2-5. Typical use of the project management and matrix organizations simulraneously. "when" and "for how much." Unfortunately the person \\tho often comes off the \\lorst in the matrix is the poor soul (at Level 111) \vho is actually doing the work. He reports to t\vo bosses-his project manager and his functional manager-which is not in itself necessarily a proble~n except when, as often happens, the project manager and functional manager are themselves in conflict (for instance, over how much should be spent on the project). Matrix structures generate considerable conflict and sufTer frcm constantly changing boundaries and interfaces (25). The rclative mcrits of the matrix organization vis-a-vis the full-fledged project organization is one of those hardy perennials of project management. Various writers 3t various times have offered all kinds of reasons why one or other form is better. Three points seem to stand out, however. First, the full project management role-with a project manager in overall command of the project-does offer stronger leadership and better unity . ~ is better for achieving the big challenge. Second, the of c o n ~ m a n d It matrix organization is more economical on- resources. For this reason alonc it is offen almost unavoidable on large or complex projects. Third, it is qbite common to find a full-fledged project manager sitting on top of a matrix structure (Figure 2-5)-the two forms are not incompatible but in fact fit rather well together: the top project manager (Level I) providing ' Note, then. that if one were to list the integrating devices in terms o f ascending project management authority. the last three forms would change order to become "project coordinator, matrix organization. full project fonn" (26). the leadership and ultimate decision-making authority, the matrix providing maximum middle management (Levels I1 and Ill) integration. The challenge in moving through this range of liaison devices is very clear. Achieving greater integration requires increased attention to interfacing parties. Interfaces tend to become increasingly difficult to manage as one moves through the continuum. Let us look now at somc experience of managing project interfaces. MANAGING PROJECT INTERFACES Interface Management is not, it must be admitted, a well-developed theory of management well supported by a tight body of research and experience. It is more a way of looking at project management which is particularly useful especially on large, complex, or urgent projects. The insights which are offered below are therefore illustrative rather than comprehensive in their exposition of Interface Management. Keep Static Inferfaces Clearly Defined On pr,ojects, problems require solutions within short : ; ~ n frames, e organizational conflicts abound, and compromises are inevitable. In such an environment, boundaries can blur. It is therefore a fundamental principle of Interface Management to maintain the static interfaces clearly defined. In the Apollo program there was a constant need to reinforce organizational boundaries. When General Phillips was appointed director of the Apollo program in 1963, he found that the program was organized entirely nlong project grounds: one group for the Lunar Excursion Module, one for the rocket, and so on. This created a number of problems, particularly with the wide geographic dispersion of the program. The program was therefore reorganized to stress its functional and geographic needs as well as its project requirements. Five functional divisions wcre created-systcms erigineering, checkout and test, flight operations, reliability and quality assurance, and program control-with project officcs in Houston, I-Iuntsville, and Cape Canaveral. A matrix organization was thus created which rcported to a strong but small program office in Wasl~ington,D.C. (Figure 2-6). This oflice, of only about 120 persons, managed a program which consisted at times of upwards of 300,000 pcrsons. It did so by very clc;rrly dcfining lines of responsibility and a u t h o ~ity and program interface relationships, and by insisting that work bc dclcgated and accounted for strictly in accordance with these lines and procedures (27). Organizational khecks and balances also help kecp organization interfaces clearly dcrnarcated. There are four groups which must always be I Apollo proqrm office Wxhinqron. D.C. I Figure 2-6. Apollo program organization. organizationally distinct on projects: project management, project control, the functional groups, and subprojects. Project management should be separate since its role as an integrator requires it to maintain an independent viewpoint and power base. The Project Control Ofice should be independent since its job is to report accurate and objective data on project progress. If positioned as pad of another group, say project management or construction, there will be a tendency to downplay poor performance because management will inc\sitably hope that things will irn- provc. Functional groups-engineering, contracting, production, testing, rclinbility, contracts, etc.-represent the "cnsine room" of he project: the place wherc technical progress is accomplished. On a large project or program which is divided into subprojects there is usually a number of in~portantschedule interlinkages and competition for scarce resources between t h e subprojects. Often budget and personnel are swapped between the subprojects. Subproject boundaries should be clearly defined and their intcrfaccs closely monitored by senior management if subproject performance is to be properly controlled. T h e organization structure used for the $3.5 billion Aqominas steel nlill project shows these four PI-incipalgroups very clearly (Figure 2-7), a s does the Apollo organization shown in Figure 2-6. (The Aqominas case is described in more detail below.) Early Firm Control of Technical Definition Is ESSENTIAL to Project Success. Research has. shown that time and again projects fail because the technical content of the program is not controlled strictly enough or early enough (28). Software devclopmcnt projects (parlicularly large, complcx ones) are extremely difficult to manage at the best of timcs since their work content (residing for most of the project in the project team's heads) is not as tangible as in other projects. Unless the system design is very carefully dcfined and communicated, the system often ends up technically inadequate, late, and very costly. The software devclopmcnt life cycle consists of five basic phases: concept definition, design, development, evaluation. and operation. The first phase involves problem definition and feasibility study; t hc second, specification of user functions and technical syslcm design; the third, coding, integration testing, user documentation, etc.' Many software projects rush the first two phases and move too quickly into coding. Subsystem interfaces are then wrongly designed and code is inappropriately written. Project managemenl techniques are now being incrcasingly applied to software projects. Configuration managemeqt is being used to help specify the technical content of the system a s it devclops and t o control all changes as they arise. Software developnlent techniques a r e also now emphasizing the careful, top-down evolution of the system's design and programming. Figure 2-8 shows a model of the building process life cycle (29). Interestingly, in most building projects there is in practice no obvious checkpoint a t t h e interface bc~wccnSketch Design and Detailed Design. Thcre ' An "exp:~ntied" vdrsion of !he inform;rlion syslems devclnpmcn! life c),clc i s trcii~edin Chaplcr 12. -- -0 .Y, B r o t Ukz r r c. Kk "I" 1 Z & -E .$ff g5.g L a cl m z f 1 7- ' b L 1 ."e~ -g a 5 .z 2 ~9 s 2" I V I 1 E-2 .Eri g -. 22: a:t;< ' r .- I L / Sketch design - Time Detail design Figure 2-8. Model o f the "oflicial" British building process. are no clearly recognized discontinuilics in either technology, territory, o r , ~ are there any major organizational changes. time at this i n t e ~ f a c e nor Yet the outline design developed in the Sketch Design phase determines the character of the building, and thus lays the iollndations for many of the technical problems uthich may subsequently arise during the project. If this dynamic interface is not properly controlled-and sadly, often it is not-the danger of design errors cropping up unexpectedly later in the project is high (31). Churchill Falls is often lauded as an example of a successful project. The praise invariably centers on its tiglit control of design. An ambitious project in northern Canada, the project consisted of retaining the flow of the Chltrchill River through a series of vast reservoirs and dams over a 250-square-mile basin in Upper Labrador. The project was begun in carnest in 1966 for a budget of $550 million and was completed eight and a half years later on budget and ahead of schedule. Thc project was marked by an early and very intense coordination bctween project management, engineering, construction management, and finance (including insur- ' Miller has shown that.boundan'es should be formed where there are discorrtinui~icsin lime, technology, or territory (30). ance)-each of which was conducted by quite separate companies. At the time of arranging project financing the state of project documentation was such that there were "virtually no questions unanswered" (32). Following this exhaustive initial design there was continuous close review by construction management of the engineering design as it developed, and intensive engineering t o achieve cost savings wherever possible. One way of achieving firm control of design is through configuration management. Configuration rnanagcment documcnts the fechnical design of the project. cnsures regular dcsign reviews, rigorously checks the technical cost and schcdule impact of all changes bcfore approving them, and ensures that all parties working on the project arc using cp-to-date documentation. Configuration rnanagcment has bccn used oriniarily in the U.S. aerospace and information systems scctors; i t is only slowly being applied to other types of projccts. It is not used in the building and ci\lil enginecring industries, for example, though its potential applicability on the larger and more complex of such projects is quite strong. The Skills Required in M a n a g i n g D y n a r ~ l i cInterfaces V a r y Depending on the M a n a g e m e n t Level a n d S t a g e of t h e Project The Trans Alaskan Pipcline (TAPS) r.cmains onc of the largest and most ambitiol~sof recent major projccts. Although constructed in the thrce years bct\vccn 1974 and 1976 (at a cost of approxirnatcly $8 billion). the project had in fact bccn on and off since 1968. Scnior manngcmcnt was requircd to concentrate on a' series of stratcgic issucs of startling variation: firstly, on enginccring (how to prcvcnt hot oil d;ininging 111: Alaskan permafrost and how to dcsign for scismic dani;lgc), and thcn moving through political support (the projcct managcr actt~;tllymoved to Washly in the ington, D.C. to advise the political cffort tliat c v c n t ~ ~ a l rcsultcd 1973 TAPS Act), ir-lfr.astrtrcturc dc\~cIopnlcnt (trnnsportation, camps, cquipmcnt slrpply, and union n c g o t i n ~ i o n ~organi7;rtion ), issues (the devclopmcnt of a higlily dcccntralizcd m a ~ r i xo~.gnniz;ltiononce construction began), cnvironmcnt~lregulations, and firinlly, cnginccring and construction aghin. Thc scquence 01' issucs is intcl-csting: firstly, achicvc agrecnicnt on the technical conccpt and political support for thc project; secondly, assllrc adequate infrastructure and organization; thirdly, resolve environmcntnl, construction, and cnginccring issues as the projcct is built. This is csscntially the institutional, stratcgic, and tactical sequence already noted a s typical for all projects. At the middle managemcnt le\~el,howcvcr (where managers were responsible for anything up to $4 billion of work!), activity cerltcrcd either on resolving enginccring and construction problen~so r on issues of organization and Icadcrship. No- where was the organizational concern more clearly evident than in the change at about 15% offhe way through construction from a 9-tiered, centralized, functional organization to a 4-tiered, decentralized, matrix organization (Figure 2-9). The result was a. highly flexible construction organization relying, like Apollo, on a small cadre of senior managers (33). Emphasis was on leadership, horizontal and informal communication, simple structures, and tight reporting relationships-and getting the job done. The TAPS pipeline matrix organizatio'n tells of an experience very simi- President '-7-J Figure 2-9. The TAPS matrix. to the project management teams: a s the project moved from its design phase to production, there was a "swing" of thematrix towards a greater "project orientation (34). The TAPSIAcominas organization development leads to three important observations about the development of projects. . Typically, Large Projects Require a Decentralized Organization During Production with Centralization Before and After. Both projects 'exhibited the same pattern of centralization-dccentralization-centr;llization. (The final centralized phase during Tlrrn-Over rrrid SfnrtUp has not been discussed here.) The initial, design pl~aserequires unified strategic decision making. During production the volume of work becomes s o great that responsibility must be delegated: the organizi9tion becomes decentralize J under the project and functional matrix control. Finally, at rut-11-Over,the volume of \vork decreases whilc ~ h need c for unified inlegration \vilh Operations' Stclr.t-Up creates the nced for cenlr~lizationonce again. 2. The Project Organization Must Change According to the Needs of t h e Project's Size, Speed, and Complexity. The Acorninas matrix was pln~~ned to change at !he onset of production-about one and a half years after it was set up, \vhich fits with the time it usually takes to " g 1 . 0 ~ "a matrix (35). Research suggests that wthile the timing of the organization change is a function of the project's schedule, the severity of the change depends on the project's size, speed, and complcsit y (36). Once Decentralized, Projects Require a Substantial Management Superstructure t o Effect the Necessary Coordination. Projects decentralize, essentially, to ease the pressure on decision making. Once decentralized, informal controls and comnlunication tend to proliferate and there is a rapid growth in the number of meclings and committees. With this growth in informal decision making there is more need than ever for careful configuration management and budgelary and schedule control. Formal reporting will clearly Ing actual cvcnts considerably, but in an informal org:~nization there is a danger of zissuming that things are happening when in fact they may not be. (Also, informal reponing will tend to concentrale on strategically in~portnntitems only; formal reports should provide a regular update on all :rspcct~of project progress.) I t is, thcrefore, vital to ensure full, regular reporting during this decentralized phase. The c;nracter of a project rhus varies both at different stages of its developnient and at different levels of management. The skills which are required in managing the project's evolution vary depending on the'level of management and stage of the project. Each Major Project Change Point Requires Its Own Distinctive Total Management Changing from one major life-cycle phase to another-from PrcfitrsibilifylFi*n.ribilityto Design, Design to Prodrrcfion, and Prodrrcriorr to TrrrrrOver ( ~ n dSfnrf-Up-is a major event. The Prefeasi6ilit)tlFeasibilify-Dcrigrr transfer is econonrically the most important step in the project's life. Major federal acquisitions have long placed great importance on the need for very thoroirgh feasibility studies. Thorough agency needs-analysis and exploration of alternative systems is now mandatory federal practice. Yet while the importance of a thorough feasibility study is now generally recognized, it is surprising how many proJ'ects do become conrmitted to and move into Dcsigrl on the basis of a totally inadequate feasibility study. Two of the most notorious of recent projects exhibit this clearly. Concorde was conceived almost entirely by the British aircraft establishment, largely on the wings of technological fascination \vith only Ihe minimum of financial analysis (37). The proposal was championed ardently by one 01 two senior British ministers who effectively resisted Treasury pressure to review the financial assumptions. Once the French government joined the project the political momentum became virtually unstoppable. Final commitment to thc project was made on the basis of a twenty-page report which was "little more than a sketch" (38). At this stage the research and development was estimated to cost f 150 million to f 170 nrillion; the final cost is some f 2 billion. Likewise, the Sydney Opera House was committed to on the basis of a totally sketchy design backed by strong political support. New South Wales* prime minister, John Cahill, saw the Opera House as an imaginative political act (39). A design competition was held and Utzon's design was selected as winner. The design was, however, little more than diagrammatic. There was little evidence of strt~cturalfeasibility and no cost estimate. A quantity sur\!eyor was therefore asked to prcparc one, which hc did "undcr tlurcss in a few hours" (40) arriving at a figure of $A 7 million. The final cost, after drastic redesign (resulting in so reducing the scencry spacc that opera cannot be fully staged in the building) was $A 102 million. The transition from Design to Prodrrcfion is less clear-cut than that from P~~cfcasibilifylF~asibili~y to Design. It is also much broader in scope and involves much fullcr management attention. At this interface, management must be fully active in all the major project subsystems: project definition, organization, environment, and infrastructure. The overriding preocct~pationshould be that the strategic parameters are properly set as the interface is crossed, since once Prodrrctiort begins the scale and pace of events increase dramatically. So important is this interface to project implementation success that it requires "total" management attention: planning, organizing, directing, and controlling. Contracting-the key interface activity in fact-offers a good example. The contracting process must be supported by project management through integrated planning, thorough negotiating, and through close monitoring. Often contracting is not managed but just happens, thus swamping the project with \vork and delaying it considerably. (A~ominashad to sign 400 contracts during an 18-24-month period. Accomplishing this was a major management achievement in its own right.) Trrrn-Over aud Sfart-Up probably receives considerable non-project planning and control but generally from places other than within the project team. The meshing of the two imporlant phases of project construction and operations is complex and has large cost implications. All 1,evel I subsystcrns must be complete and activated as soon as Turn-Over occurs. Despite the obvious importance of smoothly transitioning from Prodrrction to Turn-Over, it appears rare, especially in the construction industries, that a project planning group prepares and monitors integral..-;Iplans for crossing this interface. This might be partly because of the subsfantial differences in Stc~rt-Upbetween aerospace programs and large ncvf .:,qpital expansion facilities (factories, telecommunication systems, ports, sic.) which may havc dcpressed the e\tolulion of dcfense/aerospace program management ideas on this important area of \vork. Pljnning Must Be Phased to the Stage of the Project Life Cycle The differing nature and requirements of the various project life-cycle phases require that different issues be addressed as the project unfolds. Project planning cannot be done comprchensi\~ely,once and for all, ng !he beginning of the project. The uncertainty during the early stal;c.:. (IC a project is too great. Instead, planning must be incremental (41). !rl,tral planning must concentrate on building viable planning bases for each principal subsystcm, detail being added later in phase with the project schedule (Figure 2-1 1). The Apollo Mission's "Phased Project Planning" explicitly recognized this (42)-major life-cycle phases bcing identified and planning rc a#;..-.v checkpoints positioned along the life cyclc (Figure 2-12)-as have subse- Time A.2.1 A. 1 1 C Project 8 A.1.1 Time ;A.1.2 ___C , 0.1.Parcels 0 2.Parcels BJParcelr fr 0.2 1 B.4.Parcels 1 C 1 1 I 1 Figure 2- 1 1. Project planning development. quent U.S. aerospace manuals (c.g., DOD 3200.9, DOD 5000.1, AFSC 800-3, to mcntion just three). Apollo was fortunate in that it had nearly a decade for N A S A to develop its systems and program planning. Many projects are less fortunate. Early North Sea oil projects, for example, were i n ~ ~ l c m c n t cwith d great urgency, and made worse by the short summer weather window for towing and positioning platforms in the North Sea. Oil companies found is wonh while o r how it i s to be achieved. Ifyou choobc a sc~lurior:, ~ i t h c ~ u t - 8 PROJECT MNNIW) M RASIBIUW Y SMTECY NRHOVER DUlGN PRODUCTION & SfART-UP ECONOMIC EVALUATION Benefits Risk Continue Appraisal with View lo Changing Project Specifiulions i f Necuury Impact on other Dusinclrs Functions Auused Adjustments Made as Necessary Auur Project Cost for Product Pricing " PROJECT DEFINITION System Spea Ease Technology S blimate Project Schedule Outline Design Configuration Dcfinition Budget by Major Arcas Milestone Schedule Detailed "Planning" Schedule Futher development or Outline Design. Schedule and Budget Detailed Contract Specs and Drawings Ovcnll Schedule Requirements Dctrilcd Budge~/Contnct Bids Operating Manuals Training Primay Malerials Preparation HandovrSchedule Test 'Schedules Mmc FINANCE Potential Sources Principl SourMajor Payments Schedule ~etailcdSdureclr hforc Dc~ailedCjsh Requirements Dcleilcd Pnymcnts Schcdule by Crcditon B Currency Annual financial Operating Plan Initial Impact Assessment Definition of Environmcntal Impact Statement National & Larl Govcrnment Suppon Schcdulc of Approvalr Required Government or Community Supporl Groups Identified Permit ExpeditingSystem Expediting Schcdulu Pubic Relations Markctina Personnel Invcn~oryPlanning Safety Procedures Ou~tandingLegal luues ENVIRONMENT 1 Attitude Autucd Supplier Situation Assased' 'I"p?y3S sru~isAg Urld J 2 ~ o d l J r ~ xulr).y n!i!(!q!suodux ~ w u o u a di s r o r d P "-W"!fi lm%d uqia!ur%o Ix!OJd • dn-uns pur i u m d o l u ~ UO!la!u~8~0 ~qtu+do \O rr) d PJaPJ0 PO u q WoS klq!md • suo!an3s!a uo!un paua!p ~ X J I -ug) ~ o r ar u~q Urld 8 u ! l ~ ! l 0 8 3i~~ ~ l u o 3 SIV"~~W w r l d s u y l . 1 ~PaI!Vl¶Q ~=l!rlaa uo!la!utho r a u ~ suo!l!p 4103pur S U J ~l ~~ ~ l u o 3 P=Y!1 -uapl [ a ~ u w a dLax p ? y ! ~ l p lSUJ¶I~(S UO!lWJOJUI 1 0 f 1 ~ pw!uuaaw n!i!l!q -!suodnn 1 t d 1 3 u ~ d m ~ n - rlru -airw q ~ o q q U O ! ~ ~ ~ J ~ S L'~011 IO~ -mug y ua!aasa!O -¶lrJlS J O l 3 ~ J l U O 3 - :JOJ ndmuo-~l l r ~ a r o ¶U!Ilno 133r0~dlr!1!ul thcmselves having to develop a new generation of rigs, which involved the use of new technology, working in a harsh and poorly documented environment, without adequate codes of practice or regulations-all within a very tight schedule. A slippage ofjust a few weeks could result in the delay of the whole program for nearly a year. To speed up the program, projects were often sanctioned on the basis of preliminary designdata and manufacturing was overlapped with design as much as possible. While these early projects could probably have been managed more efficiently if there had been a longer start-up time to acquire environmental design data, develop codes, and put project management systems and procedures in place (as has been suggested (43)), the economic pressure on distributing North Sea oil as soon a s possible effectively precluded this option. In these projects, the project life cycle not only determined the sequence and dcgrce of planning appropriate at a given stage, it set an absolute limit on the time available for planning. Ensure Full Working o u t of the Static Subsystems a t Each Stage of the Project Life Cycle "Static" project subsystems (technical definition, organization, environment, and infrastructure) must be fully worked out at each phase of the project's life cycle. Unfortunately this does not always happen, often because the habits of an industry o r organization have institutionalized a culture of neglecting certain "essential" subsystem considerations. Movie pr.oductions are notorious, for example, for overrunning budget. The most common reason for their doing so is that there is a culture of allowing the director to work out how thc film will develop as he shoots it-there is neither "design" nor "schedule." (No one on the set, including the director, knew how Cnsablar~cawas going to end until the cnd of shooting; Francis Ford Coppola did not have the ending of Apocalypse Now worked out even as he entered the editing room.) Sometimes, the effects of urgency are so great that it is not possible to \{pork out fully the static subsystems before moving on to the next phase. Many defense projccts cannot avoid modifying their performance requirements (because of a changed Threat Analysis) yet must still keep to schedule (44). This is the situation known as concurrency (45): there is a substantial body of knowledge available showing that concurrency is invariably associated \vith difficulties and overruns (46). Concurrency has proved unavoidable on several occasions in the nuclear power industry, basically when changes have been required at an advanced stage of construction- o r during commissioning because of rcgulalory changes o r technical PI-oblcms(47). The.tendency for architectural design to dominate the building proccss has already been commented upon. Figure 2-U compares two basically well managed, large, complex building projects (48). Thc first (A) was managed by architects working under the RlBA1s Plan of Work (49). Since the Plan of Work assumes compcritive bidding by the main contractor, it does not mention the use of any form of construction management advice prior to construction bidding. Thus the architects did not schedule the project or seek any form of production advice during the early design phases. Project B, however, used a large U.S. AIE firm which was familiar with project management practice and employed both systems design techniques and a management contractor. As a result there was early project scheduling and production advice so that the technical definition subsystem was fully worked out early in the project. Control Needs Vary Depending o n t h e Level of Control and Stage of t h e Project As the tasks of the different levels of project management vary, so do their control needs. Level 111 management uses frequent, often daily, control of key performance factors such as earih moved, concrete poured, pipe laid, vessels installed, tests completed, together \(lit h basic cost data. At Level I1 further data are required in addition to the Level 111 "key drivers": for example, inventories, drawing approvals, transportaPreliminary Design Review (PDR) - First A n i c l e Customer Design Certification Configuration Acceptance Inspection Revim Rcadinm Review critical (FACI) (CARR) (DCR) Fliaht Readiness ~erign Rwiew Rwim Phase I Phaw II P h r w Ill (FRR) CARR CARR CARR (CDR) - - r. Requirement def initi o n - Oesiw Manufacluring d ~ v t l o p ~ t hign s t ~ t - Design requirements bawlin established - V a i f i c r t i o n of manufacturing n des* specilication and assessment of t a t rcsult~ P~oduct c o n f i i r a t ion Drawing b a u l i ~ established bulb mablishcd. Certify Determine f lipht readinem 0' flight a d ~ . ~ u ~ a f t rele.~ . Orou,,d and for launch f w manned flight figure 2-13. Comparison of two British building projects. (For Subsystem Coding refer to Figure 2-8.) tion, camp capacity, security, accidents, contract management, changes pending and approved, contingency reserves, These are d d a not needed on such a frequent basis as the "key drivers" data. Level I interests are broader still: exception reporting on progress (i-e., report problems and poor trends only), interface relationships with other subsystems (e.g., between subprojects, between operations planning and project progress), training, cash flows, etc. (50). "Control" has a meaning which is greater than merely monitoring. It is used in the broader context of setting standards, monitoring, and correcting for deviations between actual and planned performance. This more complete interpretation of control is the one used in cybtrnefics. (The word "cybernetics*' itself derives from the Greek "to steer".) The nature of control during PrefeasiBili~lFeasibilityand Design is different from that during Prohrcfion and Turn-Over. As has been already noted, the need during the early stages of a project to plan, design, and estimate correctly is very large. The costs during these early phases are small compared with the total project cost, and so the need to monitor them (at least from a project as opposed to design point of view) is correspondingly S'mall. Later, during Production, the crucial cgntrol need is the monitoring of performance to ensure that quality is being achieved and resources are being deployed on schedule and in budget. Hence the nature of control changes during the project life cycle from predicting to monitoring. "Control" in this broader sense also varies depending upon the interests and objectives of the persons doing the cofi~rolling.Different groups, with their different interests, will have different control needs. A project owner, for example, will want to monitor the economic worth of the project as much as its technical, cost, and schedule performance. A contractor might be particularly concerned with cash flow control. Personnel Issues Will Vary, Again Depending on the Level of Control and the Stage of the Project Conflict is .inherent in every project, not least since the primary project objectives-quality, cost, and 'schedule-are themselves in conflict. Quality costs money and requires time; working more quickly costs money. Also, projects often engender contractual and community conflict. Studies in 1he mid-1970s (51) have shown that the pattern of conflict varies during the project life cycle (Figure 2-14): schedule and priorities dominate thc carlx phases, with technical issues coming to the fore later (and with cost as a consistently low-conflict item). One should not a.ssumc that this pattern applies for all projects, however-one would normally V At tt,e project f o r m t i o n Stan At the early During the main project project phase8 Proicct life Toward the end of the project * F inid Figure 2-14. Relative intensity of conflict over the project life cycle. expect greater conflict over technical issues earlier in the project; the conflict pattern might -vary by type of project (and contract type); cost pressures may be generally more domipant than they were on the projects studied; and personal issues are probably stronger on matrix and overseas projects. Dczpite such necessary cavcats, the findings arc extremely valuable: they provide solid evidence that the type of conflict varies according to the stage of the project life cycle. Similar rescarch ( 5 2 ) has also studied how the factors which are (a) most important and (b) most inhibiting to project success vary with the life-cycle stage (Figure 2-15). I t too has shown that personal issues vary with the stage of project life cycle. The nature of conflict also differs according to the level of management. All ~ h rcscarch c on project conflict undcrlcrkcn to datc cither concentrates on Level IIIIII management o r has h e n in the aerospace industries, which have typically been more sheltered fronr cxtcrnal pressures. Conflict n t Levcl I is usually totally different, requiring other modes of resolution. Most of (he bchaviorial work in projects to date assumes a normative, mechanistic view of the world: a world where people make rational % significant 0 0 U 0 I I I -0 H1 I TI 5 Unmotivated tram I Poor 1 W . 0 1 Penonrl ambition/rnotivation Financial limits Technologcul advantage - 1 Team motivation and coowration 5. - I Personal ability and motivation Poar smiw mrnwJemmt Technical problems J P a n financial sum& 1 - 7I " m Senlor manaaement s u ~ w n Technical ex~ertise I I 1 Team motivation E gg Lnd.nhip probirms -- Technical p o b l m c. Poor financial wpport Clirnt Poor finrncirl support Objeclivn Scntor management support F~nancialsuomn I I decisions based on trade-offs of costs and' benefits, and where open dialogue between men of goodwill leads to amicable solutions. While this approach is largely valid for most ir~t~a-project conflict and behavioral issues, it is often inappropriate for dealing with outside-world issues. The Level I manager ill as often as not find himself having to deal with people having cgmpletely different \lalue systems from those of the project's. Hence, while the more mechanistic approach to personal issues is appropriate to 1,evcls I1 and 111, Level I often requires a more political approach. Many practitioners conlment on the importance of leadership to project success. Strangely, however, the research evidence on !.his is at best skimpy (53). There would appear to be little doubt, Ilowever, that leaders' abilities are associated with success or failure, although it is generally difficult to isolate the effect of leadership from other factors such as top management support and authority. Leaders have a particularly important role to play in creating the context for success. Leaders create climates: their personality can sway the ability to ev;tluate a project objectively; they can often negotiete or communicate things which otherwise ulould not be agreed. These abilities are particuhrly important at Level I in working with outside parties. RELATION TO PROJECT SUCCESS To what extent are thes; findings relevant to that most important issue, the question of project success? The first point to note in addressing this question is that the notion of project success is, as has already been observed, relative. In recent research on this question, I used four distinct measures: functionality-does the project function in the.way the sponsors expected?; project management-was the project completed on schedule, in budget, to technical specification?; contractors' long-term commercial success; and, where appropriate, termination efficiency (54). This said, there would seem to he considerable evidence that several of. the issues raised in this chapter bear strongly on the chances of success. Firstly, the discipline of Interface Management itself was seen to be important whcre there is significant interdependence on projects (55). Organizational issues have been shown to relate to project success (56). Issues of design management, concurrency, and planning have a polentially enormous impact on projects if not nccomplished effectively. The evidence on the relationship between human factors and project success is, by comparison, considerably softer, as was just noted (57). Insofar as Interface Management relates to clear planning and the orderly management of relevant project systems and dimensions, it is almost self-evident that it must contribute to project success. The insights related in this chapter are, almost without exception, proved through quoted experience or academic research to have an effect o n the final outcome o f a project. The twist, however, is that the definition of success varies from individual company to individual company; and that as one moves from Level 111 to Level 1, the diversities and uncertainties o f the many groups operating in the project's environment are likely to make judgments about the causes and effects of project success increasingly difficult to make. REFERENCES I. Katz, D. and Kohn. R.-L. 71re Sociijl Psyclrolog~Of Orgonizatiotrs (Wiley. New York. 1966), pp. 14-29. 2. Kast. F. E. and Rosenzweig. J. E. 0rgtrtri:trtinn n r ~ dAfotrtlgetrrrtrr. A Systctrrs A p proach (McGraw-Hill. New York. 1970). 3. Lorsch, J. W. and Lawrei-re, P. R. Srtrdies In 0rgtltri:rrrion Drsign (Itwin Dorsey. Homewood, Ill., 1970). 4. Putnam W. D. "The Evolution of Air Force System Acquisition hfanagement." RAND. R-868-PR, .Santn hfonica, Calif., 1972. 5. Beard, E. Dcvrlopitrg rlre ICBM (Columbia University Press. New York. 1976). 6. Sapolsky, H. Tlrr Poluris System Devrlopmmr: Blrreurrcrnric nnd Progrr~trrtnnricSlrccrss in Govertrt~rttrr(Harvard University Press. Cambridge. Mass., 1972). 7. Peck, M. J. and Scherer, F. M. 7/rc Wrripotts Acqrri~itionProcrss; A n Econottric Analysis (Harvard University Press. Cambridge, Mass., 1%2). 8. Acker, D. D. "The Maturing of the DOD Acquisition Process." Defrtrsr S ~ s r e m s Vol. Mnt~ngettrcnrR E V ~ C I O , 3(3) (Summer. 1980). 9. Kelley, A. J. "The New Project'Environment" in N e w Ditrrtttsions of Projrcr Afntroget n m ? (Lexington Books. Lexington, Mass., 1982). 10.- Wearne, S. H. Principlrs of E~rginfcringOrgnttizurion (Ed\vard Arnold. London. 1973). I I. Morris, P. W. G. "Interface Manage-An Organization Theory Approach to Project Management." Project A4unngrmctr1 Q~rarrrrly.Val. lW2) (June, 1979). 12. Kerzner, H. Projrcr A4unngrttrm!: A Sys~rmsApproaclr t o Plntrtring. Schrdrrling orrd Cotrrrollitig (Van Nostrand Reinhold. New York, 1982). 13. The organistic/mechanistic classification of organization types was developed by Bums, T. and Stalker, G. M. in Tlre Afanogrnrenr of ltrnovn~ion(Tavistock. London, 1961). 14. Parsons, T. Srrticrure and Procrss in Modern Socittics (Free Press. Glencoe, Ill., 1960). 15. Morris, P. W. G. and Hodgson, P. "The Major Projects Association and Olher MacroEngineering Societies: Their Activities and Potenrial Contribution to the Development of Project Management," in 81lr H'orld Congress on Projcct Mntrngrtrrrnt (Internct. Rotterdam, 1985). 16. See, for example: Metcalf, J. L. "Systems Models. Economic Models and the Causal Texture of Organizational En*:ironments: An Approach to Macro-Organizational Theory." Hlttrron Rrlrrions, Vol. 27 (1974). pp. 639-663. Also;Emery, F. E. and Trist, E. L. "Sociotechnical Systems," in Systrtns T/rinking, ed. Emery. F. E. (Penguin. Harmondsworlh, 1969). pp. 24 1-257. 17. hforris. P. W. G. and Hough, G. Ji. Preconditions of Success and Failure in Afajor P r o j r c ~ s(Major Projects Association, Templeton College. Oxford, September, 1986). 16. Miller. E. J. "Technology, Territory and TTme: The Internal Differentiation of Complex Production Systems." Htrnian Relations. Vol. 12(3) (1959). pp. 270-304. Also. Miller. E. J. and Rice, A. K. Sysrcms of Organizarion. Tlie Corirrol of Tnsk arid Scnrienf Borrndnrics (Tavistock. London, 1%7). 19. Lawrence, P. R. and Lorsch, J. W . 0rgarti:arion and Enuirontricn~;Ifnrragirrg Differcnriarion and Jnrrgrclrion (Harvard University Press. Camhridge, Mass.. 1967). 20. Morris, P. W. G. "Organizational Analysis of Project Managemen't in the Building Industry." Brrild Jrrtcraorio~rnl.Vol. 6(6) (1973), pp. 595-616. 21. Ibid. 22. Thompson, J. D. Org;atii:n~iovsit; Action (h!cGraw-Hill. New York, 1967). 23. This list is based on Ga!braith, J. !?. 0rgnni:nrion Drsilpn (Addison-\\'csley. Reading, Mass.. 1973). . 24. Davis, P. and Lawrence, P. R. Mmrrix (Addison-Wesley. Reading, hfass., 1977). 25. Ibid. 26. Youker, R. "Organizational Alternatives for Prgject Management." Projrcr Afrlnagernrnr Qrmrrerl~,Vo!. 6(1) (1977), pp. 18-24. 27. See for instance Baumgartner, J. S. "A Discussion with the Apollo Program Director. General Sam Phillips." in Sysrtnis .!ftrnngmtmr, ed. Baumgartner. J . S. (The Bureau of National Affairs. \Vashington. D.C.. 1979). 28. Alexander, A. J. and Nelson, J. R. "Measuring Technological Change: Aircraft Turbine Engines" Rand Corporalion, R-IOI?-ARPNPR, Santa Monica. Ca (June 1972); Cochran, E. G., Pat). A. L. and Rowe, A. J. "Concurrency and Disruption in Hew Product Development" Cnli/ornia Manogrnicnt Rruie~c*(Fall. 1978); Department of Energy. "North Sea C rsts Escalation Study" (Her hfajesty's Stationery Oflice, London, 1976); General Accsunting Office, "Why Some \\'capons Systems Encounter Production Problems \Vlile Others Do Not: Six Case Studies" GAOISSIAD-85-34, \Vashinglon DC,24 MaY 1985; Harman A. J. assisted by Henrichsen S "A Methodology for Cost Factor Comparison and Prediction" Rand Corporation, R-6269-ARPA, Santa Monica, Ca (August 1970); Marshall, A. W. and Meckling. W. H. "Prediclability of the Costs. Time add Success of Development" Rand Corporation. P-1821. Santa Monica, Ca (December, 1959); Merrow. E., Chapel. S. W. and Woflhing. C. A. "A Review of Cost Estimation in New Technologies: Implications for Energy Process Plants" Rand Corporation, R-2481-DOE, Santa Monica, Ca (July, 1979); National Audit Onice "hfinistry of Lkfence: Control and Management of the Development.of hfajor Equipment". report by L'te Comptroller and Auditor General (Her Majesty's Slationery Office, London. I2 August 1986); Murphey, D. C.. Baker. B. N. and Fisher, D. "Dctermin3nts of Project Success." National Technical Information Services, Springfield. Va.; Myers. C. W. and Devey, M. R., "How Management Can Affect Project Outcomes: An Exploration o f t h e PPS Data Base." RAND, N-2106, Santa Monica, Calif., 1984; Perry, R. L. el at. "System Acquisition Experience" Rand Corporation, Rhf-6072-PR. Santa Monica, Ca (November, 1969); Pugh, P. G. '"Who Can Tell What Might Happen? Risks and Contingency A l l ~ \ ~ ~ a n c e sRoyal ." Aeronautic Society Management Studies Group. 1985. 29. Royal Inrtitule of British Architects, Plori of Work. Hundbook of Arclrirrcrrrrnl Prncricr and Afatia~tmcnt(RI BA. London. 1963). 30. Miller, E. J., op. cit. (18). 31. Morris, P. W. G. "Syslems Study of Project Management." Buildini. Val. CCXXVl(6816 and 6817) (1974). pp. 75-80 and 83-88. 32. Warnock, J. G. "A Giant Project Accomplished-Design Risk and Engineering Management," in Successf~rllyAccotnplislting Giant P r j r r r s . ed. Sykes. A. (Willis F a k r . London. 1979). pp. 31-61. 33. Moolin. F. P. and hfccoy, F. "The Organization and Management of the Trans Alaskan Pipe!ine: The Significance o f Orgenizational Structure and Organization Changes." Proc.errlirrgs of rlrc Projeer Alntrngcr~rrrrrIr~srirrrrrCorr/i~rcrrcr.Arlur~rrr.1980 (Project Managenirnt Institute. Drexel Hill. Pa.. 1960). 34. Reis de Carvalho, E. and Moms. P. W. G. "Project Matrix Organizations, Or How To D o The Matrix Swing." Procrdirrgs of rlrr Projrcr Alrrtrrrger~rrtrrIn~rirrtrrCot?/croice. 1-05 Arrgrlcs, 1979 (Project Managenlent Institute. Drcxel Hill, Pa., 1979). 35. See, for instance Davis, P. and Lawrence. P. R. h.ftrrri.r, op. cit. (24) Also, Whitmore. K. R. hfo1ri.v O i ~ g ~ ~ r r i ~ oi ft ri oC~IIU~III~~II~I~ ~is h f r r ~ ~ ~ r f i r ~ ~ ~ ~ r r i r ~ g - hCor~rp(rnirs, f t ~ r L e ~ ; n gM.S. Thesis (Sloan School o f hfanngement, MIT. Cambridge, htass., 1975). 36. See Rcis de Carvalho. E. and Morris. P. W. G., op. cit. (34). 37. Ifall. P. Grcnr Plu~rrtitrgDisnsrcrs (H'eidenfeld and Nicolson. London, 1980). pp. 87108; Morns. 36. See Hall, P., op. cit. (37); and Edwards. C. E. Co~rcorde: Trn Yrtrrs c ~ r r da Billion c r Press. London. 1972). f'orrrrds L ~ r ~ (PIUIO 39. Ko~rzmin,A. "Building the New Parliament House: A n Opera House Revisited?" 1~1111run Fr'r,trrrcs, \'ol. 31 1) (Spring 1960). pp. 5 1-74. 40. Ilall. P.. op. cit. (37). p. 141. 41. lionr.itch, M. "Designing and hfanaping Large-Sci~le, I'uhlic-Pri\*a~e Tcchnolopical Enirr Sorirry. Vol. 1 (1979). pp. 179tcrpri-es: A State of the Art Revie\r+." Trclrrrolo~)~ 192. 42. Scamans. R. and Ord\t,ay, F. I."The Apollo Tr;jdi~ion: An Object Lesson for he htanngcrncnt of Large Scvlc Technological Endenvors." I r r r r r d i ~ r i p l i ~ t t rSiirrrce ~ - ) ~ Reui~bcl.\'01. 2 (1977). pp. 270-304. 43. Depnnmcnt o f Energy, op. cit. (28). 44. htorris, P. W. G. and Hough. G . H.. op, cit. (17). 45. Acker. D. D..op. cit. (8); also Iiarvcy. T. E. "Concu~rcncy Today in Acquisition hlanagcmcnt." Drfrrtsr S~srerrrsAfnnri~r~trcrr~ Rcuict18. \'ol. 3 ( 1 ) (\Vin~cr.1980). pp. 14- IR. 46. C o c t ~ r ~ n E.. G..Patz. A . L. and Ro~ve.A . J. op, cit. ( 2 8 ) ;Gcnerlrl :\ccounring Office. ibid; Patz, A . L . . Irrrroutrrio~rPirfolls crrrd Aftrtrtrgc~r~rorr S ~ ~ l r r ~ i ior r~Iiiglr r c Trrlrtrolog)~ Irrdrrrrrirs (University o f Southern California. I.os Angclcs. Calif.. 1984). 47. Kutr~er,S. "The Impact of Rcgula~oryAgencies on Supcrprc<ccts." in f l ~ r ~ t ~ r i Enging. rrccrirrg rrnd Corr,c~rrrcrirrgrltr Slrprrprojrcrs, American Socic~yof Civrl Engineers' Conference, P;~cihcGrove. 1970 (ASCE. New York, 1978); hfonopolics and Mcrgcrs Commission, Ccnrral Elccrricity Gerrcrnri~rg~ h o r d A. Rcporr 011 rltc Opcrrr~iotrofrhe Dourd of 11s Sysrrrn for rlrc Gorrr-trrio~rrrnd Srrpply of EIPCII ,c ;I). i r r RrrlA ( l j e r Stajcsty's Stationery Office. London. 1981). 48. Morris, P. W. G. op. cit. (20. 31). 49. Royal Instit~rteo f British Archittcts, op. cit. (29). c 50. Morris, P. W. G. "The Use and hfnnagcment o f Projccr Control Sys~cmsin ~ h 60's." f'rojrcr Afrrnngrrtrrrrr Qrrurrrrly, Vol. XI(4) (Deccmhcr. 1980). pp. 25-28. 51. Thanihain, 13. J. and Wilcmon, D. I-. "Conflict hf;tn;~g~.mcntin Projcct Life-Cycles." Slootr dfnrrrrgt~rrr~~rrr Rcuic,tv (Summer, 1975). 52. Dugnn, 11. S., Thilmhain, H. J . and \\'ilemon, D . I-. "hfi~nngingChange Through Project hfanagcment." Pr.occrdingr of rlre h o j r c r Af~rtrrr~crtrc~rrr I ~ r \ r i r r tCorrfcrcrrcc. r~ Arlarrro. 1980 (Projcct M;~nngcment Insti~utc.Drcxel tiill. Pa.. 1980. 53. Gcnrniil, C . and Thamhnin, H. J. "Project Pcrfo~mnnceas a Fi~nctionof the Leadership Sl).les of Project hfnnagcrs." Projrcr hloirogr~ttrirrlrrsrirrrre Conferrtrcc. 1972 (Project Sfanagemcnt Institute. Drcxel 11111,Pa.); lionvdle. G . and \'an Snnt, J . I~trplc~nr~rrnrion for Sus~rrinabiliry.Lcssotisfionr Itircgrn~cdR~rrulDcvelopnirtir (Kumarian Press. West Hartford, Conn., 1985); Murphy, D. C. et at., op. cit. (28); Myers, C. W. and Devey, M. R.. ibid; Rubin. I. M. and Sellig, W. "Experience as a Factor in the Selection and Performance of Project Managers." IEEE Transucrions otr Etipiriccririg A/unrrptr?icnr, Vol. EM-131(35) (September, 1967). pp. 131-135; Ruskin, A. M. and Lerner, R: "Fore- 54. 55, 56. 57. casting Costs and Completion Dates for Defense Research and Development Contracts" IEEE Trt~nsctctiorison Ertgitrtcring AI~~tinpcnient, Vol. Ehi-19(4) (November, 1972), pp. 128-133. Morris. P. W. G. and Hough, G. H., op. cit. (17). Ibid. Pnulson, B. C., Fondahl, J. W. and Parker, H. W. "Dc\*clopment of Research in the Constr~~ction of Transportation Facilities." 'Technical Report NO. 223. The Conslruction Institute, Department i,f Civil Engineering, Stanford University, Slanford. Ca.. 1977. Gcmmil, G. and Thamhain, H. J.. op. cit. (53).