Key points are not available for this paper at this time.
It has been 4.5 years since the establishment of the Design Science (DS) department in the Journal of Operations Management (JOM). Even with the recent name change to the department of Intervention-based Research (IBR), its mission remains as discussed in the initial essay (Van Aken, Chandrasekaran, Hevner, March, Park, Wieringa, 2014; Baskerville, Baiyere, Gregor, Hevner, Rai, 2017). Solutions to problems of information systems, for example, often involve a design artifact such as an algorithm or process model—a man-made solution that can then be made available more widely. Designed artifacts usually make a practical rather than a theoretical contribution. Given JOM's emphasis on theoretical contributions, we discovered that most DS articles were not a good fit with the journal. As we reflected on these published articles, we realized that the element of their contributions that benefited from being incubated in the review process was that which stemmed from the active involvement of the researchers in deploying operations management (OM) theories and tools, along the lines described by Simon (1996) in distinguishing between natural sciences and the sciences of the artificial. We thus made the decision to change the department's name to IBR to reflect this evolution in thinking at the editorial level. The change was announced during the 2019 Academy of Management meetings. Oliva (2019) drew on Checkland's (1985) approach to provide direction in the use of intervention as an OM research method. Interventions are defined as the use of a method (M) to apply a basket of theories (T) to bring a problem situation (S) to S*. If the intervention succeeds in moving the problem situation from S to S*, then the theory has been confirmed. However, a successful move from S to S* often requires adjustments to either T or M. If the adjustments are to M—as applies to a traditional DS contribution—what is learned tends to be practical in nature and is likely to be atheoretical. If, however, the required adjustment concerns T, the resulting insights about T may produce a theoretical contribution. We have also witnessed occasions where what emerges is that S*—where everyone had agreed to head—turns out to be the wrong place to go. Answering the question “Why did we end up in the wrong place?” also offers interesting theoretical insights. Adding guidance to calls that researchers step outside their ivory tower (e.g., Van de Ven Van Mieghem, 2013), IBR guides how to formalize the kind of surprise that such mixing with practice is able to provide and how to form that surprise into a theoretical contribution. To illustrate our points, we reflect on a few of the DS papers published in JOM and explain what was interesting from these works when observed through the lens of IBR. Consider the study by Groop et al. (2017) that set out with the objective to improve the delivery of home healthcare to the aging population in Finland, with S* being an expected improvement in service for given resources. The problem was framed along the lines of the well-known traveling salesperson optimization that seeks to minimize the travel time of caregivers (e.g., home health nurses). The authors were able to convince the review team of the reasonableness of this formulation. The relevant theory was translated into methods that were implemented, forming the intervention. Rather than bringing the situation to the expected S*, the intervention instead reduced system performance. This unexpected result triggered abductive logic as researchers and decision makers worked together to figure out what had gone wrong. They realized that improving the efficiency of caregiver travel only improved system performance if the freed-up capacity was available when capacity was actually needed. However, the traveling salesperson formulation increased resource availability at a time when it was not needed. The research team working with the decision makers from the home health agency reformulated the problem as an inventory problem and developed a second intervention that then improved system performance and was rolled out to other cities in Finland. Because of the plausibility of the traveling salesperson formulation, a failed intervention was needed to unlock the theory-reframing exercise. Interventions can also result in the refinement of an initial basket of theories, as exemplified by Akkermans et al. (2019). These authors studied the change in service firm performance when output controls were added to outsourcing contracts. The theory that formed the basis for the intervention predicted that supplier performance would improve (see Ouchi 1979), but—contrary to expectations—it declined. The abductive logic triggered by the unexpected results led to further exploration, which revealed that supplier performance depends on buyer as well as supplier performance in a service-outsourcing relationship. Adding conventional output controls made it more difficult for firms to develop proper mechanisms to manage these relationships. The authors then explored whether this unexpected finding also held in other service-outsourcing relationships and discovered that it did. The researchers then collaborated with a group of service firms to develop a contracting mechanism that added key performance indicators for the buyer to those specified for the supplier. The updated method, referred to as “collaborative service contracting,” was then tested and refined by being applied to a second service provider. This intervention thus led both to a refinement of the theory basket and of the method, and so, it meets the criteria for IBR and DS research, with the impact being the expected S to S* improvement. Kuntz and Van Wassenhove (2019) investigated how changes in fleet sizes affected the ability of relief organizations to support their missions. The authors used data from the fleet management unit of the Office of the United Nations High Commissioner for Refugees to show that both centralizing fleet sizing at the fleet management level and decentralizing it at the country office level resulted in inefficiencies (too many or too few vehicles across county locations). A proposed intervention based on traditional vehicle management approaches was seen to be inappropriate for this context, triggering a reevaluation of the applicable theory basket. The authors worked together with the Fleet Management Unit to develop simple prediction models, discovering that three independent variables sufficed to provide a highly accurate estimate of fleet sizes. The unsuccessful application of traditional theory—that control would be either centralized or decentralized—also led the research team to recognize the need to facilitate communication between country offices and the Fleet Management Unit. The realization that the common assumption—that an efficient fleet size will stem from either centralized or decentralized decision-making—is incorrect, and the resulting model—that combines prediction with facilitated communication—are now being rolled out to hundreds of country offices. In contrast to the above examples, Brusset and Bertrand (2018) demonstrated a traditional DS approach in their development of weather-based financial instruments to improve supply chain coordination. Along the lines described in the 2016 essay by Van Aken et al., these authors extended a tool used in finance for the supply chain context to solve a problem in one chain. After the initial implementation was successful, they applied the tool in two other kinds of supply chains. In addition to describing the designed tool, the manuscript also deepens understanding of the role of weather hedging in supply chain management. Although the first three papers (Groop et al., 2007; Akkermans et al., 2019; Kunz & Van Wassenhove, 2019) were accepted by the DS department editors, they diverged from the traditional DS view of creating and testing an artifact as exemplified by Brusset and Bertrand (2018). Rather, all three papers described research that began from a failed attempt to apply theory to a problem, resulting in a new understanding of the problem or an expansion of the set of theories used to solve the problem. The department remains receptive to the traditional DS approach (e.g., Ilk et al., 2020). The feedback process that results from confronting a theory with reality has now been shown to produce knowledge that may well meet the requirements for publication in JOM, so we expect to see more papers being published in JOM that use the IBR approach than DS. The contribution of IBR is both distinguished and generated by the shift to abductive reasoning that interventions may put into motion when they do not go as planned. Much inductive and deductive logic builds on heuristics that define the world and make a given problem tractable. IBR comes into its own when that initial set of assumptions—often unquestioned—does not lead to an appropriate problem formulation. IBR projects may not start as such but may rather take on their IBR identity when the problem being studied goes from seemingly well understood to ill defined (as described by Simon (1996)). Well-defined problems can be addressed without the need for an active framing exercise. As unexpected events reveal the ill-defined nature of the problem, the shift to abduction is accompanied by reframing activity. The switch to abductive logic facilitates a recalibration of the theory basket, ideally bringing the problem back to being well understood, albeit in a new way. In Table 2, we position IBR in the ecosystem of common research methods organized according to the dominant mode of reasoning used and the typical ordering of the research cycle. Modes of reasoning include deductive, abductive, and inductive (Van de Ven, 2007). Deductive reasoning typically starts off with a logical theory or hypothesis that is then examined to reach a specific conclusion. Inductive reasoning, instead, starts off broadly from a specific observation, which is then analyzed for patterns resulting in propositions and theoretical generalizations. Abductive reasoning starts with an observation similar to inductive reasoning but then uses preexisting explanations similar to deductive reasoning to group observations for the best possible explanation. While all three modes of reasoning may exist in all forms of OM research, one is more likely to dominate the others depending on the research cycle. In Table 2, we also three common research seen in research approaches from the problem but in the of forming a finding and insights. The most common approach for OM first in Table concerns the where the problem leads to which then guides the of which then leads to insights. This referred to as traditional and data Table some of research using this along with the of knowledge whether it is or and (1996) insights. After inventory and in the researchers information and optimization to create an approach to estimate demand and support decision-making for This was tested using data from a The implementation of the theory the expected and insights. Researchers developed from theory and used supply chain from firms to how supply chain affected They did not the design of these and insights. Researchers about process that they then tested using data from an implementation in on factors that affected was Groop et al. (2017) Researchers with the stakeholders from the homecare systems to the design and develop a different approach to their initial application of theories to et al. & risk and communication help and manage insights. Researchers observed that projects in organizations between and This research began from observation of a potentially interesting and the objective of the research was more to this than to solve a problem. and (2017) their operational and customer by a insights. Researchers used natural on from and studied customer with other forms of data (e.g., to derive insights on and & theory insights. Researchers and other data on and their that they used to develop insights about the supplier collaboration practices. The research in the of Table starts by an problem or with from the which then leads to theory development and insights. We that this approach is used in research in theory Table also some of this approach from the OM literature. The of Table a research that is common in OM the problem leads to an between theory and IBR comes into its own when to apply existing theory to a problem are by are made based on the understanding that has from T is using to move S to S*. This is However, the unexpected at instead of S* abductive This abductive logic to The unexpected at for the step on the knowledge this unexpected and the need for that it have all too often led to the research being as it not fit the research of our objectives in IBR research at JOM is to OM not to be by the unexpected to but rather to whether to abductive logic would create an opportunity to new The of different research methods in Table is not to be but rather to a to our understanding of how these research methods fit is some between the methods and the reasoning that is not in Table As an example, traditional research uses deductive reasoning, by abductive reasoning when the are not by learning comes in from the insights from the analysis. It may that a from the application of accepted the process of T applied to change would results to a of results have been difficult to publish in an the research and data were not out at the of the The of IBR at JOM the to such projects as as the theoretical contribution is for interesting research, and that call has the of management As we what emerges from it has come to our that one of the of this approach is to create an that allows the of research that are interesting in of the interesting propositions examined were to involve the between and between the of and the of interesting was always the of an accepted interesting may also as or of accepted The intervention in IBR from an accepted providing the opportunity for its or through between the have about the on the one and the have about the on the which began in a with an of the and by of an with its own to a that will be interesting by as well as experts with the of this the one will experts only if it the of their the other will only if it a of the since the of experts is a of a of will that which experts it their will not it their and What will be interesting to the one will be to the We with the that IBR research will lead to interesting theories. We researchers to go into organizations and theories into either or the that abductive logic and insights. of the of theories from implementation or works are likely to fit well in such as that our understanding and provide may provide for a publication in a theory and we that this editorial will authors in this The authors feedback from Ketokivi, and on of this
Chandrasekaran et al. (Fri,) studied this question.