«Derek Hillison Matt Wimble Brian Pentland Eli Broad College of Business Eli Broad College of Business Eli Broad College of Business Michigan State ...»
Why Rosetta Can’t Dance: The Combinatorics of
Standardizing Inter-organizational Service Processes
Derek Hillison Matt Wimble Brian Pentland
Eli Broad College of Business Eli Broad College of Business Eli Broad College of Business
Michigan State University Michigan State University Michigan State University
firstname.lastname@example.org email@example.com firstname.lastname@example.org
Interorganizational processes have been described as an intricately choreographed dance. Standardized workflow serves as a medium for participants to exchange complimentary moves that comprise the ‘dance’.
Attempts to develop open process standards, such as Rosetta Net, have had limited success in finding wide-spread adoption. We argue that standardization of interorganizational business processes, previously viewed as aiding interorganizational collaboration, can impede collaboration in service contexts. We show that the modeling methods used to standardize inter-organizational processes fails to account for the impact such standardization has on managerial flexibility.
Introduction In this paper, we consider the problem of how to support inter-organizational routines using network-based e-commerce technology. Computing standards organizations, like Rosetta Net, have tried for years to develop inter-organizational process standards.
Unfortunately, they have met with little success outside the IT manufacturing industry that initially established the idea. So, while Rosetta Net (and others) have attempted to choreograph the dance of inter-organizational collaboration, most organizations still can’t dance.
We argue that inter-organizational choreographers, such as Rosetta Net and ebXML, are likely to be impeded by a variety of problems. First, like people, organizations may have different expectations and ideas about how to “dance.” While e-commerce technologies provide detailed “choreography” for the pattern of messages and documents that must pass between participants, these patterns may not mesh well with the organizations internal expectations. In short, the lack of standardization and agreement among participants may impede collaboration.
Second, due to novel combinations of actors, actions and sequences of action, interorganizational routines can entail vast numbers of possible performances. The extent to which these variations occur is an empirical question, but as we show here, their possibility is mathematical fact. We argue that this variability is a good thing – it embodies the highly valued flexibility or agility that organizations seek when they participate in inter-organizational networks. Unfortunately, limitations and biases in modeling methods and managerial norms lead participants to systematically reject or underestimate this variability. As a result, inter-organizational technology that promotes standardization may have the undesirable effect of undermining flexibility and agility.
Our century-long tradition of scientific management would suggest that the way forward is to push standards that organizations should adopt. Without question, standardization has been an essential foundation of the Internet and the WWW. However, when it comes to inter-organizational relationships and routines, we need systems that encourage diversity and variability. We begin by developing some key concepts, and applying them to the problem of inter-organizational routines and technology.
Conceptual development Inter-organizational routines We define inter-organizational routines as repetitive, recognizable patterns of interdependent action between multiple organizations. This definition mirrors that of Feldman and Pentland (2003) for organizational routines, except that each actor or participant is an organization. We also adopt the Feldman and Pentland’s (2003) view that routines cannot exist without two inter-related aspects. The ostensive aspects of a routine embody the expectations or dispositions of the participants; they allow participants to guide their actions and account for their actions. The performative aspects are simply the actual events that occur.
As Pentland and Feldman (2005) point out, different participants do not necessarily share the same ostensive understanding of a routine. A person in marketing may understand the new product development process differently than people in engineering. In an interorganizational routine, the problem is potential even greater. As the level of analysis shifts from within an organization to between organizations, achieving a shared understanding of what the routine becomes more difficult because of differences in organizational structure, culture, practices, interests, goals, and so on.
Figure 1 gives an example of how three organizations can overlap and share some definitions of the ostensive aspects of the routine. If organizations are not able to achieve this overlap, collaboration would not be possible. Standardization of work process allows organizations to define the routine in a way that facilitates sharing understandings of the routine, and also allows the workflow technology to encode some of these aspects into the model of work.
Figure 1: Ostensive understandings may not agree
Nevertheless, organizations do collaborate. The authors of Rosetta Net use the term “choreography” to describe their standardized model of inter-organizaitonal business processes. The idea of “routines as dance” (Feldman & Rafaeli, 2002) offers a useful metaphor for inter-organizational collaboration, and for the pitfalls of having a partner who dances to a different beat.
Standardization versus the need for flexibility Process standardization often has a number of benefits. By standardizing processes a set of firms can all try to implement best practice procedures. Suppliers and businesses know what to expect. Those carrying out the processes can concentrate on improving the process using a standard quality management toolset. Research has found that process standardization is necessary prerequisite for business improvement through software (Maidantchik, et. Al., 1999).
At the same time, there are enormous benefits to flexibility. While inter-organizational collaboration often involves the flow of goods (as in a supply chain), each organization is providing a service to its partner. Services differ dramatically from traditional manufacturing because of the nature of production in a service context is inherently different from production in a manufacturing context. Services exhibit the characteristics of intangibility, inseparability, and heterogeneity. Intangibility refers to the idea that services cannot be inventoried, are not readily measured, and they do not even consume physical space (Shostack, 1977). Inseparability refers to the idea that the consumption of a service and the production of a service often occur simultaneously (Carmen and Langeard, 1980). Service production is often inseparable from consumption to such a degree that the consumer rises to the level of co-production (Parasuraman, et. al., 1985).
Heterogeneity refers to the idea that services often vary from day to day and customer to customer (Parasuraman, et. al., 1985).
Co-production of output that is common in services necessitates a high degree of cooperation between consumer and producer. In service industries the production process is highly contingent upon the specific interactions of consumers and producers, which implies far greater uncertainty a priori in the sequence of events necessary for production of services. As a result high degree of uncertainty results from the coproduction found in services (Argote, 1982; Jones, 1987).
The heterogeneity inherent in service processes manifests as variety that can be seen as a sign of the flexibility that is necessary for high quality (Feldman, 2000). The ability of a service provider to deal with a wide variety of situations is a mark of high customer service (Zeithaml, et. al., 1990; Cronin and Taylor, 1994) and a key factor in retaining customers in service environments (Keaveney, S., 1995).
Co-production of output in services is what necessitates the dance between consumer and producer. In service industries the choreography of this dance-of-production is highly contingent upon the specific interactions of consumers and producers, which implies far greater uncertainty a priori in the sequence of events necessary for production of services.
Choreographing this dance-of-production in services requires a fundamentally different modeling framework, which incorporates the flexibility that service production necessitates.
Process modeling frameworks in use today evolved out of the modeling of manufacturing processes, which have the explicit goal of having as little variety as possible. Initially these modeling frameworks were adapted to model services in a project-oriented context. Projects, such as construction services or software design, are by definition unique. Constraints imposed by traditional manufacturing modeling frameworks are not excessively restrictive in a project environment because the sequence of production was customized for each instance (Meredith and Mantel, 2002). When these modeling frameworks are moved out of a project context into an ongoing service context, where the model is explicitly designed to handle multiple instances, the limitations that the framework places on flexibility becomes especially relevant. By failing to address inherent differences the production process between services and manufacturing, the use these modeling frameworks to form standards results in poor adoption because of the loss of necessary flexibility.
The role of technology Just as in dance, the partners of an inter-organizational routine need a common action language or a ‘lingua franca’ that allows them to communicate. Standard-creating organizations, workflow vendors and proponents of service-oriented architectures have developed ontologies or ways of representing the language of actions in specific cases.
The increasing integration of workflow technologies into inter-organizational routines highlights the need to discuss its impact on the flexibility of processes.
Workflow Systems are intended to limit variety Increasingly, workflow management systems are used to embed (or enforce) an particular ostensive understanding on a process. Thus, these technologies have some to define work processes in service industries. However, workflow technology breaks processes down into sequential steps for the precise aim of limiting task variety (Fletcher, et. al., 2003). Current process modeling technology, which is based on traditional manufacturing, also has known limitations. Using Petri-nets it has been shown that all of the commercial process-modeling frameworks are unable to model some known workflow patterns (van der Aalst, et. al., 2003). When modeling or supporting processes, workflow technology tends to be biased toward defining the “happy path” – the most desirable or familiar variation.
RosettaNet is an example of such technology. It offers standards for implementation frameworks, business message schemas, process specifications using workflows, and data dictionaries. The process standards used by RosettaNet follow the standard workflow framework established by the Workflow Management Coalition (Piccinelli, et. al., 2002).
The Workflow Management Coalition (WfMC) which models processes in sequential steps (Holligsworth, D., 1994). While RosettaNet has been primarily adopted by electronics manufactures, Cluster 6 in RosettaNet is explicitly designed for customer service activities (RosettaNet Consortium).
Technologies like RosettaNet (include EDI and ebXML) allow system and process designers to model the business processes, based on the underlying paradigm and ontology. This digital representation of the business process serves as the bridge between the ostensive and performative aspects of an interorganizational routine. Some form of the ostensive understanding of the routine is captured in the language of the model, and is expressed as constraints on the possible performances of the routine. Figure 2 shows this relationship in graphic form. The integration of technology as the main artifact into the ostensive-performative aspects of routines represents a further contribution of this paper.
Figure 2: Workflow as the connection between the ostensive and theperformative
The most common paradigm for workflow design starts by specifying a single most likely task sequence. After specifying this most likely task sequence, the process modeler relaxes this fully constrained model by allowing for exceptions to the most likely task sequence. A production sequence is referred to as fully constrained when there is one and only one task sequence, this is opposed to a fully open model where the task can occur in any order. It will be shown that relaxing constraints on a fully constrained model, rather than constraining a fully open model results in an unexpected loss of necessary flexibility.
Analysis of Combinatorics In order to understand why a more flexible approach to modeling better it is necessary to examine how inter-organizational process standards can lead to either a) the combinatorial nature of service process makes it is nearly impossible for organizations to find the sequencing and assignment of tasks on which they agree or b) even if they can find agreement the adoption of process standards destroys value. Process standards inherently forces managers to commit apriori to both who performs a given task and in what order it is to be done. By adopting inter-organizational process standards organizations greatly reduces one of the largest sources of organizational value, managerial flexibility. The magnitude of the reduction of this flexibility is not obvious until one realizes that both task-sequencing and task-assignment1 gives rise to an explosion in the amount of production options available to managers. Interorganizational process can be viewed as containing two problems 1) a task-sequencing component and
2) a task assignment component. The two components can be viewed together as in figure 2.