The direct and interaction effects of flexibility dimensions on NPD project performance in dynamic environments

1

The direct and interaction effects of flexibility dimensions on NPD

project performance in dynamic environments

Master BA - Strategic Innovation Management

University of Groningen - Faculty of Economics and Business

January 2019

Victor Laan

S2558408

Supervisor: dr. J.D. van der Bij Co-assessor: dr. W.G. Biemans

Word count: 9121

Abstract

This research contributes to NPD literature, by examining the direct and interaction effect of flexibility dimension imposed by Biazzo (2009) on NPD project performance. By means of a survey, empirical data is gathered and added to a database consisting a total of 58 firms. The regression shows that overlapping-task strategy moderated by technological turbulence is positively related to efficiency. The interaction effect of overlapping-task strategy and design iterations moderated by technological turbulence is positively related to both efficiency and effectiveness. Market turbulence moderating overlapping-task strategy with formalization is positively related with efficiency.

2

Introduction

Today’s society is more changing than ever, making new product developments (NPD) necessary for organizations to remain competitive. In this context, there is a growing importance for the role of flexibility. Flexibility seems inevitable in order to anticipate

changing customer needs or technological innovations. It is interesting to have a closer look at the different dimensions of flexibility and how they (possibly) interact, also keeping in mind dynamic environments. What is their cumulative influence on the NPD project performance? Under what conditions are flexibility measures enhancing or detrimental? This research aims to contribute to the NPD literature by examining the direct effects of flexibility dimensions and their interaction terms on the NPD project performance, within dynamic environments.

Authors in the NPD literature are not in agreement how organisations should structure their NPD project to enhance the performance. Cooper (1990) argues there is a ‘one best way’ for developing new products. In contrary, Iansiti (1995) MacCormack and Verganti (2003) stress that flexibility plays a crucial role in NPD processes in order to adapt to changes in customer demand. Biazzo (2009), describes two forms of product development; anticipation and reaction (Verganti, 1999; Kalyanaram & Krishnan, 1997). The anticipation strategy concerns an early and sharp definition of the product. A well-known example is the Stage-Gate model of Cooper (1990). A reaction strategy involves moving the freeze point of the product definition as close to market launch as possible. These conflicting perspectives are a point of discussion for organizational structures, where organizations consider balancing between flexibility and structure (Eisenhardt et al., 2010). Biazzo (2009), addresses the dichotomy between the flexibility and structure and proposes a three dimensional framework.

In this study three types of NPD project flexibility (Biazzo, 2009) are included, namely; structural flexibility, informational flexibility and flexibility in time. Respectively, these types concern how strictly the NPD process is organized, the intersection between product

formulation and product solution and whether the NPD projects are performed simultaneously (sequential, overlapping, concurrent). For managers in charge of, or to some extent

responsible for an NPD project, knowing how each type of flexibility is influencing the performance can be essential, especially in highly volatile industries.

In former NPD studies, variables related to flexibility and their relationship with NPD project performance were examined (Fischer, 2017; Borgeld, 2018). The outcomes remain

3

examine the interplay between flexibility dimensions. The one dimensional meta-studies lack in covering all the performance variables. For now, only one project performance variable is examined in most studies: the time to market. In the rapidly changing environment we are now facing, this is an important factor, however, it covers just a part of the total NPD project performance.

There is also a lack of empirical studies measuring the variables representing the three flexibility types mentioned earlier, especially on project level (Biazzo, 2009). In addition, Biazzo (2009) argues for a contingency approach taking into account the moderating effect of dynamic environments, such as market- and technological turbulence. To address the above mentioned literature gaps, the following research question is formulated;

RQ1 What is the influence of the flexibility dimensions on NPD project performance within dynamic environments?

Previous research focused on the definition of the three flexibility types, however, this study takes it a step further. By examining the interrelationship between the three flexibility types and their cumulative influence on the NPD performance, this study attempts to address this gap in the current NPD literature. Empirical data about the flexibility practices in real-life can help managers to choose the appropriate measure taking into account the environmental conditions. To address this gap, the second research question is formulated:

RQ2 What is the influence of the interplay between the flexibility dimensions on the NPD project performance, within dynamic environments?

To get a better view of the different variables regarding the types of NPD project flexibility and NPD project performance, there is a collection of primary data about real NPD projects in this study. Based on former research, a selection is made of multiple variables from which a correlation matrix is established. In addition, these results are analysed with factor and regression analyses. Previous research is examined together with the ‘organizational information processing theory’ (OIPT) (Galbraith, 1974), which is the foundation for the research hypotheses. The OIPT will be further elaborated in the literature review. As

4

Literature Review

The success of NPD projects depends on which value system is used to evaluate the project. It is generally agreed that projects should be appropriate in terms of strategic fit and it should meet the organizational objectives (Van Wyngaard et al., 2012). Hence, the innovative performance is based on the process and outcome of the project. In order to measure the performance, the expectations before the project were initiated are compared with the outcomes, instead of just looking at the absolute numbers. To measure the success of new products, Griffin and Page (1996), made three groups; customer based success, financial success and technical performance. In this research, the performance of firms will be evaluated judging from their relative performance on; product development cost, product quality, technical performance, time to market, market share, overall profitability and overall commercial success of the product (Schleimer & Faems., 2016; Ahmad et al.,2013).

Organizational Information Processing Theory (OIPT)

The organizational information processing theory entails, that solving uncertainty is a central part of organizational design. When environments are uncertain, more information is needed to deal with this uncertainty, within or across organizational (sub)units (Galbraith, 1974). When an organization has a clear description of tasks and performance objectives in advance, the structure can be set beforehand, to address the needed resources. In contrary, when there is high uncertainty, more information is required throughout the process to enable processes to anticipate on changes in resource allocation, planning and objectives. Uncertainty is limiting the ability of organizations to plan in advance or to allocate resources before execution. The strategies are chosen based on their relative costs. Meaning, when there are variations in the degree of uncertainty, the integrated mechanisms; coordination by rules or programs, hierarchy and coordination by targets and goals, should cope with this change (Galbraith, 1974).

The capability to handle non-routine events will determine the success of handling with uncertainty. When uncertainty arises, the organizational structure has to adapt accordingly to cope with the change. Galbraith (1974) suggests that organisations can improve their

information processing capacity by investing in vertical information systems and creations of

lateral relations. Organizations can also reduce the information processing requirement

5

Tushman and Nadler (1978) built on the OIPT theory of Galbraith (1974) with the

‘information processing model’. This model argues that need for information processing is determined by the units’ level of uncertainty, consisting of; 1) unit task characteristics, 2) unit task environment and 3) inter-unit task interdependence. Consequently, this results to various degrees of information processing requirements. The authors propose there is a ‘fit’ when the information processing requirements and the information processing capabilities are balanced (Tushman & Nadler, 1978).

Flexibility in NPD projects

Biazzo (2009), states: ‘flexibility in an NPD project can be defined as the ability to embrace

environmental turbulence rapidly, adapting to new technological and market information that becomes available over the course of the project.’ The relation between flexibility measures

and NPD project performance is complex, therefore Biazzo (2009) argues, flexibility should be treated as a three dimensional framework covering the organizational, informational and temporal dimension.

Organizational flexibility refers to the structuration of the process. Biazzo (2009) states,

‘process structuration consists of formal segmentation of the temporal progression in stages by identifying a series of predefined decision-making points’. This will contribute to better

understand the complex characteristics of structure in product design. Regarding the project structure, there can be a number of formal stages and milestones. The extent to which these stages are formalized are related to the degree of flexibility. In previous studies, multiple variables are assigned to the organizational dimension, namely; formalization,

cross-functional integration, hierarchy of authority, centralisation and professionalism (Mushbah et al., 2016). There are mixed results concerning the formalization dimension and the impact on NPD performance. Due to this controversy, this study hopes to give more clearance on this matter in dynamic environments.

Formalization is ‘the degree to which the firm utilizes rules, procedures and written

documentation to structure the behaviour of individuals or groups within the organisation’

6

stating the opposite and arguing the need for less rules and procedures. Kamoche & Cunha (2001), find that in turbulent environments, too much control hinders the creativity and consequently hinders the innovation process. Innovation requires a degree of flexibility to explore new opportunities. Organizational flexibility and freedom, as a result of less formalization, can facilitate more room for these explorations. Employees will have more autonomy and are able to discover more, when breaking the organizations’ routines (Cosh et al. 2012). However, when lowering the degree of formalization there is a point at which the organisation lacks the basic formalization to inform all employees about the main objectives. So in order to facilitate the flexibility and freedom in all units, a certain amount of

formalization is required to set the needed rules and procedures.

The OIPT states that rules and procedures are most effective when enough information is available to predict an outcome (Galbraith, 1974). In that context, an organization has a clear description of tasks and performance objectives in advance. Setting the right structure will be more adequate in addressing the right resources. In case of dynamic environments, there is not much information and certainty about the course of technology or customer needs. Strictly predetermined work processes and work roles, will not be satisfying when market or technological changes occur. With a high degree of formalization, integrating a number of predefined rules and procedures, it can be very difficult and time-consuming to process all the required information. According to Tushman & Nadler (1978), there is a ‘fit’ when the

information processing requirements and the information processing capabilities are balanced. Under dynamic environment conditions the formalized organisation will lack information processing capabilities. Therefore, in line with the OIPT (Tushman & Nadler, 1978; Galbraith, 1974), the following hypothesis is proposed:

H1: Formalization is negatively related to NPD project performance within dynamic environments.

7

the two stages for a longer period of time. This interaction will enhance the flexibility of the definition and design activities. Another way of looking at the informational flexibility, is the ability to freeze the concept as close to market launch as possible. Previous research argues that experimental orientation is essential, especially in uncertain environments, meaning more design iterations and delaying the design freeze (Jin, 2000).

Within the informational variable, design iterations are an important component during the development process within projects (Biazzo, 2009). By testing the design of the product, technical issues can be recognised and dealt with. In addition, the customer feedback can give an update of the current needs and whether the product is still matching the customer needs. On the other hand, increasing the design iterations can postpone the marketing and industrial design decision, making it more time consuming to launch the product (Veryzer, 2005). However, by postponing the marketing and industrial design decisions, the concept is more accurate in serving customer needs. Also, it is essential to gain formal recognition and support.

Looking from the perspective of the OIPT, gaining more customer insights by increasing the number of iterations, more market information is gathered. As a result, the uncertainty will decrease and the product will be more adequate when entering the market. Especially when the environment is rapidly changing, it is important to verify whether the product is still meeting customer demands. Technological turbulence is one other factor contributing to uncertainty. With enough design iterations installed, technical issues and challenges can be dealt with in time, to prevent errors later in the process or when the product is launched.

Incorporating slack resources by the design iterations, is a way to deal with these market and technological uncertainty (Galbraith, 1974). Design iterations can help an organization to cope with unforeseen events. When organizations do not know how to anticipate on new information or lack the resources to do so, slack resources can function as a buffer to fill up the potential gaps or to fix problems that arise. With an increase in design iterations, more (slack)resources are needed due to the increase in number of stages, because each stage will contain enough resources to anticipate when changes occur. In other words, the number of design iterations will enhance the NPD performance, thus:

8

Flexibility in time, concerns which tasks are simultaneous executed within the NPD project. Flexibility in time refers to task scheduling and the interrelationships between multiple tasks within the organization. There is a division of three execution strategies: sequential,

overlapped or concurrent (Joglekar et al., 2001). Sequential strategies follow a step-by-step path with clear division of stages. There is overlap when different clusters of people execute tasks simultaneously. Iansiti (1995), states that flexible processes ‘go beyond concurrent

engineering’ through ‘stage overlap’. Lin et al. (2008), found that for intense competitive

industries, overlapping development processes are becoming increasingly important to enable firms developing products in less time.

In order to reduce the NPD process time, the advantages of the overlapping-task strategy should outweigh the extra resources and time that comes with the overlapping tasks. Unintended side-effects during concurrent engineering can cause extra time and costs for cooperations and rework when downstream processes are initiated earlier (Repenning, 2001; Chakravarty, 2001). However, multiple studies have shown the benefits of the overlap strategy, agreeing that overlap can substantially reduce the product cycle time (Smith and Reinertsen, 1995; Helms, 2004) and is related to the reduction of uncertainty of NPD projects (Eisenhardt and Tabrizi, 1995; Terwiesch and Loch, 1999). By sharing the information of developments in the downstream stages, issues that arise upstream can be explained and action can be taken for the corresponding activities. Also, the organization can better anticipate on unexpected events hereby reducing the uncertainty.

This reduction of uncertainty due to an increase of information sharing, is also central in the organizational design with regards to the OIPT (Galbraith, 1974). Overlapping can reduce the lead time in development, especially for projects in uncertain, dynamic environments

(Terwiesch and Loch, 1999). Thus:

H3: The overlapping-task strategy has a positive influence on NPD project performance in dynamic environments.

9

independent from the informational dimension. He mentions the interplay between

organizational and temporal dimension, but leaves their impact on NPD project performance undisputed.

As hypothesized before, it is expected that formalization (organizational) is negatively related to the NPD project performance. However, it is interesting to examine what happens with this relationship, when formalization is combined with the overlapping-task (temporal) strategy. Following the findings imposed by Biazzo (2009) it is possible that negative effect of

formalization can be temperate when incorporating an overlapping-task strategy. In order for the overlapping tasks to function properly, a number of rules and regulations must be set. While implementing these rules and regulations it is essential that the individuals carrying out the tasks are able to communicate swiftly and without any interruptions. Problems can arise when employees of different stages are not on the same page or are not speaking the same language. Formalized measures ensure every overlapping task can work efficient.

Looking from an OIPT perspective, the overlapping-task strategy benefits from formalization. To facilitate the information sharing between two overlapping task, there are now rules and regulations about the coordination. When, where and who will communicate? Investing in vertical information system is a method to transfer information amongst vertical layers (Galbraith, 1974). This makes it easier to share (explicit) knowledge. During the process, many unexpected events can occur, therefore it is essential to have an adequate information system ready. According to Tushman & Nadler (1978) there is a ‘fit’ when the information processing requirements and the information processing capabilities are balanced. By

incorporating rules and regulations the stages are capable to process the required information. To sum up the following hypotheses is formulated:

H4: The overlapping-task strategy will temper the negative influence of formalization on the

NPD project performance so, there is a positive interaction effect.

10

taken into account when designing the process in the temporal dimension. In other words, temporal and informational dimensions are two interdependent process design levers and therefore can be a risky combination between flexibility and simultaneity. The connection between the problem formulation and problem solving would influence the impact of overlapping activity.

Whether the interplay between temporal and informational flexibility practices has a positive or negative outcome for the NPD project performance, is still up for debate. To measure the cumulative influence, a corresponding variable from each dimension is chosen. Terwiesch et al. (2012) argue that overlapping of tasks in combination with iterations, carries risks

concerning reworking engineering orders, therefore not always the best strategy to follow. However, taking an OIPT perspective, uncertainty makes the iterations inevitable in order to anticipate within dynamic environment (Galbraith, 1974). Biazzo (2009) comes up with the question; ‘is the level of efficiency of iteration-based flexibility worsened by an overlapped

execution strategy compared with a more traditional sequential execution?’. Terwiesch and

Loch (1999) argue that during projects, overlapping activities can decrease the development lead time in environments with fast uncertainty resolution.

In case of overlap, information is extensively shared amongst development stages, speeding up the NPD process (Lin et al., 2010). In case of iteration-based flexibility design and

overlapped execution strategy there are various scenarios possible. It is interesting how these two dimensions interplay and what the consequence is for the NPD project performance. When, for example, task A is characterized as overlapping with task B and task B is iteration-based, rework of task A is likely as task B is making several changes during the process. Also, the information sharing amongst task A and B can put pressure on the expenditures and can negatively influences the NPD project performance. See figure 1 below.

Figure 1.

11

In the case that with each iteration there is overlap between task A and task B, the coexistence of iteration-based and overlap strategy might be good (see figure 2 below). However, this will depend of the type of product and the environmental conditions. In dynamic environments it can be good to incorporate many iterations accompanied by overlapping tasks to update the product current needs. Nowadays, an agile approach in project management is popular amongst organizations. Agile is an iterative approach were teams work on small increments instead of betting on one ‘big launch’ (Highsmith, J. R., 2009). Teams continuously evaluate initial plans and current states in order to cope with changes, quickly share information across horizontal and vertical business units.

Figure 2.

Iterations can facilitate this information sharing where employees communicate the current status of the product. Also, changes in the environment can be elaborated on and coped with by implementing new alternatives. Organisations should be cost cautious, combining

iterations with overlap, as rework puts pressure on time and budget (Sanchez & Mohoney, 1996). Here the context is of the essence, since in stable environments a high degree of iterations would be extremely inefficient, because of the time wasted. In accordance with the OIPT the greater the uncertainty, the greater the need for information that should be

communicated between participants during the execution of tasks (Galbraith, 1974). In addition, to accommodate the information sharing, enough iterations should be installed to ensure that all changes in the environment are dealt with. The following question is

formulated:

Q: Do overlapping development tasks in combination with a high number of design iterations

positively influence the NPD project performance within dynamic environments?

Overlap Overlap Overlap

12

Conceptual Model

Figure 3 shows the conceptual model composed from the hypotheses.

Figure 3: Conceptual model

Methodology

In prior NPD literature, an overview is provided regarding the flexibility dimensions and corresponding variables. However, Biazzo (2009) argues for an empirical approach to develop a better understanding of the (interaction) effect of flexibility on NPD project performance in turbulent environments. In this research, there is a collection of primary data concerning various NPD projects in order to run multiple regression analyses.

Sample

Together with a group of 8 students, a database has been established containing a total of 58 companies operating in the Dutch market. Using the ORBIS list of company’s worldwide, additional or missing information can be addressed. The database consists of a variety of industries ranging from small businesses up to multinationals with 112.000 employees. There is a great mix of sectors such as; health care, insurances, manufacturing and high tech.

Data collection

To collect the empirical data, two surveys were conducted; one for the senior manager and one for the project leader or employee, in order to reduce common method bias (Conway &

13

Lance, 2010). To ensure that the NPD project performance could be measured, the project had to be launched at least one year ago. In addition, the project had to be developed in-house, so not in alliance with other firms. The presence of an organizational structure is important, therefore we aimed to target companies with >25 employees. Also, for each company there was one NPD project chosen for which at least two or more employees were actively

involved. The surveys were distributed via e-mail containing a link, transferring participants towards the survey program; Qualtrics.

Measurements

Overlapping tasks, as discussed by Zirgir & Hartley (1996), relates to the extent tasks are carried out simultaneously. This construct is measured by asking respondents to scale the overlap between tasks during their NPD project, ranging from 1 = fully sequential through to 7 = fully concurrent engineering (simultaneous).

Formalization, is adapted from Despandé & Zaltman (1982). Respondents are asked whether they agreed to five statements regarding the extent the NPD project is characterized by formalization. (1) Written procedures were available for dealing with any situation. (2) Rules and procedures occupied a central place in the project organization. (3) Written records were kept of everyone’s performance. (4) Project members were hardly checked for rule violations. (5) Written job descriptions were formulated for positions in the project team. A 7 point Likert scale is used ranging from 1 = completely disagree through to 7 = completely agree.

Iterations, derived from the research of Eisenhardt & Tabrizi (1995), are measured by asking respondents about (1) how far the project was completed, in %, when the project definition fixed. (2) The number of iterations during the project (an iteration is a change of more than 10% in a design of one product alternative).

Moderating variables

14

the change, (3) the extent the technological turbulence supports ideas for new products and (4) whether technological developments in the industry are limited.

Market Turbulence, regards the changes in customer demand (Jaworski & Kohli, 1993). This construct is measured with a total of five question, scaling to what extent (1) customers are changing in demands, (2) whether customers are continuously searching for new product developments. In addition, respondents indicate (3) how many customers are new to their business, (4) to what extent new customers have different demands compared to existing customers regarding products and (5) delivery of product to customers whom they had done business with in the past. A 7 point Likert scale is used ranging from 1 = completely disagree through to 7 = completely agree. From the four questions, only market turbulence 1 and 2 had significant impact on NPD project performance and are separately integrated in the regression models.

Dependent variable

NPD Project Performance is measured by comparing the manager expectations before the project with the actual outcome. Therefore, this is considered a subjective measure based on the relative performance. Fisher (2017) argues, performance should be measured in a multi-dimensional way. Together with the three-stage model of Mallick & Schroeder (2015), Schleimer & Faems., (2016) and Ahmad et al., (2013), there are a number of items included in this research to measure the performance, namely; product development cost, product quality, technical performance, time to market, market share, overall profitability, overall commercial success. A 7 point Likert scale is used ranging from 1 = significantly worse than expected through to 7 = significantly better than expected.

Control variables

15

Analysis

Before running the regression models, the dataset had to be cleaned and prepared. The inversely coded items were reversed in order to avoid mistakes. After performing the factor analyses, the moderating variables are mean-centred to reduce the chance of multicollinearity.

Construct reliability and validity

In order to check the construct validity, an exploratory factor analysis is conducted to determine whether the items used, measured the right construct intended. Items could only load on one factor. If an item would load on multiple factors or were smaller than .40, they were deleted. The first step was to verify if the multi-item scales actually measure the construct they were supposed to measure, by means of an exploratory factor analyses. As a result, the construct ´performance´ could be divided into two constructs. Hence, performance is too broad to be measured with just one variable.

To adequately measure performance, two constructs are developed in order to yield

significant outcomes in the regression. On the one hand, there is ‘efficiency’, loaded with the corresponding variables; development cost and the time to market. On the other hand, there is ‘effectiveness’ concerning; quality, market share, technical performance, commercial success

and profitability (Table 1a). In addition, the factor loadings of the following multi-item

16 Descriptive statistics

In table 2, the correlations between all variables used in the regressions are calculated. The correlation between efficiency and effectiveness is evident, as they are made off a split in the comprehensive variable; performance. However, the different correlations with efficiency and effectiveness towards other variables can be interesting in the following chapter, when the regression is run. For example, continuous changes in customer demand (M.Turbulence 1) and customer searching for new product developments (M.Turbulence 2) not only correlate with each other, but also negatively or positively correlate with efficiency, effectiveness and iterations. These signs can possibly lead to significance in the regression model

Table 1b: Factor Loadings

Technological Turbulence Market Turbulence Formalization T.Tur. 1 .792 .161 .068 T.Tur. 2 .709 .070 -.060 T.Tur. 3 .645 -.134 .037 T.Tur. 4 .541 .058 -.168 M.Tur 1 .216 .686 .146 M.Tur 2 -.065 .666 -.075 Form 1 .067 .064 .801 Form 2 -.022 .011 .782 Form 3 -.084 .032 .468 C. α 0.746 0.616 0.765

Extraction method: principal axis factoring Table 1a: Factor Loadings

Efficiency Performance Effectiveness Performance Development costs (PERF1) .892 .041 Time to market (PERF4) .767 .284

Market share (PERF5) .088 .879

Profitability (PERF6) .282 .726

Commercial success (PERF7)

.136 .890

Cronbach’s α 0.612 0.806

17

Results

Multiple hierarchical regressions are executed in order to test the hypotheses. First, only the control variables are included. Next, the independent variables; overlapping, iteration and formalization are added to examine their direct influence on efficiency and effectiveness. Lastly, moderators and the interaction between flexibility dimensions are included.

Table 2: Descriptive Statistics

Mean Standard Deviation 1 2 3 4 5 6 7 8 9 10 11 1.Efficiency 3.67 1.27 2.Effectiveness 4.63 1.04 .332* 3.Overlapping tasks 4.34 1.37 -1.80 .048 4.Iterations 6.07 12.78 -1.20 -.127 .081 5.Formalization 3.259 1.43 -1.23 .235 .121 .065 6.Tech. Turbulence 5.290 1.11 .026 .131 .169 .031 -.026 7.Changes in Customer needs 4.73 1.33 .146 .293* .041 -.280* .088 .268* 8.Customers looking for new products

4.91 1.31 -.377* .073 .066 -.001 -.020 .056 .445** 9.Radicalness 4.71 1.71 -.042 -.131 .404* .010 -.084 .007 -.063 -.143 10.Management Support; goals explicitly formulated 4.91 1.35 -.167 .071 -.182 -.079 .317* -.013 .056 .106 -.284* 11.Management Support; enough resources available 6.03 .991 -.046 .050 .353** .130 .130 .143 .045 -.096 .016 .146 12.B2B 4.11 1.28 -.216 -.088 -.136 .043 -.041 .054 -.179 .005 -.218 .006 .250

* Correlation is significant at the 0.05 level ** Correlation is significant at the 0.01 level

18 Table 3: Hierarchical Regression Results

A1 A2 B1 B2 C1 C2 D1 D2 E1 E2 F1 F2

Controls

Management Support; goals explicit .158 .949 .131 .663 .129 .900 .339 .504 .461 .879 .416 .777 Management Support; enough resources .778 .966 .833 .935 .818 .839 .955 .955 .785 .974 .896 .979

Radicalness .338 .273 .795 .588 .683 .314 .777 .466 .443 .378 .543 .356

B2B .141 .620 .114 .496 .143 .653 .054† .735 .070† .681 .129 .660

Changes in customer needs (MT1) .025** .091† .059† .204 .016* .079† .023* .355 .054† .083† Customer demands new products (MT2) .001* .595 .057† .454 .001** .468 .031* .385 .004** .754

Technological Turbulence .977 .676 .865 .552 .507 .763 .724 .878 Main Effects Overlapping (OVER) .426 .982 .710 .717 .800 .309 .271 .896 .494 .916 Formalization (FORM) .616 .381 .709 .210 .739 .204 Iterations (ITER) .734 .605 .067† .114 Moderators Technological Turbulence (TT) .227 .475 .667 .646 Market Turbulence 1 (MT1) .042* .116 Market Turbulence 2 (MT2) .008** .654 Interaction Terms TT x OVER .017* .500 TT x OVERITER .043* .033* MT1 x OVERFORM .042* .838 MT2 x OVERFORM .070† .260 R2 .292 .122 .301 .199 .355 .132 .356 .227 .393 .158 .380 .183 F 2.654* .897 1.681 .969 2.629* .729 2.155† 1.322† 2.416* .698 2.288* .834 † = p < .10, * = p < .05, ** = p < .01. N = between 52 and 58, due to missing data

19

The regression shows, the direct effect of formalization and iteration does not significantly influence either efficiency or effectiveness within dynamic environments. No support is found to suggest formalization is negatively related to NPD project performance. In addition, no support is found for the hypothesized positive relationship between the number of design iterations and the NPD project performance. Therefore, H1 and H2 are rejected. The regression analysis does show some support for hypotheses 3 indicating overlapping-task strategies, moderated by technological turbulence, have a positive influence on NPD project efficiency (p = 0.017 < 0.05). However, the same does not apply for effectiveness, therefore, H3 is partially accepted.

Model D shows the interaction effect of overlapping tasks and design iterations. Moderated by technological turbulence, they are positively related to efficiency (p =.043 < 0.05) and effectiveness (p = .033 < 0.05). Therefore, the question (Q) can be answered in a positive way. Although, no negative direct relationship between formalization and NPD project performance is found, regression models E and F show significant interaction effects with the overlapping-task strategy. With moderation of market turbulence 1 (continuous changes in customer demand), overlapping-task strategy and formalization have a positive interaction effect on the efficiency (p = .042 < 0.05). However, there is no significance noted for the influence on effectiveness. Model F shows the moderation effect of market turbulence 2 (continuous search for new product developments by customers) on the interaction terms of overlapping-task strategy and formalization. Again, only a positive relationship with

efficiency is found (p = .005 < 0.05). Therefore, H4 is partially supported.

Discussion and Conclusion

This research aims to contribute to the NPD project literature by examining the direct effects of flexibility dimensions and their interaction terms on the NPD project performance, within dynamic environments. By taking a multi regression method, the flexibility dimensions’ most influencing corresponding variables are identified. In the next part, the first research question will be answered regarding the direct influence of the flexibility dimensions on NPD project performance. The second and third part of this chapter will cover the second research

20

Technological Turbulence x Overlapping-task Strategy

Overlapping-task strategy being positively moderated by technological turbulence is in line with multiple studies. Biazzo (2009) argues, technological turbulence comprises the degree of novelty and the knowledge base of design solution. By means of information sharing and integrated problem solving, the turbulence can be coped with. Task overlapping provides a collective learning, concerning system and technology knowledge exchange. The moderation effect is elaborated by Iansiti (1995) who states, organizations are overlapping technological integration tasks in order to react to most recent technological developments. The

performance measures taken into account are lead time and productivity.

These measures are similar to the significant performance measure ‘efficiency’ in the regression of table 3 consisting of ‘product cost’ and ‘time to market’. In other words, efficiency is a more process related performance measure. On the other hand, ‘effectiveness’ (market share, profitability and commercial success), is related to the outcome and result of the overall performance. Looking from the perspective of this study, organizations within technological turbulent environment, can benefit from emphasizing on overlapping-task strategy to speed up lead time and improve productivity. While there is significance towards efficiency, the effectiveness does not seem to be related. Executing the overlapping-task strategy puts pressure on resources, therefore the positive significance with efficiency does not automatically results in a positive outcome in terms of effectiveness.

For the NPD literature this means that, until now, only ´time to market´ was examined as a dominant performance measure. To split up the performance variable, the relationship between the overlapping-task strategy and efficiency within technological turbulence is empirically tested. Although direct effects are already examined, by means of

one-dimensional meta studies, this research contributes the NPD literature by taking a look at the efficiency. As Fischer (2017) noted, performance is too complex to measure as one. This study has made a begin with empirical findings, by examining two performance measures

efficiency and effectiveness which showed to be differently effected by the flexibility

21

Technological Turbulence x Overlapping x Iterations

Another gap in NPD project literature concerns the interaction between flexibility dimensions and their corresponding variables. This study aims to answers the question imposed by Biazzo (2009), whether the interaction between overlapping-task strategy and design iterations has a positive influence on the NPD performance. In addition, this part will cover the second research question regarding the interaction between flexibility dimensions.

Recent literature points out that concurrent engineering brings a trade-off between time gained and cost caused by rework (Hoedemaker, Blackburn, and Van Wassenhove, 1999; Loch and Terwiesh, 2000; Yassine, Chelst and Falkenburg, 1999). The moderation of technological turbulence is in line with the product architectures imposed by Thomke and Reinertsen (1998). To foresee in future environmental changes, emphasis on inherently flexible product technologies is advised. Iansiti (1995) adds on this by arguing, organizations use a system-focused approach towards technological integration, within turbulent

environments. Under these circumstances organizations should emphasize on design iterations, in order to verify and experiment whether technological changes are necessary.

These findings bring us a step closer in the field of NPD project literature to find the right formula for coping with technological turbulence. Already known in the NPD literature was the danger of additional costs that come with combining the design iterations with

overlapping of task. These empirical findings show the interaction terms between overlapping-task strategy and design iterations are complex. Organizations require to

carefully structure their process in order to benefit from a combination of flexibility measures. As suggested by Biazzo (2009) it is advised to take a contingency approach.This research shows that in technological turbulent environments combining two flexibility measures; overlapping-task strategy and design iterations, can be beneficial to both efficiency and effectiveness of NPD projects.

22 Market Turbulence x Overlapping x Formalization

No support was found for a negative significance of formalization on NPD project performance, within dynamic environments. In contrast with previous research (Fischer, 2017), neither a positive relationship was found. This insignificance could be explained by the balance organizations tend to make, between strict rules and regulation and the freedom and flexibility (Eisenhardt et al., 2010). There should be a certain degree of formalization to make sure there is enough structure. However, the formalization should be just enough in order to not hinder creativity and innovation.

Model E shows an interesting role for formalization in combination with overlapping tasks. The interaction terms of formalization (organizational) and overlapping tasks (temporal) moderated by continuous changes in consumer demands (MT1) is an example of a concurrent engineering. Biazzo (2009), states formal entities within organizations are intended to spot opportunities in the market for implementing recent technological advantages. Therefore, with the moderation of market turbulence an organization can benefit from a degree of

formalization. Guidelines and rules concerning how to coordinate the overlapping tasks within the organization seem logical, especially within turbulent markets. Information regarding customers continuously change their demands (MT1) and their search for new product developments (MT2), need to be communicated throughout the entire organization. This asks for rules and regulation, because rework and failures puts pressure on the efficiency of the product development process.

This study contributes to NPD project literature by empirically showing the role of

formalization. There is discussion whether formalization is positively or negatively related to NPD performance. However, this research shows that formalization can be a way to make another flexibility measure; overlapping-task strategy, work. Although the direct effect of formalization is not significant, the combination with the two other flexibility dimensions simultaneously, is an interesting topic for future NPD research.

23

to further examine the influences of flexibility variables on the effectiveness. To find significances, using an objective performance measured is advised to be able to measure the outcome of an NPD project.

Managerial Implications

Managers can use the findings of this research in their advantage in several ways. They can distinguish different flexibility dimensions and their corresponding variables using the empirical data from this research. Hereby, managers can evaluate the current state of their NPD process and decide whether they see opportunities to improve their NPD project

performance. Taking a contingency approach, means managers need to consider what kind of environment they have to deal (Biazzo, 2009).

Nowadays, both technological and market turbulence are almost always present, making good choices about flexibility measures essential. Within technological turbulent environments, an overlapping of tasks can improve the organizations efficiency by means of information sharing. Furthermore, increasing design iterations in combination with overlapping tasks can help organization to implement technological changes in the NPD process, by catching up with technological innovations. Market turbulent environments characterized by ‘continuous changes in customer demand’ and ‘customers searching for new product developments’, positively moderate the interaction effect of overlapping-task strategy and formalization on NPD project efficiency. Formalization can help in fostering the coordination between overlapping tasks, therefore acknowledging the interaction effect of both organizational and temporal dimension is advised, within market turbulent environments.

Limitations and Future Research

Although this research has a great mix of different sectors, the sample size of 58 firms, is too small to make generalizations. In addition, there was some data missing, thus reducing the statistical explanatory power and the likelihood of finding significant results. Future research about NPD project performance should aim to enlarge the database to run regressions with. As the relation between flexibility and NPD project performance is complex it, would be advised to focus on a specific sector in order to draw conclusions.

24

years ago, can influence the outcome, as they could not recall everything. Future research can take into account objective measures, for example comparing the performance with

competitors or set a deadline for respondents to fill out the survey about past projects. Thirdly, the flexibility dimensions’ corresponding variables are selected based on previous literature. However, by choosing one variable per dimension, there is a lack in the

representation of the entire construct, while running the regression analysis. Also, yet undefined variables could play a role in finding significant results, as the NPD project environments we are now facing are changing rapidly. This means, researchers can use the variables represented in this research and review the steps taken, as a guideline for future research.

Lastly, this research made a start with identifying interaction effects of flexibility dimensions. As the direct effects have different and mostly insignificant influence on NPD project

performance compared to interaction terms, future research can further examine the influence of flexibility measures on all performance variables. After examining the interaction between two flexibility dimensions it could be of significance to look at three dimensions

simultaneously. What is their cumulative influence on NPD project performance? When is it beneficial to be flexible in all dimensions? What is the role of moderating variables such as technological, market turbulence and radicalness? This research can function as a guideline in making choices for future research.

Appendix

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.

Appiah-Adu, K. and Singh, S. (1998). Customer Orientation and Performance: A Study of SMEs. Management Decision 36(6):385–94.

Biazzo, S. (2009). Flexibility, structuration, and simultaneity in new product development.

25

Bodewes, W. (2002). Formalization and innovation revisited. European Journal of Innovation

Management, 5(4), 214-223.

Chakravarty, A.K., (2001). Overlapping design and build cycles in product development.

European Journal of Operational Research. 134 (2), 392–424.

Conway, J. M., & Lance, C. E. (2010). What reviewers should expect from authors regarding common method bias in organizational research. Journal of Business and Psychology, 25(3), 325-334.

Cooper, R. G. (1990). Stage-gate systems: a new tool for managing new products. Business

horizons, 33(3), 44-54.

Cosh, A., Fu, X., & Hughes, A. (2012). Organisation structure and innovation performance in different environments. Small Business Economics, 39(2), 301-317.

Damanpour, F. (1991). Organizational innovation: A meta-analysis of effects of determinants and moderators. Academy of management journal, 34(3), 555-590.

Daniel J. Fernandez & John D. Fernandez (2008) Agile Project Management —Agilism versus Traditional Approaches, Journal of Computer Information Systems, 49:2, 10-17

Eisenhardt, K. M., Furr, N. R., & Bingham, C. B. (2010). CROSSROADS—Micro

foundations of Performance: Balancing Efficiency and Flexibility in Dynamic Environments.

Organization Science, 21(6), 1263–1273.

Eisenhardt, K.M., Tabrizi, B.N., (1995). Accelerating adaptive processes: Product innovation in the global computer industry. Administrative Science Quarterly 40, 84–110.

Galbraith, J.R. (1974) Organization design: an information processing view, Interfaces 4, pp. 28–36.

Griffin, A. and Page, A.L. (1996). PDMA Success Measurement Project: Recommended Measures for Product Development Success and Failure. Journal of Product Innovation

Management 13(6):478–96.

Helms, R., (2004). Framework for releasing preliminary information in product development.

Advanced Engineering Informatics 18, 231–240.

26

Iansiti, M. (1995). Shooting the rapids: Managing product development in turbulent environments. California Management Review, 38(1), 37-58.

Jin, Z. (2000). How Product Newness Influences ‘‘Learning and Probing’’ and the Linearity of Its Development Process. Creativity and Innovation Management 9(1):21–45.

Joglekar, N.R., Yassine, A.A., Eppinger, S.D., and Whitney, D.E. (2001). Performance of Coupled Development Activities with a Deadline. Management Science 47(2):1605–1620

Jørgensen, B., & Messner, M. (2009). Management control in new product development: The dynamics of managing flexibility and efficiency. Journal of Management Accounting

Research, 21(1), 99-124

Kalyanaram, G., & Krishnan, V. (1997). Deliberate product definition: Customizing the product definition process. Journal of Marketing Research, 276-285.

Kamoche, K., & Cunha, M. P. E. (2001). Minimal structures: From jazz improvisation to product innovation. Organization studies, 22(5), 733-764.

Ledwith, A., with O’Dwyer, M. (2008). Product Launch, Product Advantage and Market Orientation in SMEs. Journal of Small Business and Enterprise Development 15(1):96–110.

Lin, J., Chai, K. H., San Wong, Y., & Brombacher, A. C. (2008). A dynamic model for managing overlapped iterative product development. European Journal of Operational

Research, 185(1), 378-392.

Loch, C.H., Terwiesch, C., (1998). Communication and uncertainty in concurrent engineering. Management Science 44 (8), 1032–1048.

MacCormack, A., & Verganti, R. (2003). Managing the sources of uncertainty: Matching process and context in software development. Journal of Product Innovation Management,

20(3), 217-232.

Marchewka, J.T. (2006) Information technology project management: Providing measurable organizational value. 2nd ed. Hoboken, NJ: John Wiley & Sons, Inc.,

Musbah, A., Cowton, C. J., & Tyfa, D. (2016). The role of individual variables,

27

Newell, M. W. & Grashina, M. N. (2003) The project management question and answer book. New York, NY: Amacom

Repenning, N.P., 2001. Understanding firefighting in new product development. The Journal

of Product Innovation Management 18, 285–300.

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.

Shepherd, C., & Ahmed, P. (2000). NPD frameworks: A holistic examination. European

Journal of Innovation Management, 3(3), 160-173.

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

Thomke, S., & Reinertsen, D. (1998). Agile product development: Managing development flexibility in uncertain environments. California management review, 41(1), 8-30.

Van Wyngaard, C. J., Pretorius, J. H. C., & Pretorius, L. (2012). Theory of the triple constraint—A conceptual review. In Industrial Engineering and Engineering Management (IEEM), 2012 IEEE International Conference on (pp. 1991-1997).

Verganti, R. (1999). Planned flexibility: linking anticipation and reaction in product development projects. Journal of Product Innovation Management, 16(4), 363-376.

Veryzer, R.W. (2005). The Roles of Marketing and Industrial Design in Discontinuous New Product Development. Journal of Product Innovation Management 22(1):22–41

Research Measures

The survey is translated from Dutch to English and only items used in the regression analysis are shown.

Dependent variable

 NPD project Performance

28

1. Product development costs (1=significantly worse than initial expectations, 2=worse, 3=somewhat worse, 4=about the same, 5=somewhat better, 6=better, 7=significantly better)

2. Product Quality (same anchor)

3. Technical performance with respect to specifications (same anchor) 4. Time to market (same anchor)

5. Market share (same anchor)

6. Overall profitability of the product? (same anchor) 7. Overall commercial success of the product (same anchor)

Independent variables

The following questions have to do with the formalization in and around the project.

1. Whatever situation arose during the project, written procedures were available for dealing with it. Anchor: 1=completely disagree, 2=disagree, 3=somewhat disagree, 4=neutral, 5=somewhat agree, 6=agree, 7=fully agree.

2. Rules and procedures occupied a central place in the project organization. (same anchor)

3. Written records were kept of everyone’s performance. (same anchor) 4. Project members were hardly checked for rule violations. (same anchor)  Overlapping-task strategy:

Each project consists of several tasks that are executed by different (groups) of people. This question is about the extent the different tasks have been executed simultaneously or

sequentially.

- The different tasks during the project were carried out: 1=fully sequentially, 2=almost sequentially, 3=some overlap, 4=overlap, 5= a lot of overlap, 6=almost concurrent engineering (simultaneous), 7=fully concurrent engineering (simultaneous)

 Design Iterations:

What is the number of iterations during the project (an iteration is a change of more than 10% in a design of one product alternative)

Moderators

 Market Turbulence:

The following statements are about market changes in the environment:

1. In our industry, the needs of our customers are constantly changing. Anchor: 1=completely disagree, 2=disagree, 3=somewhat disagree, 4=neutral, 5=somewhat agree, 6=agree, 7=fully agree.

29  Technological Turbulence:

The following four statements are about technological change in the environment:

1. In our industry technology is changing rapidly. Anchor: 1=completely disagree, 2=disagree, 3=somewhat disagree, 4=neutral, 5=somewhat agree, 6=agree, 7=fully agree.

2. In our industry, technological change offers great opportunities. (same anchor) 3. In our industry, many ideas for new products have emerged from technological

breakthroughs (same anchor)

4. Technological developments in our industry are limited. (same anchor)

Control variables

 Management Support

The following questions concern the supported provided by the management:

1. Goals regarding the project were very explicit formulated. Anchor: 1=completely disagree, 2=disagree, 3=somewhat disagree, 4=neutral, 5=somewhat agree, 6=agree, 7=fully agree.

2. Enough resources were available in supporting the project, making it a success. (same anchor)

 B2B_B2C:

- The ratio of B2B to B2C turnover in the whole business was:

(1)= 0 – 100, (2)= 25 – 75, (3)= 50 – 50, (4)= 75 – 25, (5)= 100 – 0

 Radicalness: