Target-oriented prototyping in highly iterative product development T. Gartzen Laboratory for Machine Tools and Production Engineering (WZL) of RWTH Aachen University, Steinbachstraße 19, 52074 Aachen, Germany
C. Reuter Laboratory for Machine Tools and Production Engineering (WZL) of RWTH Aachen University, Steinbachstraße 19, 52074 Aachen, Germany
F. Brambring Laboratory for Machine Tools and Production Engineering (WZL) of RWTH Aachen University, Steinbachstraße 19, 52074 Aachen, Germany
F. Basse Laboratory for Machine Tools and Production Engineering (WZL) of RWTH Aachen University, Steinbachstraße 19, 52074 Aachen, Germany
[email protected]
Abstract For the development of radical product innovations agile processes are essential. These are characterized by a highly iterative production of prototypes. The paper presents a systematic method of physical prototyping in order to gradually reduce the high degree of market and technological uncertainty associated with radical innovations. Keywords: Physical Product Development, Agile Product Development, Highly Iterative Prototyping
INTRODUCTION Short product life cycles and continually increasing customer requirements characterize today’s markets. The majority of established manufacturing companies have difficulties to launch radically innovative hardware products at high speed - an essential ability to succeed in this dynamic environment (McDermott and O'conner, 2002; O'Conner and Rice, 2001). The dominant approach in physical product development, the sequential Stage-Gate process model, is overstrained with the realization of radical product innovation (Schuh et al. 2015). Companies must rethink and restructure their processes to meet the requirements for a successful development of radical products (Meyer 2012). Based on these thoughts, in this paper, we present a prototyping approach suitable for a highly iterative product development processes.
1
The Characteristics of Radical Innovations Radical innovations are characterized by a high degree of both market and technological uncertainty (Lynn and Akgün, 1998). This fact makes it harder to control them compared to evolutionary and incremental innovations with their medium and low degree of uncertainties, respectively. The high degree of market uncertainty of radical innovations results from a lack of clarity about the target market and the market potential. Customer requirements are unknown and the willingness to pay cannot be easily estimated (Lynn and Akgün, 1998; Walcher 2007). Technological uncertainty can be classified in terms of product and process. Concerning the product, a lack of information regarding technical specifications and technical feasibility exists for radical innovations. Furthermore, technical challenges are often unclear and the cost of development is difficult to predict. Regarding the process, a variety of alternative production processes and unknown production costs are characteristic for radical innovations (Lang 2013; Lynn and Akgün, 1998). In Figure 1 a) the three types of innovation mentioned above are positioned in a graph according to their specific level of uncertainty in the dimensions market, technology (product) and technology (process).
Development Processes as a Function of Uncertainty In his article What’s Next?: After Stage‐Gate Cooper (2014) states that “one size should not fit all”: The product development process should be adjusted to the specific degree of uncertainty. In the case of incremental innovations which are based on existing market knowledge and technological know-how, it is appropriate to make use of approaches which systemize the process of innovation. Therefore, Cooper developed the Stage-Gate model, which structures the innovation process into separate phases (Cooper 1990). A further reduction of the time to market and the development costs could achieved by a parallelization and forward displacement of activities, as with Concurrent Engineering (Shina 1993) and Front-Loading (Fujimoto and Thomke, 2000). In the case of radical innovations, however, a learning-oriented approach is recommended (Herstatt and Verworn, 2007). Contrary to the clear hierarchical structures that are commonly used to manage daily operations, a high degree of agility and flexibility is necessary to give development teams more creative freedom (Meyer 2012). Being open to new ideas, as well as working target-oriented rather than process-oriented, are the basic enablers for successful market innovations (Meyer 2012). Concerning evolutionary innovations a mixture of agile and sequential processes is suggested. Figure 1 b) visualizes the suggested types of development processes for the specific types of innovation.
2
Technological uncertainty (Product)
Radical
Evolutionary Produkt Inkrement
…
Incremental Produkt Inkremen t
Market Uncertainty
Technological Uncertainty (Process)
Figure 1 – a) Position of radical, evolutionary and incremental innovations according to their level of uncertainty (left); b) Visualization of the suggested development processes for the three types of innovation (right)
Prototyping in Agile Process Structures (Scrum) In the development of software, sequential approaches have been identified as too bureaucratic and “heavyweight” due to an extensive documentation and strict project role divisions (Gao and Rusu, 2015). For this reason, the popularity of agile development processes has increased significantly: A comprehensive assessment of industrial surveys of agile software development shows that the usage of agile methods worldwide is reported at about 55% (Stavru 2014). Among various agile frameworks, Scrum is the most commonly used one (Stavru 2014). It structures the work in a development project along so called Sprints. These Sprints are of fixed duration and take place one after the other. Every Sprint starts with a Sprint Planning in which a cross-functional team selects the top-prioritizes items of the Product Backlog. The Product Backlog is a list of items containing short descriptions of all functionalities desired in the final product. The items can range from specifications and requirements, to use cases, epics, user stories, or even bugs and chores. The team commits to complete a specific amount of items by the end of the Sprint. During the Sprint, the chosen items and the overall aim do not change. At the end of it, the Product Increment, the sum of all the Product Backlog items completed, is reviewed with all stakeholder interested in the Sprint Review. The feedback obtained can be incorporated into the next Sprint, again in the form of items. (Sutherland 2013) Scrum emphasizes a functional product at the end of the Sprint that meets a predefined Definition of Done. Within the context of software, this means a code that is integrated, fully tested and potentially shippable (Sutherland 2013). Physical product development, however, is different from software development. As software consists of lines of code which can be arbitrarily broken down into software increments, it is almost infinitely divisible. The development of a physical product cannot be incrementalized this way, and thus the notion of short time-boxed Sprints has to be adjusted to the characteristics of physical developments (Cooper 2014). In physical development with Scrum the deliverable of each Sprint should be something physical that can be demonstrated (Cooper 2014): A prototype. Corresponding to the 3
development of functional code in software development, the rapid implementation and testing of prototypes enables an agile development of physical products. With the help of prototypes the development team can learn and incorporate obtained information into the next development cycles. The potential customers and stakeholders can be integrated already in the early stages into the development process, in order to provide important information about their needs and requirements. Furthermore, prototypes represent milestones and act as a channel of communication between the various stakeholders in the development process. (Ulrich and Eppinger, 2012). Figure 2 shows the Scrum framework in the context of the development of radically innovative products of physical nature. The high uncertainties in all dimensions can be reduced in steps by the realization of prototypes (blue dots) in various Sprints (arrows from dots to dots) in order to arrive at a point of minimum uncertainty (green dot) where the series production should begin.
Technological uncertainty (Product)
Product Backlog
Sprint Planning
Daily Scrum
Sprint
Market Uncertainty
„Inspect and adapt!“ Product Increment
Technological Uncertainty (Process)
Sprint Review
(Prototype)
Figure 2 – The Scrum framework in the context of physical product development
CHALLENGE As shown in Figure 2 by the stepwise approach of a point of minimum market and technological uncertainty, the development process is a process of reducing uncertainty (Burmeister and Rüggeberg, 2008). The construction and testing of prototypes within the Sprints is a key enabler in continuously reducing the high degree of uncertainty of radical innovations in a process of learning. However, approaches of selecting the most suitable prototype in the complex environment of physical products are mostly unsystematic. Prototyping has always been an established part of the product development cycle, but the underlying motivation is rarely explicitly formulated and therefore not taken into account when selecting prototypes (Kohler et al. 2013). With the prototype being at the center of the agile development process as a means of interaction with the customer, this can be seen as a major obstacle to the implementation of this approach.
4
In the Scrum framework, the motivation for a specific prototype results from the top items to be implemented within the next Sprint as derived from the prioritized Product Backlog. The prioritization of the Product Backlog items and the selection of the top items should thus control the development and the prototyping strategy. In this context, the question of how to prioritize the items of the Product Backlog emerges. Different prioritizations of items lead to a different design of prototypes. The potential differences of the design of the second prototype (framed) based on a different prioritization and selection of items is portrayed in Figure 3. The effect is a difference in the remaining degree of uncertainty in the three dimensions. The result of the varying prototyping strategies are numerous possible paths along which prototypes are supposed to lead to the final target of a minimum amount of uncertainty remaining. Technological uncertainty (Product)
? Sprint Planning
Product Backlog
Prioritization and selection of top items 1.
5.
2.
6.
3.
7.
4.
…
?
?
?
Market Uncertainty
Technological Uncertainty (Process)
Figure 3 – Numerous possible prioritizations and selections of the top items to be implemented within the Sprints give rise to various development paths
With the following methodical approach, we would like to propose a model of prioritizing the items of the Product Backlog in order to follow the most efficient development path in the development of radical products. Furthermore, a model describing the characteristics of prototypes is presented. Together, these models are supposed to define the most suitable prototypes within the Sprints of an agile product development process. This forms the foundation of a methodology aiming at a gradual reduction the high degree of market and technological uncertainty associated with radical physical innovations.
METHODICAL APPROACH Prioritization of the Product Backlog We propose that the comparably high degree of complexity in the development of physical products caused by the high degree of uncertainty in the three dimensions makes a more analytical approach necessary to prioritize and select the top items compared to the experiencebased approaches in the software industry (in software, technological uncertainty concerning the process is basically absent). For a methodical and uncertainty-reducing prioritization the basis is 5
a comparison between the systematically derived benefit (maximum output in terms of reduction of uncertainty) and effort (minimum input in terms of invest resources) connected to the potential implementation of each item. At the beginning of the development, in the case of radical innovations there is a high degree of uncertainty in all three dimensions mentioned above. A systematic prioritization of the Product Backlog items at a given point in time is suggested to be based on their specific ratio between potential benefit and effort. The items that show the highest ratio are top priority and will thus be implemented in the upcoming Sprint. In our approach, the benefit of an item at a certain point depends on the one hand on the specific amount of uncertainty reduction in the three dimensions through the item’s realization. On the other hand, the reduction of uncertainty in a specific dimension can be currently of more interest than the reduction of uncertainty in the other two: For each dimension there is an estimation of the benefit generated by the reduction of its uncertainty. In this context, we suggest a target-oriented prioritization approach that addresses the three dimensions of uncertainties in specific sequences: Intuitively, as a first step, the development team should focus on the market to analyze its potential and needs to ensure that the product meets the market requirements. At this stage the principle of "Fail early and cheaply" applies – the aim should be the generation of the highest possible output information in form of the reduction of market uncertainty with the least possible resources. Often, cost-efficient prototypes can be used to integrate the customer in the development process to uncover their latent needs. The market information obtained serves as a framework for the further prototype development, which should focus on the reduction of the technological uncertainty in terms of the product. The customer needs must be systematically translated into technical specifications and be technically realized. Furthermore, the functionality of the product technology needs to be understood. In this step, the development and testing of prototypes can provide the development team with important information about the final product. Late prototypes should focus on reducing the technological uncertainty in terms designing the most efficient and economic manufacturing process for the series product. Figure 4 shows the sequential reduction of uncertainties suggested in our approach. Technological uncertainty (Product)
Benefit Market Uncertainty
Technological Uncertainty
Product
1.
Technological Uncertainty
Process
Uncertainty 100%
3.
t
Market Uncertainty
2. Technological Uncertainty (Process)
Figure 4 – The sequential reduction of uncertainties in the three dimensions helps to assess the benefit of the items of the Product Backlog
The effort of the implementation of an item has to be considered for a target-oriented prioritization, as well. Depending on the company’s capabilities, the implementation of an item needs a specific amount of effort which needs to be assessed individually. The analysis of benefit 6
and effort for every item is the basis for a systematic prioritization of the Product Backlog. For each Sprint the information on the potential amount of reduction of uncertainty in the three dimensions, the importance of the three dimensions according to their sequence as proposed above and the required effort has to be updated.
Characterization and Design of Valid Prototypes As mentioned above, the prioritization and selection of the items of the Sprint Backlog forms the motivation for the prototype to be developed in the upcoming Sprint. There is a need for a generic descriptive model of prototypes, in order to ensure the selection of a prototype design compatible with the prioritized items. There are different characterizations of prototypes in literature. While Gebhardt (1996) makes a simple distinction of six different physical model types, Lim et al. (2008) suggests a multidimensional approach in order to characterize software prototypes in a generic way. Arising from the multidimensional approach, Kohler et al. (2013) present a filter-fidelity-model based on the context of human computer interaction products. With the target of adequately characterizing prototypes used in physical product development, we have adjusted this filter-fidelity-model. We propose to characterize a prototype along the dimensions appearance, functionality, interactivity and topology. With each dimension, productspecific attributes needs to be defined. Prototypes can be described by evaluating the degree of fidelity in each attribute. In order to select a valid prototype design for a Sprint, the motivation based on the selected items and the filter-fidelity-profile has to match (Kohler et al. 2013). Table 1 contrasts a full production model and a physical design prototype. While the full production model shows a high fidelity in all attributes, the prototype shows only a high fidelity in the attributes which are relevant according to its underlying motivation. The motivation for the prototype characterized in this example is to test and validate the appearance of the product to be developed. For this motivation, a valid prototype must have a high fidelity in the attributes that are connected to the dimension appearance. The prototype does not fulfill every function of the final product and the necessity to interact with the prototype is limited. The topological attributes have a high fidelity, since it is a fully physical prototype and not a virtual one. Hence, the presented filter-fidelity-profile is a reflection of the prototype’s characteristics.
7
Table 1 - A possible prototype profile according to items aiming at the testing and validation the product’s appearance Degree of Fidelity Appearance Dimension ο ο ο ο ο XX Color ο ο ο ο ο XX Shape ο ο ο ο ο XX Weight ο ο ο ο X X Hardness ο ο ο ο X X Haptics ο ο ο ο X X Sound ο ο ο ο X X Functionality Functional Depth ο ο ο ο X X Functional Width ο ο ο ο X X Interactivity Action Reaction Topology Physical Tangible
ο X
ο ο
X ο
ο ο
ο ο
X X
ο
ο
ο
ο
ο
XX
ο
ο
ο
ο
ο
XX
X Prototype X Full Production Model
CONCLUSION Incremental innovations have been successfully developed for many decades with the help of traditional sequential process structures. However, agile approaches that literature suggests for the realization of radical physical innovations still deserve thorough study. The approach presented in this paper is a contribution to existing research in this field. It serves as a basic framework for an efficient Sprint Planning within an adapted Scrum process. In Scrum, experiencable prototypes possess a crucial role. So far, however, prototyping often lacks a specific prototyping strategy in practice. Addressing this shortcoming, our approach follows a sequential reduction of the high uncertainties associated with radical innovations as a basis for a target-oriented prototyping strategy. This way, the most suitable prototypes for the various Sprints can be derived. In the Sprint Planning, the items of the Product Backlog with the highest priority are implemented in the upcoming Sprint. The prioritization follows the assignment a specific benefit/effort ratio to each item of the Product Backlog. The sequential addressing of the uncertainties in the dimensions of market and technological uncertainty guarantees that the prototyping strategy focusses on the most beneficial items. At the same time, the effort for the implementation of the items is estimated on the basis of the team’s capabilities. The motivation formed by the prioritized items is systematically translated into a filterfidelity-profile. The derived profile provides a structure for the design of the most suitable prototype for the upcoming Sprint. The generic character of the profile leaves enough degrees of 8
freedom to prevent a limitation of the team’s creativity in the search for solution.Future work must address various aspects of the models presented in this paper to permit their use in practice. For example, research is necessary in the context of the assessment of benefits connected to the items of the Product Backlog. On the one hand, an operationalization of the degree of uncertainty in the dimensions market and technology is necessary. The development of key performance indicators on the basis of aspects affecting the corresponding uncertainties is seen as suitable here. On the other hand, a method of assessing the amount of uncertainty reduction in the three dimensions through the implementation of an item remains to be investigated. Also, the number of items from the prioritized list selected for the upcoming Sprint must be determined based on the specific amount of time at disposal for the realization of the prototype. Further research remains in the identification of the correlation of the nature of items and the dimensions of the filter-fidelity-model in order to derive a valid fidelity-profile for the prototype.
ACKNOWLEDGEMENT The new approach of “Target-oriented Prototyping in Highly Iterative Product Development” is being examined by the Laboratory of Machine Tools and Production Engineering (WZL) within the publicly funded (German Research Foundation, DFG) University graduate training program “Interdisciplinary Ramp-Up” (Graduiertenkolleg Anlaufmanagement).
9
Bibliography Burmeister, K., H. Rüggeberg. 2008. Innovationsprozesse in kleinen und mittleren Unternehmen. Working Papers 41: 35. Cooper, R. G. 1990. Stage-Gate Systems A New Tool for Managing New Products. Business Horizons 33(3): 44-54. Cooper, R. G. 2014. What’s Next?: After Stage‐Gate. Research Technology Management 57: 20-31. Fujimoto, T., S. Thomke. 2000. The Effect of "Front-Loading" Problem-Solving on Product Development Performance. Journal of Product Innovation Management 2: 128-142. Gao, S., L. Rusu. 2015. Modern techniques for successful IT project management. Business Science Reference An Imprint of IGI Global, Hershey. Gebhardt, A. 1996. Rapid Prototyping. Hanser, München. Herstatt, C., B. Verworn. 2007. Management der frühen Innovationsphase. Gabler, Wiesbaden. Kohler, K., T. Hochreuter, S. Diefenbach, E. Lenz, M. Hassenzahl. 2013. Durch schnelles Scheitern zum Erfolg: Eine Frage des passenden Prototypen?. Usability Professionals 2013 1: 78-84. Lang, H. 2013. Forschungskooperationen zwischen Universitäten und Industrie: Kooperationsentscheidung und Performance Management. Gabler Verlag, Wiesbaden. Lim, Y.-K., E. Stolterman, J. Tenenberg. 2008. The anatomy of prototypes: Prototypes as filters, prototypes as manifestations of design ideas. ACM Transactions on Computer-Human Interaction 15(7): 1-27. Lynn, G. S., A. E. Akgün. 1998. Innovation Strategies Under Uncertainty: A Contingency Approach for New Product Development. Engineering Management Journal 10(3): 11-18. McDermott, C. M., G. C. O'Connor. 2002. Managing radical innovation: an overview of emergent strategy issues. Journal of product innovation management, 19(6): 424-438. Meyer, J.-U. 2012. Radikale Innovation. BusinessVillage, Göttingen. O'Connor, G. C., M. P. Rice. 2001. Opportunity Recognition and Breakthrough Innovation in Large Established Firms. California Management Review 43(2). Schuh, G., T. Gartzen, F. Basse, E. Schrey. 2015. Enabling radical innovation through highly iterative product expedition in ramp up and demonstration factories. Procedia CIRP. Shina, S. G. 1993. Successful Implementation of Concurrent Engineering Products and Processes. Van Nostrand Rheinhold, New York. Stavru, S. 2014. A critical examination of recent industrial surveys on agile method usage. Journal of Systems and Software 94: 87-97. Sutherland, J. 2013. The Scrum Handbook. Scrum, Inc., Cambridge. Ulrich, K. T., S. D. Eppinger. 2012. Product Design and Development. McGraw Hill Higher Education, Boston. Walcher, D. 2007. Der Ideenwettbewerb als Methode der aktiven Kundenintegration. Deutscher Universitätsverlag, Wiesbaden.
10
