Co

Ra

b

a

AA

KOSSC

1

tmccomtdamtGLok

AT

0d

Journal of Operations Management 29 (2011) 434–448

Contents lists available at ScienceDirect

Journal of Operations Management

journa l homepage: www.e lsev ier .com/ locate / jom

ross-functional alignment in supply chain planning: A case study of sales andperations planning

ogelio Olivaa,b,∗, Noel Watsonb

Mays Business School, Texas A&M University, College Station, TX 77843, USAZaragoza Logistics Center, C/Bari 55 PLAZA, Zaragoza 50193, Spain

r t i c l e i n f o

rticle history:vailable online 25 November 2010

eywords:perations interfaceales and operations planningupply chain planningase study

a b s t r a c t

In most organizations, supply chain planning is a cross-functional effort. However, functional areas suchas sales, marketing, finance, and operations traditionally specialize in portions of the planning activi-ties, which results in conflicts over expectations, preferences, and priorities. We report findings from adetailed case analysis of a supply chain planning process that seemingly weathers these cross-functionalconflicts. In contrast to traditional research on this area, which focuses on incentives, responsibilities,and structures, we adopt a process perspective and find that integration was achieved despite formalfunctional incentives that did not support it. By drawing a distinction between the incentive landscapeand the planning process, we identify process as a mediator that can affect organizational outcomes.

Thus, organizations may be capable of integration while functions retain different incentives and orien-tations to maintain focus on their stakeholders’ needs. Through iterative coding, we identify the attributesof the planning process that can drive planning performance—information, procedural, and alignmentquality—but also find evidence that achieving alignment in the execution of plans can be more impor-tant than informational and procedural quality. In addition to process attributes, we also identify socialelements that influenced the performance of the planning process and place the information processing

er soc

attributes within a broad

. Introduction

In most organizations, supply chain planning—the administra-ion of supply-facing and demand-facing activities to minimize

ismatches, and thus create and capture value—requires aross-functional effort (Braunscheidel and Suresh, 2009). Suchross-functional collaboration facilitates an assessment of the statef the supply chain, of the needs of the organization, and the deter-ination of an approach for creating and sustaining value based on

hat collaborative assessment. Such an approach usually involvesetailed evaluation, planning, and execution at the strategic, oper-tional, and tactical levels (Anthony, 1965). Both the operationsanagement and organizational behavior literatures refer to this

ype of collaboration as integration (Barratt, 2004; Ellinger, 2000;

riffin and Hauser, 1996; Kahn, 1996; Kahn and Mentzer, 1998;awrence and Lorsch, 1986). In some cases, this integration mustvercome the specialization of functional areas, such as sales, mar-eting, finance, and operations, and their focus on their own portion

∗ Corresponding author at: Mays Business School, Wehner 301C - 4217, Texas&M University, College Station, TX 77843-4217, USA.el.: +1 979 862 3744; fax: +1 979 845 5653.

E-mail addresses: roliva@tamu.edu (R. Oliva), nwatson@zlc.edu.es (N. Watson).

272-6963/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.jom.2010.11.012

ial and organizational context.© 2010 Elsevier B.V. All rights reserved.

of the planning activities. Such specialization or differentiationis notorious for generating conflicts over differing expectations,preferences, and priorities with respect to how the matching ofdemand and supply should be accomplished (Shapiro, 1977). Withincreased competition and globalization creating new opportuni-ties and challenges for supply chain planning (Raman and Watson,2004) and fostering further differentiation within the organization,it is clear that firms will struggle even more with supply chain inte-gration as they attempt to manage and respond to the increasingcomplexity of markets, suppliers, and investors.

We expect integration in supply chain planning in a highlydifferentiated organization to require quite an explicit and broadcross-functional reach. Although particular cross-functional inter-faces have been developed—e.g., marketing and logistics (Ellinger,2000; Stank et al., 1999), and purchasing and manufactur-ing (Fawcett and Magnan, 2002)—very few organizations haveachieved the broader-reaching integration that consistently devel-ops multi-functional plans that are executed in a coordinatedfashion (Barratt, 2004; Fawcett and Magnan, 2002). Furthermore,

very little empirical research has been done on functioning inte-gration approaches (Malhotra and Sharma, 2002) and a detailedunderstanding of interdepartmental integration based on micro-level data has yet to be established. Therefore, a comprehensiveunderstanding of cross-functional integration is lacking in the lit-

ration

ecta

cpttTmoiittsps

foidasi1ptpc

pr

2

insaiprPweee

oei(obtpbotis

R. Oliva, N. Watson / Journal of Ope

rature (Pagell, 2004). Given the lack of detailed frameworks forross-functional integration, we decided to use case-based researcho explore how a functionally differentiated organization couldchieve such integration for supply chain planning.

We identified a highly differentiated organization with a supplyhain planning process that, on initial observation, seemed to beroviding support for cross-functional integration. As we mappedhe formal incentives in the organization, we found a typical collec-ion of different goals and orientations for each functional group.his diversity of incentives and objectives had generated misalign-ent in planning and execution until the firm instituted various

rganizational changes. Although unobserved dynamics during themplementation of the changes in the organization could havenfluenced operational improvement, the organization reported lit-le change in the organization’s formal incentives. This suggestedhat some contemporary organizational mechanisms, such as theupply chain planning process, could still be at work providing sup-ort for integration. How could a planning process act as an ongoingupportive mechanism for cross-functional integration?

To answer this question, we adopted a process perspectiveor our analysis, focusing on the sequence and interdependencyf activities designed to achieve a goal. Through iterative cod-ng, we identified the attributes of the planning process thatrive planning performance. The constructs resulting from thisnalysis—information, procedural, and alignment quality—shareome characteristics with distinctions made in decision making andnformation-processing theories (Daft and Lengel, 1986; Galbraith,973; MacKenzie, 1984; Simon and Newell, 1972). In addition torocess attributes, we also identify social elements that influencedhe performance of the planning process and place the information-rocessing attributes within a broader social and organizationalontext.

In the following sections we review the relevant literature androvide motivation for our research (Section 2), and describe ouresearch site and methodology (Section 3).

. Literature review

Integration occupies a central place in several domains—ncluding management, strategy, organization theory, eco-omics, operations management, marketing, and informationystems—with each discipline focusing on different organizationalctivities or components. These disciplines share a perspective onntegration as distinct and interdependent organizational com-onents constituting a unified whole, while displaying requisiteesponsiveness to each other and to the environment (Barki andinsonneault, 2005). In operations and supply chain management,ithin-firm and across-firm integration has been shown to influ-

nce positively firm performance (Baker and Sinkula, 1999; Droget al., 2004; Fugate et al., 2009; Narver and Slater, 1990; Vickeryt al., 2003).

Theoretically, one stream of operations management researchn integration/coordination takes its cue from the economics lit-rature, which explores coordination in terms of how incentives,nformation flows, and hierarchy affect the allocation of resourcesCachon and Lariviere, 2005; Weng, 1995). This approach assumesptimal system objectives to which allocation decisions shoulde aligned and are generally modeled as a buyer–supplier rela-ionship (e.g., Cachon, 2003), with little attention paid to thelanning processes within the firm or the supply chain beyond the

uyer–supplier dyad. Under this perspective, lack of coordinationccurs when decentralized decision makers ignore the optimal sys-em target because they have incomplete information or conflictingncentives (Narayanan and Raman, 2004). Coordination, however,hould be considered different from integration in that where coor-

s Management 29 (2011) 434–448 435

dination takes the target for granted, integration often involvesdetermining this target simultaneously with the aligning of allo-cation decisions.

Much coordination research then, concerns how actors shouldbe compensated given the informational and hierarchical structure(see Eliashberg and Steinberg, 1993; Sahin and Robinson, 2002;Whang, 1995, for surveys). Coordination mechanisms for inter-nal alignment identified by this research include accounting-basedcost schemes (Celikbas et al., 1999; Porteus, 2000; Watson andZheng, 2005), improved contract design (Chen, 2005; Gonik, 1978;Li and Atkins, 2002), decision making hierarchies such as first-movers (Kraiselburd and Watson, 2007; Li and Atkins, 2002), andinternal markets (Kouvelis and Lariviere, 2000). Many researchers,however, observe that an incentive-compatible scheme or aninformation scheme may not induce the actors to implementsystem-wide optimal behavior (Chen, 1999; Porteus, 2000; Watsonand Zheng, 2005). In practice, operations managers are limitedby their decision making capabilities and may commit errors intheir replenishment decisions (see Croson et al., 2005; Sterman,1989, for evidence of poor replenishment decision-making perfor-mance even under conditions of reduced complexity). To addressthese limitations, the recommended coordination mechanisms arebroadened to include assessment of the cognitive burden imposedby the evaluation and incentive systems (Kouvelis and Lariviere,2000; Porteus, 2000; Watson and Zheng, 2005); support for com-plex decision making, whether from quantitative models (Yano andGilbert, 2003) or decision support systems (Crittenden et al., 1993);and outsourcing planning to competent third parties (Troyer et al.,2005).

Recognizing the influence of organizational factors on integra-tion, empirical researchers have striven to identify organizationaland environment antecedents of integration, how integration influ-ences firm performance, and mediators and moderators of theeffect of integration (Braunscheidel and Suresh, 2009; Fugate et al.,2009; Jaworski and Kohli, 1993; Pagell, 2004; Slater and Narver,1994; Vickery et al., 2003; Wong and Boon-Itt, 2008). The market-ing literature has given special attention to issues of learning andinnovation, and their relationship with integration and firm perfor-mance (Baker and Sinkula, 1999; Hurley and Hult, 1998; Sinkula,1994; Slater and Narver, 1994), and the operations managementliterature has explored the influence of enterprise informationsystems on integration across or within organizations (Bendolyand Cotteleer, 2008; Bendoly and Jacobs, 2004). However, withinthe operations management literature, beyond the recognitionof a link between process and information technology, we findinsufficient attention paid to a process perspective on drivers ofintegration. Even when the above recommendations are consid-ered to have some implications for the process dimension, theyare usually only directionally suggestive, rather than appropriatelyprescriptive, with respect to process. Specifically with respect towithin-firm integration, very little empirical research has beendone on functioning organizational or supply chain planning inte-gration approaches (Malhotra and Sharma, 2002) as most of theempirical research focuses on supplier–buyer dyads (Wu and Choi,2005). As a result, a detailed understanding of interdepartmentalintegration based on micro-level data has yet to be established(Griffin and Hauser, 1996; Kahn, 1996; Kahn and Mentzer, 1998)and a comprehensive understanding of cross-functional integra-tion is still lacking in the literature (Holweg and Pil, 2008; Malhotraand Sharma, 2002; Pagell, 2004).

Within the organizational behavior literature, the focus on gen-

eral integration within firms has a longer and better-establishedtradition that more explicitly incorporates the behavioral dynamicsof the key actors. Classic research suggests that the effort requiredto achieve integration increases with the level of differentiationin the organizational environment (Galbraith, 1977; Lawrence

4 ration

afemvtrt

tHa(tisca1attr1Tptmo

bublemsrog

2

assaiatbmigboo

sts

icp

36 R. Oliva, N. Watson / Journal of Ope

nd Lorsch, 1986; Lorsch and Allen, 1973; Thompson, 1967), dif-erentiation being defined as “differences in the cognitive andmotional orientation of managers in different functional depart-ents” (Lawrence and Lorsch, 1986, p. 11). Differences amongst

arious functions’ cognitive and emotional orientations—not onlyheir goals and incentives but also their perspectives on time andelationships—create short-term conflicts and deemphasize long-erm organizational goals.

The organizational behavior research on integration has concen-rated on the responsibilities and structures supporting integration.ere, “responsibilities” refers to the distribution of decision rightsmong participants in the collaborative effort. Lawrence and Lorsch1986), for example, recommend for highly differentiated settingshe role of integrators for coordinating functional efforts. Thesentegrators act as translators, mediators, and integrative goal-etters, helping guide the various function, which have differingognitive and emotional perspectives, into collective efforts (Brownnd Duguid, 1991; Hargadon and Sutton, 1997; Orlikowski et al.,995; Yanow, 2000). “Structures,” in this literature, refers to theccompanying formal (and social) systematic arrangements, rela-ionships, and infrastructure that regulate the interaction amonghe participants in the collaboration effort. Examples of structuralecommendations include the formation of work groups (Galbraith,977) and the use of boundary objects (Carlile, 2002; Star, 1989).his literature, however, also pays little attention to the processerspective. Even in the case of work groups or groups whose iden-ity is conceivably based on what they do and how they interact,

ore attention is focused on the fact that they act and interact thann how they act and interact (Brown and Duguid, 2001).

In both the operations management and the organizationalehavior literatures, therefore, process is one of the lesser-nderstood components of integration. For the organizationalehavior literature, with its broad organizational overview, the

ack of focus on this context- and operations-specific dimension isxpected. Although processes are a touchstone of the operationsanagement community, recommendations for integration have

imilarly focused on responsibilities and structures or throughefining the organizational behavior’s broad organizationalverview to more operational themes, yet still taking processes forranted.

.1. A process perspective on integration

By process, we mean a sequence and interdependency ofctivities designed to achieve a goal. Processes systematize andtandardize organizational learning at the micro-level (i.e., deci-ions and actions) in ways that are not easily matched bypproaches based on responsibilities and structure or by contract-ng or market-based interventions (Cyert and March, 1963; Nelsonnd Winter, 1982). Thus, a process perspective could complementhe macro-level focus of those approaches from the organizationalehavior and operations management literatures. This comple-entarity could materialize in scenarios where all approaches,

ncluding the process approach, are directly supportive of inte-ration. However, given process’ potential intermediate positionetween, on one hand, macro-level explanations and, on the other,rganizational performance, it could act as a mediator of the effectsf these macro-level interventions on performance.

We believe that the process perspective therefore, can shedome much-needed light on the challenges of functional integra-ion in supply chain planning and, in so doing, extend the focus in

upply chain management from coordination to integration.

We, however, are not the first to affirm a process perspectiven this way. It is arguable that a focus on the effect of pro-ess on the integration of R&D and manufacturing in the newroduct development literature has revolutionized both the aca-

s Management 29 (2011) 434–448

demic field and practice (Wheelwright and Clark, 1992). Similarly,a focus on processes and their implications for organizationaldesign has already been recommended in the information tech-nology literature. Malone and Crowston (1994) and Malone et al.(1999) emphasize the management of interdependencies amongresources and activities and seek to develop a coordination theoryby characterizing various kinds of interdependencies and iden-tifying the mechanisms that can be used to manage them. Thatperspective does not, however, capture the traditional focus on theactors and their natural differentiation seen in the work of orga-nizational theorists (e.g., Lawrence and Lorsch, 1986; Thompson,1967).

The collaborative planning processes we examine in our casestudy are referred to in the practitioner literature as sales andoperations planning (S&OP) processes (Bower, 2005; Lapide, 2004,2005). Among the primary roles of S&OP processes is to facilitatemaster planning, demand planning, and the flow of informa-tion between them. Master planning is primarily concerned withthe coordination of the supply side of the organization andseeks the most efficient way to fulfill demand forecasts over themedium term (Stadtler, 2005), facilitating finer levels of plan-ning such as purchasing and materials requirements, production,and distribution planning. Demand planning is concerned withthe customer-facing side of the organization, predicting futuredemand from scheduled customer orders or extrapolating demandfrom prevailing market conditions or from the demand-influencingactivities (e.g., promotions and new product launches) of the orga-nization or its competitors. A basic S&OP process facilitates thetransfer of information from demand planning to master planning.Practitioners and academics alike argue that this transfer processcan move beyond the superficial synchronization of master anddemand planning to sophisticated joint planning (Chen et al., 2006;Lapide, 2005; Van Landeghem and Vanmaele, 2002).

The fact that little empirical micro-level data exists for sup-porting the development of a process perspective on supply chainplanning sets the expectation that, at least initially, such a perspec-tive should be based on empirical studies such as ours. Furthermore,processes such as the S&OP process, which are the objects of ongo-ing research on their potential integrative capabilities but are alsopractitioner-inspired, make good candidates for empirical scrutiny.Finally, given that other disciplines such as organizational behavioralso have a rich perspective on integration, we expect the processperspective may need to draw on insights from these disciplines.

3. Research methodology

As argued in the previous sections, how a process integratessupply chain planning across multiple highly differentiated areasis understudied. As such, our research is exploratory in nature andrequired qualitative methods to identify the relevant constructsand develop propositions about their interactions (Eisenhardt,1989). For the theoretical elicitations of constructs we followedthe methodology proposed by Glaser and Strauss (1967) for thedevelopment of grounded theory.

3.1. Research site

Our research site, Leitax (the firm’s real name has been dis-guised), is a consumer electronics firm with headquarters innorthern California and with a global sales presence. Leitax sellsprimarily through retailers such as Best Buy and Target and has

distribution centers in North America, Europe, and the Far East.Production is handled by contract manufacturers with plants inAsia and Latin America. Leitax maintains seven to nine models inits product portfolio, each with multiple SKUs. Product life rangesfrom fifteen to nine months and is getting shorter.

ration

pcwocviacetrthaibe

sfctamVr(wicia

3

pAeosasiimapS

wpaacatoaa

twa

R. Oliva, N. Watson / Journal of Ope

Our initial contact with Leitax was for an unrelated researchroject. However, as we realized the nature of the S&OP pro-ess, we requested permission to study it. This planning processas considered at least reasonably successful by many of the

rganizational participants (although tensions with the processonsistent with differentiation were revealed in in-depth inter-iews), and its current forecasting accuracy and other operationalndicators (e.g., inventory turns, obsolescence) corroborates thisssessment. Our initial analysis found that a forecasting pro-ess, together with the supporting mechanisms for informationxchange and elicitation of assumptions, was capable of managinghe organizational conflict and the informational and procedu-al shortcomings that emerged from the forecasting biases ofhe functional areas (Oliva and Watson, 2009). This analysis,owever, did not make clear whether performance and forecastccuracy improved solely because of the improved forecast-ng methodology, or also because the whole organization waseing more effectively managed through a coordinated integrationffort.

Three reasons make this site interesting from a research per-pective. Prior to the implementation of the S&OP process, Leitaxaced challenges in cross-functional integration which seem quiteommon to supply chain planning across a wide range of indus-ries with an active sales force driving demand (Shapiro, 1977). Inddition, the S&OP process is an increasingly common approach foranaging supply chain planning (Chen et al., 2006; Lapide, 2005;an Landeghem and Vanmaele, 2002). These two reasons make ouresearch site a suitably representative case for a single case studyYin, 2003). Finally, because Leitax implemented its S&OP processithout changing its seemingly unsupportive incentive landscape,

t provides an opportunity for unusual revelation about how pro-ess can foster integration given a seemingly hostile context; its therefore a critical case that also justifies the single case studypproach (Yin, 2003).

.2. Interview protocol

Since we had heard about the S&OP process at Leitax from arevious visit, we were able to develop an explicit protocol (seeppendix A) for the two visits we scheduled to explore its differ-nt elements as suggested by Eisenhardt (1989). Given the naturef the research, interviewees were not required to stay within thetandard questions; an interviewee who seemed to be exploringfruitful avenue was permitted to continue in that direction. This

emi-structured protocol changed over time as each subsequentnterview was used to triangulate the responses from previousnterviews and expanded the list of questions as we uncovered

ore elements of the planning process. This continuous expansionnd improvement of the protocol after each interview is a normalart of the process of grounded theory development (Glaser andtrauss, 1967).

The intent of each interview was to understand the intervie-ee’s role in the S&OP process and his or her perception of therocess and to explore the orientations of the different actorsnd functional areas. To assess these orientations, we explicitlysked interviewees about their incentives, goals, internal work pro-esses, and relationships to other actors and functional areas. Whenppropriate, we asked interviewees about their own and other par-ies’ sources of power, i.e., the commodity through which theybtained the ability to influence an outcome—e.g., formal authority,ccess to important information, external reputation (Checkland

nd Scholes, 1990).

Finally, part of the protocol also included direct observation ofhe main planning meetings and extensive debriefing time after-ards. This allowed us to observe the behavior of the different

ctors in the planning process (as opposed to just hearing their own

s Management 29 (2011) 434–448 437

accounts of that behavior) and to obtain explanations for observedbehavior during the meetings.

3.3. Informant selection and data collection

The S&OP process that we studied at Leitax, although still evolv-ing, was already fully operational at the time of our study. Data werecollected through 25 semi-structured, 45- to 90-min interviewswith leaders and participants from all the functional areas involvedin the S&OP process, as well as with heads of other divisions affectedby the process. Although the initial set of interviews was arrangedwith all the participants in the planning process (those who sat atthe main planning meetings) the list of informants was expandedthrough interviewees’ referrals. If an interviewee mentioned, forinstance, a report that she took into the planning meeting but thatshe did not prepare, we asked for access to the person directlyresponsible for preparing that report. Additionally, at the end ofeach interview we asked the informant if there was anyone elsewe should interview in order to gain a richer understanding of theprocess. We had access to all the informants we requested. Our listof interviewees included four Sr. VPs, of which three participateddirectly in the process (Product Marketing, Global Operations, andProduct Business Unit); two Sr. Directors (IT, and Demand Planningand Replenishment Management); four departmental directors(Demand Management Organization, Product Planning and Strat-egy, and two regional directors of sales); three production planners;two facility managers (a factory and a distribution center); andthree analyst that supported the planning process. Some key actorswere interviewed more than once. Minimum interviewee tenure atLeitax at the time of the study was three years.

We conducted most of the interviews in Leitax’s northern Cal-ifornia facility, but some follow-up interviews were conductedby telephone. All interviews were recorded. Several participantswere subsequently contacted and asked to elaborate on issuesthey had raised or to clarify their comments. We triangulatedeach interviewee’s responses with answers from other actors anddocumentation of the process and its outcomes provided by theorganization, and used follow-up interviews to clarify differences.

Interviews were supplemented with extensive reviews ofarchival data—such as worksheets and presentation slides thatformed the bulk of the information shared between depart-ments and the supply chain, financial performance data, annualreports, and written communications among the participants of theforecasting and planning process—and direct observation of twoplanning and forecasting meetings. Although we were not autho-rized to tape the meetings, we were allowed to take notes and hadaccess to all the participants right after the meeting to clarify anyexchange that had taken place during the meeting.

3.4. Data analysis

In analyzing the data, we controlled for the effects of our a prioribeliefs regarding integration in a variety of ways. Prior to catego-rizing or coding, we summarized our field notes in the form of adetailed case study that relates the story of the initiative and its cur-rent challenges (Watson et al., 2005). This narrative was primarilydetailed from one researcher’s field notes; the other interviewer’snotes were used for corroboration. The recorded interviews wereused to help reconcile discrepancies. Finally, the company partic-ipants were asked to review the entire case, not only their own

quotations, for accuracy. To minimize researcher biases during thedata analysis, researchers alternated between independently cod-ing data based on categorizations and then jointly assessing the fitof the current categorization scheme and making refinements tothe scheme.

438 R. Oliva, N. Watson / Journal of Operation

Table 1Key constructs supporting cross-functional integration.

Constructs Definition

Information quality The degree to which a process enables theinformation used for decision making to beappropriate, both in content and in form, for thedecision maker and the decision

Procedural quality The degree to which a process continuouslyensures that the rules of inference used tovalidate information, and to make decisionswithin and across functions, are sound

Alignment quality The degree to which a process ensures thatorganizational and functional goals are supportedand that resulting actions are synchronized

Constructive engagement Active involvement by relevant participants ineffectively collecting, validating, and processing

tatcihtimtogbdovmcspstrstywcdaoac

tootaieitThcc

and Milgrom, 1994). High-powered incentives represent explicit or

information and in voicing and defending theirinterpretations

The purpose of our investigation was to try to explain whyhe S&OP process provided benefits despite the maintenance ofseemingly unsupportive set of functional incentives and orien-

ations. Our first step was to map the sequence of activities thatonstituted the planning process and the information flows withint. We then moved to analyze how the process was enacted andow it affected individuals’ behaviors. Our analysis used an even-ually rich coding structure to categorize various factors at workn the organization that influence supply chain planning perfor-

ance. We used this coding structure to identify key constructshat could explain supply-chain planning performance. Althoughur focus was on process, other categories were needed to distin-uish other organizational features. For an initial coding, we usedroad categories such as process, integrative behavior, functionalifferentiation, and infrastructure. Given the exploratory naturef the research, initial constructs from the literature did not pro-ide complete guidance. Our categorization, therefore, requiredultiple iterations as we attempted to be both parsimonious and

omplete (Glaser and Strauss, 1967). Related to the process dimen-ions, we first explored the role of information processing in thelanning process, which resulted in a focus on coding categoriesuch as information collection, sharing and interpretation. Wehen broadened our perspective to include decision-making thatesulted in additional coding categories such as problem framing,election of evaluation criteria, and solution evaluation. Focus onhe social and political dimension of organizational integration,ielded coding on items like, consensus reaching, and engagement,hile integrative behavior items focused on information sharing,

ollective validation, and coordinated execution. For organizationalifferentiation, we focused on coding categories such as incentivesnd general functional orientation. Infrastructure covered typicalrganizational structural mechanisms used for information sharingnd organizational decision-making such as meetings, informationhannels and information packages.

This iterative coding eventually yielded a representation ofhe process in which the logic of integrated planning was morebservable. Additional coding and analysis of the process and itsrganizational environment resulted in three constructs relatedo process attributes—information quality, procedural quality, andlignment quality—and a construct related to the nature of thenteractions between the participants in the process—constructivengagement. These constructs, together with the logic embeddedn the planning process, provide a parsimonious explanation of whyhe process was effectively sustaining cross-functional integration.

hat is, through the lens of these constructs, it is possible to explainow the sequence of activities of the planning process supportedross-functional integration (see Table 1 for a description of theseonstructs).

s Management 29 (2011) 434–448

While process attributes—i.e., information, procedural andalignment quality—are the result of designed process character-istics, and hence relatively stable, constructive engagement is adynamic element that needs to be initiated and sustained. Thus,we further expanded the scope of our inquiry to account for thedrivers of this social engagement.

In the next section, we describe the organizational context ofthe observed planning process and the tensions created by theexisting functional incentives and orientations at Leitax, thus estab-lishing one of our claims about the research site. The next threesections follow the logic of our inquiry. Note, however, that ourline of inquiry during the interviews was not about individual’sperceptions of a particular construct, i.e., we did not rely on ourinterviewees’ assessment of integration or their understanding ofhow integration was achieved by the process. Instead, we focusedon obtaining an accurate description of the implemented process,and how the process specifications affected the actors’ involvementin it. As such, the evidence that we present here is based more ondescriptions of the process and examples of the actions taken bythe actors (either observed or elicited from the interviews), ratherthan the more traditional reporting of case studies based on quotesfrom the actors. In Section 5 we present a description of the sup-ply chain planning process and its performance up to the time ofthe study. The focus of this presentation is to emphasize the logicdriving the planning processes and the roles that different activi-ties play on creating effective plans. It is this detailed descriptionof the actors, their activities, and their motivations that providethe grounding for the theoretical developments that follow. Sec-tions 6 and 7 present our analysis and generalization to theoreticalprinciples and propositions (Yin, 2003). In Section 6, we presentthe constructs that describe the process attributes that explain theprocess’ performance. Since the developed constructs have solidtheoretical underpinnings on planning and decision making pro-cesses, in this section we focus on providing evidence of how theseprocess attributes are manifested in the observed process and howthey interact. Finally, in Section 7 we present the mechanism thatexplains the initiation and sustenance of constructive engagementand examine priority for alignment within planning in the broadercontext of competing organizational dynamics. We close the paper(Section 8), by discussing the implications of our findings for prac-titioners and researchers interested in supply chain integration.

4. Incentive landscape

In this section, we describe the organizational context in whichthe observed planning process operates. We first map the incen-tives and orientations, what we refer to as the incentive landscape,across different groups within the organization at Leitax. This map-ping highlights the natural tensions in the landscape, which cancreate integration challenges. These challenges are particularly vis-ible in the absence of the implemented S&OP process.

The hallmark of a differentiated organization is the diversityof roles and responsibilities that are distributed among func-tional areas, and create the need and challenges of integration. Wedefine an incentive landscape as the formal incentives that compli-ment these roles and responsibilities and the objectives/mandatesthat define these roles and responsibilities. As such, the incentivelandscape captures the formal mechanisms that the firm has toinfluence the behavior of groups and individuals and includes bothso-called high-powered and low-powered incentives (Holmstrom

implicit contracts that can be based on measures of performance(e.g., commissions), while low-powered incentives reward a par-ticular role/job as opposed to performance in that role (e.g., fixedcompensation). Most firms provide a combination of explicit high-

ration

phettipcht(so

ilpsL

tettcfifizpwctrttstpgpihbtgenaflbtmop

tpFtictgct

R. Oliva, N. Watson / Journal of Ope

owered and low-powered incentives; the latter are used whenigh-powered incentives are either absent or where attempts toxploit them result in highly dysfunctional outcomes due to con-ract incompleteness (Holmstrom and Milgrom, 1994). Note thathis definition of incentive landscape does not include the socialncentives or intrinsic motivators that an individual obtains forerforming a job (Ryan and Deci, 2000), as these are generallyonsidered not as easily determined by the firm. Our definition,owever, captures: (a) the mechanisms that the organization haso align the efforts of individuals to the goals of functional areas,b) the variety of internal objectives that can emerge from a diver-ified organization, and (c) the consistency or complementarity ofbjectives across functional areas.

The incentive landscape at Leitax can be mapped by consider-ng the cross-functional differences in objectives and orientation:ong-term goals that naturally define functional purpose, functionalreferences for planning horizons and the handling of relation-hips, and short-term direct monetary incentives (Lawrence andorsch, 1986).

The differences in functional objectives at Leitax are theraditional ones (Shapiro, 1977) and therefore support the gen-ralizability of our findings. The sales function’s main goal waso drive wholesale sales to resellers and retailers given organiza-ional and internally determined sales targets; finance managedash flow, monitored financial viability and communicated withnancial stakeholders; while operations managed contract manu-

acturers, and distribution of products. This natural differentiationn responsibilities drove orientations with respect to planning hori-ons, and formality of structure and relationships. With respect tolanning horizons, sales emphasized current sales opportunitieshile operations focused on medium- to long-term inventory and

apacity planning requirements. Differences in formality of struc-ure were observed in terms of reporting relationships, criteria forewards, and control procedures. Operations, for example, in ordero communicate manufacturing requests to its contract manufac-urers needed more established routines and specific details thanales needed to manage its sales accounts. With respect to incen-ives, generally a mix of low- and high-powered incentives wereresent for most functions with the sales force having the most toain from their performance based incentives. The sales force wasartly compensated through commissions that were based on sell-

n—i.e., the sales from distribution centers to resellers. Operationsad part of their compensation based on inventory performanceut were also accountable for avoiding shortages or excess inven-ory and for managing relationships with suppliers. The financeroup and other participants such as product planning and strat-gy were mainly compensated with low-powered incentives. Theet result of this incentive landscape was both a perception of biasnd propensity for bias in decision-making that could lead to con-icts. For example, the sales organization was credibly perceived toe biased in favor of ensuring higher inventory levels than neededo avoid any stockouts, while the operations group wanted more

easured and stable inventory levels, preferring smoothed extrap-lations from historical orders versus seemingly enthusiastic salesrojections.

The potential for conflict and lack of coordination are evident inhe absence of a formal planning process. Before 2001, the demandlanning and master planning processes at Leitax were ill-defined.or new product introductions and midlife product replenishment,he sales directors forecasted sell-in sales and disseminated themnformally to the operations and finance groups, sometimes via dis-

ussions in the hallways. These shared forecasts were to be used byhe operations group to guide the supply chain and by the financeroup to guide financial planning and monitoring. These sales fore-asts, however, were often mistrusted or second-guessed whenhey crossed into other functional areas. In response, for example,

s Management 29 (2011) 434–448 439

the operations group often generated its own forecasts in order tominimize the risk of inventory shortages, its primary responsibil-ity. Similarly, marketing would devise its own forecasts when thatgroup suspected that the sales forecast had been distorted by pro-motions. Sales, for its part, believed that finance exerted too muchpressure on the forecasts by, for example, urging sales to increaseforecasts that did not meet the company’s financial goal. A managerin operations described the range of orientations as they influencedforecasting as follows:

From finance forecasting to financial plans that met their expec-tations “start with a revenue number and then lets see what kindof products can build to the number,” passing through the salesand product planning forecasts, each with its own biases, to ourown operations forecast that ignored any market informationand just looked at past builds as the most stable and desirableforecast.

Although the incentive landscape naturally encouraged mis-alignment, the lack of formality in the planning process exacerbatedits impact on planning performance. Data relevant to forecastingwere usually inaccurate, incomplete, or unavailable and the lack ofobjectives and monitoring mechanisms for the planning processmeant that process improvement could not be managed. A his-tory of poor demand and supply planning fostered complacencyin the preparation of the sales force’s forecasts. Faulty forecastprocess features included forecasting capacity shortages insteadof unconstrained demand and little mitigation of subjective andquantitative biases. It was not unusual for sales to arrange deals toextend the production of products for which an end-of-life deci-sion had already been approved and the supply chain had beendepleted. Support for supply management was equally ill-definedas master production schedules were both sporadic and unreliable,and suppliers had learned to mistrust them.

This inefficiency and lack of coordination, previously hidden bybooming growth in the consumer electronics sector, caught up withLeitax in 2001 when poor planning and execution resulted in aninventory charge of roughly 15% of revenue for FY 2001–2002. Theinventory write-offs were followed by major changes during thefall of 2001, including the appointment of a new CEO and severalvice-presidents.

The new SrVP of Global Operations soon initiated a redesign ini-tiative to formalize the planning process. Throughout this redesign,however, the incentive landscape remained largely unchangedwith high- and low-powered incentives being retained and func-tional orientations maintained as a result of the natural roles andresponsibilities of these functions. Yet, at the time of our study,three years later, there was evidence that the S&OP process wasfacilitating integrated supply-chain planning despite the unsup-portive incentive landscape.

5. Planning at Leitax

In this section, we document the planning process introducedat Leitax, as well as the responsibilities and structures created tosupport it. This description of the process was generated frominterviewees’ descriptions, archival documents, and our directobservations of the process. We use a representation based on theconstruct of validation of plans that arose from our refinement ofour coding of the process features. This representation character-izes the cross-functional planning at Leitax as the simultaneous

validation of an initial organizational plan by multiple functionsand stakeholders, spurring the plan’s revision until it reachesa certain level of acceptance. The section concludes by report-ing on the performance and social outcomes of the implementedprocess.

440 R. Oliva, N. Watson / Journal of Operations Management 29 (2011) 434–448

ration

5

auuraonaw

oppBtdoiacdtirtutacopffisw

5b

nmacdp

Fig. 1. Sales and ope

.1. Sales and operations planning process

Each month, a forecasting group that included the sales directorsnd the VPs of marketing, product strategy, finance, and prod-ct management generated a consensus sales forecast that wassed to drive all supply management, demand management, andelated financial planning decisions. Ownership of the forecastingnd planning process was assigned to the demand managementrganization (DMO), which was responsible for synthesizing theecessary data, managing the planning process, resolving conflicts,nd creating and disseminating demand projections to keep paceith worldwide operations.

Before describing the S&OP process in detail, we provide anverview (see Fig. 1). The first step was the cross-functionalreparation and dissemination of a planning-related informationackage known as the business assumptions package (BAP). TheAP contained the core details of the product offering and promo-ions plan: decisions on product offerings, launch and end-of-lifeates for each product line, price offerings, promotions, and detailsf business deals with customers. The BAP was discussed in detailn a formal meeting with all the members of the forecasting group,nd, after a signoff, was carefully validated via (1) functional fore-asts; (2) a consensus forecast, which determined the potentialemand based on the product offering and promotion decisions inhe BAP; (3) financial validation, which revealed the financial viabil-ty of the plan; and (4) operations (supply chain) validation, whichevealed whether there was sufficient capacity to accommodatehe demand. If there were capacity or financial concerns, the prod-ct offering and promotion decisions in the BAP were changed andhe process of validation repeated. Additional planning, feedback,nd learning were facilitated by deliberate interaction within theonsensus forecasting meeting. The whole process was repeatednce per month, with the BAP and the Consensus meetings takinglace in the first and third Wednesdays of the month. The functionalorecasts were performed between those two meetings, and thenancial and operational validation taking place after the consen-us meeting but before the following month’s BAP meeting. Belowe describe each of these activities in detail.

.1.1. Information collection and product offering planning: theusiness assumptions package

The BAP integrated the plan for product offerings (price plans,ew product introduction dates, and end-of-life dates) and pro-

otions, and information reflecting the market environment, such

s marketing strategies and intelligence about market trends andompetitors’ products. Each month, the entire BAP was updated,iscussed, and agreed upon by the forecasting group. The productlanning and strategy (PPS), marketing, and DMO groups used the

s planning process.

market information to assess the market’s impact on future busi-ness performance and entered their recommendations (explicitlylabeled as such) into the BAP.

5.1.2. Validation: functional forecastsOnce the BAP was agreed upon, it was used as a starting point to

elaborate three different forecasts at the product-family level witha focus on sell-through (the quantities sold by resellers). These threeforecasts were the work of separate functional areas:

Product planning and strategy (PPS): This three-person groupsupported all aspects of the product life cycle from launch toend-of-life, assessed competitive products and the effects of pricechanges on demand, and prepared a top-down forecast of antici-pated global demand for Leitax products. The PPS forecast derivedproduct- and region-specific forecasts from a worldwide estimateof Leitax’s product demand and from historical and current trendsin market share.

Sales directors (SDs): Leitax’s five SDs used a bottom-up approachto generate their forecast, aggregating their account managers’knowledge about channel holdings, current sales, and expectedpromotions with their own knowledge. Their forecast was firststated as a sell-in forecast, then translated into a sell-throughforecast by maintaining a maximum level of channel inventory(inventory at downstream DCs and resellers) for each SKU.

Demand management organization (DMO): The DMO prepared aforecast of sell-through by region, entirely on the basis of statisticalinferences from past sales. This was primarily intended to providea reference point for the other two forecasts. If either of these fore-casts deviated significantly from the DMO’s statistical forecast, theresponsible group was called upon to investigate and justify itsassumptions.

These three functional forecasts ignored capacity constraintssince it was common at Leitax at this time for forecasts to be affectedby perceptions of present and future supply chain capacity andthus to become self-fulfilling prophecies. For example, even if moremanufacturing capacity were to become available in the future,deflated forecasts might have positioned insufficient quantities ofraw materials and components.

5.1.3. Validation: consensus forecastsThe three groups’ forecasts were submitted on Excel templates

and merged into a proposed consensus forecast using a formu-laic approach, devised by the DMO, which gave more weight to

the SDs’ bottom-up forecast in the short term and increased theweight of the PPS’s top-down forecast as the consensus forecastwent out to the future. Each month, the forecasting group evalu-ated the proposed consensus forecast and the three independentforecasts.

R. Oliva, N. Watson / Journal of Operations Management 29 (2011) 434–448 441

y turn

sdtafiokfcsssvp

5

wfepwtdgdfiacsu

ocrcmCIcb

5a

T

Fig. 2. Inventor

Whereas the early consensus meetings had sometimes con-umed an entire day and had been characterized by heatediscussions, by the fall of 2003 the meetings were lasting betweenwo and four hours and conversations were cordial. The reasonsnd assumptions underlying diverging forecasts would be identi-ed and discussed and the proposed consensus forecast revised bypen conversation. When justified on the basis of the SDs’ intimatenowledge of upcoming sales deals or prospects, bottom-up salesorecasts that were slightly higher than the PPS’s or DMO’s fore-asts were often accepted. The finance group, although it did notubmit a forecast, voiced opinions and concerns based on its under-tanding of the revenue potentials at stake. With little functionaltake in the outcomes of the meetings, the PPS group tended to beocal, objective, and unemotional about the forecasts and demandlanning.

.1.4. Validation: financial and operational assessmentsThe final consensus forecast was sent to the finance department

here, in conjunction with pricing and promotion informationrom the BAP, the forecast in units was converted into its revenuequivalent. Forecasted revenues were compared with the com-any’s financial targets. If gaps were identified, the finance groupould first ensure that the sales group was not underestimating

he product’s market potential. If revisions made at this point stillid not result in satisfactory financial performance, the forecastingroup would return to the BAP and, together with the marketingepartment, revise the pricing and promotion strategies to meetnancial goals and analyst expectations. These gap-filling exercises,s they were called, usually occurred at the end of each quarter andould result in significant changes to the marketing plans and, con-equently, the forecasts. The approved forecast was released andsed to generate the master production schedule (MPS).

Operations validation of the final consensus forecast was anngoing affair. Over time, suppliers responded to the increasinglyonsistent and reliable MPSs by providing Leitax with more accu-ate information about the status of the supply chain and theirommitments to produce orders. In addition, more reliable MPSsade the suppliers better prepared to meet expected demand.

apacity issues were also discussed in the consensus meetings.n essence, the operations feedback to the planning group syn-hronized the demand and supply plans and ensured compatibilityetween them.

.1.5. Other roles of the consensus meetings: tactical planningnd learning

Consensus forecasting meetings were also planning meetings.he forecasting group considered new product introductions and

s performance.

determined initial inventory for product launches. As a launch datedrew near, the consensus forecasting meeting was used to reportthe expected inventory status at launch, revise regional requests,and seek consensus on regional allocations, taking into considera-tion any expected shortfalls.

Products to be discontinued in the current quarter were also dis-cussed during the consensus forecasting meetings. Proposed dropdates were presented together with sales to date, predicted sales forthe rest of the product’s life cycle, and available inventory. Consen-sus was sought on how a product’s end-of-life should be managed.

For ongoing products, sell-in and sell-through rates and channelinventory were analyzed and compared to expected sales. Financeaggregated channel inventory and sales data from resellers’ weeklyelectronic data interchange (EDI) reports. Discrepancies betweenreported inventories and inventories calculated from sell-in andsell-through data were resolved and consensus sales forecasts wereupdated according to the latest information. Promotion and price-change decisions were also revised in light of these data.

Finally, the consensus forecasting meeting was a source offeedback on forecasting performance, particularly on biases exhib-ited by previous functional forecasts. The DMO, being responsiblefor forecast accuracy, continuously monitored the accuracy of allthe forecast streams and used this information to improve itsalgorithms and heuristics and to help functional areas improvetheir forecasting processes. At some point, for example, the DMOpresented evidence that sales forecasts tended to overestimatenear-term sales and underestimate long-term sales; the forecast-ing group used this information to better interpret and weight thesales forecasts and the SDs used this information to improve theirforecasting processes.

5.2. Operating and social performance

By the fall of 2003, Leitax had attained good performance lev-els in forecasting accuracy. Three-months-ahead forecast accuracy(ABS[sales-forecast]/forecast) for sell-through (sell-in) was 88%(84%)—up from 58% (49%) in the summer of 2002 (see Oliva andWatson, 2009, for a description of Leitax’s efforts to improve fore-cast accuracy). Operational effectiveness was also at good levels:inventory turns in Q4’03 were 26 (up from 12 the previous year),average on-hand inventory was $23 million (down from $55 mil-lion) (see Fig. 2), and on-time delivery of orders was 75% (up from

35%) (see Fig. 3). Excess and obsolescence costs were practicallyzero in fiscal year 2003 (down from an average of $3 million forfiscal years 2000–2002).

We also observed that all of the functions were actively engagedin each step of the process. In such a collaborative-based process,

442 R. Oliva, N. Watson / Journal of Operations Management 29 (2011) 434–448

delive

adooamoimmobia

abBptlbwibotn

tttaonwtsfwp

Fig. 3. On-time

ny function that takes a passive stance could see its goals overrid-en by the more active or vocal participants—a constant criticismf consensus approaches (Armstrong, 2001). In Leitax, however, webserved genuine constructive engagement—active involvement byll the participants in collecting, validating, and processing infor-ation and in voicing and defending their interpretations. Our

bservations of current meetings and the recollections shared dur-ng our interviews revealed that the “heat” and length of the early

eetings was a valid measure of the active participation. Further-ore, the contrast between the orderliness of the meetings we

bserved and the “heat” and length of the early meetings describedy our interviewees was generally attributed to an increase in trust

n the process and not to any increase in apathy. One of the man-gers responsible for the S&OP process stated:

. . . we try to be very participative and everyone tries to throw intheir two cents and try to explain their point of view. We havepeople take strong stances because they are convinced they aredoing the right thing for the firm and because they know theycan influence outcomes.

A final noteworthy result out of this process was the level ofgreement reached by the forecasting group and, through the mem-ers of the group, by all the functional areas involved in the process.ecause of their involvement in and understanding of the process,articipants reported a higher level of commitment and complianceo the resulting plans, though participants differed on whether theevels of commitment from individual functions could/should stille improved. In fact, during our analysis of the forecasting process,e found it difficult to determine whether the current performance

n forecast accuracy was the result of the forecasting processes or ofetter adherence to the agreed plans. While this question remainspen, process participants and observers consistently reported thathe implemented process had achieved levels of support for orga-izational plans not previously seen at Leitax.

These observations were remarkable in light of the fact thathe incentive landscape had remained virtually unchanged. Recall,hat seeming evidence of unsupportive incentive landscape washe inherent mistrust that functional areas had of each other’sctions, i.e., they had to protect themselves against the biases of thether functional areas. The strongest evidence of the unchangedature of the incentive landscape was that the planning processas described by participants as not being completely immune to

he various functional preferences and incentives. For example, theales group was still suspect for manipulation when sharing theirorecasts and pushback from finance during gap-filling exercisesas perceived in some cases to be excessive. As one of the seniorarticipants of the consensus meetings expressed it:

ry performance.

Sales has a very well-known biases. Sales likes to sell with a fullbag and they are not worried about how much is in inventory.They are worried that when a customer orders something theyhave it. Sales always does its forecast high because they don’tpay a penalty if there is too much inventory and they do get apenalty if there is not enough.

Indeed, the maintenance approach to the process by its facili-tators was one of constant vigilance as they felt that participantscould not be completely trusted to not take advantage of the sys-tem where possible — even despite improvements in planning thathad already been achieved.

6. Analysis: quality of the planning process

In this section, we present the results of our analysis of theplanning process at Leitax. We complement the validation-basedrepresentation of planning presented in the previous section byidentifying process attributes that support its functionality andexplore the role of the observed constructive engagement in sup-porting these process attributes. We ground these constructs ininformation processing and decision-making theory and provideevidence for them by elucidating the process mechanisms thatfacilitated each feature. Observing the process through the lens ofthese constructs allowed us to explain the overall functionality ofthe planning process, and more specifically, how each function’sengagement with the process had an impact on its effectiveness.

6.1. Attributes of the planning process

Recall that supply chain planning requires cross-functional col-laboration to first, assess the state of the supply chain and the needsof the organization, and then determine and execute an approachfor creating and sustaining value based on the assessment of theinformation. The first part of the planning process—assessmentof current and desired state—relies on the information-processingcapabilities of the organization, while the second part of theprocess—selection of plan and execution—relies on the decisionmaking capabilities and the synchronization of activities within theorganization. Evidence from Leitax suggests that these informationprocessing, decision making, and synchronization requirements in

turn require three corresponding process attributes: informationquality, planning procedural quality and alignment quality. Theseconstructs are supported by information processing and decisionmaking theories which are traditional approaches in the organi-zational behavior literature (see MacKenzie, 1984).

ration

nW1iasmoswuititoBcistfgmtfc

6

pipwsaustecobDacoppf

tacefoaa

icot

R. Oliva, N. Watson / Journal of Ope

An organization’s decision making and information-processingeeds are affected by information uncertainty and equivocality.hereas uncertainty reflects an absence of information (Galbraith,

973), equivocality reflects an absence of clarity even when there isnformation (Daft and Lengel, 1986). Many organizational decisionsre characterized by noisy data that do not always lend them-elves to straightforward interpretation or to use in a decisionaking process. In such circumstances, especially in the presence

f equivocality, the quality of information and the quality of deci-ion making cannot be taken for granted. By information quality,e mean the degree to which a process enables the informationsed for decision making to be appropriate, both in content and

n form, for the decision maker and the decision. A second threato the quality of decision making and of the resulting plans arenconsistent decision making procedures or procedures subject tohe cognitive and social limitations, influences, and idiosyncrasiesf individuals and groups (e.g., Bowman, 1963; Sterman, 1989).y procedural quality, we mean the degree to which a processontinuously ensures that the rules of inference used to validatenformation, and to make decisions within and across functions areound (March and Simon, 1993; Simon and Newell, 1972). Finally,he cross-functional nature of planning and execution implies thatunctional decisions and actions need to support organizationaloals and synchronization through time, which we refer to as align-ent (Daft and Lengel, 1986). By alignment quality, we refer to

he degree to which a process ensures that organizational andunctional goals are supported and that resulting actions are syn-hronized (Bendoly and Jacobs, 2004; Sahin and Robinson, 2002).

.2. Ensuring information quality at Leitax

The forecasting group’s emphasis on the business assumptionsackage (BAP)—the main instrument for collecting and aggregat-

ng data for the planning process—initially pointed us towardsrocess information quality. It is interesting to note that the BAPas not initially part of the planning process and that early con-

ensus meetings proved difficult to manage. Varying assumptionsbout product price changes, product offering, promotion sched-les, competitors’ actions and general market conditions were aignificant source of conflict, strongly suggesting that informa-ional inefficiencies or poor informational quality were significantnough to hamper Leitax’s integration efforts. These inefficien-ies also affected the quality (or the perception of the quality)f the functional forecasts and plans. Realizing that transparentusiness assumptions were crucial to the overall process, theMO devoted considerable attention to developing and refiningpackage that summarized such information about Leitax’s and itsompetitors’ products. The BAP provided not only a common setf data, but also a common interpretation of the data (for exam-le, the PPS’s assessments of the threats posed by competitors’roducts), which helped create a common problem space for theunctions.

The manner in which the BAP was generated also had a posi-ive effect on integration efforts. Retaining functional orientationsnd incentives while requiring the functions to participate in aollaborative effort, in which no particular function could keep rel-vant information to itself, motivated the participants to becomeully engaged. As mentioned before, during the BAP meetings webserved constructive engagement in that all of the functions werectively involved in developing and assessing the product offeringnd promotions plan.

Constructive engagement, in turn, improved information qual-ty in two ways. First, it led to more thorough informationollection. As more functions became engaged in the BAP’s devel-pment, the document included more of the necessary informationhat had previously been available only to a particular functional

s Management 29 (2011) 434–448 443

area. Norms discouraging private information and favoring coop-erative interpretation of public information were established andreinforced. The second way in which constructive engagementaffected information quality was through compelling the indi-vidual functions to make their information not merely availablebut also accessible. Functional idiosyncrasies in the submission ofdata that hampered widespread dissemination of information wereaddressed, as were idiosyncrasies in the receipt and interpretationof the data.

A second mechanism to surface data not fully articulated inthe BAP, either because it was too hard to explain or becausethe functional area was not aware that it was relevant, was theuse of separate functional forecasts. By using and challengingthe functional forecasts, the forecasting group was able to obtaininformation that previously might have only been visible to thefunctional area preparing the forecast.

6.3. Ensuring procedural quality at Leitax

Procedural quality of the planning process refers to the appro-priateness of the perspectives and the soundness of the rules ofinference and judgment used for developing and validating theplans. Procedural quality can suffer from undue influence from theincentive landscape, since incentives and priorities can bias therules of inference used in the assessment of a plan’s validity. Recog-nition of this dynamic leads the forecasting literature to argue, forexample, for the separation of decision making from forecasting(Armstrong, 2001). We found that procedural quality was enhancedat Leitax by mechanisms that promote overall soundness of indi-vidual inferences, and the use of explicit and extensive validationacross the organization.

Leitax’s S&OP process included specific mechanisms that pro-moted the soundness of the rules of inference and judgment thatwould be used to validate the information in the BAP and theresulting forecast. These mechanisms included the combination ofmultiple forecasts in the consensus forecasting process, a focus onsell-through instead of sell-in, forecasting at an aggregate level,and the use of statistical forecasts to spur discussion about theassumptions behind the forecasting. In the forecasting literature, ithas been established that combining forecasts, even through sim-ple averaging, can improve accuracy (Lawrence et al., 1986). Theemphasis on forecasting sell-through provided a reality check forsell-in forecasts and shifted the focus away from sales incentivesthat could compromise forecast accuracy. Mechanisms for promot-ing procedural quality in financial validation included using BAPdata to convert the forecasts in units into their monetary equiva-lents. Mechanisms for promoting procedural quality in operationsvalidation included publishing production requests to suppliersmore frequently.

The S&OP process also included explicit and extensive valida-tion across the organization, which in turn increased each function’sawareness of, and therefore its responsiveness to, the needs andperspectives of other functions’ stakeholders. The separate andexplicit validation steps ensured that function-specific concernswere given individualized attention so that they could be col-lectively planned for, rather than being overemphasized by onefunction or underemphasized by the others. For example, feed-back from operations and finance validations directly promptedchanges to the product offering and promotions plan in the BAPand then indirectly prompted changes in the forecasts, rather thaninappropriately affecting the forecasts directly.

Finally, support for both overall soundness of individual infer-ences and the use of explicit and extensive validation acrossthe organization, came from the learning that took place duringconsensus meetings. Feedback on forecasting and planning per-formance linked with previous validation attempts would at least

4 ration

ca

asespcrfec

6

aflgarmttorTtsnwmtptp

pieCntmpaponltpc

6

adtlpvd

44 R. Oliva, N. Watson / Journal of Ope

ontinually encourage or generate support for improvement of thepproaches to validation.

Constructive engagement in the validation steps of the processlso contributed to improvements in procedural quality. In the con-ensus forecasting meetings, the attending functions were activelyngaged in reconciling differences in the functional forecasts. Byurfacing the private information (or private interpretations ofublic information) that motivated objections to the proposedonsensus forecast, these discussions strengthened the procedu-al quality of the forecasting step. Open discussion of a particularunction’s forecasting logic served to filter out poor rules of infer-nce. Constructive engagement, by its very nature, ensured that theoncerns of different stakeholders were at least partially addressed.

.4. Ensuring alignment quality at Leitax

Finally, we explore the process characteristics that ensuredlignment at Leitax—that is, support to organizational goals acrossunctions and synchronicity of actions across time. While the col-aborative approach of Leitax’s S&OP process sets the stage foroal alignment, it is the constructive engagement that enables thislignment to be realized. Constructive engagement in validationequired every function to focus on the product offering and pro-otions plan, which could stimulate increased alignment with

hese plans through various mechanisms. For example, each func-ion’s allocation of resources to validate the plan would help createperational momentum for the plan and made the allocation ofesources to alternative plans (and their validation) less likely.he involvement of the individual functions in the validation ofhe product offering and promotions plan resulted in a greaterhared understanding of constraints, which translated into orga-izational plans that were easier for each function to execute andith which it was easier for that function to align. Because engage-ent increased the imprint of each function on the strategic plan,

here was more explicit and collective ownership of the plan, whichromotes alignment. Finally, engagement encouraged participantso trust that the other participants would adhere to the plans, whichromoted alignment.

With respect to synchronicity of actions, consensus meetingserformed double duty as tactical planning meetings, in that,

n addition to reaching a consensus forecast, detailed aspects ofvents such as product introduction and end-of-life were discussed.onsensus meetings, beyond validation, ensured the coordinationeeded for execution, since they allowed for the timely dissemina-ion of coordination signals to relevant participants. The consensus

eetings also gave participants constructive feedback on processerformance by relating it to specific process changes or to devi-tions for which participants had been responsible. This reducedrocess deviations, either by promoting voluntary conformancer by demonstrating the need for additional constraining mecha-isms. For example, feedback to the sales force revealed short- and

ong-term biases in its forecasting. If the sales force had acceptedhis feedback but been unable to mitigate its own bias, which wasrobably due to its short-term orientation, there would have beenollective recognition that the process needed to be modified.

.5. Implications of process attributes perspective

The above examples suggest that fulfillment of the three processttributes was encouraged despite an incentive landscape whichid not seem to support it. To explain this, we drew a distinc-

ion between the degree of alignment reflected in the incentiveandscape and the mediating effect of process characteristics onlanning and execution. We found an S&OP process imbued to aisible degree with consistency of information flows, quality of theecision making, and an ability to transform decisions into action-

s Management 29 (2011) 434–448

able plans that all functional areas would ultimately support. Thiscross-functional support was achieved because the plans not onlysatisfied organizational objectives but also, at minimum, respectedfunctional goals and priorities. These characteristics match ourthree attributes for a successful planning process. By contrast,the planning process before the implementation of the S&OP pro-cess lacked many of these characteristics. For example, havingthe sales force provide forecasts to the other functional groupswas inherently ineffective because it was a one-way transfer ofinformation with little formal procedure for reconciliation (poorinformation quality). But its ineffectiveness was exacerbated by thefact that forecasts communicated by the sales group were some-times inaccurate (poor procedural quality) or at least mistrusted(poor alignment quality), not only because of the known biasingeffect of sales’ incentives, but also because of insufficient moti-vation to generate more accurate forecasts (poor engagement), afaulty forecasting process (poor procedural quality), and haphazardchanges to end-of-life and product introduction schedules (pooralignment quality).

From a conceptual perspective, Leitax had two options foraddressing the integration problems that plagued it: either reducethe level of incentive misalignment between the functions suffi-ciently to enable integration to be achieved through the existingplanning approach, or develop a new planning approach that couldeffect better integration among the functions despite their func-tional misalignment. As mentioned above, some of the literaturein supply chain planning focus on reducing the negative effects ofdifferentiation among organizational units through incentive align-ment. But this approach has its limits. For managers, functionaldifferentiation is a natural response to a limited span of surveillancethat forces them to focus on only a portion of their total environ-ment or the needs of only certain stakeholders. Losing this focus onlegitimate stakeholders’ needs could be detrimental to organiza-tional performance. This partly justifies Leitax’s decision to retainfunctional orientations—despite their inherent misalignment—andtry to improve integration by implementing changes in process,organizational structure, and reporting mechanisms. This leads toour first proposition regarding the management of cross-functionalplanning:

Proposition I. For organizational supply chain planning, changesin incentives are not the only solution for what appears to be incen-tive misalignment. In a highly differentiated organizational context,that is, a context susceptible to incentive misalignment, a process thatpromotes informational, procedural, and alignment quality can be amediator in achieving integration.

We also observed active participation, or constructive engage-ment, and hypothesized that it supported the three information-processing attributes of the planning process, particularly align-ment in the execution of plans. In fact, the benefits of constructiveengagement at Leitax were numerous, including more completeand accessible information for the planning process, rules of infer-ence sharpened through debate, more accurate and validatedforecasts, efficient and coordinated functional plans, and organiza-tional plans that reflected the interests of the multiple stakeholdersin the organization. Constructive engagement also opened theS&OP process itself to objective scrutiny and continuous improve-ment. Thus, we propose that:

Proposition II. Constructive engagement can have a direct positiveimpact on the process attributes—i.e., informational, procedural, andalignment quality—and an indirect positive impact on cross-functional

integration.

We have emphasized a process perspective on integration forsupply chain planning, an approach that complements the exist-ing focus on more macro-level interventions such as structure

R. Oliva, N. Watson / Journal of Operations Management 29 (2011) 434–448 445

plann

adoaaedto

7a

ectiti(Ot

7

tgdispcttco

mtpeispqtt

Fig. 4. Structural determinants of

nd responsibilities. From this process perspective, the quality ofemand and supply planning can be roughly related to the qualityf the information used, the quality of the inferences made fromvailable data (e.g., forecasts and plans), and the organization’slignment—its conformance to and support of its own plans. How-ver, our observations also highlighted organizational behavioralynamics surrounding constructive engagement and alignmenthat hold additional implications for operations management the-ry and practice.

. Behavioral dynamics: constructive engagement andlignment

In this section, we focus on two of the key constructs thatxplain the performance of the implemented planning process:onstructive engagement and alignment. Specifically we explorehe mechanisms through which constructive engagement wasnitiated and, more importantly, is continuously sustained. Wehen examine the dynamics surrounding alignment and determinets priority as a process attribute. We use causal loop diagramsSterman, 2000) to structure our findings (see Keating et al., 1999;liva, 2001, for additional examples of using causal loop diagrams

o structure findings from case studies).

.1. Achieving constructive engagement

It is a key point that Leitax did not achieve this construc-ive engagement by reducing differentiation among the functionalroups involved in demand and supply planning, rather groupifferentiation empowered the groups’ constructive engagement

n organizational planning and left each group focused on itstakeholders’ needs. The S&OP process was open, transparent, andarticipatory; it not only enabled all participants to influence out-omes, but it also motivated them to do so in order to have accesso all the relevant information and agreed resolutions, and to serveheir stakeholders’ needs. It, therefore, explicitly confronted theonflicts between participants rather than trying to smooth themver.

In addition to these two drivers of constructive engagement—isaligned incentives and an open transparent process—much of

he reported motivation to engage in the process came from theerceived results of the planning process. As plans became moreffective drivers of the firm’s activity, and achieved more precisentegration among functional areas and the firm’s customers and

uppliers, participants updated their perception of the planningrocess’ quality (better information, procedural, and alignmentuality) and saw more and more reason to engage in it in ordero influence organizational plans. (Recall the first quote in Sec-ion 5.2: “[they participate] because are convinced they are doing

ing process quality and outcomes.

the right thing for the firm and because they can influence out-comes.”)

Fig. 4 portrays the reinforcing structure of this virtuous cyclein which participant engagement yields better process quality andoutcomes, which in turn motivates participants to engage in theprocess even more. Note that the initial drivers of participant’sengagement are the lack of incentive alignment, thus making itmore important for the participants to engage, and the opennessand transparency of the process. Once the process is given an oppor-tunity to show results, the positive results further pull participantsinto greater engagement (Keating et al., 1999; Shiba et al., 1990).

Our third proposition formally closes the feedback loop in Fig. 4:

Proposition III. The quality of the attributes and outcomes of a plan-ning process can have a positive impact on participants’ constructiveengagement in that process.

Since Leitax did not attempt to align incentives, it is ironicthat the incentive structure, which certainly did not seem to sup-port integration, indirectly encouraged integration when mediatedby the social and organizational dimensions of the new S&OPprocess. Open debate and explicit accommodation of conflictingfunctional goals moved the S&OP process from a coordination andinformation-sharing process (Dougherty, 1992) to a highly inte-grated collaborative process (Pinto et al., 1993).

7.2. Additional significance of alignment

With respect to deliberate choices concerning design fea-tures of the process, both our evidence and feedback frommembers of the organization suggest an operational distinctionbetween the benefits that would accrue from a logical and effi-cient information-processing algorithm—i.e., from informationaland procedural quality—and those obtained from organizationalalignment supported by the organization’s ability to engage par-ticipants. Specifically, it can be argued that a significant fractionof the reported benefits were less the result of informationaland procedural quality than of the alignment resulting from aneffective planning process. Alignment yielded two important ben-efits: first, as action plans gain in credibility, the organization’sreputation and trustworthiness grows in the eyes of customers,suppliers, employees, and investors, giving it even greater lever-age with which to execute its stated plans. Second, the abilityto execute stated plans is the key to continuous improvement,because predictable processes are the first requirements for reliablyinterpreting historical data and making inferences for learning and

improvement (Spear and Bowen, 1999). Thus, a lack of alignmentcan significantly compromise a process graced with informa-tional and procedural quality, while alignment alone, even inpursuit of suboptimal targets, might reduce the level of uncer-tainty in an organization by virtue of the predictability and

4 ration

i1

tsrffreofttrfsotii

Pct

8

sootBipaasida

tBpbotnsoti

elacfp

tpai

46 R. Oliva, N. Watson / Journal of Ope

ncreased potential for coordination it engenders (Daft and Lengel,986).

Given the benefits of alignment and the resulting commitmento the developed plan, an organization might even be willing toacrifice information accuracy or procedural efficiency in order toetain a process that delivers cross-functional alignment or con-ormance to plans. In fact, the DMO at Leitax had evidence that,or the second half of 2003, the statistical forecast was more accu-ate than the consensus forecast approved by the group. Althoughliminating the consensus forecasting process would save the costf the management time consumed in lengthy BAP and consensusorecast meetings, the DMO leadership was reluctant to streamlinehe process if it meant eliminating opportunities to engage par-icipants through confrontation and validation of the forecast andesulting plan. In addition, without a participatory process, a majororecasting error could subject the statistical forecasts to the samekepticism to which the sales forecasts had been subject in the “badld days.” The value of organizational alignment revealed throughhis study suggests a new dimension in the design of coordinat-ng/integrative planning systems that seek to go beyond simplenformation-sharing and coordination of action. Thus,

roposition IV. The pursuit of alignment in organizational planningan be more important than achieving particular levels of informa-ional or procedural quality.

. Discussion and conclusions

The purpose of this case study is not to argue for a specificolution but rather to develop an explanation for an interestingbserved phenomenon which seems to have implications for the-ry and practice (Yin, 2003). Still, it is fair to ask how much lighthis (or any) explanation can shed on a given set of problems.y characterizing the supply chain planning context as exhibit-

ng functional differentiation and by characterizing Leitax’s initiallanning approach as being complicated by functional mistrustnd poor inter-temporal coordination, our research potentiallyddresses a range of planning dysfunctions that may not alwayshow up specifically as they did at Leitax, but may spring from sim-lar causes. We believe, and conversations with management fromiverse industries have confirmed, that these unhelpful dynamicsre not only prevalent, but also persistent in industry.

Similarly, when we examine Leitax’s S&OP process, we aimedo generate an explanation of why that particular process worked.y drawing a distinction between the incentive landscape and thelanning process, we conceptually recognize process as a mediatoreyond the structural mechanisms and responsibility approachesf the extant supply chain and organizational behavior litera-ures. Recognizing a specific characteristic of this process approach,amely constructive engagement and its relationship to the ten-ions in the incentive landscape, we provide a credible descriptionf the mechanism by which a seemingly unsupportive incen-ive landscape can nevertheless indirectly support cross-functionalntegration in planning.

Through the synthesis of our observations on these relevantlements of the planning process, observations on Leitax’s assimi-ation of the S&OP process and the priorities that emerged from it,nd credible deductions linking the process approach to the suc-ess of the planning approach, we provide sufficient evidence forour propositions concerning the management of cross-functionallanning.

The empirical and theoretical grounding of our proposi-ions suggest implications for practitioners and researchers. Forractitioners, the Leitax case is, first, a proof-of-concept thatn S&OP-based process can do more than simply coordinatenformation flow; it can fulfill both the information-processing

s Management 29 (2011) 434–448

requirements and the collaborative-assessment and problem-solving requirements of simultaneous demand and supplymanagement. Furthermore, although the primary site for supplychain planning at Leitax was within the company, by retain-ing functional differentiation and the integrity of representationof stakeholders external to the organization—e.g., suppliers, cus-tomers and investors—the process can, in principle, support theplanning processes that span organizations and have a more diverseincentive landscape. In addition, a consensus planning system, withall its embedded advantages for buy-in and integration, was shownto be capable of making prompt and responsive planning adjust-ments in a dynamic and challenging supply chain environment.Finally, the details of Leitax’s approach make it clear that it takesmore than the implementation of an efficient information-sharingtool to achieve true integration.

For supply chain management researchers, the Leitax caserenews interest in the dimension of process—a touchstone forclassic operations management researchers and a germane con-cept for organizational behavior researchers—as part of a solutionto a problem that had been largely classified as structural. Par-ticularly promising is the fact that process specifications play amediating role between the incentive and structural choices cur-rently proposed by the literature, and the firm performance. Finally,determining when our propositions hold remains an open researchquestion, beyond the scope of a single case study. Addressingthis question calls for a contingent analysis of the effectivenessof process approaches and for an assessment of the significanceof the information-processing framework—information quality,procedural quality, and alignment—and engagement dynamics inexplaining organizational integration.

Acknowledgments

We are grateful to Laura Black, Amy Edmondson, Benito Flores,Zeynep Ton, an anonymous Associate Editor, and the participantsof the Harvard Business School Technology and Operations Man-agement Seminar and the Behavioral Operations Conference forvaluable feedback on earlier versions of this work.

Appendix A. Interview protocol

1. Personal background1.1. Can you tell us a bit about yourself? Education, previous

experience.1.2. How long have you been with Leitax? In what capacities?1.3. What is your current position and what are your main

responsibilities?2. Functional area

2.1. So you work for the department/functional area, whatare the main functions and responsibilities of the functionalarea?

2.2. What do you consider are the main objectives of that func-tional area?

2.3. What are the main objectives of the functional area in rela-tion to the Planning and Forecasting process?

2.4. If not clear already, what is the process that you follow to ful-fill your responsibilities? What inputs do you need? How doyou communicate your results? Who do you need to coordi-nate with? As far as you can tell, in what ways is this processdifferent from the processes of other functional areas within

the organization?

2.5. Is your performance evaluated in relationship to thoseobjectives? Is compensation linked to those objectives?

2.6. Is compensation of other members of the functional arealinked to those objectives?

ration

3

4

5

R

A

A

B

B

R. Oliva, N. Watson / Journal of Ope

. History (if the person was in Leitax prior and during the imple-mentation of Project Redesign)3.1. Could you describe the forecasting and planning mecha-

nisms that were in place in Leitax prior to Project Redesign?3.2. How effective were those mechanisms? What were the

major issues you saw in them?3.3. Were there conflicts with other functions during forecast-

ing and planning? Were there conflicts within your functionduring forecasting and planning? What seemed to you to bethe source of those conflicts?

3.4. How did these problems impact your personal or your func-tional area’s performance?

3.5. How did you work around those problems?3.6. So, the Project Redesign initiative comes along, can you tell

us about the implementation process? How did they goabout it? How did it impact your functional area? How didit modify your responsibilities?

. Planning and forecasting process4.1. Could you please describe your understanding of the Plan-

ning in Forecasting Process as is currently executed? Whatactivities does it entail? For each of the activities,4.1.1. Inquire about the rationale for that activity: Why do

they do that? What is it that the process is trying toachieve through that activity?

4.2. How do you participate in the Planning and Forecasting Pro-cess? For each of the activities,4.2.1. Inquire about the rationale for that activity: Why do

you do, or are required to do, that?4.2.2. Inquire about the intentionality of that activity: In

performing that activity, what are you trying toachieve? What are your goals with respect to influ-encing the process?

4.3. The two consensus meetings in the planning process play asignificant part in determining the outcome of the planningprocess, and the discussion during those meetings seems tobe open-ended and very fluid, how do you attempt to influ-ence the outcomes of the meeting? What kind of leverage(or bargaining chips) do you (or your functional area) haveto influence the outcome of the meeting?

4.4. How do you decide when to stop pushing for your (or yourfunctional area’s) agenda during those meetings?

4.5. How is the process (or your activities) addressing some ofthe issues you had prior to the implementation of ProjectRedesign?

4.6. What is your assessment of the performance of the Planningand Forecasting Process?

. Follow up5.1. Is there anything else that you think we ought to be aware

of regarding how the S&OP process is working and yourinteractions with the other functional areas in the process?

5.2. Is there anyone else you would recommend we contact inorder to get a better understanding of the objectives andinner workings of the S&OP process?

5.3. Could we contact you in the future for clarification questionsin case we uncover some gaps or inconsistencies as we gothrough our notes?

eferences

nthony, R.N., 1965. Planning and Control System: A Framework for Analysis. Har-vard Business School Press, Boston.

rmstrong, J.S., 2001. Principles of Forecasting: A Handbook for Researchers and

Practitioners. Kluwer Academic Publisher, Boston.

aker, W.E., Sinkula, J.M., 1999. The synergistic effect of market orientation andlearning orientation on organizational performance. Journal of the Academy ofMarketing Science 27, 411–427.

arki, H., Pinsonneault, A., 2005. A model of organizational integration, implemen-tation effort, and performance. Organization Science 16, 165–179.

s Management 29 (2011) 434–448 447

Barratt, M., 2004. Understanding the meaning of collaboration in the supply chain.Supply Chain Management: An International Journal 9, 30–42.

Bendoly, E., Cotteleer, M.J., 2008. Understanding behavioral sources of processvariation following enterprise system deployment. Journal of Operations Man-agement 26, 23–44.

Bendoly, E., Jacobs, F.R., 2004. ERP architectural/operational alignment for order-processing performance. International Journal of Operations and ProductionManagement 24, 99–117.

Bower, P., 2005. 12 most common threats to sales and operations planning process.The Journal of Business Forecasting Methods & Systems 24, 4.

Bowman, E.H., 1963. Consistency and optimality in managerial decision making.Management Science 9, 310–321.

Braunscheidel, M.J., Suresh, N.C., 2009. The organizational antecedents of a firm’ssupply chain agility for risk mitigation and response. Journal of OperationsManagement 27, 119–140.

Brown, J.S., Duguid, P., 1991. Organizational learning and communities-of-practice:toward a unified view of working, learning and innovation. Organization Science2, 40–57.

Brown, J.S., Duguid, P., 2001. Knowledge and organization: a social-practice perspec-tive. Organization Science 12, 198–213.

Cachon, G.P., 2003. Supply chain coordination with contracts. In: de Kok, A.G., Graves,S.C. (Eds.), Supply Chain Management: Design, Coordination and Operation.North Holland, Amsterdam, pp. 229–340.

Cachon, G.P., Lariviere, M.A., 2005. Supply chain coordination with revenue-sharingcontracts: strengths and limitations. Management Science 51, 30–44.

Carlile, P.R., 2002. A pragmatic view of knowledge and boundaries: boundary objectsin new product development. Organization Science 13, 442–455.

Celikbas, M., Shanthikumar, J.G., Swaminathan, J.M., 1999. Coordinating productionquantities and demand forecasts through penalty schemes. IIE Transactions 31,851–864.

Checkland, P.B., Scholes, J., 1990. Soft Systems Methodology in Action. Wiley, Chich-ester, UK.

Chen, F., 2005. Salesforce incentives, market information and production/inventoryplanning. Management Science 51, 60–75.

Chen, F.G., 1999. Decentralized supply chains subject to information delays. Man-agement Science 45, 1076–1090.

Chen, J.M., Chen, L.T., Leu, J.D., 2006. Developing optimization models for cross-functional decision-making: integrating marketing and production planning. ORSpectrum 28, 223–240.

Crittenden, V.L., Gardiner, L.R., Stam, A., 1993. Reducing conflict between marketingand manufacturing. Industrial Marketing Management 22, 299–309.

Croson, R., Donohue, K., Katok, E., Sterman, J.D., 2005. Order Stability in SupplyChains: Coordination Risk and the Role of Coordination Stock, Sloan School ofManagement Working Paper, Cambridge, MA.

Cyert, R.M., March, J.G., 1963. A Behavioral Theory of the Firm. Prentice-Hall, UpperSaddle River, NJ.

Daft, R.L., Lengel, R.H., 1986. Organizational information requirements, media rich-ness and structural design. Management Science 32, 554–571.

Dougherty, D., 1992. Interpretive barriers to successful product innovation in largefirms. Organization Science 3, 179–202.

Droge, C., Jayaram, J., Vickery, S.K., 2004. The effects of internal versus externalintegration practices on time-based performance and overall firm performance.Journal of Operations Management 22, 557–573.

Eisenhardt, K.M., 1989. Building theories from case study research. Academy ofManagement Review 14, 532–550.

Eliashberg, J., Steinberg, R., 1993. Marketing-production joint decision making. In:Eliashberg, J., Lilien, G.L. (Eds.), Handbooks in Operations Research and Manage-ment Science, 5: Marketing. Elsevier Science, Amsterdam, pp. 827–880.

Ellinger, A.E., 2000. Improving marketing/logistics cross-functional collaboration inthe supply chain. Industrial Marketing Management 29, 85–96.

Fawcett, S.E., Magnan, G.M., 2002. The rhetoric and reality of supply chain integra-tion. International Journal of Physical Distribution and Logistics Management32, 339–361.

Fugate, B.S., Stank, T.P., Mentzer, J.T., 2009. Linking improved knowledge man-agement to operational and organizational performance. Journal of OperationsManagement 27, 247–264.

Galbraith, J.R., 1973. Designing Complex Organizations. Addison-Wesley, Reading,MA.

Galbraith, J.R., 1977. Organization Design. Addison-Wesley, Reading, MA.Glaser, B.G., Strauss, A.L., 1967. The Discovery of Grounded Theory: Strategies for

Qualitative Research. Wiedenfeld and Nicholson, London.Gonik, J., 1978. Tie salesmen’s bonuses to their forecasts. Harvard Business Review

56, 116–123.Griffin, A., Hauser, J.R., 1996. Integrating R&D and marketing: a review and analysis

of the literature. Journal of Product Innovation Management 13, 191–215.Hargadon, A., Sutton, R., 1997. Technology brokering and innovation in a product

development firm. Administrative Science Quarterly 42, 716–749.Holmstrom, B., Milgrom, P., 1994. The firm as an incentive system. American Eco-

nomic Review 84, 972–991.Holweg, M., Pil, F.K., 2008. Theoretical perspectives on the coordination of supply

chains. Journal of Operations Management 26, 389–406.Hurley, R.F., Hult, G.T.M., 1998. Innovation, market orientation, and organizational

learning: an integration and empirical examination. Journal of Marketing 62,42–54.

Jaworski, B.J, Kohli, A.K., 1993. Market orientation – antecedents and consequences.Journal of Marketing 57, 53–70.

4 ration

K

K

K

K

K

L

LL

L

L

L

M

M

M

M

MN

N

N

O

O

O

P

P

P

R

R

48 R. Oliva, N. Watson / Journal of Ope

ahn, K.B., 1996. Interdepartmental integration: a definition and implications forproduct development performance. Journal of Product Innovation Management13, 137–151.

ahn, K.B., Mentzer, J.T., 1998. Marketing’s integration with other departments.Journal of Business Research 42, 53–62.

eating, E.K., Oliva, R., Repenning, N., Rockart, S.F., Sterman, J.D., 1999. Overcomingthe improvement paradox. European Management Journal 17, 120–134.

ouvelis, P., Lariviere, M.A., 2000. Decentralizing cross-functional decisions: coor-dination through internal markets. Management Science 46, 1049–1058.

raiselburd, S., Watson, N., 2007. Alignment in Cross-functional and Cross-firm Sup-ply Chain Planning, Harvard Business School Working Paper 07-058, Boston.

apide, L., 2004. Sales and operations planning. Part I: the process. The Journal ofBusiness Forecasting Methods & Systems 23, 17–19.

apide, L., 2005. An S&OP maturity model. Journal of Business Forecasting 24, 15–20.awrence, M.J., Edmundson, R.H., O’Connor, M.J., 1986. The accuracy of combining

judgmental and statistical forecasts. Management Science 32, 1521–1532.awrence, P.R., Lorsch, J.W., 1986. Organization and Environment: Managing Differ-

entiation and Integration. Harvard Business School Press, Boston.i, Q., Atkins, D., 2002. Coordinating replenishment and pricing in a firm. Manufac-

turing & Service Operations Management 4, 241–257.orsch, J.W., Allen, S.I., 1973. Managing Diversity and Interdependence: An Organi-

zational Study of Multidivisional Firms. Harvard Business School Press, Boston.acKenzie, K.D., 1984. Organizational structures as the primal information system.

In: Chang, S.K. (Ed.), Management of Office Information Systems. Plenum, NewYork, pp. 27–46.

alhotra, M.K., Sharma, S., 2002. Spanning the continuum between marketing andoperations. Journal of Operations Management 20, 209–219.

alone, T.W., Crowston, K., 1994. The interdisciplinary study of coordination. ACMComputing Surveys 26, 87–119.

alone, T.W., Crowston, K., Lee, J., Pentland, B., Dellarocas, C., Wyner, G., Quimby,J., Osborn, C.S., Bernstein, A., Herman, G., Klein, M., O’Donnell, E., 1999. Toolsfor inventing organizations: toward a handbook of organizational processes.Management Science 45, 425–443.

arch, J., Simon, H., 1993. Organizations. Blackwell, Cambridge, MA.arayanan, V., Raman, A., 2004. Aligning incentives in supply chains. Harvard Busi-

ness Review 82, 94–102.arver, J.C., Slater, S.F., 1990. The effect of market orientation on business profitabil-

ity. Journal of Marketing 54, 20–35.elson, R., Winter, S., 1982. An Evolutionary Theory of Economic Change. Harvard

University Press, Cambridge, MA.liva, R., 2001. Tradeoffs in responses to work pressure in the service industry.

California Management Review 43, 26–43.liva, R., Watson, N., 2009. Managing functional biases in organizational forecasts:

a case study of consensus forecasting in supply chain planning. Production andOperations Management 18, 138–151.

rlikowski, W.J., Yates, J., Okamura, K., Fujimoto, M., 1995. Shaping electroniccommunication: the metastructuring of technology in the context of use. Orga-nization Science 6, 423–444.

agell, M., 2004. Understanding the factors that enable and inhibit the integrationof operations, purchasing and logistics. Journal of Operations Management 22,459–487.

into, M.B, Pinto, J.K., Prescott, J.E., 1993. Antecedents and consequencesof project team cross-functional cooperation. Management Science 39,1281–1297.

orteus, E.L., 2000. Responsibility tokens in supply chain management. Manufactur-ing & Service Operations Management 2, 203–219.

aman, A., Watson, N., 2004. Managing global supply chains. In: Quelch, J.A., Desh-pandé, R. (Eds.), The Global Market: Developing a Strategy to Manage AcrossBorders. Jossey-Bass, San Francisco, CA, pp. 263–287.

yan, R.M., Deci, E.L., 2000. Intrinsic and extrinsic motivations: classic def-initions and new directions. Contemporary Educational Psychology 25,54–67.

s Management 29 (2011) 434–448

Sahin, F., Robinson, E.P., 2002. Flow coordination and information sharing in sup-ply chains: review, implications, and directions for future research. DecisionSciences 33, 505–536.

Shapiro, B.P., 1977. Can marketing and manufacturing co-exist? Harvard BusinessReview 55, 104–114.

Shiba, S., Graham, A., Walden, D., 1990. A New American TQM: Four practical Revo-lutions in Management. Productivity Press, Portland, OR.

Simon, H.A., Newell, A., 1972. Human Problem Solving. Prentice-Hall, EnglewoodCliffs, NJ.

Sinkula, J.M., 1994. Market-information processing and organizational learning.Journal of Marketing 58, 35–45.

Slater, S.F., Narver, J.C., 1994. Does competitive environment moderate the markerorientation-performance relationship? Journal of Marketing 58, 46–55.

Spear, S., Bowen, H.K., 1999. Decoding the DNA of the Toyota production system.Harvard Business Review 77, 96–106.

Stadtler, H., 2005. Supply chain management and advanced planning—basics,overview and challenges. European Journal of Operational Research 163,575–588.

Stank, T.P., Daugherty, P.J., Ellinger, A.E., 1999. Marketing/logistics interaction andfirm performance. International Journal of Logistics Management 10, 11–24.

Star, S.L., 1989. The structure of ill-structured solutions: boundary objects andheterogeneous distributed problem solving. In: Huhn, M.N., Gasser, L. (Eds.),Distributed Artificial Intelligence, vol. 2. Morgan Kaufman, Menlo Park, CA, pp.37–54.

Sterman, J.D., 1989. Modeling managerial behavior: misperception of feedback in adynamic decision making environment. Management Science 35, 321–339.

Sterman, J.D., 2000. Business Dynamics. Irwin McGraw-Hill, Boston.Thompson, J.D., 1967. Organizations in Action. McGraw-Hill, New York.Troyer, L., Smith, J., Marshall, S., Yaniv, E., Tayur, S., Barkman, M., Kaya, A., Liu, Y., 2005.

Improving asset management and order fulfillment at Deere & Company’s C&CEdivision. Interfaces 35, 76–87.

Van Landeghem, H., Vanmaele, H., 2002. Robust planning: a new paradigm fordemand chain planning. Journal of Operations Management 20, 769–783.

Vickery, S.K., Jayaram, J., Droge, C., Calantone, R., 2003. The effects of an integrativesupply chain strategy on customer service and financial performance: an anal-ysis of direct versus indirect relationships. Journal of Operations Management21, 523–539.

Watson, N., Oliva, R., Winig, L., 2005. Leitax (A). Harvard Business School Case 606-002. Harvard Business School, Boston.

Watson, N., Zheng, Y.-S., 2005. Decentralized serial supply chains subject to orderdelays and information distortion: exploiting real time sales data. Manufactur-ing & Service Operations Management 7, 152–168.

Weng, Z.K., 1995. Channel coordination and quantity discounts. Management Sci-ence 41, 1509–1522.

Whang, S., 1995. Coordination in operations: a taxonomy. Journal of OperationsManagement 12, 413–422.

Wheelwright, S.C., Clark, K.B., 1992. Revolutionizing Product Development. Wiley,New York.

Wong, C.Y., Boon-Itt, S., 2008. The influence of institutional norms and environmen-tal uncertainty on supply chain integration in the Thai automotive industry.International Journal of Production Economics 115, 400–410.

Wu, Z.H., Choi, T.Y., 2005. Supplier–supplier relationships in the buyer–suppliertriad: building theories from eight case studies. Journal of Operations Manage-ment 24, 27–52.

Yano, C.A., Gilbert, S.M., 2003. Coordinated pricing and production/procurementdecision: a review. In: Chakravarty, A.K., Eliashberg, J. (Eds.), Managing Business

Interfaces: Marketing, Engineering, and Manufacturing Perspectives. Springer,New York, pp. 65–104.

Yanow, D., 2000. Seeing organizational learning: a “cultural” view. Organization 7,247–268.

Yin, R., 2003. Case Study Research: Design and Methods, 3rd ed. Sage, Beverly Hills,CA.