The Business Process Design Space for exploring process redesign alternatives

2.1 Existing process redesign approaches

Process redesign refers to the intentional change of processes within and beyond organizational boundaries (Davenport and Short, 1990) and is an essential part of the BPM discipline (Dumas et al., 2018). Process redesign methods have been highlighted as a critical success factor (Rosemann and vom Brocke, 2015) and provide guidance in a step-wise manner (Mendling et al., 2020). Techniques assist in performing one or more method's activities (Brinkkemper, 1996) and are thus narrowly focused on a specific outcome. For instance, a technique might focus on the diagnose or redesign of a process, two consecutive phases in a process redesign project (Gross et al., 2019; Kettinger et al., 1997).

A wide range of process redesign methods and techniques exists (Brocke et al., 2020; Dumas et al., 2018; Gross et al., 2019). These redesign approaches can be characterized based on their nature (which can be analytical or creative), perspective (inward-looking or outward-looking) and ambition (transactional or transformational) (Dumas et al., 2018). Based on methods and techniques discussed in reviews and articles on process redesign and service design (Bettencourt et al., 2013; Bitner et al., 2008; Dumas et al., 2018, p. 306 ff; Frank et al., 2020; Lee et al., 2008; Malone et al., 1999; Rosemann, 2020; Vanwersch et al., 2015), we further distinguish seven underlying rationales of existing redesign approaches (Table 1). The identified redesign rationales are (1) problem-based, (2) imitation-based, (3) pattern-based, (4) interaction-based, (5) outcome-based, (6) customer-based and (7) alternative-based. While they are non-exclusive, we observe that one rationale typically predominates.

We describe the different rationales in more detail below.

1. Most prominently, problem-based process redesign approaches like lean management (Bortolotti and Romano, 2012) and six sigma (Kwak and Anbari, 2006) aim at continuously tweaking existing processes through the identification and elimination of process problems. These approaches are problem-based, in the sense that they aim at improving issues or bottlenecks (Van den Bergh et al., 2014; Grisold et al., 2019) and thus mainly investigate the problem space of processes. As an example, lean management classifies different types of waste to be avoided, while non-value-adding process activities should be eliminated.

2. Imitation-based approaches like benchmarking (König et al., 2019) and positive deviance (Setiawan and Sadiq, 2013) search for superior process designs with the aim to imitate (part of) the process' design, which causes this effective behavior.

3. Pattern-based approaches build on this imitation rationale by abstracting from various real-world examples to sketch out useful and proven solutions. Representatives of this group are best practices of process redesign (Reijers and Liman Mansar, 2005), explorative process design patterns (Rosemann, 2020), customer-centric design patterns (Frank et al., 2020) and the Rethink of Process (RePro) principles (Vanwersch et al., 2015). Business process reengineering (BPR) also belongs to this set of approaches. Besides its radical ambition and clean-slate approach, it builds on several principles (i.e. higher-order patterns) that have been found to improve business process work (e.g. by involving as few people as possible in the performance of a business process) (Hammer and Champy, 1993). Anti-patterns follow a similar rationale by abstracting from common mistakes (Koschmider et al., 2019).

4. Interaction-based approaches like 7FE (Jeston and Nelis, 2008) and NESTT (Rosemann, 2017) enable process redesign by providing guidance on the interaction of process redesign participants and artifacts, e.g. by using the walls of a room to visualize, organize and discover different viewpoints on a process in the NESTT method. They may also specify roles for the redesign project, as does 7FE with the specification of a redesign facilitator (Jeston and Nelis, 2008).

5. Outcome-based approaches like product-based design (Reijers et al., 2003) or the process model canvas (Koutsopoulos and Bider, 2018) follow an end-to-start logic. They predefine a desired process outcome and reverse-engineer an ideal process design for reaching this outcome.

6. Customer-based approaches like service blueprinting (Bitner et al., 2008) and the job-centric approach (Bettencourt et al., 2013) identify and analyze a process from the customer point of view. Compared to outcome-based approaches, customer-based approaches are more outward-looking, with the customer as the main point of reference.

7. Finally, alternative-based approaches like process grammar (Lee et al., 2008), the handbook of organizational processes (Malone et al., 1999) or the BPR framework (Reijers and Liman Mansar, 2005) aim to explore a wide range of possible process design alternatives without investigating prior process problems. Thus, they are opportunity-driven (Grisold et al., 2019). They build on the underlying assumption that some alternatives within this range outperform the status quo.

As indicated, a redesign method or technique can build on more than one of the previously described rationales. For instance, lean management is problem-based as well as customer-centric. Its intention to avoid non-value-adding activities from the customer point of view can yet be regarded as problem-based, which thus manifests the method's primary rationale.

2.2 Ontological and procedural aspects of process redesign

Existing process redesign approaches provide limited support during the actual creation of to-be process designs (Sharp and Mcdermoot, 2009, p. 323; Vanwersch et al., 2015; Zellner, 2011, p. 217), as we will further explore in this section. The creation of to-be process designs is a critical part of redesign endeavors and refers to the creation of one or more design alternatives. While the underlying rationales of existing redesign approaches offer different frames of reference for the redesign task, most redesign approaches do not explicate the exploration and creation of to-be processes (Sharp and Mcdermoot, 2009, p. 323). This often “happens in a black box” (Zellner, 2011, p. 217), i.e. it is left to the user to creatively generate to-be processes, before their evaluation and selection. The challenge of generating to-be processes can be viewed through two distinct yet related perspectives.

First, methods and techniques follow the underlying assumption that process redesign alternatives are available (i.e. more than one potential to-be process), but hardly develop an ontological perspective on what can be changed from a design point of view. For instance, lean management follows the idea to eliminate different types of waste (Bortolotti and Romano, 2012), but there are numerous design choices on what to change to avoid waste. The waste of motion, for example, can be countered by the integration (or exclusion) of resources, decisions, activities and events, or by adjusting the sequence or message flow, to name but a few alternatives. However, the method itself does not explicate these design choices nor the resulting alternatives. Rather, it depends on the user to find appropriate leverage points in response to an identified waste. Other approaches focus on specific design aspects of processes while neglecting others. Product-based design (Reijers et al., 2003), for instance, mainly focuses on the sequence flow, while the explorative process design patterns (Rosemann, 2020) center around the revenue generation through a process. Thus, what is missing is a holistic understanding of redesign choices, considering and placing equal emphasis on all design choices of a business process.

Second, many redesign approaches provide limited procedural guidance on how to derive possible process design alternatives. Sticking to the example of lean management, the approach provides a general structure for redesign projects but does not define a procedure to create new process designs in which a certain waste is eliminated. Some approaches provide more explicit procedural guidance but are limited in their scope. A product-based design, for instance, provides structured guidance for the identification of possible sequence flows, but no procedure for other aspects of a process. Other approaches explicitly depend on creativity for the identification of design alternatives, e.g. 7FE (Jeston and Nelis, 2008) or service blueprinting (Bitner et al., 2008). Offering an ontological perspective provides at least some procedural guidance, as iterating through explicated design elements can be used for the creation of design alternatives. However, procedural guidance does not necessarily come with an ontological perspective.

There are a few examples of more explicit guidance (Vanwersch et al., 2016). Within pattern-based approaches, a pattern offers a specific recommendation for action, e.g. Contact reduction or Task elimination (Reijers and Liman Mansar, 2005). However, pattern-based approaches often focus on specific aspects of a process, e.g. revenue-oriented value for the explorative process design patterns (Rosemann, 2020), and thereby offer only a limited ontological perspective. Their underlying design choices are also partially implicit and not explicit (Lee, 2014). For instance, the pattern contact reduction (Reijers and Liman Mansar, 2005) can be used to question the necessity of existing points of contacts, but not to explore other possible forms of interaction with the customer.

Alternative-based approaches might have the potential to respond to the previously stated ontological and procedural challenges of existing process redesign approaches, as they aim for a wide exploration of process design alternatives. Explicit procedural guidance on the derivation of redesign alternatives is provided by the process grammar technique, which describes valid structural possibilities (i.e. combinations of process elements) in a given domain (Lee et al., 2008). A predefined grammar can thereby be used to algorithmically generate possible alternatives under specified constraints (Lee et al., 2008). However, this approach does not explicate the design elements of business processes and thus leaves it to its user to specify ontological design choices. More ontological guidance is offered by the BPR framework. This framework explicates six linked elements that should be considered in BPR projects (Reijers and Liman Mansar, 2005). These elements are, for instance, the customers of the process and the products or services generated for the customer. While these elements offer a high-level ontological perspective, they may be too abstract for the creation of to-be process designs and do not explicate possible design manifestations. Table 1 gives an overview of prominent redesign approaches for each redesign rationale and summarizes their core idea, implications for redesign, together with an assessment of their ontological and procedural guidance.

2.3 Toward a design space and design dimensions

The ontological challenge of process redesign can be approached using the design space concept. A design space is a “space of possibilities” (Maclean et al., 1991, p. 203) in which design ideas are created and considered (van Amstel et al., 2016). It is a conceptual model that contains design alternatives (Dove et al., 2016). To provide structure, an artifact and potential design alternatives are thereby placed along design dimensions (Maclean et al., 1991). Design dimensions comprise characteristics (i.e. design choices) by which possible artifact manifestations differ. Thereby, they provide an ontological design perspective. A desk, for instance, has height, width and length as spatial design dimensions with any length specifications as characteristics. Other design dimensions are color (with characteristics such as white or blue), material (e.g. wood, glass) or shape of the tabletop (e.g. rectangular, round). Thus, design dimensions offer an angle to dynamically remove, add or create new artifact manifestations (Crilly and Cardoso, 2017).

The explication of and reflection on a design space increases the awareness of constraints that come with particular design choices and how these can be challenged (Dove et al., 2016). It can also make a wider range of design aspects explicit, which were previously disregarded (Dove et al., 2016). This is particularly useful for less experienced designers (Dove et al., 2016). A representation of a design space thereby does not (and cannot) capture all design aspects, but provides useful guidance during the design process (Crilly and Cardoso, 2017). As such, the design space concept has been applied in various contexts, e.g. for the design of microservice decision logics (Haselbock et al., 2018), for the design of visualization tasks (Schulz et al., 2013), for the design of deep learning approaches for entity matching (Mudgal et al., 2018) and for structuring process architectures (Lapouchnian et al., 2015).

The procedural aspect of generating and evaluating design alternatives is called exploration (Navinchandra, 1991, p. 67). Investigating the design space supports the design process by defining design choices, which in turn enable the creation of new artifacts (Maclean et al., 1991; Mose Biskjaer et al., 2017). For this purpose, Maclean et al. (1991) introduced Design Space Analysis, which assists the exploration of the design space through specific questions, options and criteria. Questions are meant to introduce design dimensions and to structure and explore possible alternatives in the form of options (Maclean et al., 1991). Design options (also referred to as characteristics) offer possible alternative answers to questions to choose from (Maclean et al., 1991). Criteria support the evaluation of alternatives (Maclean et al., 1991), e.g. according to time, flexibility, quality or costs (Reijers and Liman Mansar, 2005). A design space schema is a graphical representation of a design space that displays design dimensions and possible characteristics (Crilly and Cardoso, 2017).

While the previously mentioned approaches like Design Space Analysis (Maclean et al., 1991) and the process grammar (Lee et al., 2008) build on the design space concept for determining alternative process designs, they do not explicate the design dimensions of processes. Thus, the identification and selection of design dimensions in process redesign projects leave the involved project members to their own devices. This is especially challenging for persons with limited redesign experience (Dove et al., 2016). Other approaches claim to have broadened the design space, without having defined its dimensions (Rosemann, 2020; Wurm et al., 2019). The BPR framework (Reijers and Liman Mansar, 2005) offers high-level design dimensions, but can be too abstract for covering the BPD-Space adequately and does not come with possible design manifestations (i.e. characteristics) and guiding questions. This research aims at identifying common process design dimensions together with potential characteristics to support its exploration. Thus, it aims at providing explicit ontological guidance during the to-be process design creation to identify and explore possible redesign alternatives. It thereby contributes to the group of alternative-based redesign approaches, which are opportunity-driven in nature.

3.1 Taxonomy development

For the derivation of process design dimensions within the BPD-Space, we use the similarities between the design space concept and taxonomies. A taxonomy provides a systematic structure and organization to the body of knowledge in a certain field (Glass and Vessey, 1995; Nickerson et al., 2013). This structure builds on a set of dimensions with underlying characteristics (Nickerson et al., 2013). A taxonomy does not describe objects of interest in full detail but provides explanations that are useful about the object's nature (Nickerson et al., 2013). Similarly, process design dimensions aim to structure and organize the design space of process redesign alternatives. Process design dimensions are meant to be a useful (and not necessarily complete) collection for the purpose of exploring redesign alternatives. The design choices associated with design dimensions should be exhaustive, i.e. every process has a certain design manifestation in each dimension. This property aligns with the definition of collective exhaustion of taxonomy characteristics (Nickerson et al., 2013). To emphasize this aspect and to align with the terminology of the taxonomy development method, we will refer to the design choices within a design dimension as (potential) characteristics of a dimension.

We use these conceptual similarities to systematically derive the BPD-Space by following the method for taxonomy development by Nickerson et al. (2013). This method starts by determining a meta-characteristic that functions as the basis for the later selection of taxonomy dimensions and characteristics. The selection of this meta-characteristic is based on the purpose of the taxonomy and its intended use. Then, the taxonomy development method suggests proceeding iteratively by deriving dimensions and characteristics through a conceptual-to-empirical (deductive) or an empirical-to-conceptual (inductive) approach until the predefined ending conditions are met.

3.2 Definition of meta-characteristic and ending conditions

Starting with the taxonomy development process, we first defined the meta-characteristic. In line with our research question, our meta-characteristic is defined as changeable design elements of business processes that support the creation of redesign alternatives for business processes. Therefore, we used the internal (linked) framework elements of the BPR framework (Reijers and Liman Mansar, 2005). The BPR framework is well established in the process redesign domain. It consists of six elements to be considered in process redesign attempts and thus offers an initial set of ontological guidance on a high level of abstraction (Table 1). For this reason, the BPR framework is a suitable foundation for the purpose of this paper. Its elements cover process redesign aspects that involve the customer (internal or external) of the business process, the product/service that is generated for the customer and the business process itself (the implemented workflow and how it is executed) that generates this product/service (Reijers and Liman Mansar, 2005). The business process is linked to three more elements, namely, the participants in the business process (roles, users, groups, etc. and agents assigned to work), information used or created by the business process and technology used by the process (Reijers and Liman Mansar, 2005).

Following the taxonomy development method, it is important to determine the ending conditions that stop the development cycle. Nickerson et al. (2013) differentiate between objective and subjective ending conditions. For this research, we decided to include the following ending conditions, which will be checked together with the subjective ending conditions after each iteration. First, after the last iteration, no new dimension is identified. Second, there are no dimensions that are split or merged in the last iteration. Third, all dimensions and characteristics are unique, meaning that there exist no duplicates. To create the taxonomy, we conducted three iterations. Details are presented in Appendix 2; an overview of each iteration is provided in sections 3.3 and 3.4.

3.3 Identification of design dimensions – conceptual-to-empirical

First, we identified process design dimensions deductively. We started with the conceptual-to-empirical approach by building on the current state of literature. Traditionally, process redesign was grounded in manufacturing logic (Miles, 2008), which is production-oriented. BPM expanded this production-focused view to a more holistic understanding of processes in the organizational context. Services, on the other hand, emphasize the value co-creation between a service provider and service clients (Spohrer et al., 2007). To account for these diverse ontological perspectives on process redesign, we conducted a literature review and included relevant scientific articles from the service and process redesign domain. We included artifacts (frameworks, models, etc.) that are used to describe, analyze or communicate the design of services and processes. Hence, these artifacts provide design elements of processes that are relevant for process redesign.

We considered the following artifacts. From the process literature, we include the BPR framework (Reijers and Liman Mansar, 2005), the work-centered analysis (Alter, 1999) and work system framework (Alter, 2008, 2009), process meta-models (Dumas et al., 2018, p. 7; Laguna and Marklund, 2013, p. 9), the process modeling framework (Melao and Pidd, 2000), the process profile (Dumas et al., 2018, p. 52; Wagner and Patzak, 2015, p. 393) and the process canvas (Koutsopoulos and Bider, 2018). From the service domain, we included the service innovation capability framework (den Hertog et al., 2010), the digital product and service innovation framework (Nylén and Holmström, 2015), the service design planning model (Goldstein et al., 2002) and service blueprinting (Bitner et al., 2008). While we do not claim for completeness, we believe to have selected relevant literature integrating a rich view on process (and service) redesign. The taxonomy development method will complement this view through the next iterations.

We derived process design dimensions from the literature as follows. We extracted all design elements from the previously mentioned artifacts. For each design element, we identified real-life processes, which deviated from other existing processes with the same purpose through a specific manifestation of this design element. When at least two such processes were found, we kept the design element as a design dimension and investigated relevant literature for its characteristics. For instance, we identified the design dimension Customer channel, which we derived from Bitner et al. (2008), den Hertog et al. (2010) and Nylén and Holmström (2015). We then found two real-life processes, which differed from existing processes in this dimension. First, we found Webcam Social Shopper from Zugara, which enables customers to try out different clothing products virtually through an augmented reality environment (Bonetti et al., 2018). It thereby extends the traditional website-based customer channel of e-commerce. Second, we identified DocOnline, which is a digital health-care service provider in India. It offers online access to physicians through different modes of interaction, e.g. through (text, audio or video) live chat (DocOnline Health India Ltd, n.d.; Kulshrestha, 2017). It extends the traditional health-care access, which is usually bound to direct physical interaction. Details on the first iteration can be found in Appendix, Table A2.

3.4 Identification of design dimensions – empirical-to-conceptual

In the next step, we enriched our conceptual understanding with empirical insights. As the first objective ending condition (no new dimension has been identified) has not been met, we conducted a second iteration through the taxonomy development method. For this iteration, we went through the empirical-to-conceptual approach and inductively derived process design dimensions. Therefore, we included academic experts. We held a workshop at the end of 2019 with BPM researchers who were not involved in this research project. Five researchers participated, out of which one was a visiting associate professor and four were PhD candidates. We included PhD candidates as academic experts as all studied information systems and did their research in the field of BPM. The researchers thereby covered a wide spectrum of process-related topics and thus provided a rich academic view on business processes. In the course of the workshop, we introduced the concepts of BPD-Space and design dimensions, followed by the hitherto derived dimensions and examples from the first iteration. Based on the discussion of the existing design dimensions, we consolidated, renamed, split or deleted some dimensions. The participants also gave real-life process redesign examples, which did not fit the initial dimensions. For this reason, we added new dimensions.

We again went through the empirical-to-conceptual approach by building on the knowledge of professional process experts who are currently involved in or responsible for process redesign projects. In total, we conducted six semi-structured interviews (in one more iteration of the empirical-to-conceptual approach) at the end of 2019 with process experts from six different organizations (Appendix 1). We chose the interviewees following expert sampling (Bhattacherjee, 2012), i.e. the experts were purposefully selected from our industry network, based on their knowledge and expertise in the redesign of processes. To counter potential biases, we selected experts from various industries and different organizational sizes.

Qualitative interviews are well suited to generate rich data in explorative research, as the respondents' context is been taken into account (Schultze and Avital, 2011). All interviews were semi-structured (Myers and Newman, 2007) and consisted of four parts. In the first part of the interview, we introduced the research objective and the current state of the process design dimensions and characteristics. In the second part, we asked the experts to comment on the usefulness, understandability and completeness of the design dimensions. In the third part, we discussed the allocation of the design dimensions to the layers. In the last part, we encouraged participants to think about real-life process redesigns that were not yet covered by the existing design dimensions. By following this procedure, we accounted for the main criteria of model evaluation (Sonnenberg and Vom Brocke, 2012). After the last interview did not add new design dimensions, nor change the allocation of dimensions to layers, we felt confident that all ending conditions of the taxonomy development have been met.

3.5 Evaluation of the artifact

We evaluated the resulting BPD-Space in two phases. In the first phase, we used the previously gained information from the semi-structured expert interviews for a formative artificial evaluation, as the artifact was still under development (Venable et al., 2016). The interviewees thereby commented on the usefulness, understandability and completeness of the artifact's current state (Sonnenberg and Vom Brocke, 2012). We used this evaluation to further improve the artifact. Section 5.1 lists a summary of the interviews regarding the evaluation criteria and the resulting implications.

In the second phase, we validated the applicability and usefulness of the BPD-Space (Sonnenberg and Vom Brocke, 2012) by performing a summative naturalistic evaluation (Venable et al., 2016) through three real-world applications. Thereby, we conducted three half-day workshops in which the design space has been applied together with professionals from different case organizations. The workshops aimed to find alternative process designs for one pre-selected organizational process per organization. We deliberately chose organizations that differed in their context (i.e. industry, size and resources) (Brocke et al., 2016). Details on context, procedure and results of the application can be found in section 5.2. The resulting evaluation regarding applicability and usefulness is available in section 5.3.

4.1 Overall structure of the Business Process Design Space

In this section, we lay out the derived process design dimensions, which collectively form the BPD-Space. By applying the taxonomy development method, we derived 19 dimensions. Figure 1 summarizes these dimensions and their allocation to six layers that reflect the linked elements of the BPR framework (Reijers and Liman Mansar, 2005). In the following, we define each layer and its design dimension, list potential characteristics, guiding questions and an illustrative example, which describes how a specific characteristic manifestation has changed the process design in real-world settings. While the dimensions are supposed to be stable over time, characteristics may be added or changed for some dimensions (e.g. through the introduction of new technology). For this reason, the listed characteristics are meant to be illustrative and not exhaustive.

4.2 Dimensions of the Business Process Design Space

The Customer layer comprises dimensions that directly address customer needs and interactions. The layer includes four design dimensions, namely, Customer segment, Customer experience, Customer value and Customer channel, which can be found in Table 2. Customer segments categorize customers with respect to different criteria. Customer experience refers to “the internal and subjective response customers have to any direct or indirect contact with a company” (Meyer and Schwager, 2007, p. 2). Customer value refers to the positive outcome of a process or service (Dumas et al., 2018). Value can also be generated through a specific experience (Bitner et al., 2008; Knutson and Beck, 2004); thus, these two dimensions are inter-related. Whereas experience is part of every interaction (Jain et al., 2017), value refers to the perceived benefit of the customer (Goldstein et al., 2002); thus, it is not necessarily delivered to or perceived by the customer. Customers interact with the process through a customer channel (Bitner et al., 2008).

The product/service layer comprises dimensions that further define what the process offers. It includes the dimensions Scope, Flow unit, Location and Temporality (Table 3). Process Scope refers to the end-to-end perspective of the process and is mainly informed by its boundaries (Maddern et al., 2014). A Flow unit has been defined as a “transient entity that proceeds through various activities and finally exits the process as finished output” (Laguna and Marklund, 2013, p. 5). Location is a key environmental context variable of processes and captures a position and its topographical information (Zhu et al., 2014). Temporality refers to the time used to specify the availability and termination of the service or to specialize its execution (e.g. through specific offers).

The Business process layer comprises dimensions that further define the operational logic of the process. It comprises three design dimensions referring to Coordination, Trigger and Outcome (Table 4). Process Coordination describes how actors, interdependent activities, goals and resources structure processes (Crowston, 1997) and thus describes how processes are set up and coordinated to meet different needs. A process Trigger is an event that starts the process execution (Dumas et al., 2018). While a channel is the medium of interaction, a trigger is a certain event communicated through this channel and detected by the recipient, which is followed by certain actions. Process outcomes refer to all legitimate ends of process instances (Koutsopoulos and Bider, 2018) and can be positive or negative (Dumas et al., 2018). While positive outcomes deliver value, negative outcomes do not (Koutsopoulos and Bider, 2018).

The Organization layer comprises dimensions that define the strategic organization of the process. It includes the design dimensions Objectives, Internal participants, Revenue model and Business partners (Table 5). Objectives are derived from the organization's vision and strategy and can be refined by performance measures (Dumas et al., 2018). Objectives depend on each other by either complementing or conflicting with each other. Process participants refer to anyone who is directly or indirectly involved in the execution of the process (Alter, 2013). Internal participants contribute to value creation as they bring in different competencies and skills and take over responsibilities in process work. The Revenue model refers to the business logic to turn customer value into revenues. Digitalization leads to new opportunities to gain revenue, e.g. through platform ecosystems (Parker et al., 2016). According to Anthony et al. (2019), digitalization enables organizations to open up additional sources for revenue that exist side-by-side the business core. Finally, Business partners are external partners that are involved in the process but are not subordinated to the organization (Dumas et al., 2018). For example, organizations outsource certain activities that are carried out by partners (e.g. shipping products).

The Information layer comprises dimensions that focus on the integration and use of information (Table 6). It includes the design dimension Information source, which refers to the origin of data integrated into processes. The second dimension, Information usage, captures how data can contribute to (new) value creation for the customer or to the business value of the organization.

The Technology layer focusses on dimensions that address the use of software and hardware to support the process. In light of the advent of sophisticated digital technologies, this layer has been receiving increasing attention (Breuker et al., 2016). As technology is an enabler for many dimensions within the design space (e.g. new customer channels), we introduce two design dimensions, namely, Infrastructure and Automation (Table 7). On the one hand, technologies can be used to provide Infrastructures enabling the execution of the process. On the other hand, technology can be used for Automation purposes (van der Aalst et al., 2018), i.e. “hardware/software configurations that perform totally automated activities” (Alter, 2013). In this context, specific activity sequences or parts of the process are being automated by technologies.

5.1 Insight from the expert interviews

Through the expert interviews, we gained the following insights. Regarding usefulness, all practitioners deemed the BPD-Space as a valuable artifact during process redesign projects. One expert specifically mentioned that the approach is “very helpful in the course of actual consulting projects.” Two of the six process experts pointed out that it is not clear which dimension to focus on during redesign projects in their respective organizations, and that the derived dimensions may be too detailed for top management considerations. Two process experts who consult external clients mentioned that the dimensions to focus on in actual consulting projects are determined by the scope of the project, which should be defined before applying the BPD-Space. Some dimensions were also deemed as rather strategic management considerations, which might not be part of every process redesign project. One expert mentioned that the dimensions to consider in redesign projects mainly depend on “the strategic importance of the redesign project within the organization and the people involved in the project.” The experts recommended that additional guidance on the application of the BPD-Space would be useful. We responded to this comment by offering application guidelines in section 6.2.

The professional process experts also evaluated the understandability and completeness of the approach. Regarding understandability, one expert proposed to change the representation of the dimensions, such that layers, dimensions and characteristics are visible in one summarizing representation. We included this consideration for our representation of the design space (Figure 1). All experts commented that the underlying idea behind the design space is understandable. One expert found the given redesign examples valuable for understanding the design dimensions. Regarding completeness, three experts proposed the introduction of new characteristics. This feedback has been incorporated into the resulting design space. Overall, the BPD-Space and its underlying idea have been found to be a novel and stimulating approach for process redesign.

5.2 Real-world applications

We conducted three real-world applications. The corresponding organizations were a financial technology startup, a large-scale medical diagnostic organization and a mid-sized transmission system operator. All organizations are headquartered in Central Europe. This set of organizations allowed us to evaluate the design space in a variety of industry contexts. We started all workshops by explaining the research context, the underlying idea behind the design space concept and the BPD-Space. Consecutively, we presented the design dimensions. One at a time, we introduced the respective dimension and applied it to the process of the case organization. The participants explored how the respective design dimension is currently represented. Subsequently, they identified other potential characteristics (i.e. design manifestations of the design dimension in focus) and used them for the creation of to-be process ideas. The first design dimension was in the customer layer, and we further proceeded by considering the participant's interest until the allocated time for the idea creation was over. In what follows below, we describe the case organizations, the process in focus and the resulting redesign ideas. Appendix 3 offers a list of additional redesign ideas that were created throughout the three workshops.

5.3 Applicability and usefulness

As for applicability, practitioners from all real-world applications found the BPD-Space suitable for the generation of redesign alternatives. Workshop participants from baningo noted that once an alternative characteristic manifestation of a design dimension has been identified, it can be challenging to move on to another potential characteristic. Furthermore, we observed that generated redesign ideas sometimes affect more than one design dimension. For instance, a new information source may also be used as a trigger. We also found that we can use design dimensions of a previously generated redesign idea to jump between the design dimensions, instead of iterating through all dimensions consecutively. The guiding questions and potential characteristics were perceived as particularly helpful to understand the direction and intention of a design dimension. Labnetic and baningo both reported that these are important to understand and apply the dimensions in a given context. In general, the practitioners found the design dimensions to be more suitable for generating redesign alternatives compared to process model representations, as it abstracts from “irrelevant implementation details” and does not require process modeling knowledge (Labnetic).

As for usefulness, baningo found the BPD-Space “very supportive” during the creation of process redesign alternatives, as it “supported the structured analysis of the process through different perspectives.” Participants from the first two workshops considered the majority of generated redesign alternatives useful and feasible for future changes to the process. This indicates that the BPD-Space helps create adequate redesign alternatives. ELECTRO noted that the BPD-Space provides a “good guideline to completely screen a process in various respects,” especially for less prominent processes in the organization. The practitioners also noted that it was challenging to find entirely new process designs for the new grid connection process. This might be due to the specific context of ELECTRO. As a monopolist, it needs to account for a high number of regulations that limit the number of alternatives within the design space. In a similar vein, participants from baningo noted that some constraints have been disregarded during the creation of redesign alternatives, which may make some redesign ideas less likely to be implemented in the near future. In general, the BPD-Space has been perceived to trigger the creation of more ambitious ideas by thinking into directions that would have been disregarded in regular brainstorming sessions. In this regard, it helps as a “mental liberation” (baningo) to break out of normal thought patterns and constraints. Labnetic noted that the design dimensions helped to shift the focus away from technology-centered redesign ideas, indicating that most of their redesign initiatives usually focus on technology-related dimensions, but tend to exclude most other dimensions.

6.1 Implications for research and practice

In the following, we will discuss the implications that follow from this research. First and foremost, we provide comprehensive ontological guidance during the to-be process creation. While existing approaches provide important reference points to look at processes (e.g. Reijers and Mensa, 2005), we zoom into the underlying dimensions and characteristics. Thereby, we direct designers' attention to various aspects that are part of a specific process but might not be explicitly considered. This can be the case, for example, when specific components of a process are taken for granted and are not questioned during redesign activities. Hence, the BPD-Space can be used to explore, question and rethink a process in various respects. This work reflects the first attempt to organize and synthesize the various dimensions and underlying characteristics involved in process redesign.

Exploring the BPD-Space also provides initial procedural guidance for the creation of process redesign alternatives. Existing redesign methods provide structure for the overall redesign project, but have hardly shown to provide directions during the to-be process creation. Exploring the BPD-Space, e.g. in a workshop setting, also comes with initial procedural guidance for the creation of alternative to-be process designs. Building on the idea of design space analysis (Maclean et al., 1991), users of the design space can iterate through its dimensions and use the guiding questions to analyze current and alternative characteristic manifestations. By combining different manifestations, alternative process designs can thereby be generated. However, while our application guidelines support the procedural perspective, a method for providing full procedural guidance is yet to be developed.

It is important to note that the BPD-Space is not competing with but complementing existing process redesign approaches. Existing methods and techniques use different rationales for the redesign endeavor (Table 1). The BPD-Space can be a valuable means to create to-be processes during the redesign phase of other redesign methods, e.g. in the course of interaction-based approaches like NESTT (Rosemann, 2017). It can also be applied in conjunction with other techniques of process redesign, e.g. with pattern-based approaches like the explorative process design patterns (Rosemann, 2020). The BPD-Space itself can be classified into the group of alternative-based redesign approaches, as it aims to identify and explore a wide range of process design alternatives. It thereby builds on and extends the BPR framework (Reijers and Liman Mansar, 2005), while it also provides a language for applying the process grammar (Lee et al., 2008).

The BPD-Space abstracts from specific process flows and representations of processes. It thereby enables an unconstrained exploration of various alternative process designs. Users of the process dimensions are not influenced (or limited) in their thinking by detailed process descriptions, as in form of process models. Research on nudging, and digital nudging in specific, has shown that human thinking and human choice is significantly influenced by the way information about choice is presented (Weinmann et al., 2016). By abstracting from the inner flow and details of a process but highlighting the design dimensions instead, it stands to reason that process designers are nudged more toward thinking ways of conducting a process that is influenced less by the current practice of conducting a process, so in other words, to find more innovative ways of conducting a process.

The BPD-Space can be used to realize explorative BPM. Following recent claims that BPM activities should become more ambidextrous (Grisold et al., 2019; Rosemann, 2020), we argue that the proposed BPD-Space can serve as a starting point for exploration, i.e. to identify new opportunities for organizations. Here, it is important to highlight the distinction between design dimensions and characteristics. The design dimensions we have identified are meant to be stable over time. However, new characteristics may emerge and change, e.g. due to new developments and innovations. Thereby, design dimensions can serve as a means to scan and categorize dynamics in an organization's environment and to translate them into concrete implications for processes (Grisold et al., 2019). This holds especially for the technology layer, where continuous research and development lead to new technological features.

Finally, the BPD-Space allows exploring design alternatives for different types of processes. Research on the role of context has found that processes differ regarding various characteristics, such as variability, repetitiveness or uncertainty (Brocke et al., 2016). While most of the process design research has focused on well-structured and repetitive processes, the need for process design in organizations, however, is rather in the non-repetitive, knowledge-intensive and creative processes. In this regard, the BPD-Space provides a means to explore design alternatives thinking out of the box across all processes types, which is highly needed in practice.

6.2 Guidelines for application

The three real-world applications of the BPD-Space demonstrated its potential for the exploration of process redesign alternatives. To support future users, we provide the following guidelines for applying the BPD-Space based on our insights from the formative and summative evaluation:

1. Ahead of iterating through and applying the design dimensions, determine which dimensions are especially useful in the context of the application (e.g. based on the goal of the redesign endeavor) and focus on these. Context-based constraints might also cause the exclusion of design dimensions.

2. Predefine specific time slots (e.g. 20 min for each design dimension) to ensure a wide coverage of design dimensions and to prevent idea fixation. Make adequate breaks to ensure mental focus and creativity.

3. When applying the BPD-Space, first allow for and encourage an unconstraint creation of redesign ideas before prioritizing these according to weighted evaluation criteria (e.g. feasibility, viability and desirability).

4. In the course of a workshop, provide a visual representation of the BPD-Space to navigate through the workshop and to provide a shared understanding and language. Allocating the layers to different walls of the room can further support the interaction during a workshop.

5. To prevent a potential anchoring effect around the first redesign idea of a design dimension, aim to apply more than one characteristic within one design dimension.

6. To prevent a potential anchoring effect and to adapt the BPD-Space to the context of its application, explore other potential characteristics (i.e. design manifestations of dimensions) for each design dimension, which are not yet mentioned in the BPD-Space.

7. For each generated redesign idea, note which design dimensions are affected to detect if a wide range of the design space is covered (or which design dimensions are underrepresented).

8. In a workshop setting, break out into sub-groups when generating redesign ideas to cover a wider range of ideas, if group size allows.

6.3 Limitations and future research

Our research comes with limitations. First, design dimensions do not aim for completeness but have to be useful for an intended purpose and audience, similar to dimensions in taxonomies (Nickerson et al., 2013). Although the derived BPD-Space has been evaluated as useful by professional process experts and in the course of three real-world applications, there might be application scenarios in which additional design dimensions could be helpful. In line with the extendibility requirement of taxonomy development (Nickerson et al., 2013), the design space should be continuously challenged and further developed by future applications and research. Second, the derived design dimensions are implementation-independent, i.e. rather abstract design considerations. The actual implementation (on a process model level) of a specific manifestation of design dimensions can take various forms and is again a creative task. We find this abstraction useful to detach from process model-specific constraints and biases. However, the interrelation of design dimensions and process models is not yet researched.

We see two main avenues for future research. On the one hand, descriptive research is required to explore how and under which circumstances the use of design dimensions can lead to innovative design outcomes. For example, it should be investigated if there are contexts where the use of these dimensions is more valuable than in others or if there are dominant strategies to combine or reconfigure specific dimensions. On the other hand, despite our application guidelines, prescriptive research should evaluate and propose how these dimensions should be used in redesign projects. While the BPD-Space provides a comprehensive ontological perspective on the creation of process redesign alternatives, it only offers moderate procedural guidance. Therefore, it should be investigated how a method can systematically guide through the BPD-Space and which dimensions should be emphasized taking into consideration the context of the redesign project. This also includes guidance on how to integrate the BPD-Space into existing process redesign approaches. Additionally, evaluation criteria should be defined to prioritize the high number of resulting redesign alternatives (Nygren et al., 2017). Overall, we call for future research in the area of systematic creation of process design alternatives to foster improvement and innovation of processes.