European Journal of Information Systems (2007) 16, 628–642 & 2007 Operational Research Society Ltd. All rights reserved 0960-085X/07 $30.00 www.palgrave-journals.com/ejis

Manchester United Football Club: developing a Network Orchestration Model Duncan R. Shaw1 1

Nottingham University Business School, Nottingham, U.K. Correspondence: Duncan R. Shaw, Nottingham University Business School, Jubilee Campus, Nottingham NG8 1BB, U.K. Tel: þ 44 (0)115 84 67756; E-mail: [email protected]

Abstract This paper investigates a particular type of coordination role called ‘network orchestration’. It uses a revelatory case study of the very large network orchestrated by Manchester United Football Club, a global sporting and entertainment brand that is partnered by some of the most successful consumer brands in the world and has an estimated 70 million fans (MUFC, 2007). We use business process modelling and systems theoretical concepts to investigate the complex horizontal and vertical relationships between partner firms and then develop a multi-level model of network operation, sustainability and governance. Inter-organisational networks are open systems that are sustained far from equilibrium by the constant flow of materials, energy and information that we have called a ‘value flow system’. Here we model the flow of commercial ‘value’ through the network that sustains it in a far from equilibrium state. The contribution for managers of orchestrator firms is an architectural model of the properties and mechanisms of network orchestration that aids value flow for network building and maintenance. The implication for coordination researchers is a development of Malone and Crowston’s coordination theory (1994) via the novel introduction of Hierarchy Theory to this domain. European Journal of Information Systems (2007) 16, 628–642. doi:10.1057/palgrave.ejis.3000702 Keywords: Manchester United; business process modelling; cultivation; network; orchestration; inter-organisational; B2B; open systems; Hierarchy Theory

Introduction

Received: 8 November 2004 Revised: 21 September 2005 2nd Revision: 8 June 2006 3rd Revision: 17 April 2007 Accepted: 8 August 2007

In the last few decades organisations have been vastly changed as a result of the possibilities enabled by information and communications technologies (ICTs) (Nolan, 2000). These changes are compounded by the ability to link together the information systems of different organisations to thus enhance the operation of inter-organisational business networks across space and time. Focal business networks can be viewed as entities that process material and information inputs (raw materials) into material and information outputs (products and services). This is consistent with the systems theoretical concept of an open system which is a system that maintains itself in a state far from equilibrium by processing external inputs into ejected outputs (Checkland, 1999). On a lower level an example is a firm whose total value is deemed to be greater than the sum of the disposal value of its assets. Its value is increased by what it does, that is, its processes, in addition to what it constitutes. In an inter-organisational network the member firms process environmental inputs into outputs that are then composed into the network’s meta-service (or meta-product, although henceforth only ‘services’ will be mentioned). An example is a single car

Manchester United FC’s orchestrated network

that is produced by an automotive supply chain made up of many tiers of suppliers. Organisational business processes transform inputs into outputs, where a business process is defined as ‘a specific ordering of work activities across time and place, with a beginning and end, and clearly defined inputs and outputs: a structure for action’ (Davenport, 1993, p. 5) or a ‘collection of activities that takes one or more kinds of output and creates an input that is of value for a customer’ (Hammer & Champy, 1993, p. 38). This is purely a reconfiguration of inputs which spans: (i) informational levels of complexity and (ii) nuclear or chemical material levels. By reconfiguration we mean that business processes compose and decompose material and informational inputs. This involves the coordination of separate outputs from suppliers and inputs to customers. Malone & Crowston (1994) define coordination as managing dependencies between activities and here we focus on the activities or transactions between network members, for example, inter-firm business processes. Current coordination theory literature is fragmented theoretically and biased towards empirical description and quantitative practice; see Olson et al. (2001). More recently, a 10-year retrospective suggests ‘More work could be done’ in organising the coordination mechanisms studied so far (Crowston et al., 2004, p. 28). Johnston & Gregor (2000) present a description of the concepts of distributed coordination, at a firm level, and centralised coordination, at a network level. These are based upon a goal-orientated or ‘intentional’ action view of coordination with decisions being made at two different levels of analysis (Johnston, 2001). In the literature intentional action is broadly divided into ‘deliberative action’ theories and ‘situated action’ theories. In ‘deliberative action’ theories coordination decisions are made for the whole network according to a whole network model and in ‘situated action’ theories coordination decisions are made by each firm according to its model of its local environment (ibid). If these models include a temporal-forecasting capability, then these concepts are consistent with Haeckel & Nolan’s (1993) ‘plan-do’ vs ‘sense and respond’ strategies for complex business environments (1996). In fact, Haeckel and Nolan advocate a ‘sense and respond’ (i.e. situated) strategy for boundedly rational firms in increasingly complex business environments (Simon, 1997). Intuition would suggest that in an inter-firm network there are both distributed and centralised network coordination mechanisms. Distributed change that is planned on a firm level appears as emergent change on a network level. However, here we focus upon the centralised network coordination function of a single firm, that is, planned change, on a whole network level that is led by one firm. One example of a centralised network coordination mechanism is Ofgem, the U.K. electricity industry regulator, which coordinates the customer– supplier interactions for the whole industry via a very complicated set of dataflow diagrams (Shaw et al., 2004).

Duncan R. Shaw

629

Rice & Hoppe (2001, p. 52) call these dominating network members ‘channel masters’ but we prefer Hagel’s term ‘orchestrator’ as it more strongly denotes the influencing of many other organisations and it can apply ¨ ller et al. also use the term to many channels (2002). Mo ‘orchestrator’ (2002) and Andrew & Sirkin list it as one of three approaches to influencing a product developer network (2003). Hagel calls the network of organisations that is influenced by the orchestrator a ‘process network’ (2002). Such process networks, supply chains or value chains compose business processes into higher level products or services. The object of this study is to start to investigate the orchestrator’s role within an inter-organisational network and how such an orchestrated network is sustained. We seek to understand the process mechanisms of network governance and network operation. To do this we seek to develop an analysis model that encompasses both the whole network level and the firm level of the network. Specifically, we ask what the process mechanisms that maintain the viability of this open system are and how the orchestrator firm accomplishes this in its relationships with normal member firms. The research problem from an orchestrator’s perspective is how to manage the complexity of inter-organisational networks. This translates into a development of coordination theory by asking how is a dyadic relationship coordinated, at the firm level, and at the whole network level. First, we assemble a conceptual lens from the systems theoretical literature and business-to-business (B2B) process modelling and then we describe our research approach. Then we introduce our case study of a specific orchestrated B2B network and analyse it using the conceptual lens that we have developed. Finally, we discuss the novel analytical perspectives generated and synthesise a model that describes sustainable relations between network members across multiple levels and organisational boundaries.

Literature review A meta-service is a network level output that is made up of the sub-services of more than one member firm. Its composition therefore requires two things. Firstly, firms must link together, which they accomplish using business processes. Secondly, it requires the coordination of these inter-firm business processes in order to link together the correct firms and business processes. Next we develop these two concepts and then assemble them into our analysis framework.

Relations: the multi-level nature of inter-firm business processes The linking of firms via business processes is a businessto-business (B2B) interaction that occurs on multiple ‘levels’. Figure 1 shows two network levels, which consist of a supplier level and a customer level. The network’s environment incorporates the customer’s customer and the suppliers’ suppliers. The single supplier ‘pyramid’ on

European Journal of Information Systems

630

Manchester United FC’s orchestrated network

B2B interaction

Meta-service to customer’s customer

Composite service

Composite service

Many sub-services

Many sub-services

Composite processes

Composite processes

Many sub-processes

Many sub-processes

Integrated applications

Integrated applications

Humans & machine m applications Human & technology infrastructure

Supplier

Humans & machine applications Human & technology infrastructure

Customer

A supplier process servicing a customer process (e.g. translated documents). Other processes (e.g.payment) go diagonally in the opposite direction.

Figure 1: Multi-level view of supplier processes composed into customer’s processes which produce a service for the customer’s customer (based upon Kalakota & Robinson, 2003).

the left of Figure 1 represents many suppliers interacting with a single customer. Figure 1 shows how each supplier’s service is composed together with other services (internal and external) to meet a customer’s specific need and how the customer then goes on to supply its own customer. Each of the arrows is itself a two way exchange process, that is, a request for service, the provision of a service, and a payment. The activities of the customer and the supplier can be viewed as sub-processes, processes, sub-services and services depending upon the level of analysis. These business processes are concerned with the exchange of both physical materials and information. The cascading and enabling levels of enactment and meta-property composition within a firm’s business processes is described more fully in Kalakota & Robinson (2003), from a business strategy–business process perspective, and in Shaw et al. (2007) from a business process–technology infrastructure perspective. The levels in Figure 1 can also be conceptualised as different perspectives or points of view. This analysis is justified theoretically by the systemic property of emergence (Checkland, 1999) and, also, the concept of bounded rationality (Simon, 1997). Emergent phenomena are only perceivable on the level at which they emerge or on higher levels. The properties of meta-services are only perceivable, or even possible to contextualise, after component-services have been composed. The phrase ‘you can’t see the wood for the trees’ applies to an unavailability of wood-level concepts at the tree-level, not just the masking of most trees by nearer trees. Bounded rationality is a fundamental constraint upon human and machine data capture and processing that can be compensated for by taking multiple observations. This provides several different perspectives of the same subject system. Both emergence and bounded rationality require humans to use multi-level analyses in order to perceive complex phenomena.

European Journal of Information Systems

Duncan R. Shaw

Evolution: coordinating the composition of inter-firm business processes Just as services are produced by business processes, they can also be conceptualised as producing service-needs. These are requirements for the output of another process. Using a systems theoretical perspective, the justification for this concept of ‘service symmetry’ is that process composition joins two lower level processes into a single meta-process, that is, a single higher level process system. We conceptualise a service-need as one of the two lower level processes viewed from the perspective of the other process and before composition. As a firm enacts business processes that are designed to realise its business goals it produces service-needs. An example of this is a firm that purchases electricity. The service is the provision of electrical power and the service-need is the sum of requirements for electrical power that the firm’s business processes produce as they are enacted. This description of business process composition is also conceptually supported by the Process Philosophy literature, which views reality as processual in nature (Rescher, 2000; Styhre, 2002). Thus inter-firm business process composition can be thought of as the ‘docking’ of one firm’s service with another firm’s service-need. This is a coordination process and in an orchestrated network of firms it is managed by the network orchestrator so as to achieve a successful fit. The concept of fit is an intersection between design theory and systems theory. It is a match between context and structure (Alexander, 1964; Drazin & Van de Ven, 1985). In terms of problem solving it is the fit between the architectures of the problem system and a solution system that join into a larger system when the solution is implemented. If the solution does not fully fit the problem then the resulting system is sub-optimal. The solution system may not address a subsystem in the problem system, for example a product portfolio that contains only desktop PCs fails to address a customer’s mobile computing requirements. The concept of architectural fit is further developed by introducing the two concepts of system hierarchy and system span. Complex systems are organised into hierarchical levels (Simon, 1969; Checkland, 1999), for example the network level and the firm level. System levels are asymmetrical and the asymmetries take the form of emergent properties. Hierarchical level asymmetry is a concept from Hierarchy Theory (ibid; Allen & Starr, 1982; Allen et al., 1984; Salthe, 1985; Wilby, 1994) and we incorporate it into our analysis because the perception and context of process properties depends upon the level of the observer as well as the observed. The second concept, span, is related to the sum of the properties of subsystems on any one level (Simon, 1969). The span of a system level is a function of the number of subsystems in the system on that level. Span and hierarchical level are architecturally orthogonal concepts. But they both describe phenomena that exhibit near-decomposability (ibid), which is enabled by loose coupling (Weick,

Manchester United FC’s orchestrated network

1976). Span describes the extent of subsystems on the same level that are loosely coupled and level referees to entities on different hierarchy levels that are loosely coupled. Inter-subsystem span relationships are symmetrical whereas inter-level relationships are asymmetrical. Inter-level relationships are ‘driven’ and enabled by lower level phenomena and filtered and constrained by higher levels (Wilby, 1994). Span includes all the process outputs, from different firms, that are composed together into a single higher level meta-process. The meta-process’ properties are emergent and therefore inconceivable upon the level of the firms themselves. An example of this is when firms supply services to customers without understanding what the services are actually used for. Some suppliers, especially successful suppliers, fully understand the use of their services from the ‘point of view of’ or ‘in the context of the value system of’ their customers. This is only possible when the supplier moves its understanding up to the level of the use of the service by using customer relationship management (CRM) methods. CRM enables the supplier to understand how a customer values the supplier’s service (or doesn’t), that is, how the service fits into the customer’s value system. Value systems are themselves organised in asymmetrical levels such as Keeney’s means objectives that facilitate fundamental objectives (1999). This process-oriented use of span and hierarchical level is consistent with some of the more recent literature on organisational dependencies and coordination mechanisms (Crowston, 2003). For a service to fit a service-need, when it ‘docks’, the service must architecturally fit with the service-need. The architecture of the service must mirror the architecture of the serviceneed in terms of both number of hierarchical levels and degree of span. The successful docking of service-needs with services requires a composition meta-process, which we conceptualise as a search by the orchestrator for a service that fits a known service-need. Without knowing the service-need an orchestrator would be unable to measure or even contextualise the value or fit of any subsequent offers of service. Even if the orchestrator actually purchased a service without a service-need it would be impossible for it to conceptualise a service-need by using the available conceptual building blocks, that is, the concepts that make up the service. Examples of this are the early conceptions of the future types and extent of information technology applications. The reason for this of course is that service-needs are made up of higher level concepts. In fact they are composed of concepts that make up the component services plus some additional emergent concepts, which are not available on the service production level.

Duncan R. Shaw

631

ships. Firstly, an analysis of multi-level systems, such as inter-firm processes, requires a multi-level analysis. Table 1 contrasts the different perspectives of the customer and the supplier and also shows how they vary on different levels. The number of levels shown in Table 1 could be extended to any number of levels of supplier. However, we have limited it so as to reflect our unit of analysis, which is the focal network of the orchestrator and its direct suppliers. This is because our research question is concerned with how one organisation, the orchestrator, relates to a network of organisations. Specifically we ask ‘how do a network member and the network orchestrator relate to each other?’ and ‘why can this sustain the network?’ This is a multi-level question. It forces us to integrate phenomena from both the network level and the firm level. Thus our unit of analysis is the orchestrated network itself, which includes the orchestrator organisation and the entities that it directly interacts with. This framework can be used to assess architectural fit between the services and service-needs of different system entities. Secondly, the above framework does not include the start or end of the relationships between firms. Starting with a service-need the orchestrator would search for a fitting service in a similar process to Simon’s observation: ypeople solve problems by selective, heuristic search through large problem spaces and large data bases, using means–ends analysis as a principal technique for guiding the search (Simon & Associates, 1986, p. 20).

Heuristics significantly hasten the generation of alternative potential services. These are then compared with the service-need and the degree of fit, in terms of both span and level, is used to prioritise subsequent search iterations. Metaphorically, process connection can be likened to a manual telephone switchboard operator using a wire with a jack plug on each end to connect an incoming call to the correct telephone extension. The incoming caller is the customer firm presenting a serviceneed, the telephone extension owner is the service supply firm and the switchboard operator is the composition meta-process. Such a meta-process would include at least the following sub-processes: finding, assessment, connection, service production, and then disconnection. Our argument is supported empirically by its congruence to the Web Services technology standards WSDL, UDDI and SOAP. These technologies enable web services to automatically describe themselves, publicise their location, connect, and then produce the actual service. We describe the functions of these web services technologies with respect to a composition meta-process in Table 2. We have also added two additional functions, assessment and disconnection, that are decisions for the service consumer rather than a web service standard.

Analysis framework Next we develop the two main components of our analysis framework. These focus on how firms relate to each other and how they start and end such relation-

Research method Our investigation of Manchester United’s network is a multi-actor as well as a multi-level study so we take an

European Journal of Information Systems

632

Manchester United FC’s orchestrated network

Table 1

Duncan R. Shaw

Proposed multi-level analysis framework for inter-firm process interaction

Perspective/level

Description

From the customer’s customer

This level is outside the customer–supplier system, in the environment, and here it is possible to perceive the meta-product of the whole customer–supplier interaction system.

From the ‘whole’ customer

This level perceives the composition and creation of the meta-service but it does not perceive the metaservice’s context or value as its customer (above) does. Only the customer controls the configuration of suppliers’ services into a meta-service because suppliers do not perceive this level and the customer’s customer has ‘outsourced’ it. Suppliers compete for orders from the customer and supplier-generated services are only composed on the customer level so competition is actually on the level of the customer.

From the ‘whole’ supplier

This level perceives the composition and creation of its service but it does not perceive its service’s context or the value placed upon it by its customer. The supplier controls the configuration of its sub-process into its service (as the customer does with the suppliers’ service one level up). Direct supplier–supplier interactions, such as the transfer of customer data after a customer changes supplier, exist on this level but are not part of the process that produces the ‘whole’ supplier’s service so we do not include them here.

From a sub-process of the customer

This level perceives and accepts elements of the supplier’s whole service but not the whole service. These elements are composed into its own sub-sub-processes.

From a sub-process of the supplier

This level perceives and produces elements of the supplier’s whole service but not the whole service. These elements are composed from its own sub-sub-processes.

From a human or application in the From a human or application in the

This agent enacts a sub-process of the customer or supplier, for example, a human activity or a machine programme (Curtis et al., 1992). If this is a human then it is the actual agent of perception.

machine customer firm machine supplier firm

From the supplier’s supplier (not shown in Figure 1)

Table 2

This level is outside the customer–supplier system, in the environment, and here it is possible to perceive the inputs to the whole customer–supplier interaction system.

Web services technologies and their functions in a composition meta-process (based on Muschamp, 2004)

Web services technology standard

Function

UDDI (Universal Description, Discovery and Integration)

Enables the firm with the service-need to find the web service.

Assessment

Assesses the available services produced by all current network members against the service-need and chooses one or more if they fit, that is, a filter function

WSDL (Web Services Description Language)

Enables the firm with the service-need to connect to the web service.

SOAP (Simple Object Access Protocol)

Enables service to be produced.

Disconnection

Planned or unplanned termination of the relationship between service consumer and service producer.

interpretive stance, because of the subjective nature of human interaction, and iterate around a hermeneutic circle, between network and organisational level perspectives so as to consider an interdependent whole (Klein & Myers, 1999; Chalmers, 2004). The novelty of using hierarchy theoretical concepts in the inter-firm network

European Journal of Information Systems

domain points to a qualitative approach because our investigation is concerned with initial questions of ‘how’ and ‘why’ rather than of ‘how many’ and in seeking to answer ‘how’ and ‘why’-type questions we follow Yin’s recommendations (2003) and use a case study approach. This is fitting because we are concerned with the

contemporary phenomena, which we have no control over, of business relationships between many different firms from many different sectors. It is also a valid approach for information systems research (Benbasat et al., 1987; Lee, 1989) and the inter-firm relationships are ultimately mediated via business information systems. We use a single case study because it is what Yin calls a revelatory case (2003). Yin defines a revelatory case as one displaying ‘phenomenon previously inaccessible to scientific investigation’ (p. 42) and while many organisations work closely in networks with other organisations, far fewer organisations significantly influence the direction of their network. But the case phenomena of Manchester United is revelatory because it is a special type of European football club due to its financial success, the extent to which it works with external partners and the variety of partners that it works with. Manchester United is one of the first football clubs to do this and also one of the most successful at the time of the study. Our use of a single case has external validity implications, that is, generalisation implications (Lee, 1989), but a single case is justified at the outset of theory generation (Benbasat et al., 1987) and although it may limit statistical generalisation is does not degrade analytic or theoretical generalisation (Robson, 2002). This is consistent with the theory building objectives of this study. We are concerned with dynamic phenomena so we have used different data collection methods and different sources (Eisenhardt, 1989). Over a three and a half year period interviews ranged from 15-min informal conversations to semi-structured meetings and a formal presentation of early conceptual models to senior management. Most interviews were with the company’s then Director of Commercial Enterprises because he was the specific person responsible for inter-firm relationship building, for example, he played a key role in creating the relationships with Nike and Vodafone as well as developing links with the supporters’ clubs around the world. Overall, we used triangulation to converge evidence, analysis and synthesis upon the same phenomena at the dyadic relationship level and at the network level. Specifically, and according to Patton’s four types of triangulation (in Yin, 2003), (i) we used data triangulation by accessing multiple sources of evidence on the same relationship phenomena; (ii) we did not use investigator triangulation because there was only one investigator, although findings were exposed to the perspectives of colleagues and students as well as those of United’s managers; (iii) we used theory triangulation by incorporating theoretical concepts both from the extremely broad-based Systems Theoretical literature as well as introducing novel Hierarchy Theoretical concepts from different disciplines such as Biology and Ecology; and (iv) we used some limited methodological triangulation in the different data evaluation methods used (i.e. the different types of investigator interactions with participants, and the period of the EJIS revise and resubmit cycles, as well

633

Duncan R. Shaw

as evaluations of findings by participants and academic colleagues). The prolonged relationship with the case managers also helped to reduce validity reactivity and increase trust as well as disclosure. Data sources included meeting notes, telephone conversations, archival data, company reports, commercial market research reports and the website content of two of the company’s websites (www.manutd.com and www.manumobile.com) and those of partner organisations. An early version of the theoretical model (Figure 4) was shown to the CEO for validation and an earlier version of this paper was validated by the then Commercial Enterprises Director. A supplementary interview, market research data and web content data were also obtained at a later stage to answer reviewers’ points and update the case. Finally, this version of the paper was validated and commented upon in detail by United’s Director of Commercial Enterprises and the Director of Marketing, at the time, and the current CEO who was also the CEO when the research was first undertaken.

Case background One example of an orchestrated network is the group of firms that collectively produce commercial value by using the brand equity of the U.K.’s Manchester United Football Club (MUFC), a company that was floated on the London Stock Exchange in 1991 and taken into private ownership by Malcolm Glazer in May 2005. Manchester United was formed in 1878 and until the mid-1990s derived most of its income from match day ticket sales, catering and some merchandising. However, the then ground-breaking showing of U.K. football matches on satellite television stations, such as Sky, not only opened up new revenue streams from television but more importantly it hugely amplified the popularity of the sport. The possibility of new revenue streams for football clubs and the greater popularity of the sport itself produced the possibility of yet more ICT-based revenue generation channels. Between 1999 and 2004 United increased its annual revenues from d88 million to d169 million (Figure 2) (Keynote, 2002; Keynote, 2005; MUFC, 2004). Only a small part of this increase came from the traditional revenue channels of match tickets sales,

£ millions

Manchester United FC’s orchestrated network

180 160 140 120 100 80 60 40 20 0 1999

2000

2001

2002

2003

2003

2004

Year

Figure 2: MUFC revenues 1999–2004 (Keynote, 2002, 2005; MUFC, 2004).

European Journal of Information Systems

634

Manchester United FC’s orchestrated network

merchandising and memorabilia; or from conferencing and catering. The increase over the last few years came from sponsorship deals, charging television companies for the right to screen live matches and other media royalties. The clubs in the U.K.’s top league, the ‘Premier League’, increased their income from d305 million in 1992 to d670 million in 1997 and then d1760 million in 2004 (Keynote, 2002, 2005). In their 2002/2003 financial year United earned 41% of revenues from match day activities, 32% from media rights and 27% from other commercial uses of their brand (MUFC, 2004). In 2003/ 2004 they earned 36% of revenues from match day activities (Keynote, 2005). However, the costs associated with purchasing and employing top quality players have increased hugely over the last few years. Fees paid from the purchasing club to the vendor club for transferring a single player are measured in tens of millions of pounds sterling. For example Manchester United signed Rio Ferdinand from Leeds United in 2002 for d30 million and Real Madrid signed Zinedine Zidane from Italy’s Juventus in 2001 for d42 million (Keynote, 2002). In July 2002, United calculated players’ salary costs were increasing at a compound annual rate of over 40%. Player wage growth across the industry has reduced from a high of 25% p.a. in that year (Keynote, 2005), but in the first six months of 2004 total wage costs (i.e. including non-players) accounted for 41% of United’s revenues (MUFC, 2004). Very strong player salary inflation means that United has had no choice but to gain revenues from new sources, if it is to attract world-class players in order to continue its on-pitch success and remain popular with the general public. United’s strategy is to produce value from its brand equity. This is the value placed upon a branded product that is in addition to the value of an equivalent nonbranded product as perceived by the consumer, that is, the fan (Bradley, 1995). Based upon its financial and sporting performance United is commonly thought of as one of the most successful football clubs in the world at deriving value from its brand. The club calculates that it has over 70 million fans globally and holds database records for over 1.9 million fans. (When commenting upon this version of the paper the club’s directors noted that this figure was from 2004 and it has since risen to approximately 3.5 million fans (MUFC, 2007)). These records hold information that includes fan contact details and consuming preferences. In the next few years the club’s goals are to vastly increase both the fraction of its total fanbase that it holds full records for and the value derived per fan, what it calls the ‘return on customer equity’ (MUFC, 2007). However, the sheer scale of the task and the complexity of a fan population distributed across most countries on earth is a significant barrier to these goals. The problem for the club is in three parts. Firstly, how can it capture data on a sizable proportion of its global fan-base? Secondly, how can it understand what products and services they want and how they prefer to

European Journal of Information Systems

Duncan R. Shaw

access them? Thirdly, how can it supply all the diverse requirements of a potential customer base scattered around the whole world? Essentially, by late 2001 United recognised that it had this huge challenge. It had millions of potential customers, for whom the MUFC brand could be a significant influence on their buying behaviour, but it did not know what products and services they wanted or would want in the future. Also, as a football club, it was unlikely that United could supply these needs itself unless they were directly related to its historical core business of running a football team. United’s solution to the supply part of this challenge is to work with other firms who supply the services that are outside United’s core business. It has constructed a partner network made up of organisations that it considers to be ‘the very best operators in each field’ (MUFC website, 2004). United does not partner with firms that only contribute sponsorship cash. They see themselves as converting fans into customers by following a process for each fan that starts with IDentifying who the fans are; Understanding more about them and their needs and values; then Engaging with them by finding an appropriate channel to start Interacting with them in twoway communications; and then ultimately Transacting with them in a paid service. This service relationship them repeats the process cycle by enabling United to better Understand the fan and better Engage them in further Interactions, which leads to repeat service Transactions and new service Transactions. United only recruits partners that can contribute to accelerating this cyclical process and they call this their ‘Accelerator Principle’ (MUFC, 2007). Manchester United’s main partners and their ‘accelerant’ services are shown in Table 3. ‘Accelerant’ services hasten United towards its goal of transacting more services with more fans. One example of an accelerant is the Pepsi ring pull competition in 2002, which generated over 1 million entrants from three Chinese cities. When they entered the competition each fan gave personal data to United via Pepsi. Partners may also contribute sponsorship cash but accelerating United’s number and depth of relationships and transactions with fans is their primary purpose. For example, Sharp was the shirt sponsor for 18 years but it was replaced by Vodafone because Vodafone could help United to engage and then interact with fans using its mobile communications technology (MUFC, 2007). Vodafone’s ‘Vodafone live!’ service to United’s fans delivers content around the world to fans’ mobile handsets. This content consists of club news, downloadable games, ringtones, pictures, game fixtures and results, squad info and video downloads for 3G phones. Availability varies by country, for example ringtones are available in the U.K., Australia, New Zealand and Sweden but pictures are available in six countries and games are available in 12 countries including the U.S.A. Unitedthemed graphics cost up to d1.50 plus the cost of a text message and pictures of the club’s crest and different players vary in price. Fans pay for all these services but

Manchester United FC’s orchestrated network

Table 3

635

Duncan R. Shaw

Manchester United’s partners and their services (MUFC, 2004, 2007; Keynote, 2005)

Partner

Accelerant services

Description

Vodafone

T, E, I

Developer of mobile services such as live match and news alerts; and affinity mobile games, images, and screen savers in a d36 million deal over 4 years (Keynote, 2005)

Nike

T, E, I

Produces United replica kits, training wear and all other United merchandise operations. Manages the stadium’s ‘Mega Store’ at the Old Trafford football ground in a d303 million deal over 13 years

Terra Lycos

ID, U, E, I

Website partner provision and development services (1.2 million unique website users per month as of late 2004)

Ladbrokes

T, I

Stadium and online betting services

Dimension Data

ID, U

ICT systems such as the fan relational database

Fujifilm

T, I

Global media rights manager and developer of branded imaging products

BSkyB

ID

Satellite and terrestrial television broadcasters supporting MUTV

Granada

ID

Terrestrial television broadcasters supporting MUTV

MBNA, Barclaycard, Endsleigh, The Britannia Building Society, Halifax Bank of Scotland

T

Credit card services, insurance services, personal financial services

Audi UK

E, I, T

Official car supplier, for example, personal transport for senior team executives and high profile players

Pepsi

ID, U

Global brand, for example, a ring pull competition generates customer data and engagement

Budweiser

ID, U, E, I

Association with Budweiser’s supply chain in China gave visibility and fan data

Wilkinson Sword

ID, I, T

Created a licensed male grooming range that gave United visibility in China, a royalty from each sale and fan data

Century FM

E

Regional UK Radio station that attracted fans to interact with United

Note: ID ¼ identify; U ¼ understand; E ¼ engage; I ¼ interact; T ¼ transact.

they are all designed to enable the fan to experience more of the club than a non-member of ‘Vodafone live!’ For example video downloads give the fans access to clips during and after a match, analyses of match performances and the reactions of players and other fans; the ring tones are recordings of real crowd chants; and the news service can update them before non members (MUFC, 2004; Vodafone, 2004).

Case analysis Following our research question, next we assemble the theoretical concepts from the literature into an analysis framework and then link them to the appropriate case phenomena. In this analysis the focal network does not include the fan. The fan is part of the environment and thus a supplier of inputs to the open system and a

receiver of outputs. We have grouped the analysis into three sections: (i) Network Structure, (ii) Network Behaviour and (iii) Inter-level Network Behaviour (Table 4). Network Structure focuses upon the basic network elements. The focal system is the orchestrated network that is made up of elements that include Manchester United and its partners. Here the unit of analysis is United’s orchestrated network plus its immediate environment. This environment can be decomposed from perspectives that include the Commercial environment (fans’ financial resources), the Informational environment (fans’ service-needs), the Legal environment (U.K. and global commercial law and regulatory rules), the Technical environment (ICTs such as network infrastructure, browsers, the internet, the WWW, home PCs, mobile phones, data processing and storage, B2B data

European Journal of Information Systems

636

Manchester United FC’s orchestrated network

Table 4

Duncan R. Shaw

Inter-level behaviours that cross levels and processes that reconfigure the network (evolution)

Inter-level behaviour of focal network – system evolution Concepts

Case phenomena

Service composition: docking service-needs with services

United’s ‘filtering process’, which sorts among prospective partners for the most appropriate fit with the service-needs of United (which accelerate the cycle of understanding fans and turning them into customers), acts as a proxy for the fans’ service needs

Business process composition

The actual joining of United’s business processes with those of a partner. Stage 1, design: negotiation of legal, contractual and technical agreements. Not just service-need outcomes but how these are produced and technologically enabled, for example, intellectual property agreements, service level agreements and data sharing standards. Stage 2, implementation: actual process instances enacted each time these processes are joined, for example, single payment to United from sponsoring partner

The orchestrator’s composition meta-process

From Manchester United’s perspective

Finding

Partner firms are specifically targeted by, or approach, the club. Potential partners register interest and show their potential for fulfilling United’s service-needs (i.e. service partners to outsource to) via business proposals

Assessment

United use a detailed checklist to ensure fit with potential partner’s brand, corporate values and business strategy. In addition to ensuring that the service and the contractual conditions fit United’s service-need. Checks: business plan quality; the look and feel of potential partner’s service; financial returns; commercial rights secured (e.g. over any intellectual property produced); time and effort required; and adequacy of business plan funding

Connection

Partners already partially connected to United in the finding sub-process: includes other connections required for the production of an emergent service to fit United’s service-need – separate systems joining into one newly formed system, that is, a standards agreement process for factors such as ICTs; data semantics and syntax; service levels; and synchronisation (e.g. as in Just In Time manufacturing)

Service production

Partners connect to fans by using fan data from United in their sales and marketing processes. There is a detailed mixing of business processes between the partners and United across all firm boundaries, for example, stadium areas and ICT channels. This is a newly formed system starting to enact its own existential processes, that is, the B2B transport of physical or informational resources (represented by arrows in Figure 1)

Disconnection

The termination of information and physical resource flows so that the service-need is no longer fulfilled. Some relations are preserved to varying degrees, for example, with building contractors for occasional stadium expansions and other project work.

sharing standards) and the Physical environment (the city of Manchester, the U.K. and the world depending upon the physical location of elements and their relations). Another main structural characteristic of the United’s network is its Hierarchy of Levels, that is, different person and firm elements exist upon different levels (see Figure 3). Also, each level has the structural characteristic of Span since the system elements are of different types and produce different services, for example, Vodafone is a telecommunications company and MBNA is a financial services company. Different types of elements on the same level logically require a span that is greater than one. Network Behaviour focuses upon the relationships between elements on the same level, that is, the element

European Journal of Information Systems

Partner firm Environment Network level

MUFC Staff member

Firm level Human level

Figure 3: Hierarchical representation of the MUFC commercial system (based upon Wilby, 1994).

urations and swaps new firms in, and old firms out, of the network. These are processes that reconfigure lower level processes, that is, they are meta-processes. They can be compared to the role of strategic human resource management in managing staff turnover so as to enhance the mix of a firm’s overall skills and capabilities. Process reconfiguration is different to process change or reengineering because actual processes may remain the same but their inputs and outputs may be connected to different customer or supplier processes. In this case it is represented by United finding and then starting to work with new partners, such as when Vodafone became United’s new main advertising sponsor. The business processes of Sharp, the old sponsor, were not necessarily changed, nor stopped. But from United’s perspective it was not just that a source of sponsorship changed because Vodafone could also contribute highly developed competencies in mobile communications technology. The sponsorship revenues certainly increased but the old sponsor presumably was still able to produce the service of sponsorship for another club. It was an example of higher level process reconfiguration rather lower level re-engineering. Judging change to be reconfiguration rather than re-engineering is a matter of hierarchical level and it depends upon the observers’ perspective. For example, United’s merchandising operations, including managing the Mega Store under the Old Trafford stadium, were outsourced to Nike. This was also a reconfiguration of higher level processes from the perspective of United’s senior management. But from the perspective of people that worked in the store, who may have changed their role or changed their job, this was an example of process re-engineering. The network that is made up of United, its fans and its partners is shown in Figure 4 together with the physical and informational resources that flow between the network members. In order to focus on the network’s organisational members, and their connections to fans, Figure 4 does not include stadium-based services that

bra

h cas

d at a

h

n atio oc i ass i ce erv t/ s duc

nd

nd

pr o

pr o

Focal system’s environment

Focal system

a da t

ass oci duc atio t/ s n erv i ce da t a

Partner

cas

functions and inter-element processes that produce emergent services on the network level (meta services). Here the processes of several firms upon the same system level combine to produce a composite service. The Interelement relations are the flows of information and physical goods between elements, that is, they are characterised by the business processes that are enacted by United and its partners. At the boundary of the focal network Input flows to this open system, from the environment, are represented by commercial payments from fans and data on their service-needs. Output flows from the open system to the environment are represented by services produced for fans by Manchester United and its operational partners. These Input and Output flows maintain the open system in a state of non-equilibrium. For example, United’s player salary bill totals much more than the revenues from the traditional revenue channels of stadium activities and merchandising. The players’ salary bill consumes value that ultimately comes from an external source, the fans. The fans’ Service-needs are represented by their individual requirements and the context with which they individually judge service quality, for example, the need for ‘up to the minute’ information on football game progress generates a feeling of closer association with United. The Services themselves are the outputs of business processes, for example, a text message to fans that immediately alerts them, wherever they are, when United scores a goal. Services can be joined into a Composite Service when several component services on the same level are combined to produce an emergent service on the next level up. Component services, on level zero, produce an emergent service on level one, which may or may not fit a service-need on level one. For example, two successive text messages may give some information about a football game but together they may not produce a feeling of close association for the fan, that is, they may not fulfil the higher level service-need. Combined together all the services that the fan consumes constitute the Network Meta-service. This is represented by a perceived better quality of life that is caused by feelings of closer association, by the fan, to United. Conceptually this is on a higher level than a single service, for example, a fan’s happiness may be caused by a football score but happiness is irrelevant and unknowable on the lower level of mobile phone texting technology. The Network Meta-service is hopefully designed to fit the service-needs of the Environmental meta-process. This is the sum of service-needs of all the processes enacted in the environment. One component of this is each fan’s life process which the fan constantly enacts by existing. Other meta-processes are concerned with other environmental actors, for example, government taxation processes. Inter-level Network Behaviour focuses upon phenomena that cross levels such as business processes that coordinate service and process composition as well as develop the network (evolution). Here the orchestrator actually connects firms together in process new config-

637

Duncan R. Shaw

br a

Manchester United FC’s orchestrated network

Fan

MUFC data Flow of physical or informational resource

Figure 4:

Focal system boundary

The value flow system of United’s process network.

European Journal of Information Systems

638

Manchester United FC’s orchestrated network

flow directly to fans from United or the large cash flows that are players’ salaries. The football players are not shown in Figure 4 because they are part of a sub-process that is concerned with the ongoing production of the United brand. So from the perspective of a partner firm they are a sub-process of MUFC. From this higher partner firm level perspective phenomena associated with reengineering football team processes, like picking different players to play, is filtered out. In Figure 4 we can see how United captures data from fans, and partners, and exports it to partners in return for cash, which originates from the fans. The partner organisations produce services for both fans and United in return for cash from the fans and highly refined fan data from MUFC. The fans provide marketing data to United and the partners in return for associating themselves with United’s brand via the partners. Fans also interact directly with United via focus groups and fan forums. We have chosen to use the term ‘association’ because this is a common term for the sponsors or partners using a brand in their marketing process. We extend it here to fans desiring to associate themselves with a brand as part of the fulfilment of their own service-needs, that is, the needs stimulated by their personal value systems.

Discussion Operating value flow systems for sustainability Business-to-business (B2B) relationships have long been thought of as flows of diverse informational or physical resources (Van De Ven, 1976). In Figure 4 we represent the composition of United’s business processes with those of its partners as the flows of various resources between the elements of the focal network (MUFC and its partners) and environment (the fan). These flows represent linked changes in the properties that we associate with different entities in the system. These properties are diverse and include: financial value (e.g. cash); physical products (e.g. football kits), services (e.g. credit cards), data (e.g. fan buying preferences) and brand association. These properties do not consist of the same substances, nor even of any substances in most cases. In fact they actually exist on different hierarchical levels. The physical products exist upon one level and other resources are informational resources about them, that is, they are ‘meta-resources’. Similarly, the physical products themselves are made up of a lower level of components and from which they are a ‘meta-product’. If we use ‘value’ as a conceptual ‘common denominator’ we can abstract away the complexity of the different levels of resources by thinking of the flows as flows of ‘value’ that circulate around a ‘value system’. ‘Value’ is the quality that is exchanged, the compensation for effort, or the production in return that relates the activities of one firm to those of another and to those of a fan. ‘Value’ flows across system boundaries and are ‘in eye of the beholder’, that is, it is very much an interpretist concept and consistent with Hammer and Champy’s definition of business

European Journal of Information Systems

Duncan R. Shaw

processes that create ‘an input that is of value for a customer’ (1993, p. 38). Modelling a network’s value flow system in this way enables an understanding of relationships at different hierarchical levels and from different firm perspectives. This is required for maintaining the sustainability of open systems. For example, there is no actual flow of brand association from United to the fan. Indeed, where partnerbased services are concerned, the only direct flows to the fan are from the partners (Figure 4). The significance of this is twofold: firstly, the only business processes that service United’s fans (neglecting games and other stadium activities) are enacted by its partners. Thus the club is dependent upon the service quality of its partners’ business processes. Secondly, the major factor in the fans’ choice of these services is a desire to associate themselves with the MUFC brand. If the fans are dissatisfied with the service quality that they receive then the brand will be damaged in their perception. This may subsequently reduce their desire to associate themselves with the MUFC brand. Thus, there is a possibility of blocking the only flow that keeps the fans wanting to remain members of this network, in contrast to fans that buy other nonfootball branded products. Blocking this perceived flow of ‘association’ from United to the fan’s life experience has the potential to destroy the viability of the network. However, Manchester United is extremely sensitive to fan satisfaction levels and annually surveys 10,000 fans in a Fan Satisfaction Study, which divides its business into 20 areas ranging from stadium toilets to website content.

Governing value flow systems for sustainability In the last section we focused upon the smooth operation of the network’s value flows in order to maintain it far from an equilibrium state. Here we look at metaoperations like swapping out old firms and recruiting new ones. To do this the orchestrator has to focus upon a firm’s role in the ‘big picture’ of network activity. To do this we develop our concept of the virtual network member into a model of a single firm’s relationship with the rest of the network that we call the Network Orchestration Model (NOM). Figure 5 is a 2  2 matrix with the columns divided into Single firm and Other firms in the network and the rows divided into Firm and Network levels of perception. Thus the top left quadrant is empty because the single firm does not exist on the network level. The NOM takes the single organisations view of its relationship with a network and simplifies it to a B2B dyadic abstraction. The coordination function of a network orchestrator removes the complexity of modelling a singles firm’s relationship with each firm in the network. The simplification is that the network is treated as a single firm. This could also be applied to any network member such as a firm’s relationship with this network of other firms. A network level view plus a firm level is required for any multi-level network analysis but in an orchestrated network an additional filter is present. In an orchestrated

VNM Coordination Function ice rv ic rv se

e

Firm level

Single firm’s view of other firms

se

Single firm’s view of the whole network

Network level

Manchester United FC’s orchestrated network

Single firm

Other firms in network

Figure 5: The Network Orchestration Model (NOM) a single firm’s connection to a network as a dyadic connection to a Virtual Network Member (VNM).

network the prospective member firm assesses a much reduced subset of all the network members’ services for fit with its service-needs because the orchestrator has performed a pre-assessment. The single firm just requests a service and then is offered a service or a selection of services, that is, the orchestrator passes to the single firm a filtered offer of services. This is how it coordinates the docking of services with service-needs with the goal of maximising fit. Also it is consistent with the hierarchy theoretical filtering of signals transmitted from lower hierarchical levels by the contexts (e.g. service-needs) of higher levels. This occurs at the level boundaries, that is, the zones of maximum interaction strength variability or ‘surfaces’ (Allen et al., 1984). For example, some of the service needs that are generated by a credit card partner, like Barclaycard, include a requirement for data on fans that can be used to support the credit card firm’s marketing processes. Instead of connecting to all fans, or all the partners in the network, the credit card firm receives filtered data from United, which has been cleansed and checked using E.piphany CRM software. Another example is the outsourcing of the Mega Store to Nike. From the senior management perspective this just looks like an example of process reconfiguration since a service that was received from an internal department is reconnected to Nike. But from the worker perspective this looks like an example of re-engineering, that is, a much more detailed change, since the workers role or job will change. The composition meta-process in Table 4 is produced from the perspective of United but it could also be produced from the perspective of a partner firm or a fan. For example, one service produced by United for a partner firm is the MUFC brand whose use enables a partner to behave in some small way, to a fan, as though it is the football club itself. Another service that United

Duncan R. Shaw

639

produces for partners is access to high-quality aggregated and filtered data on fan segmentation. This comes from other partners and fans themselves. A fan’s viewpoint could also be substituted for the single firm view in Figure 5. This is because although fans, partners and orchestrators (Figure 3) exist on different organisational levels the levels are separated by the same structural composition mechanism of emergence. This implies that our model can be used to understand the inter-level relationships of many types of network members on many different levels. This is valuable because these interlevel relations are made up of multiple orthogonal dimensions. Two such dimensions are: (i) several suppliers composing services into a higher level service and (ii) the orchestrator processes of new service composition and changing the actual suppliers of the services. The NOM simplifies the one-to-many relationship by modelling it as a dyadic business process relationship with a virtual network member. The virtual network member is theoretically composed of the rest of the network plus a coordinator that we call the network orchestrator. The dyadic relationship is abstracted from the complexity of a one-to-many relationship by filtering the properties of the one-to-many relationship through the orchestrator. This gives the single firm, or fan, a subset of all options, which is much more practical to choose from for boundedly rational actors.

Lessons learned The value flow systems of inter-organisational networks are open systems that exist far from equilibrium and so require maintenance. However, they are complex because value flow maintenance requires a fit between serviceneeds and services and this fit is entity dependent, for example, a partner firm needs the particular subset of fan data that supports the marketing of their particular services to fans, such as consumption preferences for that type of service rather than any other. Thus fit depends upon the perspective of the firm that consumes each service. The NOM is an architecture for simplifying network analysis and management problems such as value flow maintenance. It can be used to assess each firm’s fit with its role or ‘ecological niche’. This is done simply by listing all the services that the network members collectively produce. Then each firm’s serviceneeds are compared with the services that it could receive or currently does receive from the rest of the network. The NOM of a network contains the architectural rules that govern the value flows in business-to-business interaction. These rules are firstly based upon the simple concept of connecting a service-need to a service in a ‘switchboard’ effect. Secondarily, the addition of the concept of Virtual Network Member means the single dyadic connections can be assessed across a whole network. This generates multiple perspectives across network span and levels which can be used to coordinate overall system maintenance.

European Journal of Information Systems

640

Manchester United FC’s orchestrated network

Implications for practice and theory The logic of conceptualising such a coordination mechanism as a ‘multiple switchboard’ is demonstrated if we consider an empirical example of this switchboard concept. Search is the basic service of Google. It connects users with a prioritised subset of possible websites and also advertisers with a prioritised subset of possible customers. For users this reduces time invested and for advertisers it increases average revenue per advertisement. However, by constraining partner-to-partner relations to be via United the complexity of being a partner is damped but so is the agility of the network. This agility includes the speed of response to exogenous shocks and also the diversity of compositions of partners’ services. The NOM is neither span nor level dependent and so can be applied to complex networks of people and informational entities as well as firms. Search is increasingly being used to manage internal complexity (e.g. Sarbanes–Oxley compliance and other content management) and external complexity (e.g. manual web searches and web services). The NOM can be used like the switchboard effect of Search to connect service-needs to services and thus assemble and then sustainably manage the value flows between network members. The NOM allows managers to manage the value flows of existing networks or to design new sustainable networks from the perspectives of different members. The model does this by encouraging the model user to look at each member from a member and also from a ‘whole network’ perspective when assessing the appropriateness of current or new members. The implications for theory are in the called for development of coordination theory (Olson et al., 2001; Crowston et al., 2004) by newly incorporating the hierarchy theoretical concepts of level, perspectives and filtering (Allen & Starr, 1982; Allen et al., 1984; Salthe, 1985; Wilby, 1994). These concepts also have implications for the discussions around setting the vertical and horizontal boundaries of the firm and their filtering effects (Afuah, 2003).

Conclusions The contributions of this paper are for an audience of managers of orchestrator firms and coordination researchers. We use a revelatory case study of Manchester United Football Club and its network of partner firms to investigate a particular type of coordination role called ‘orchestration’. In our analysis of Network Structure, Network Behaviour and Inter-level Network Behaviour we conceptualise United’s network of firms as an open system that is sustained in a far from equilibrium state by input and output flows of material and information. These flows we call ‘value flows’ because they are dimensionally orthogonal, in that they do not mutually interact, but they do support business processes within firms that support each firm’s business objectives. Two such far from equilibrium states include United’s very high commercial value, in contrast to its physical assets,

European Journal of Information Systems

Duncan R. Shaw

and its operational capabilities that are actually sited within other firms. Such far from equilibrium states contrast constitution and behaviour, that is, in process philosophical terms United is what it does (Rescher, 2000). They also are an example of holism since in systems terms United’s network is more than the sum of its parts. United’s continuing commercial success increasingly comes from how it conducts its relationships rather than from its internal elements. Network Behaviour and Inter-level Network Behaviour both sustain far from equilibrium states by maintaining value flows. Network Behaviour is concerned with operating flows between elements on the same level, that is, inter-element business processes that produce emergent services on the network level (meta-services). Inter-level Network Behaviour is concerned with governing flows that cross levels such as business processes that coordinate service and process composition and develop the network (evolution). We have demonstrated that sustaining such a network of value flows is fundamentally based upon the fit between each service-need and each service. For sustainability each service-need is operationally generated by each consumer firm in the network and operationally fulfilled by each service producer. Each partner firm is an open system and so both produces and consumes services. On an operational level service-needs and services can radically change but the actual service consumers and producers are fixed. On a governing level service consumers and producers are reconfigured by the orchestrator to produce different operational level productions of the same meta-service or different meta-services. Thus one contrast between an orchestrator and normal firm, that is, one difference in what being ‘on a higher level’ signifies is that orchestrators introduce firms to other firms in a sort of ‘dating agency’ role or ‘search engine’ role. Normal partner firms only have the power to modify the service/ service-need relationships with firms that they are introduced to by the orchestrator. For example, United’s Accelerator Principle highlights orchestrator’s role as a governor rather than an operator because it applies to the provision of indirect services that generate, filter and use customer data rather than to the provision of direct fan services. This fits the bounded ability of partner firms to find the service producers that they need since orchestrators act as filters between partner firms and potential new network members but they also ‘go out’ and search for new members. In this case study new potential network members usually approached United but it does highlight a service produced by the orchestrator for the partner firms. We have used this analysis to propose the NOM as a tool for investigating the relations between one network member and the rest of the network. The NOM can be used to simplify a complex one-to-many relationship into a series of dyadic relationships. The NOM does not explicitly model any specific types of business processes, it only models relations that connect to partners via the

Manchester United FC’s orchestrated network

orchestrator and it may be limited to orchestrated networks. But it can be used to generate a view of the network’s operation from the multiple perspectives of the different network members. By assessing the fit between service-needs and potential services for different network members the NOM can be used to check the sustainability of the network’s value flows and to assess the designs of new network configurations. Thus the NOM itself is a model of the orchestrator’s role, which together with the introduction of Hierarchy Theory to this domain is a novel development of Malone and Crowston’s Coordination Theory (1994) that uses systems theoretical and business process modelling concepts. Our use of a single case limits the empirical generalisation of our findings (Lee, 1989) for other networks, but this is justified at the outset of theory generation (Benbasat et al., 1987) and it does not degrade analytic or theoretical generalisation (Robson, 2002). This is consistent with the theory building objectives of this study which further case studies of networks in different

Duncan R. Shaw

641

sectors and of different scales would also support. Finally, the introduction of hierarchy theory to this research domain generates a theoretical avenue for the introduction of temporal concepts. This calls for further research which is not pursued here but a temporal perspective has obvious relevance to a domain that requires B2B process synchronisation.

Acknowledgements We would like to thank Manchester United Football Club and especially Ben Hatton, the Director of Commercial Enterprises at the time; Peter Draper, the Director of Marketing at the time; and David Gill, the CEO, for the information that they have provided and for reviewing different versions of this paper. We would also like to thank Debra Howcroft and Richard Phillips, as well as the EJIS Associate Editor and the reviewers for their invaluable comments and advice. We gratefully acknowledge the support of the U.K.’s Engineering and Physical Sciences Research Council (EPSRC) in funding the Flexible Business Integration project.

Author the author Duncan Shaw is a lecturer in Information Systems at Nottingham University Business School. His research interests are inter-organisational network cultivation, orchestration and coordination, the management of complex business environments and Business Model Theory. Duncan’s business background includes Customer

Satisfaction Manager for Motorola’s EMEA region and consultancy to Coca-cola, Xerox, Danone and other organisations across Europe. He has also consulted to the U.K.’s Office of the Deputy Prime Minister, local government and the private sector on Customer Relationship Management and Business Process Reengineering.

References AFUAH A (2003) Redefining firm boundaries in the face of the Internet: are firms really shrinking? The Academy of Management review 28(1), 34–53. ALEXANDER C (1964) Notes On The Synthesis Of Form. Harvard University Press, Cambridge, MA. ALLEN TFH, O’NEILL RV and HOEKSTRA TW (1984) Interlevel relations in ecological research and management: some working principles from hierarchy theory. USDA Forest Service. General Technical Report RM-110, 11pp. Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado, http://www.botany.wisc.edu/allen/alleinte. pdf, last accessed 26 May 2006. ALLEN TFH and STARR TB (1982) Hierarchy: Perspectives for Ecological Complexity. University of Chicago Press, Chicago. ANDREW JP and SIRKIN HL (2003) Innovating for cash. Harvard Business Review 81(9), 76–83. BENBASAT I, GOLDSTEIN DK and MEAD M (1987) The case research strategy in studies of information systems. MIS Quarterly 11(3), 369–386. BRADLEY B (1995) Marketing Management. Prentice-Hall, Englewood Cliffs, NJ. CHALMERS M (2004) Hermeneutics, information and representation. European Journal of Information Systems 13(3), 210–220. CHECKLAND P (1999) Systems Thinking, Systems Practice. John Wiley & Sons, Chichester. CURTIS B, KELLNER MI and OVER J (1992) Process modelling. Communications of the ACM 35(9), 75–90.

CROWSTON K (2003) A taxonomy of organizational dependencies and coordination mechanisms. In Tools for Organizing Business Knowledge: The MIT Process Handbook (MALONE TW, CROWSTON K and HERMAN G, Eds), MIT Press, Cambridge, MA. CROWSTON K, RUBLESKE J and HOWISON J (2004) Coordination theory: a ten year retrospective, Working Paper, http://crowston.syr.edu/papers/ coord2004.pdf, last accessed 26 May 2006. DAVENPORT TH (1993) Process Innovation: Reengineering Work Through Information Technology. Harvard Press, Boston, MA. DRAZIN R and VAN DE VEN AH (1985) Alternative forms of fit in contingency theory. Administrative Science Quarterly 30(4), 514–539. EISENHARDT KM (1989) Building theories from case study research. Academy of Management Review 14(4), 532–550. HAECKEL SH and NOLAN RL (1993) Managing by wire. Harvard Business Review 71(5), 122–132. HAGEL III J (2002) Leveraged growth: expanding sales without sacrificing profits. Harvard Business Review 80(10), 68. HAMMER M and CHAMPY J (1993) Reengineering the Corporation: A Manifesto for Business Revolution. Harper Business, New York. JOHNSTON RB (2001) Situated action, stucturation and actor–network theory: an integrative theoretical perspective. In Proceedings of ECIS 2001 the 9th European Conference on Information Systems, Bled, Slovenia, pp. 232–242. JOHNSTON RB and GREGOR S (2000) A theory of industry-level activity for understanding the adoption of interorganisational systems. European Journal of Information Systems 9(4), 243–251.

European Journal of Information Systems

642

Manchester United FC’s orchestrated network

KALAKOTA R and ROBINSON M (2003) Services Blueprint: Roadmap For Execution. Addison-Wesley, Reading, MA. KEENEY RL (1999) The value of internet commerce to the customer. Management Science 45(4), 533–542. KEYNOTE (2002) Market Report, Football Clubs and Finance, 2nd edn, September, Keynote Limited, Middlesex, UK. KEYNOTE (2005) Market Report, Football Clubs and Finance, 3rd edn, Keynote Limited, Middlesex, UK. KLEIN HK and MYERS MD (1999) A set of principles for conducting and evaluating interpretive field studies in information systems. MIS Quarterly 23(1), 67–93. LEE AS (1989) A scientific methodology for MIS case studies. MIS Quarterly 13(1), 33–50. MALONE TW and CROWSTON K (1994) The interdisciplinary theory of coordination. ACM Computing Surveys 26(1), 87–119. Mo¨LLER K, RAJALA A, SVAHN S and TUOMINEN M (2002) Network management as a set of dynamic capabilities. In Proceedings of the 18th Annual IMP Conference, Dijon, France, September 5–7 September (http://www.escdijon.com/download/imp/pdf/143_moller-svahn-rajalatuominen.pdf, last accessed 26 May 2006). MUFC (2004) Interviews, personal communications and company websites: www.manutd.com (last accessed 9 June, 2004), www. manumobile.com (last accessed 8 November 2004). MUFC (2007) Follow up interviews, personal communications and further paper validation. MUSCHAMP P (2004) An introduction to web services. BT Technology Journal 22(1), 9–18. NOLAN RL (2000) Information technology management from 1960– 2000. In A Nation Transformed by Information (CHANDLER AD and CORTAD JW, Eds), Oxford Press, Cambridge, UK. OLSON GM, MALONE TW and SMITH JB (2001) Coordination Theory and Collaboration Technology. New Mahwah, NJ: Lawrence Erlbaum Associates, Mahwah, NJ. RESCHER N (2000) Process Philosophy: A Survey of Basic Issues. University of Pittsburgh Press, Pittsburgh, PA.

European Journal of Information Systems

Duncan R. Shaw

RICE JB and HOPPE R M (2001) Supply chain vs. supply chain – the hype and the reality. Supply Chain Management Review 5(5), 46–54. ROBSON C (2002) Real World Research, 2nd edn, Blackwell, Oxford. SALTHE S (1985) Evolving Hierarchical Systems: Their Structure and Representation. Columbia University Press, New York. SHAW DR, HOLLAND CP, KAWALEK P, SNOWDON R and WARBOYS B (2007) The elements of a business process management system: theory and practice. Business Process Management Journal 13(1), 91–107. SHAW DR, SNOWDON R, HOLLAND CP, KAWALEK P and WARBOYS B (2004) The viable systems model applied to a smart network: the case of the UK electricity market. Journal of Information Technology 19(4), 270–280. SIMON HA (1969) The Sciences of the Artificial. MIT Press, Cambridge, MA. SIMON HA (1997) Administrative Behavior. A Study of Decision-Making Processes in Administrative Organizations, 4th edn, The Free Press, Simon & Schuster, New York. SIMON HA and Associates (1986) Decision Making and Problem Solving, Research Briefings 1986: Report of the Research Briefing Panel on Decision Making and Problem Solving. National Academy Press, Washington, DC, http://www.dieoff.org/page163.htm, last accessed 26 May 2006. STYHRE A (2002) How process philosophy can contribute to strategic management. Systems Research and Behavioural Science 19(6), 577–587. VAN DE VEN AH (1976) On the nature, formation and maintenance of relations among organizations. Academy of Management Review 1(4), 24–36. VODAFONE (2004) www.vodafone.com (accessed 8 November 2004). WEICK KE (1976) Educational organisations as loosely coupled systems. Administrative Science Quarterly 21, 1–19. WILBY J (1994) A critique of hierarchy theory. Journal of Systems Practice 7(6), 653–667. YIN RK (2003) Case Study Research: Design and Methods Applied Social Research Series, 3rd edn, Vol. 5, Sage, Thousand Oaks, CA.