Corner stones of NPD flexibility and their impact on NPD performance – a contingency approach.

Corner stones of NPD flexibility and their impact

on NPD performance – a contingency approach.

supervised by dr. J.D. (Hans) van der Bij*, co-assessed by dr. K.J. (Killian) McCarthy* *Department of Innovation Management & Strategy, University of Groningen, The Netherlands

__________________________________________________________________________________ ABSTRACT

Today, organizations face increasing technological complexity and fast changing consumer demands due to the globalization of economy. Phased models, in which new product development processes (NPDs) were considered as a predictable series of steps that can be planned and implemented in advance, can barely cope with those pressing challenges in many settings. Therefore, literature puts more and more focus on flexible approaches for NPD projects. Even though past literature found a variety of factors to define project flexibility, a holistic and systematic analysis is missing. Biazzo (2009) attempts to fill this gap by categorizing NPD flexibility across three dimensions: (organizational, informational and temporal) in order to explain their impact to NPD success under certain contingencies. However, the paper remains conceptual and no empirical testing of those relationship was conducted. This paper aims to fill this gap by drawing on a survey of 120 firms on the individual project level in the northern Netherlands using the theory of contingencies, information processing, and organizational learning. Four studies within the scope of the dimensions and an overall consideration show partly surprising findings on two levels of NPD performance categories. This approach questions the current understanding of product flexibility and adds new knowledge and research opportunities to the body of literature.

__________________________________________________________________________________

TABLE OF CONTENTS

1.

INTRODUCTION ... 4

2.

THEORETICAL BACKGROUND ... 6

2.1. Project flexibility ... 6

2.2. Contingency Theory, Environmental Uncertainty and Modularity ... 7

2.3. Information Processing Theory ... 8

2.4. Organizational Learning Theory ... 9

3.

STRUCUTRE OF THE PAPER ... 10

4.

FIRST STUDY: ORGANIZATIONAL DIMENSION ... 10

4.1. Theory ... 10 4.1.1. Introduction ... 10 4.1.2. Structuration ... 10 4.1.3. Formalization ... 11 4.1.4. Centralization... 12 4.2. Methodology ... 14

4.2.1. Data collection and sample description ... 14

4.2.2. Measurements ... 15

4.3. Analysis and results ... 17

4.3.1. Descriptive Statistics and Correlations ... 17

4.3.2. Factor and Reliability analysis ... 17

4.3.3. Regression Results ... 18

4.4. Dimension Discussion ... 20

5.

SECOND STUDY: INFORMATIONAL DIMENSION ... 21

5.1. Theory ... 21

5.1.1. Introduction ... 21

5.1.2. Late Freeze of product definition ... 22

5.1.3. Number of design iterations (change of at least 10% of product) ... 23

5.1.5. Customer feedback causes change in product definition ... 25

5.2. Methodology ... 25

5.2.1. Measurements ... 25

5.3. Analysis and results ... 26

5.4. Dimensional Discussion ... 27

6.

THIRD STUDY: TEMPORAL DIMENSION ... 29

6.1. Theory ... 29

6.1.1. Introduction ... 29

6.1.2. Task-Overlap ... 29

6.1.3. Gate-Conditionality... 30

6.1.4. Time between milestones ... 31

6.2. Methodology ... 31

6.2.1. Measurements ... 31

6.3. Analysis and results ... 32

7.

OVERALL DISCUSSION ... 34

8.

CONCLUSION ... 35

9.

References ... 37

10.

Appendix ... 42

10.2.1. Factor Analysis independent variables ... 43

10.2.2. Factor Analysis dependent variable... 44

10.3. Reliability Analysis ... 45

1. INTRODUCTION

responsiveness. They call for flexible iterations through system specification, detailed component design and system testing.

Despite the growing interests of research flexibility in NPD processes the phenomenon still remains under-researched (Mishra & Mishra, 2019; Kettunen et al., 2015; Fantazy & Salem, 2016; Buganza et al., 2010). Current research fails to adequately address the relationship between flexibility and performance of NPD projects due to the lack of using hard performance metrics and grounding theory (Eisenhardt & Tabrizi, 1995; Cooper & Sommer, 2016). In his conceptual paper Biazzo (2009) calls for a contingent approach for the design of NPD processes since a normative perspective could lead towards decontextualized best practices. He points out the contrasting views regarding the identification of success factors for NPD processes in dynamic environments. With his study, Biazzo (2009) aims to provide clarity regarding the dichotomy between Stage-Gate and flexible models by conceptualizing flexibility parameters across three dimensions. Thereby, he broadens the understanding by considering new factors to distinguish flexible projects from inflexible projects like product design freeze, frequent design iterations, concurrent engineering and task overlap for NPD flexibility. This thesis adds to the literature of NPD flexibility by providing empirical evidence for the proposed relationships of Biazzo (2009) and partly of Eisenhardt & Tabrizi (1995) in regard of NPD performance. In doing so, it provides implications for the dichotomy between Stage-Gate and flexible NPD approaches. From a managerial perspective, effective methods can be derived based on the tested relationships between the parameters and NPD success.

The structure of the thesis unfolds as follows: First, a theoretical background is provided to give an overview of the existing stream of research of product flexibility, contingencies and underlying theory that guides the line of reasoning. Second, the paper is divided into four sub-studies based on the three dimensions of flexibility by Biazzo (2009). In each dimension, variables are derived and their expected impact on NPD success is discussed based on information processing theory, organizational learning theory and contingency theory. Afterwards, constructs are developed by taking into account existing research and empirically tested within each dimension. Finally, they are combined into an overall model. Third, after a brief discussion of the findings in each dimension, a more comprehensive discussion including all dimensions is done and a conclusion is presented. By taking into account certain contingencies significant positive relationships were found with NPD performance for centralization, structuration, postponing design freeze and parallel product designs. Moreover, mixed findings are found for the impact of customer feedback in product design and a negative relationship was found for gate-conditionality under certain contingencies. In contrast to the applied theory, a direct negative effect of task-overlap on NPD performance was found. The following two research question guide the study:

I. Which parameters define the relationship between NPD flexibility and NPD success?

2. THEORETICAL BACKGROUND

2.1. Project flexibility

Within the stream of modern NPD literature, Biazzo (2009) questions the use of decontextualized best practices and calls for a contingent approach for deciding on an NPD-process. He stresses the importance of external environment in order to find the most suitable process. By doing so, he distinguishes between the phased Stage-Gate process and a flexible process based on the ideas of Iansiti (1995). He conceptualized the Stage-Gate as a highly structured and sequential process that aims to anticipate external change by a clear separation of problem-formulating and problem-solving activities as well as an early and sharp product definition. In contrast, flexibles processes rely on rather organic organizational structures, characterized by overlapping or concurrent tasks, and a product definition that remains fluid until the late phases of the process to constantly react on environmental change. In a similar vein Eisenhardt & Tabrizi (1995) make the distinction between the compression and experiential strategy to explain how to speed up NPD processes. The compression strategy holds that product development is a certain and well-understood process with a series of predictable steps that can be squeezed together or compressed and is therefore similar to Biazzo’s Stage-Gate process. In contrast, the experiential strategy describes that product development is a rather uncertain and non-linear path in an uncertain and changing environment which requires rather flexible approaches with limited structure. They state that “this approach is thus more a response to uncertainty than certainty, more iterative than linear, and more experienced-based than planned” (Eisenhardt & Tabrizi, 1995, p. 88). This would assume that product innovation is a crucial adaptive process. Biazzo states that due to the inaccurate description of the phenomenon of project flexibility, contradictory results arose. Therefore, he aims to develop several constructs within a framework to analyze NPD processes. In order to distinguish flexible and structured processes, Biazzo introduces three dimensions of flexibility: the organizational dimension which deals with the formal segmentation and progression of the process, the informational dimension which classifies activities into problem formulation and problem solving and finally the temporal dimension which contains the execution strategies of development tasks.

hybrid model should bring most benefits since both models work best under certain circumstances. However, they also criticize the lack of performance metrics, especially harder metrics like productivity improvements, time-to-market and success rates of products in recent project flexibility research.

2.2. Contingency Theory, Environmental Uncertainty and Modularity

Biazzo (2009) also uses contingency theory to explain that only under specific conditions flexibility leads to project success. In a similar vein Eisenhardt & Tabrizi (1995) stress out that both compression and experiential strategies can accelerate product development, but in different ways and under different conditions. Contingency theory argues that good management will have a different character based on situational variables. Usually a distinction is made between environmental contingencies which refers mainly to the stability of the firm’s environment and internal contingencies, such as size of the firm or R&D budgets. In this study, market and technological uncertainty are used as the environmental contingency whereas modularity is treated as an internal contingency.

Biazzo also describes NPD processes as means of uncertainty reduction. According to Milliken (1987), environmental uncertainty is an individual's perceived inability to predict something accurately. Therefore, uncertainty refers to the inability to predict the consequence of a response and thus limits the ability for decision-making. When uncertainty is high this may lead to overstraining R&D team member’s information-processing abilities when facing complex and unpredictable tasks. Biazzo takes a view on NPD processes as a set of activities to reduce uncertainty in regards of market needs and technological choices about the product in his consideration about dynamic environments. Therefore, he considers two kinds of uncertainty: market uncertainty which relates to the difficulty of understanding and translating customer needs into functional product characteristics and technological uncertainty which refers to the degree of novelty in the knowledge stock. Biazzo highlights that by continuous problem solving and improvements in the knowledge stock, uncertainty successively decreases during the development process while at the same time costs for unforeseen reworks become more distinct.

changes lead to a snowball effect that initiates comprehensive compensating changes for all other related components. By doing so Biazzo (2009) makes the distinction between iteration-based flexibility (one single product design changing over time) and duplication-based flexibility (maintaining several product designs at the same time). Several case studies suggest that iteration-based flexibility in a low modularity context may be incompatible with overlapping execution strategies due to slower uncertainty resolution regarding product specifications and due to the comprehensive compensating changes described earlier in high flexibility contexts. Also, modularity reinforces the effect of probing due to enabling the demonstration of critical elements of a product even if some component modules are not entirely completed.

2.3. Information Processing Theory

units and thus fewer factors have to be considered when confronted by an unexpected event. This will save information processing efforts but at the cost of increasing the resource stock which could mean increasing the lead time and in the worst case delayed time to market.

Galbraith (1974) stresses out that an organization should apply the least costly strategy in regard to the environmental context. In support for this, Tatikonda & Rosenthal (2000) found empirical evidence for a positive impact of resource flexibility on project execution success.

2.4. Organizational Learning Theory

3. STRUCUTRE OF THE PAPER

A statistical pre-test showed that combining all factors considered by Biazzo (2009) into an overall single model does lead to insignificant results. This is not very surprising because he states that flexibility is not a one-dimensional construct. Therefore, flexibility is considered across three different dimensions by Biazzo (2009) which will be the point for further analysis in this study. Despite their intersections with each other, processes can be flexible in one dimension while entailing more Stage-Gate characteristics in another dimension. Thus, conflicting results arose when combined into one model. In order to take this into account and improve clarity and logic of this paper, it is divided into three sub-studies based on the flexibility dimensions. In the organizational dimension structuration of the process, centralization of decision-making and formalization will be considered. In the informational dimension the late freeze of the product definition, frequent design iteration, parallel product designs and design changes induced by customer feedback are discussed. For the temporal dimension the overlap of tasks, number of milestones, time between milestones and gate-conditionality are reviewed. In each dimension, the contingencies of market uncertainty, technological uncertainty and modularity were considered as moderators. Finally, the findings of all three sub-studies are combined and discussed within an overall model.

4. FIRST STUDY: ORGANIZATIONAL DIMENSION

4.1. Theory

4.1.1. Introduction

Biazzo (2009) explains that process structuration consists of two factors. The first one deals with the formal segmentation of the temporal progression in consecutive intervals (stages) and predefined decision-making points (gates). The second one refers to how precisely the activities are defined in each stage and therefore captures the degree of the microlevel of activities.

4.1.2. Structuration

Taking the view of information processing theory, highly structured projects containing stages with pre-defined activities to be done provide a possibility to canalize information processing efforts of team members towards the given goal. This may help team members to focus on their given tasks and spare cognitive resources by operating within a limited set of factors to be considered. In a similar way, dividing projects into fixed stages and gates provides a channel for targeted knowledge exchange. Since gates only bring together team members affected by the decisions to be made, interpersonal information processing is reduced to a minimum. However, this should be only effective in settings characterized by low levels of uncertainty where activities can be pre-planned and the goal is clear from the beginning. If in contrast uncertainty is high, activities can’t be pre-planned since unforeseen environmental events would cause frequent changes in the plan of action which would overload the information processing capabilities of the decision makers at some point. In such situations Biazzo (2009) recommends setting targets and goals. If this mechanism is combined with providing team members slack resources, responsibility is shifted to the level of the team member by giving them enough freedom to decide how to accomplish the given goals. As long as goal accomplishment proceeds on plan the necessity of continuous communication among subunits is eliminated and information processing is reduced to the individual level. This should also limit exception making and thus relieve decision makers at higher levels. Applying modularity in the development process should significantly reduce information processing efforts since systems’ components are combinable. This reduces complexity in decision making since fewer factors need to be considered for the individual decision maker but should also make mutual adjustment between decision-makers redundant.

Taking the perspective of organizational learning theory, gates provide a regular opportunity for systematic knowledge exchange. This gives each team member the chance to distribute relevant knowledge and discuss progress which in turn stimulates interpretation of the shared information and thus mutual understanding. However, this would assume that only little unforeseen events would occur during task execution and the gates would be a sufficient channel for knowledge exchange and learning within the team. If, however the degree of turbulence in the environment is that high that it makes constant knowledge exchange necessary, structuring a process may hamper organizational learning. Assigning pre-defined tasks restricts learning of the individual since deviating from the given path will cause some sort of punishment. Thus, learning through try and error, which is vital for uncertainty reduction, is hampered. Also, major environmental changes could make spontaneous possibilities for comprehensive knowledge exchange necessary and waiting to do that until a gate takes place could cause disastrous consequences.

4.1.3. Formalization

highly organic processes neglect the importance of structure and focus in uncertain situations. In a similar vein Sheremata (2000) considers formal structures and improvisation capabilities as mutual reinforcing factors with a positive effect on quality and quantity of ideas and knowledge stock. Jansen et al. (2006) found a positive effect of formalization on exploitative innovation which suggests a complementary relationship when uncertainty is low. They argue that formalization enhances existing routines in a predictable environment and codification of new knowledge into best practices. They also proposed a negative relationship on exploratory innovation (high uncertainty) since reliance on rules and procedures hampers experimentation and ad-hoc problem solving but found only insignificant results. Tatikonda & Rosenthal (2000) also found a positive impact of formalization on project execution success. However, Tsai (2002) found a negative effect of formal structures on knowledge sharing and argues that informal social interaction as a main driver for knowledge exchange is hampered.

Taking the view of information processing theory, it is expected that formalization narrows the scope of the information to be processed since rules and procedures are already existent for certain decisions to be made. The logic here is that information is already processed at an earlier point and codified into some sort of best practice which in turn spares information processing at a later point for an indefinite number of individuals. In the case of predictable activities to be performed rules enable an efficient execution of tasks and avoid the need for inter-unit communication and thus information processing. However, as the frequency of unexpected events increases, exceptions have to be made for each deviation from the formal procedure. This could overload the decision maker and cause information overload of the project leader. Also, the scope of search for new information may narrowed so much that individual team members become unable to find, locate and process essential information.

Considering the organizational learning theory, formalization should enhance the organizational memory because individuals are encouraged to codify acquired knowledge and thus enhance mutual understanding of certain topics. Information is thereby distributed by rules and procedures and helps especially inexperienced team members to make sense of their environment. Explicit practices and procedures developed over time by conceptualizing activities to be performed make the knowledge available to them and thus improve their productivity. However, sticking to these rules and procedures discourages individuals to deviate from a given path and therefore hampers experimentation and improvising. Thus, learning during process becomes hardly possible. Combining the point of view of both theories, formalization should show a negative effect on performance when uncertainty is high.

4.1.4. Centralization

of subunit specialization, Jones (2013) considers formalization and centralization as most important organizational design challenges (p.120). In this study centralization is therefore considered together with structuration and formalization to pay particular attention towards the autonomy of team members in their decision making regarding their tasks. Prior research done by Jansen et al. (2006) showed that centralization has a negative effect on exploratory innovation processes. They argued that centralization narrows communication channels and reduces the quality and quantity of ideas and knowledge that come from problem solving. In the case of exploitative innovation they only found insignificant results. Tsai (2002) found that the level of centralization negatively influences intraorganizational knowledge sharing since it gives subordinates less discretion in dealing with their task environment and reduces initiatives for interunit knowledge exchange. Similarly, rational models of decision making are most effective when uncertainty is low whereas decentralized decision making is advisable in turbulent environments (Fredrickson, 1984).

From a view of information processing theory, centralization of decision-making shifts information processing efforts from many individuals at a subordinate level towards fewer individuals at hierarchically higher levels of the organization. Thus, the total amount of information to be processed is reduced since the effort is pooled at a higher level. However, when uncertainty increases, situations may arise for which there are no rules and hierarchy employed at an exception basis. Each exception in turn comes at the cost of information processing for the decision maker that could therefore become overloaded. This overload can arise due to bounded rationality of decision making which explains that individuals are cognitively bounded and can thus only process a limited amount of information and have a finite amount of time to do so. Also, individuals can only take a limited number of possible scenarios into account which makes them unsuitable for uncertainty resolution (Arthur, 1994). This should mean in turn that in turbulent environments groups of people are much more suitable for decision making whereas individual decision makers should be more appropriate in less turbulent environments. Also, since serious efforts of information processing are already done before by the subordinate, information processing efforts must be again done by the supervisor which would increase the total amount of information processing.

From an organizational learning perspective applying centralization should lead to conflicting results. On the one hand decision makers at higher levels of the organization should serve as some kind of information storage from which relevant information can be retrieved each time decision making takes place. In contrast to a decentralized organization where knowledge is distributed across many individuals in centralized organizations, this knowledge is much more concentrated and thus decision paths are much shorter and clearer.

environments, where a reduced number of individuals at the top of the organization have the ability to gauge the environment and make decisions. When uncertainty, however, increases, the mere number of decisions to be made would overload the decision maker. Giving team members the freedom to plan and exercise activities towards a given goal in the way they want, may also benefit double loop learning. By giving team members the possibility to not only solve problems, but also to re-examine them, they are able to tackle the problem at the root and avoid that it happens again. This enhances their ability for knowledge acquisition. Another mechanism to acquire vital information is performance monitoring. This line of reasoning would be also consistent with the reasoning of Tushman & Nadler (1978). They explain that organizations, where decision making is exercised at the team member level, are relatively independent of each individual and thus less prone to information overload.

4.2. Methodology

In this section first the data collection process, empirical sample description and the dependent variable will be discussed. Afterwards the organizational and control variables as well as the choice of the method to test the presumed relationships will be presented. This discussion will be only held in this dimension and is also valid for the informational and temporal dimension. This is followed by the analysis and discussion which is done separately for each dimension. As to be discussed in the analysis part at a later point, formalization did not lead to significant results and is therefore not included in the measurements part.

4.2.1. Data collection and sample description

4.2.2. Measurements

Dependent Variable

Project performance (related to expectations)

In this study performance is measured at the individual project level. The remarks of Griffin & Page (1996) were taken into account to develop measurements that are clearly separated from those at the program level. Therefore, respondents were asked in the instruction of the survey to keep a single NPD project in mind when answering the survey. Moreover, they were asked to only consider projects that were recently launched, the initial market performance is already known and that has been carried out within the firm (so not within an alliance) in order to limit the heterogeneity in the dataset and to guarantee measurability. According to Schleimer & Faems (2016) measures of NPD performance should include the characteristics of the developed product and the success of the product in the market. Therefore, performance in relation to quality, cost-efficiency, achieving customer satisfaction, providing value for customers and keeping current customers are suggested on a 7-point Likert scale. They rely on expectation fulfillment measures which refer to the extent to which initial expectations are met after the project. In a similar vein Ahmad et al. (2013) also used expectation fulfillment measures including market share, technical performance relative to specifications, return on investment, time to market, R&D budget, and overall commercial success on a 7-point Likert Scale. This study relies on adapted measures based on the previously describes papers measuring seven items on a 7-point Likert scale (1 = substantially worse; 2 = worse; 3 = slightly worse; 4 = about the same; 5 = slightly better; 6 = better; 7 = substantially better).

Independent variables

(Project / Process) – Structuration

Centralization

As a part of the organizational dimension, centralization of decision making is measured as suggested by Jansen et al. (2006) using five items on a 7-point Likert scale. One example question referred to extent of which small matters had to be referred to someone higher up for a final decision.

Technological Turbulence and Market Turbulence

Since Biazzo (2009) defines uncertainty as an important factor to explain the conditions under which project flexibility may lead to project success it is measured in two constructs. For market and technological turbulence the measurement of Jaworski & Kohli (1993) is adapted to this survey. On a 7-point Likert scale four items were used. One question asked whether technological changes provide big opportunities in the respondent’s industry.

Modularity

According to Biazzo (2009), modularity enables loosely coupled product creation units that can work autonomously and concurrently on a given goal. Therefore, it is considered as an important moderator, especially for the informational and temporal dimension. It was simply measured by asking whether the modularity was high or low on a 7-point Likert scale.

Control Variables

4.3. Analysis and results

4.3.1. Descriptive Statistics and Correlations

In table 1 descriptive statistics including the correlations matrix, means, standard deviations as well as minimum and maximum values are displayed. The combined Valid N was 109 out of the total data sample of 120 indicating missing values in the dataset. The mean of the project duration is 17.36 (months) with a standard deviation of 20.285 which implies a heterogeneity of projects in the dataset. The same counts for the number of employees in the firm with a mean of 2469.98 and standard deviation of 1312.08. Furthermore, the analyzed companies showed an above median value of 4.81 for customer cooperation. Also, the level of environmental turbulence was moderately high with a mean of 5.24 for technological and 4.69 for market turbulence. The mean of modularity was 4.66.

Several significant relationships were found between the regression variables. However, most of them were weak with the highest value of 3.47 (p<.01) between customer feedback and project duration as well as between technological turbulence and task overlap indicating a complementary relationship. No significant negative relationships were found.

4.3.2. Factor and Reliability analysis

In order to analyze the constructs in a structured way they were grouped together in the three dimensions suggested by Biazzo (2009). Structuration and Centralization are displayed in the organizational dimension, freezing of the product design, design iterations, parallel designs and customer feedback in the informational dimension as well as gate-conditionally and task-overlap in the temporal dimension. Multiple linear regression analysis was used to analyze the presumed relationships. As suggested by Raisch & Birkinshaw (2008) multiple performance dimensions were thereby used to analyze each construct separately. All variables were mean-centered in order to test interaction effects (Aiken et al., 1991). Moreover the variance inflation Factor (VIF) test was conducted in order to check for multicollinearity (O’brien, 2007). For all parameters VIF values were below 5 which indicates no serious multicollinearity problem (Belsleyet al., 1980). The highest value that was found in the overall model was 4.46 between customer feedback and market turbulence while most values remained below two. All three control variables were included in each dimension.

4.3.3. Regression Results

Results in table 2, 3, 4 and in the overall model show that the market performance dimension is insignificant in each model with R-square of 0.264 and F-statistic (1.266) in the organizational dimension; R-square of 0.554 and F-statistic (0.907) in the informational dimension; R-square of 0.129 and F-statistic (1.651) in the temporal dimension and R-square of 0.772 and F-statistic (0.747) in the overall model. Therefore, no results in this dimension will be considered in the following analysis.

Table 1: Descriptive Statistics: Variables, Means, Standard Deviations, Minimum values, Maximum values and Correlations

*p<.05; **p<.01

Table 2: Regression Results Organizational Dimension

Process Performance Product Performance Market Performance

Intercept 0 (3.21) 0 (0.284) 0 (0.285) Structuration -0.035 (0.097) 0.091 (0.086) 0.026 (0.088) Centralization -0.028 (0.091) 0.298 (0.081) -0.031 (0.082) Market Turbulence -0.194** (0.09) 0.097 (0.08) -0.036 (0.082) Modularity 0.012 (0.062) -0.049 (0.055) -0.005 (0.056 Centralization x Market Turbulence 0.128** (0.057) 0.018 (0.051) 0.067 (0.052) Structuration x Modularity 0.095** (0.046) 0.081** (0.041) 0.055 (0.042) Project Duration -0.021*** (0.006) 0.009* (0.005) 0.004 (0.005) Customer Cooperation 0.031 (0.061) 0.04 (0.054) 0.123** (0.055) Number of Employees -0.00002 (0) -0.00003 (0) 0.000005 (0) F-Value 3.448*** 1.580 1.266 R2 0.001 0.131 0.264 *p<.1; **p<.05; ***p<.01

4.4. Dimension Discussion

centralized decision-making might be a crucial factor.

As expected, structuration has a positive effect on process and product performance when modularity is applied. In the case of process performance, modules eliminate development efforts for certain parts while also boosting time-to-market since these released resources can be allocated to the remaining components. For product performance the reasoning is less straight-forward. Modularity and structuration could be mutually reinforcing control mechanisms to ensure product quality by avoiding development mistakes that affect the functionality of the product.

5. SECOND STUDY: INFORMATIONAL DIMENSION

5.1. Theory

5.1.1. Introduction

therefore open to design changes induced by unpredictable external influences. By taking their compression strategy as a basis, Eisenhardt & Tabrizi (1995) extend the understanding of both strategies by explaining that allocating enough time to the problem formulation (project planning) stage enhances the understanding and rationalization of team members in regards of the process. This may spare unnecessary steps, improve delegation of tasks to the most suitable team members, communication and thus mutual adjustment. Also, they propose that flexible projects are characterized by multiple design iterations that speed the product design by increasing the chances for a success and problem-solving capabilities, especially in highly uncertain environments. Also, iterations improve the building of understanding about the product and avoid attachment to one particular product design. They point out that these design iterations can be simultaneous, alternative, based on a previous design or some combination of those.

5.1.2. Late Freeze of product definition

reasonable level of information processing.

Modularization should also have a positive effect. When being exposed towards an environmental change, modules could serve as an option to react quickly to these changes and solve problems. This can be crucial to save time in the implementation phase of development that is shortened by the extended design definition phase. Assuming that modules are mutually compatible the only way of information processing is directed towards suitability to solve a certain design challenge. This would save serious efforts of inter-unit communication. From an organizational learning perspective keeping the design fluid should help the development team to build some form of routine and intuition in making sense of the environment. Since they are encouraged to constantly monitor their environment for potentially useful information, they become less prone towards shocks caused by major environmental changes. Combining both theories this should mean that a late freeze of the product definition should only lead towards higher performance if uncertainty is high.

5.1.3. Number of design iterations (change of at least 10% of product)

As another concept to achieve flexibility in the informational dimension Biazzo describes an approach where the product definition is only roughly defined in the beginning of the project. It is then considered more as a starting point than an actual design goal and gets modified several times during the development project in order to adjust it to changing market needs. This variable will thereby measure the amount of these design iterations during a project. Brown & Eisenhardt (1997) note that using a variety of probes helps individuals to reduce uncertainty since it lowers the probability of being surprised by rapid environmental changes.

From the perspective of information processing theory constantly modifying the product definition should increase information processing efforts at two points of the organization. At the individual level of the team member, being encouraged to constantly modify the product design in response to environmental changes will cause serious efforts of information processing. At the team level, every time a change is made towards the product definition, team members have to adjust their activities to it while considering relevant effects to their respective tasks. Also, decision makers at higher levels have to be involved each time a substantial change is made to ensure that the product design will meet its target. Having modular product components should lessen the negative effects of design iterations since the interdependence of products’ components is reduced. This should mean that even if one component passes through a design iteration it does not necessarily induce a snowball effect of modifications for other components.

of the most ideal possibilities of being responsive to the environment organizational learning theory holds that organizations should apply a self-designing approach. By being in a constant state of change, a fluent product design helps to frequently adapt to environmental changes. By learning about a variety of product features, the process remains flexible.

Thus, from the perspective of information processing theory frequent design iterations should lead to a negative effect on performance which however could be compensated by modularization. From the organizational learning theory design iterations should have a positive effect on performance in highly uncertain environments.

5.1.4. Several Product designs maintained in parallel

Assuming that product development is an unclear path through foggy and shifting environments Biazzo (2009) argues that following several paths to solve a problem should be much more applicable than just sticking to one possible solution. Thus, keeping multiple designs in parallel during the NPD process and dropping them one by one as uncertainty dissolves could be and promising approach to avoid a dead end in product design. From the point information processing perspective this should however entail significant costs. Making sense of the environment and applying this knowledge to multiple design alternatives causes much more information processing on the individual level. Also, on the team level information processing increases because each time environmental information is processed decisions have to be made to keep or drop each of the product designs. This makes it also hard to handle slack resources due to the difficulty to prioritize certain product designs.

From an organizational learning perspective however, learning by doing is promoted through experimenting with various product designs. By doing so, team members can develop an understanding of market needs and product specifications and thereby become able to make sense of their environment. By broadening the scope of search for new information, the chances for a hit become much higher. Especially in turbulent environments with constantly changing markets and technological demands, this should help to resolve uncertainty quickly. Taking also into account information processing, the benefits of experimentation should outweigh the costs for information processing when uncertainty is high. However, when uncertainty is low, having a single product design should be much more effective.

5.1.5. Customer feedback causes change in product definition

As one possibility to sense markets, Biazzo (2009) describes the approach to obtain customer feedback through early prototypes. Customers can thus provide important knowledge on their usage behavior and demand and thus help the development team to get the product definition right. In a similar manner Eisenhardt & Tabrizi (1995) considered supplier involvement in NPD processes as an important driver to shorten development time. They argue that involvement of external partners might reduce the workload of the focal team due to sharing ideas about product design. Moreover, they argue that due to the relive in idea generation the development team can focus more on the execution of tasks. They may also help to identify future design problems early in the process and thus help to avoid costly design iterations. In a similar vein Chang & Taylor (2016) found that in technological turbulent environments firm’s existing knowledge can quickly become obsolete. Since it is a slow and expensive process for organizations to catch up with these changes engaging customers with solution related knowledge can be vital to gain access to critical knowledge.

From the theory of information processing engaging customers in the definition of the product definition could be a ‘double-edged sword’. On the one hand, information processing increases between individuals since the input of the customer must be processed and correctly applied by the development team. On the other hand, customer themselves process information when using and giving feedback on a prototype implying that they take off the effort from the development team which in turn can focus on other matters. From the organizational learning perspective customers provide a source of vital information and thus boost knowledge acquisition. Since they provide feedback from the viewpoint of a user, they also provide a channel to interpret externally acquired and internally developed knowledge. When uncertainty is low, market and technological demands should be somehow clear from the beginning of the project. Frequent changes in the product definition through consumers could therefore bring only negligible improvements and could therefore slow down the development process. If, however uncertainty is high and customer demands as well as technological developments change rapidly, customers should possess essential knowledge to resolve uncertainty.

5.2. Methodology

5.2.1. Measurements

(Late) Design freeze

The first question captures how far the project was completed when the product definition was fixed. This is done by letting the respondent fill out the completion of the project in % at the point when the product definition was fixed. Therefore, a high number implicates late design freeze.

Parallel (product-)designs

In order to capture the amount of alternative product design that were kept in parallel respondents were simply asked to indicate the amount of parallel product designs used during the process.

Customer Cooperation

Biazzo (2009) acknowledges that flexible NPD projects accelerate uncertainty resolution through early external feedback in the problem formulation stage. Therefore, it was asked how many times during the project customer feedback caused a change in the product definition.

5.3. Analysis and results

Table 3: Regression Results Informational Dimension

Process Performance Product Performance Market Performance Intercept 3.799 (0.348) 4.295 (0.289) 3.852 (0.324) Design Freeze -0.003 (0.004) -0.001 (0.003) 0.001 (0.004) Parallel Designs -0.04 (0.094) 0.091 (0.078) -0.067 (0.088) Customer Feedback -0.004 (0.008) -0.004 (0.007) -0.008 (0.008) Market Turbulence -0.65 (0.099) 0.179** (0.082) 0.099 (0.093) Technological Turbulence -0.68 (0.113) 0.09 (0.094) -0.092 (0.105) Modularity -0.032 (0.062) -0.098* (0.051) -0.023 (0.057) Design Freeze x Market Turbulence 0.005 (0.003) 0.009*** (0.003) -0.001 (0.003) Design Freeze x Modularity 0.003* (0.002) 0.001 (0.002) 0 (0.002) Parallel Designs x Market Turbulence 0.183** 0.077 -0.093 (0.064) -0.058 (0.072) Customer Feedback x Market Turbulence 0.011 (0.013) 0.031*** (0.011) 0.023* (0.012) Customer Feedback x Technological Turbulence -0.024 (0.017) -0.028** (0.014) -0.022 (0.015) Project Duration -0.015** (0.007) 0.019** (0.006) 0.012* (0.007) Customer Cooperation 0.043 (0.063) 0.046 (0.053) 0.122** (0.059) Number of Employees -0.00001 (0) 0.0000053 (0) 0.000009 (0) F-Value 2.355*** 2.763*** 0.907 R2 0.008 0.002 0.554 *p<.1; **p<.05; ***p<.01

5.4. Dimensional Discussion

6. THIRD STUDY: TEMPORAL DIMENSION

6.1. Theory

6.1.1. Introduction

The temporal dimension refers mainly to the matter of task scheduling. In order to explain the different degrees of simultaneity in task scheduling, he introduces three possible execution strategies: sequential (task performed one after another), overlapped (subsequent task already performed when the previous one is not fully completed) and concurrent (several tasks performed simultaneous) whereby sequential is the least flexible and concurrent most flexible mode. He points out high degrees of simultaneity could help to increase time to market but come at the risk of costly reworks, redundant parallel-designs and unpredictability of the NPD process. Terwiesch et al. (2002) suggest that highly concurrent strategies should be only applied when uncertainty resolution is fast (uncertainty is low), and vice versa, sequential strategies should be applied when uncertainty resolution is slow (uncertainty is high). Stalk & Hout (1990) extend the understanding by explaining that high degrees of simultaneity in the case of predictable tasks speed up the development process because the steps are well-understood in advance and waiting time between those steps can be avoided. Modularity should be a key enabler for flexible modes of task scheduling. Since modular product components are rather independent from one another, working simultaneously on different components should become possible.

6.1.2. Task-Overlap

organizational learning also downsides of concurrent engineering seem to appear. Since evaluations of progress are more or less eliminated, a systematic and periodic possibility of knowledge exchange disappears. Even though concurrent teams need to constantly exchange about their progress, important parts of information could not be exchanged because the information holder might not interpret it as important and decides not to share it. Also, when uncertainty increases this should make it much harder to interpret external knowledge which is induced by frequent changes in the environment because each team has to take different factors into account making it much harder to find consensus. This could quite quickly lead to information overload. Therefore, in line with previous research it is to expect that task overlap is best suited for projects in environments characterized by low uncertainty whereas a negative effect is expected for projects in highly uncertain environments. Modularity may help to reduce the costs that come with concurrent engineering because the different components on which several teams work on are rather autonomous from each other.

6.1.3. Gate-Conditionality

6.1.4. Time between milestones

Adding to the understanding about milestones, the timespan between project milestones also need to be taken into account. Within Stage-Gate models the time between milestones is captured in stages. When applying more extended stages one could assume that the project is divided in rather large chunks of work including more complexity and thus more uncertainty of the performed tasks. This should also induce more autonomy for the team members since they have to make decisions themselves to solve a complex task. From an information processing point of view this should increase efforts at the individual level while reducing information processing between individuals since lesser collective decision-making takes place as it would do with frequent gates. As uncertainty increases this could quickly lead to information overload since individual team members can never have such a comprehensive understanding of their environment as they would have within a team. From the theory of organizational learning long time between milestones could lead to contradicting results. On the one hand it promotes learning by doing by giving team members the freedom to experiment in order to come up with a solution. However, on the other hand frequent possibilities for knowledge exchange and developing mutual understanding are eliminated. As uncertainty increases the need for constant mutual adjustment also increases which could lead to a negative impact of on performance.

6.2. Methodology

6.2.1. Measurements

As explained later in the analysis part, time between milestones did not lead to significant results and is therefore not included in the methodology section.

Task Overlap

Gate-Conditionally

It measures the extent to which projects are approved in the case of not fully meeting the criteria on the condition that it will be met at a later point. The same measures are used in this study using four items on the suggested 7-Point Likert Scale. One question asked whether the review criteria recognized that the project followed a different development sequence and allowed the project to proceed further even when it only met criteria partially.

6.3. Analysis and results

Table 4 reports the results of the temporal dimension model. The process performance dimension has a R-square of 0.001 and the F-statistic (3.804) is significant. The product performance dimension has a R-square of 0.014 and the F-statistic is (2.679) and is also significant. Task Overlap showed a very strong direct negative relationship with product performance (β = -0.197; p < 0.01) but no relationships were found in combination with the moderators. Gate-Conditionality showed no direct relationship with both performance dimensions. However, a strong negative relationship with process performance was found under technological turbulence. No significant results were found for time between milestones which were therefore not included in table 4 to enhance the fit of the model.

Table 4: Regression Results Temporal Dimension

6.4. Dimension Discussion

Time between milestones did not lead to significant results. Even though the duration of the project was considered as a control variable, the type of industry was not taken into account. In some industries where product development is extremely costly, having long times for experimenting between milestones could lead to a quick excess of the budget. Therefore, time between milestones needs to be kept short in order to avoid wasting budget.

Theoretical considerations regarding task-overlap assumed a positive effect on performance when uncertainty is low whereas a negative effect was expected under high uncertainty. Surprisingly a direct negative effect was found on product performance only. This effect turned insignificant under both types of uncertainty. One explanation could be that although task overlap speeds up the development process the important control mechanisms become neglected. When already working on the next stage even though the previous one was not critically evaluated technical mistakes or a mismatch with customer demands could get overlooked. This could implicate that even small mistakes in an early stage of the product development (like in making the business case) could backfire because they remain undetected until the very late stages of the product or even when they are already launched. What is also surprising is that modularity showed no effect on task overlap although Biazzo (2009) had strong arguments to support a positive moderating effect. The literature gives two possible explanation for this. In their considerations about product architecture Magnusson & Pasche (2014) explain that product platforming and modularization need to be considered together in order to explain their impact on NPD. Therefore, it would be interesting to include this factor in future research. Another explanation could be delivered by Fujimoto (2014) who states that individually developed (integral) products have higher functionality whereas modular products are more cost effective in the development process. Following that he assumes that price-oriented customers tend more to modular products whereas function-oriented customers tend towards integral products. This could at least deliver an answer to the product performance dimension.

Also, somewhat surprising is the negative relationship of gate conditionality with process performance when technological turbulence is high. It seems that the fact that team members think too much in stages outweigh the benefits of frequent knowledge exchange. However, especially in technologically fast-moving environments an immediate reaction can be crucial since waiting until a gate could lead to sticking too long towards an inferior technological choice. Moreover, giving team members the possibility to experiment with recent technologies could be hampered by monitoring them via gates.

7. OVERALL DISCUSSION

8. CONCLUSION

Existing literature about project flexibility mainly focusses on the design of processes and methods in order to examine NPD flexibility. However, an comprehensive analysis including the application of grounded theory and hard performance measurements are missing (Cooper & Sommer, 2016). This thesis offers an extensive analysis based on the ideas of Biazzo (2009) by using the theory of information processing, organizational learning and contingencies. Both conflicting and supporting results were found towards existing literature.

These findings have important implications for managers or leaders of R&D teams. First, taking the organizational dimension into account, managers should centralize decision making when market turbulence is high to speed up the development process and stay within budget. Moreover, they should provide structure in the NPD process while promoting modular product designs to both improve the process and the quality and suitability of the product. Second, from the informational perspective, keeping the product design fluid until shortly before market launch should be applied when market turbulence is high to ensure a fit of the product with the market. Also, it should be combined with a modular product architecture to ensure a smooth process. Another way to enhance the development process is by experimenting with parallel designs during the project. In addition, companies should be careful in dealing with customer feedback in regards of the product design. When markets are turbulent customer feedback is beneficial for the fit of the product with the market whereas under technological uncertainty customer feedback was found to be harmful. Lastly, the temporal dimension showed that applying overlapping or concurrent task execution has a negative impact on product performance. Therefore, managers should stick to sequential models to ensure quality of the product and fit with the market. Moreover, as technological uncertainty increases, managers should reduce the conditionality of gates. As predicted by Biazzo (2009) the results how that flexibility might be beneficial in one dimension whereas in other dimension rather negative effects are to be expected. But also, within the dimensions, some variables had a positive effect on performance whereas others had negative effects. As an important contribution to the literature the analysis of the informational dimension opens a new perspective on project flexibility. Further research should focus on a more fine-grained analysis of the intersection between product definition and product implementation and take further variables into account. Nevertheless, the question remains unanswered why the variable of design iterations only produced insignificant results.

9. References

Ahmad, S., Mallick, D. N., & Schroeder, R. G. 2013. New Product Development: Impact of Project Characteristics and Development Practices on Performance. Journal of Product Innovation Management, 30(2): 331–348.

Aiken, L. S., West, S. G., & Reno, R. R. 1991. Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage Publications.

http://catdir.loc.gov/catdir/enhancements/fy0829/91002062-b.html.

Arthur, W. B. 1994. Inductive Reasoning and Bounded Rationality. The American Economic Review, 84(2): 406–411.

Bacon, G., Beckman, S., Mowery, D., & Wilson, E. 1994. Managing Product Definition in High-Technology Industries: A Pilot Study. California Management Review, 36(3): 32–56.

Bayus, B. L. 1998. An Analysis of Product Lifetimes in a Technologically Dynamic Industry. Management Science, 44(6): 763–775.

Becker, T. 2005. Potential Problems in the Statistical Control of Variables in Organizational Research: A Qualitative Analysis With Recommendations. Organizational Research Methods, 8: 274–289. Belsley, D. A., Kuh, E., & Welsch, R. E. 1980. Regression diagnostics: Identifying influential data and sources of collinearity. New York: Wiley. http://catdir.loc.gov/catdir/toc/onix04/79019876.html. Bhattacharya, S., Krishnan, V., & Mahajan, V. 1998. Managing New Product Definition in Highly Dynamic Environments. Management Science, 44(11-part-2): S50–S64.

Biazzo, S. 2009. Flexibility, Structuration, and Simultaneity in New Product Development. Journal of Product Innovation Management, 26(3): 336–353.

Brown, S. L., & Eisenhardt, K. M. 1997. The Art of Continuous Change: Linking Complexity Theory and Time-Paced Evolution in Relentlessly Shifting Organizations. Administrative Science Quarterly, 42(1): 1–34.

Buganza, T., Gerst, M., & Verganti, R. 2010. Adoption of NPD flexibility practices in new technology-based firms. European Journal of Innovation Management, 13(1): 62–80. Chang, W., & Taylor, S. A. 2016. The Effectiveness of Customer Participation in New Product Development: A Meta-Analysis. Journal of Marketing, 80(1): 47–64.

Chen, J., Reilly, R. R., & Lynn, G. S. 2012. New Product Development Speed: Too Much of a Good Thing? New Product Development Speed. Journal of Product Innovation Management, 29(2): 288– 303.

De Meyer, A., & Van Hooland, B. 1990. The contribution of manufacturing to shortening design cycle times. R&D Management, 20(3): 229–239.

Dillman, D. A. 1978. Mail and telephone surveys: The total design method. New York: John Wiley & Sons.

Du, J., Leten, B., & Vanhaverbeke, W. 2014. Managing open innovation projects with science-based and market-based partners. Research Policy, 43(5): 828–840.

Eisenhardt, K. M., & Tabrizi, B. N. 1995. Accelerating Adaptive Processes: Product Innovation in the Global Computer Industry. Administrative Science Quarterly, 40(1): 84–110.

Fantazy, K. A., & Salem, M. 2016. The value of strategy and flexibility in new product development: The impact on performance. Journal of Enterprise Information Management, 29(4): 525–548. Fredrickson, J. W. 1984. The Comprehensiveness of Strategic Decision Processes: Extension, Observations, Future Directions. The Academy of Management Journal, 27(3): 445–466.

Fujimoto, T. 2014. The Long Tail of the Auto Industry Life Cycle. Journal of Product Innovation Management, 31(1): 8–16.

Galbraith, J. R. 1974. Organization Design: An Information Processing View. Interfaces, 4(3): 28–36. Griffin, A. 1997. PDMA Research on New Product Development Practices: Updating Trends and Benchmarking Best Practices. Journal of Product Innovation Management, 14(6): 429–458. Griffin, A., & Page, A. L. 1996. PDMA success measurement project: Recommended measures for product development success and failure. Journal of Product Innovation Management, 13(6): 478– 496.

Gupta, A. K., & Wilemon, D. L. 1990. Accelerating the Development of Technology-Based New Products. California Management Review, 32(2): 24–44.

Hair, J., Hult, G. T. M., Ringle, C., & Sarstedt, M. 2014. A Primer on Partial Least Squares Structural Equation Modeling.

Huber, G. P. 1991. Organizational Learning: The Contributing Processes and the Literatures. Organization Science, 2(1): 88–115.

Iansiti, M. 1995. Shooting the Rapids: Managing Product Development in Turbulent Environments. California Management Review, 38(1): 37–58.

Jansen, J. J. P., Van Den Bosch, F. A. J., & Volberda, H. W. 2006. Exploratory Innovation,

Exploitative Innovation, and Performance: Effects of Organizational Antecedents and Environmental Moderators. Management Science, 52(11): 1661–1674.

Jones, G. R. 2013. Organizational theory, design, and change (7th ed., global ed.). Harlow [etc.]: Pearson Education.

Kamoche, K., Pina, M., & Cunha. 2001. Minimal Structures: From Jazz Improvisation to Product Innovation. ORGANIZATION STUDIES -BERLIN- EUROPEAN GROUP FOR

ORGANIZATIONAL STUDIES-, 22(Part 5): 733–764.

Kettunen, J., Grushka-Cockayne, Y., Degraeve, Z., & De Reyck, B. 2015. New product development flexibility in a competitive environment. European Journal of Operational Research, 244(3): 892– 904.

Kleinschmidt, E. J., De Brentani, U., & Salomo, S. 2007. Performance of Global New Product Development Programs: A Resource-Based View. Journal of Product Innovation Management, 24(5): 419–441.

Liu, Li, Y., & Wei, Y. 2009. How organizational flexibility affects new product development in an uncertain environment: Evidence from China. International Journal of Production Economics, 120(1): 18–29.

Magnusson, M., & Pasche, M. 2014. A Contingency-Based Approach to the Use of Product Platforms and Modules in New Product Development. Journal of Product Innovation Management, 31(3): 434–450.

March, J. G., Olsen, J. P., & Christensen, S. 1979. Ambiguity and choice in organizations (2nd ed.). Bergen: Universitetsforlaget.

Milliken, F. J. 1987. Three Types of Perceived Uncertainty about the Environment: State, Effect, and Response Uncertainty. The Academy of Management Review, 12(1): 133–143.

Mishra, R., & Mishra, O. N. 2019. Factor influencing flexibility in new product development: Empirical evidence from Indian manufacturing firms. Journal of Business & Industrial Marketing, 34(5): 1005–1015.

O’brien, R. M. 2007. A Caution Regarding Rules of Thumb for Variance Inflation Factors. Quality & Quantity, 41(5): 673–690.

Pries-Heje, L., & Pries-Heje, J. 2011. Why Scrum Works: A Case Study from an Agile Distributed Project in Denmark and India. 2011 Agile Conference, 20–28. Presented at the 2011 Agile

Conference.

Qualls, W., Olshavsky, R. W., & Michaels, R. E. 1981. Shortening of the PLC—AN Empirical Test. Journal of Marketing, 45(4): 76–80.

Raisch S., & Birkinshaw J. 2008. Organizational ambidexterity: Antecedents, outcomes, and moderators. Journal of Management, 34(3): 375–409.

Schleimer, S. C., & Faems, D. 2016. Connecting Interfirm and Intrafirm Collaboration in NPD Projects: Does Innovation Context Matter? Journal of Product Innovation Management, 33(2): 154– 165.

Sethi, R., & Iqbal, Z. 2008. Stage-Gate Controls, Learning Failure, and Adverse Effect on Novel New Products. Journal of Marketing, 72(1): 118–134.

Sheremata, W. A. 2000. Centrifugal and Centripetal Forces in Radical New Product Development under Time Pressure. The Academy of Management Review, 25(2): 389–408.

Simon, H. A. 1973. Applying Information Technology to Organization Design. Public Administration Review, 33(3): 268–278.

Smith, P. G., & Reinertsen, D. G. 1992. SHORTENING THE PRODUCT DEVELOPMENT CYCLE. Research Technology Management, 35(3): 44–49.

Song, M., Im, S., Bij, H. van der, & Song, L. Z. 2011. Does Strategic Planning Enhance or Impede Innovation and Firm Performance?* Strategic Planning, Innovation, and Firm Performance. Journal of Product Innovation Management, 28(4): 503–520.

Stalk, G. Jr., & Hout, T. M. 1990. Competing against time: How time-based competition is reshaping global markets. New York: Free Press.

http://bvbr.bib-bvb.de:8991/F?func=service&doc_library=BVB01&doc_number=002602295&line_number=0001&f unc_code=DB_RECORDS&service_type=MEDIA.

Tatikonda, M. V., & Rosenthal, S. R. 2000. Successful execution of product development projects: Balancing firmness and flexibility in the innovation process. Journal of Operations Management, 18(4): 401–425.

Terwiesch, C., & Loch, C. H. 1999a. Measuring the Effectiveness of Overlapping Development Activities. Management Science, 45(4): 455–465.

Terwiesch, C., & Loch, C. H. 1999b. Managing the process of engineering change orders: The case of the climate control system in automobile development. The Journal of Product Innovation

Management, 16(2): 160–172.

Terwiesch, C., Loch, C. H., & Meyer, A. D. 2002. Exchanging Preliminary Information in Concurrent Engineering: Alternative Coordination Strategies. Organization Science, 13(4): 402–419.

Thomke, S. H. 1997. The role of flexibility in the development of new products: An empirical study. Research Policy, 26(1): 105–119.

Thomke, S., & Reinertsen, D. 1998. Agile Product Development: Managing Development Flexibility in Uncertain Environments. California Management Review, 41(1): 8–30.

Organizational Design. The Academy of Management Review, 3(3): 613–624.

Vlietland, J., & van Vliet, H. 2015. Towards a governance framework for chains of Scrum teams. Information and Software Technology, 57: 52–65.

Wind, Y., & Mahajan, V. 1988. New product development process: A perspective for reexamination. The Journal of Product Innovation Management, 5(4): 304–310.

10. Appendix

10.1.

Overall Model

Process Performance Product Performance Market Performance Intercept 3,934 (0,356) 4,428 (0,307) 3.916 (0.353) Structuration -0,088 (0,097) 0,056 (0,084) -0,011 (0,097) Centralization -0,057 (0,092) -0,066 (0,08) -0,001 (0,091) Design Freeze -0,002 (0,004) -0,001 (0,003) 0 (0,004) Parallel Designs -0,018 (0,096) 0,095 (0,083) -0,054 (0,095) Customer Feedback -0,007 (0,009) 0,002 (0,008) -0,007 (0,009) Task-Overlap -0,011 (0,089) -0,147* (0,077) -0,053 (0,089) Gate-Conditionally -0,072 (0,115) 0,073 (0,099) 0,013 (0,114) Technological Turbulence 0,045 (0,122) 0,171 (0,106) -0,038 (0,121) Market Turbulence -0,193* (0,104) 0,133 (0,09) 0,046 (0,103) Modularity -0,010 (0,061) -0,093* (0,053) -0,006 (0,06) Centralization x Market Turbulence 0,1* (0,059) 0,013 (0,051) 0,076 (0,058) Structuration x Modularity 0,070 (0,045) 0,050 (0,039) 0,038 (0,045)

Design Freeze x Market Turbulence 0,004 (0,003) 0,008*** (0,003) -0,002 (0,003)

Design Freeze x Modularity 0,003 (0,002)

0,001 (0,002)

-0,001 (0,002)

Parallel Design x Market Turbulence -0,172** (0,077) -0,078 (0,067) -0,056 (0,077)

10.2.

Factor Analysis

10.2.1. Factor Analysis independent variables

Structuration Gate-Conditionally

10.2.2. Factor Analysis dependent variable

Process Performance Product-Performance Market Performnce

10.3.

Reliability Analysis

Factor Item Amount Cronbach Alpha