The effect of Flexibility on Performance during New Product Development Processes

(1)

The effect of Flexibility on

Performance during New Product

Development Processes

student name: Steven Schoenmaker student number: S2606380

MSc BA: Strategic Innovation Management Supervisor: Hans van der Bij

co-assessor: Wim Biemans

date: 25-06-2018

(2)

2

Abstract

(5)

5

1. Introduction

Innovation is a hot topic. Companies make use of innovation to provide customers with superior value that fits with the firm’s capabilities, which results (if done successfully) in sufficient returns on investments. Innovation can be found in all industries, for example in healthcare. In this industry, an iconic innovation that disrupted the marketplace was the introduction of the MRI-scan. By combining a series of images, or “slices”, taken from many different angles, doctors can examine detailed parts of the body individually or produce a 3-D image of that area, allowing them to identify internal trauma or irregularities quicker and more accurately than before.

Innovation has grown to represent the core renewal process in any organisation. An organisation could risk its existence or growth prospects if it does not change it offerings to the world or the way it creates and delivers those offerings (Bessant, Lamming, Noke & Phillips, 2005).

(6)

6 This research investigates the influence of various elements of flexibility of NPD projects on the performance of those NPD projects, in order to be able to contribute to this area of literature. This is done by a quantitative analysis among several companies. The results of this research answer the question:

What is the relationship between flexibility of NPD processes and the NPD project’s performance in different industries?

(7)

7

2. Literature Review

2.1 Innovation And Its Performance

Innovation can be found in all industries, and its importance is highlighted by multiple researchers. It can contribute to strategic positioning through business model innovation (Amit & Zott, 2012), it can facilitate the conversion of market-oriented business philosophy into superior corporate performance (Han, Kim & Srivastava, 1998), and an innovative corporate culture positively impacts an organisation’s performance (Stock, Six & Zacharias, 2012). Innovation is a critical activity that is vitally important for most firms to embrace in order to create and sustain a competitive advantage. Therefore, innovation processes have become a key to a successful organisation (Graham, 1995). But what is innovation? Innovation can be defined as the “successful introduction of new products and processes” (Hagedoorn, 1996). It does not only focus on the invention of new offerings or processes, but also considers the commercialisation of it. Although its importance is highly recognised, as mentioned previously, the topic of innovation continues to be a point of frustration in many companies. Initiatives to innovate often fail, and successful innovators that previously disrupted a market, regularly have a hard time sustaining their performance, as has been the case for Nokia, Kodak, Blockbuster and many others (Pisano, 2015). The performance of the innovation is dependent on a lot of factors, for example environmental turbulence, degree of collaborations with other organisations, degree of globalisation, the openness of the innovation process, but also complementary internal capacities and practices, and technological acquisitions (e.g. Radas & Bozic, 2009; Laursen & Salter, 2005; Laursen & Foss, 2003; Ahuja & Katila, 2001). Innovation can occur in many ways. An organisation could innovate in for example their processes, distribution channels, customer experience, products or services (Sawhney, Wolcott & Arroniz, 2007). This study will continue to look at the field of product innovation.

2.2 New Product Development

(8)

8 analysing NPD processes, is the Stage-Gate® model (figure 1, Cooper, 1990). It starts with an idea generation, and takes this idea through various phases until it is developed enough to launch it into the market.

Figure 1: The Stage Gate model®. Adapted from “Stage-Gate Systems: A New Tool for Managing New Products” by R.G. Cooper, 1990, Business Horizons 33(3).

As Cooper (1990) explains: “Processes are subdivided into a number of stages or workstations. Between each work station or stage, there is a quality control checkpoint or gate.” The checkpoint, or gate, is what distinguishes this model from general models. At the beginning of the phase, a number of criteria are specified for each phase, and these criteria are checked at the gate. The stages are where the work is done; the gates ensure that the quality is sufficient. After all criteria are met, the concept moves through to the next workstation.

2.2.1 Controversy about the Stage-Gate Model

(9)

9 lot of technological and market information is acquired during the NPD process, rather than before (Iansiti, 1995). This means that the so-called ‘gates’, along with their milestones, move towards the market introduction, or commercialisation phase. This should make companies be able to adapt to rapid changes in the environment, and could adjust their concept before launching the product, Iansiti (1995) concludes. Therefore, a flexible model is introduced, that takes overlap in stages into account. A contradiction arises, whereas the Stage-Gate model by Cooper (1990) is characterised by early and sharp product definition and clear separation between concept development and implementation, while the flexible models make an attempt to delay product definition and seek overlap in stages. This contradiction is underscored by Cooper in one of his later works (2017), which states that the traditional Stage-Gate model can indeed be too linear, too rigid and too pre-planned for today’s turbulent environment and dynamic projects. Therefore, Cooper proposes a new, improved models which allow for flexibility, speed and iterations in a Stage-Gate process: the next-generation idea-to-launch system (Cooper, 2014) and Agile hybrids (Cooper & Sommer, 2018).

The next-gen system is adaptive to its environment, as it has iterative development cycles incorporated into the process. These allows for iterations, testing and changes according to feedback from customers and users (Cooper, 2017). Agile can be integrated to accelerate the product development process (Cooper & Sommer, 2018). Agile addresses collaboration, response to change, and a working product by supporting adaptive planning and delivery through an adaptive approach that focusses on fast and incremental iterations and high levels of communication. Integrating Agile into the NPD process is not without disadvantages though. Agile is a very resource-intensive way of working, it is difficult to keep the working teams connected to the rest of the organisation and there is often too much bureaucracy in place (Cooper & Sommer, 2018).

(10)

10

2.3 Information Processing Theory

Gailbraith (1974) has been one of the first to document information processing capabilities in relation to organisational structures. A basic proposition from this study is that the greater the task uncertainty, the greater the amount of information that must be processed (…) during task execution in order to achieve a given level of performance. In other words, if there is uncertainty prior the execution of certain tasks, for instance the development of a new product, there will be more information needed during the execution of these tasks. Gailbraith (1974) explains further that the strategy that is used will be dependent on the need for information processing capacity relative to its costs. This study has formed the foundation of information processing theory. Tushman and Nadler (1978) build on this theory and developed a model that aims to predict the ‘fit’ between the information processing requirements that organisations face, and the information processing capabilities of said organisations. This model is described in figure 2:

Figure 2 – Information processing model. Adapted and redesigned from “Information processing as an integrating concept in organizational design” by M.L. Tushman & D.A. Nadler, 1978, Academy of management review, 3(3), 613-624.

(11)

11

 Subunit task characteristics, which is divided into task complexity and task interdependence.

 Subunit environment

 Inter-Unit Task interdependence

Task characteristics describe the complexity and intra-unit dependence of tasks. Both are sources of uncertainty, as they limit the predictability and pre-planning of the tasks. The subunit environment consists of external actors that affect and are affected by the organisation. Since these firms lay outside the firm’s locus of control, it is a source of uncertainty. The inter-unit task interdependence describes “the degree to which a subunit is dependent upon other subunits in order to perform its task effectively” (Tushman & Nadler, 1978). These three factors together form the amount of uncertainty that an organisation – or subunit – faces.

2.3.1 Information Processing Capacity

Tushman and Nadler explain that the requirements for information processing capacity are related to this degree of uncertainty. Different organisational structures have different effective information processing capacities (Gailbraith, 1974). A centralised, formalised and high-hierarchical organisational structure (mechanistic) shall have less information processing capacity than a decentralised and destandardised organisational structure with more informal relations (organismic). Organisations that find a ‘fit’ between the information processing capacity that is required and the information processing capacity of its organisational structure, will perform significantly better than organisations that do not find this fit. Having a mismatch could mean that the information processing capacity is either too high or too low for the requirements. Requirements too low lead to an inability to process information to decrease uncertainty. When the information processing capacity is too high, it is more expensive than necessary, since an organismic structure is associated with higher costs (Tushman & Nadler, 1978), due to more coordination costs, for instance.

(12)

12 Agile focusses on the speed and time-to-market (Cooper & Sommer, 2018), while other flexible models focus in adding information during the process (Iansiti, 1995), as does the Next-Gen Stage-Gate model, which allows for iteration and customer feedback during the development (Cooper, 2017). One could argue there is a need for more clarity about the concept of flexibility during NPD processes. Biazzo (2007) obeys this call for clarity, and lays the foundation with a framework regarding flexibility and its dimensions during NPD projects, with which a distinction between the different models and their proposed flexibility arises.

2.4 Flexibility during NPD Projects

Flexibility characterises NPD processes during extremely uncertain and dynamic conditions (Biazzo, 2009). Flexibility in NPD is seen by Iansiti (1995) as the ability to gather, and rapidly respond to, new knowledge about technical and market information as a project evolves. In other words, it is the capacity to quickly embrace turbulence in the environment by adapting to new technological and market information that appears during the NPD process. In order to have a high commercial performance, NPD projects must make a match between organisational capacity and the capacity requirements, according to Information Process Theory.

Since there is indistinctness between flexibility of NPD processes in models as the classic and Next-Gen Stage-Gate model, the Agile hybrids and other flexible model, there is a need for a contingent framework. Biazzo aims to propose such a framework and overcome the dichotomy between rigid and flexible models. Biazzo finds that the classic Stage-Gate model by Cooper – which is characterised by early and sharp product definition – is not necessarily contradicted by flexible models that have overlap in the stages of Cooper’s model. Biazzo identifies dimensions that are affected by the different models. The overlap of models could be related to flexibility, which may provide a contingent approach in the design of new product development processes. Biazzo states that flexibility during the NPD process consists of three dimensions of flexibility:

- Structural dimension - Informational dimension - Temporal dimension

(13)

13 More insights in this field could result in a better understanding of the nature of product innovation processes. The relation between flexibility and the performance of the NPD project is for example still a closed-book.

2.4.1 Structural Dimension

Structural flexibility is concerned with the process structuration and “consists of formal segmentation of the temporal progression in stages by identifying a series of predefined decision-making points”, as well as the definition of the tasks that should be performed and the centralisation of decision-making during each of the phases (Biazzo, 2009). This could be more specifically defined by the degree of formalisation, centralisation and project process structuration. There can obviously be a lot of variety in the structuration of these processes. Some project may know a tremendously detailed process through the identification of microlevel activities, while others may be very limited in the amount of structure. According to the information processing theory, these factors could influence the eventual performance of the NPD project. Tushman and Nadler (1978) namely propose that a better match between the information processing requirements and the capacity of the organisation, as mentioned before, results in better performance. However, since at the start of every NPD process there is always some uncertainty, it is expected that these factors negatively influence the performance of the NPD project. Due to the extensive amount of details each factor has, this study only focusses on centralisation within this dimension. Centralisation is defined by a lot of researchers. One of the most accepted is the definition by Mintzberg (1980), who sees centralisation as the extent to which power over decision making in the organisation is concentrated among its members, which is usually the top level of an organisation. In contingency with the information processing theory and the uncertainty that comes with product innovation through NPD processes, it is expected that centralisation negatively influences the performance of the innovation project. Centralisation would hinder intra-unit communication and is associated with mechanistic structuring. Hence, the hypothesis that is tested by this study is as following:

(14)

14

2.4.2 Informational Dimension

Biazzo divides the activities in the informational dimension into two sections: The problem-formulating activities and problem-solving activities. The formulating activities are used to come up with a product definition. This describes a number of specifications that fulfil the needs of a certain target customer group and the channels to reach this group, such as price, functionality, features and the amount of resources allocated to this project. The problem-solving activities consider the implementation of the previously set definitions into a detailed design. Informational flexibility is not about overlapping decisions, but more about the design of the stages and gates. There could be interaction between those. Flexibility means a high degree of intersection between the product- definition division and the problem-solving division. This means that a concept does not already have all specifications, but it still goes to the detailed designing activities. Then the product concept returns to the problem-defining activities, in order to be able to adapt to changes in the needs of the target group. This way, new information can be added throughout the development process. It is an iterative way of designing, and allows adapting to changes in the environment (Biazzo, 2009). This iterative way of adding information during the development process, could be seen as a delay in the ‘freezing’ or specifying the product definition. By delaying the concept freeze more towards the commercialisation phase, there is the ability for knowledge absorption and adaptation before the gates, if the situation calls for it (Iansiti, 1995). This delay allows for specifications to be added to the product when new information becomes available, so it enables the switching between problem-formulating and problem-solving activities. Hence, an increase of informational flexibility, in the form of a delay of defining of the product specification to a later moment point in time, is expected to have a positive effect on the NPD performance. It allows for information processing during the NPD and therefore it is possible to anticipate on changes in the environment. For that reason, the hypothesis is as following:

H2A: Freezing the product definition at a later point in time during the NPD process, increases the performance of NPD projects.

2.4.3 Temporal Dimension

(15)

15 execution strategies can be distinguished: sequential, overlapped or concurrent execution (Joglekar, 2001). Biazzo explains a noteworthy difference between the temporal and informational dimension, regarding the concepts of overlap and intersection: overlap concerns the simultaneously performance of tasks by different groups, which belongs in the temporal dimension. Whereas intersection concerns the intersection of different tasks (problem-defining or problem-solving) that is switched between. The overlap of activities could result in problems, as “the overlapping of tasks lead to significant risks in that it inevitably generates reworking through engineering change orders” (Terwiesch, Loch & De Meyer, 2002). Therefore, the overlap of tasks is not always the most effective execution strategy. Overlapping reduces NPD lead time more in projects that take place in turbulent environments; consequently, it is better to choose for sequential task performance in more certain environments (Terwiesch & Loch, 1999).

Also the criteria evaluation allows for time-related flexibility. Gates can perform a very strict criteria-evaluation, resulting in for example a Go/No-go decision, or a more flexible criteria-evaluation, resulting in sufficient (conditional) approval to let the concept go through to the next phase (Sethi & Iqbal, 2008). This phenomenon is called gate conditionality. In other words, gate conditionality is defined as the degree to which projects are allowed to proceed further into the process of development conditional on their meeting required criteria at a subsequent stage or that certain project activities are approved out of the usual sequence (Cooper 1994). Sethi & Iqbal (2008) investigated the moderating role of gate conditionality to (in)flexibility of projects. However, they did not find significant results, meaning that this has to be investigated further. It was expected nonetheless that it weakens the relationship between the strict criteria enforcement and inflexibility. Gate conditionality could result in the end in more flexibility, which could lead to better NPD performance. On the other hand, companies may keep investing in a project that will never be feasible. Concluding, there could be both a positive or negative relationship between gate conditionality and NPD performance. However, this could be attributed more to poor risk-management capacities of an organisation, than gate conditionality. A positive relationship between gate conditionality during the NPD process and the performance of the project is therefore expected.

(16)

16 dimension, and not the time order of execution of tasks. Sethi and Iqbal’s research clearly asks for further research on this little explored concept. As explained before, gate conditionality contributes to flexibility of NPD processes by allowing a concept to proceed to the sequential stage, while criteria may not have been met. This facilitates the introduction of new information that may become available in later stages. With this is mind, it is expected that gate conditionally positively influences the performance of NPD projects. Hence, the following hypothesis is formulated:

H3A: Gate conditionality increases the performance of NPD projects.

2.5 Industry Type

The type of industry and its effect on the firms have long been suggested by previous research (e.g. Bain, 1959; Hitt & Duane Ireland, 1985). Industries have been found to differ tremendously from each other (e.g. Porter, 1979; Covin and Slevin 1989; Zahra and Covin 1995; Carrillo, 2005). A lot of research has been conducted about the moderating effects of industry type. However, what the industry type could mean for the amount of flexibility that is needed during NPD processes, is yet to be uncovered. That is the literature gap this paper will try to cover.

(17)

17 So what is considered a professional service firm? Von Nordenflycht (2010) explains that there are firms that characterise on three dimensions. The first is through high knowledge intensity, which means that the output of processes relies fundamentally on complex knowledge in an individual, for example a surgery performed by a surgeon or a lawsuit filed by a lawyer. Secondly, a professional service firm can have low capital intensity, meaning production does not involve tangible assets as inventory or raw material, or even intangible assets as patents. A third characteristic of PSF’s is the professionalised workforce. This means that a particular knowledge base and the regulation and control of it are limited to one specific industry. A profession has a monopoly on the use of that knowledge. Healthcare is again a great example of this, where only hospitals as firms and doctors as professionals are allowed to provide certain services.

Regarding those characteristics, Von Nordenflycht distinguishes four categories of firm types that can be seen as a PSF, but vary in the degree of the characteristics:

- Classic PSF, such as law and accounting firms. - Professional Campuses, such as hospitals.

- Neo-PSF, such as consultancy and advertising firms.

- Technology Developers, such as bio-/medtech and R&D labs.

Due to the limited scope and resources available for this research, only the category Technology Developers will be used. This category is especially chosen out of the four categories, because it is a type of industry that is fast-growing, highly innovative, and intensely competitive. It is characterised by fast access to technology through cooperation with universities, non-profit organisations such as hospitals, and businesses; highly skilled scientists, engineers, and technicians; and good financial support for innovative start-ups through a network of venture capital and corporate partnering relationships (Rochford & Rudelius, 1997).

(18)

18 individuals have strong preferences for autonomy of their work and generally dislike supervision, directive managers and formalisation of the organisational processes (DeLong & Nanda, 2003; Greenwood & Empson, 2003; Lorsch & Tierney, 2002; Starbuck, 1992; Winch & Schneider, 1993). This makes it difficult for managers to steer the projects, and it is compared to “herding cats”. Opaque quality refers to projects that are the output of an expert, but the quality is hard to assess for non-experts, even after production and delivery of the product. This could be the case with for example complex medical technology, where the engineers and end-users or medical specialists have very specific knowledge on the capabilities of the technology, but the manager may not. As a consequence, it is difficult to guide the process for the manager, as the professionals may dislike this guidance and the manager lacks knowledge to evaluate the quality of the product. Concluding, innovation is an important aspect in professional service industries, however, the managing of it are utterly difficult (Izetbegovic et al, 2013; Von Nordenflycht, 2010). Since there is no research yet on these related fields, it is interesting to see how this type of industry relates to flexibility of NPD processes and the performance of the project. Therefore, this study aims to cover this literature gap.

Because of the knowledge-intensity, which is centred around individuals who have specific knowledge on complex situations, it could be assumed that more centralisation and formalisation in the structural dimension of flexibility in NPD could lead to a lower performance. The demands and needs for the customer and the corresponding specifications for the product could be better specified by the executive employees, instead of a manager with a higher hierarchical position. From an information processing theory point-of-view, it is expected that in a technology developer – such as a medtech producer – decentralisation would suit NPD processes more than in other industries. The following hypothesis is formulated:

(19)

19 additions to make the product suit the consumer’s demands more. More flexibility on the informational dimension could be reached by delaying the freezing of the product-definition. This allows for iterations during the process. However, whether industry types have an effect on the flexibility of NPD projects and their performance, has not been investigated yet. Accordingly, it is a gap in the literature that this research tries to cover as well.

Since the professional service industry faces challenges regarding opaque quality, delaying the product definition freezing moment to a later point in time could be beneficial: Managers may not know at the start of the NPD what they are exactly looking for, more than in other industries. This is related to the knowledge-intense characteristic of this industry type. Consequently, they are unable to specify the product to their engineers. When looking at a medtech firm, for instance, there needs to be a lot of cooperation and communication between medical specialists and product engineers, as these are often different people. This is in contrast to engineers of other industries, As explained before, in the health care industry, for example, flexibility in this dimension could mean a better suited product for a patient. Consequently, adding more informational flexibility has a greater impact on NPD projects done in a professional service industry. This leads to the following hypothesis:

H2B: The positive relationship between the freezing of the product definition to a later point in time during the NPD process and the performance of NPD projects is positively moderated for organisations that classify as technology developers.

(20)

20 approaches for capturing the full scope of user’s needs and demands in medical devices. They highlight however that this is an poorly researched field, but repeat the importance of it. the health care sector, along with other PSIs, is seen as utterly complex. Developers often struggle to integrate user requirements that go beyond clinical effectiveness (Martin et al., 2008). Agile development has already proven to be useful in medical device development, where the iterative approach and flexible criteria have led to high quality outcomes (Rottier & Rodrigues, 2008). Since gate conditionality allows for later additions to the product as it has already passed certain phases in the NPD process, it increases flexibility. Therefore, in contingence with the other flexibility dimensions, it is expected that NPD projects that take place in professional service industries benefit more from the positive relationship between gate conditionality and NPD project performance. Hence, the following hypothesis is formulated:

H3B: The positive relationship between gate conditionality and performance of the NPD project is positively moderated for organisations that classify as technology developers.

The conceptual model is described in figure 3.

Figure 3 – Conceptual Model

H1B, 2B, 3B H1A

H2A

(21)

21

3. Methodology

3.1 Data Collection & Sampling

The data is collected from a variety of firms. The search for these firms is done through the database of Bureau Van Dijk, which is also known as the ORBIS-list. This list is ought to have every firm known in the country on it. This list makes it possible to search for specific requirements this study has set on the participating organisations.

The data will be collected through purposive sampling. Numerous Dutch organisations are questioned about their practices through a survey. These organisations ought to have recently launched a new product into the market. This product development process must have taken place (mostly) in-house. Thus, distributors and sole sellers of new products are excluded from this study. Collaboration with competitors, suppliers, buyers or other organisations during the development is naturally allowed. Moreover, the firms selected are required to have more than 30 employees. A firm with a low number of employees would be not suited for this research, as smaller firms often have a more adhocracy structure (Mintzberg, 1980). Because of the literature gap on industry type and its possible moderating effects, 15%-20% of the firms participating ought to be active in a professional service industry. To specify the industry more, this research chooses to only use technology developers, as classified by Von Nordenflycht (2010). Seven researches will contact firms via telephone or email to request participation in this research. This study aims to collect 45 firms in total. 45 samples allows the use of the statistical analyses used in this research. The survey relates to the used constructs and comes from previous literature. The survey is created in, distributed by and collected by one software program, accessible through the university’s account. The survey will be send to recipient via personal email, or the survey will be held telephonically. This research uses a cross-sectional data gathering for both the dependent, independent, moderator and control variables.

(22)

22

3.2 Measurements

This section will explain how many of the questions of the survey are actually used. The questions of the survey can be found in Appendix A. Furthermore, the control variables will be mentioned here. The questions used in this study come from previous research, and are translated from English into Dutch. Most questions used a 7 point Likert-scale (e.g. the items of performance, centralisation and gate conditionality). Furthermore, company characteristics as industry type, size and number of employees are asked. Firstly, a factor analysis will be performed to see if items load their corresponding variable. Results of this will be further explained in Chapter 4. Results.

3.2.1 Performance

In order to measure performance, both the manager and project leader give their perception of the performance. However, since the independent variables will be asked to the project leader as well, the dependent variable data of only the manager is used. This will prevent the common method bias. The questions used for this variables come from the work of Schleimer and Faems (2016); and Ahmad, Mallick and Schroeder (2012).

The factor analysis excluded the first item of Performance, as it loaded multiple factors. Then, with the remaining factors, an internal consistency test is conducted. The performance items 2 till 7 have an acceptable internal consistency of α = 0.77. However, this is not the highest possible internal consistency. If item 4, which regards time-to-market performance, is deleted, the remaining items have an internal consistency of α = 0.79. Therefore, item 4 will be excluded as well. The remaining variables are computed into one performance average.

3.2.2 Centralisation

(23)

23

3.2.3 Freezing

The delay of freezing is only measured with one item. It considers the part of the project that has already been finished, when the product definition was specified. The answer is given in a percentage. The question is adapted from the work of Iansiti (1995)

3.2.4 Gate Conditionality

Gate conditionality refers to the situation where the product concept is allowed to proceed to the next stage, when not all criteria have been met yet. However, on the condition that these criteria will be met in the future. The four questions used in the survey are adapted from the paper of Sethi and Iqbal (2008). The factor analysis results exclude item 4. The residual three items have acceptable, albeit low, internal consistency of α = 0.68. Items 1 to 3 are computed into one average item that represents the variable of gate conditionality.

3.2.5 Industry type

The industry type is asked in the survey of the manager. Since the construct is asked in an open question, contextually the same answers could be stated differently (e.g. healthcare and health care). These answered are manually adapted to the same form for similar industry types. Consequently, there is a distinction made between professional service industries and non-professional service industries. In this data, all professional service firms are positioned within the healthcare industry. Categories of 1 and 0 are made, based on whether the firm is a PSF or not, in order to enable analysing this data. To measure the moderating effect, the mean-centred values of the independent variables are multiplied with the mean-centred values of the industry type. These new values allows for analysis of moderating effects.

3.2.6 Control variables

(24)

24

3.3 Analysis

The analysis of the data is done through the software program SPSS, which is accessible through the university’s portal.

First, a factor analysis is used to describe variability among observed, correlated variables in terms of a potentially lower number of factors. The number of factors is pre-set, namely 3 (for performance, centralisation and gate conditionality) and the items are evaluated based on the following criteria:

- Each measure should have a loading of >0.5. - Each measure should load the correct factor.

- Each measure must not load another factor by more than 0.4

Then, the internal consistency reliability is measured for each of the factors and its corresponding items through the Cronbach’s Alpha. Furthermore, a multicollinearity diagnostic test is conducted and shows there is no case of multicollinearity in the mean-centred regression model, as Variance Inﬂation Factors are all lower than 5.

(25)

25

4. Results

A factor analysis is conducted on multi-item constructs. The final loadings are presented in table 1 below. The question that each item represents can be found in Appendix A.

Table 1

Factor Loadings From Factor Analysis

Performance Centralisation Gate Conditionality

MPerf2 0.57 -0.21 0.2 MPerf3 0.56 0.09 0.16 MPerf4 0.57 0.31 0.28 MPerf5 0.74 -0.19 0.31 MPerf6 0.66 -0.06 0.16 MPerf7 0.80 -0.15 0.03 Central1 -0.10 0.83 -0.02 Central2 0.14 0.64 0.03 Central3 -0.01 0.94 0.05 Central4 0.12 0.88 -0.21 Central5 0.00 0.87 0.30 GateCon1 -0.18 -0.20 0.60 GateCon2 -0.32 -0.13 0.78 GateCon3 -0.39 -0.12 0.71

NOTE: the items MPerf1 and GateCon4Reversed have been deleted from the table, as they loaded no or multiple factors.

(26)

26

Table 2

Descriptive Statistics. Correlations and Cronbach’s Alpha

Variables 1 2 3 4 5 1. Performance 0.79 2. Centralisation -0.16 0.91 3. Freezing 0.18 0.12 - 4. Gate Conditionality -0.09 -0.05 0.03* 0.68 5. Industry Type 0.01 0.05 0.18 0.21 - Mean 4.72 2.67 41.56% 5.12 0.27 Standard Deviation 0.81 1.35 31.87% 0.94 0.45 *p<0.05

In order to analyse the moderating effects, all independent and moderator variables are mean-centred, as is recommended by Aiken and West (1991) for testing interaction effects. New variables are computed by multiplying the mean-centred independent variable and mean-centred moderator variable (indicated by the mc in table 3 below). To test the hypotheses proposed in the theoretical framework, a hierarchical regression analysis is used. Table 3 below shows the results of the hierarchical regression analysis with performance of the NPD project as dependent variable. Model 1 includes the four control variables. It has a R² of 0.28 and the F value (2.76) is significant (p<0.05). Model 2 includes the control variables, plus the independent variables centralisation, freezing and gate conditionality. Model 2 has a R² of 0.45 and the F value (2.94) is significant as well (p<0.05). Model 3 regards the moderator effects, and includes the control variables, the mean-centred independent variables centralisation-mc, freezing-mc and gate conditionality-mc. Furthermore, the mean-centred moderating effect industry type multiplied by the mean-centred independent variable is included in model 3 as well. The R² of model 3 is 0.55 and the F value (2.31) is significant (p<0.05).

(27)

27 albeit significant. In both model 2 and 3, the effect of gate conditionality is significant, even though it is an opposing effect of what was expected. The moderating effect of the industry type is included in model 3. As can be seen in table 3, both moderating effect of industry type on the relationship between centralisation and performance and freezing and performance are insignificant (p>0.1). H1B and H2B are not supported. The moderating effect of industry type on the relationship between gate conditionality is positive and significant (p<0.1). H3B is thus supported.

Table 3

Hierarchical regression results

Model 1 coefficient Model 2 coefficient Model 3 coefficient Control Variables B2B/B2C -0.30 -0.33† -0.30† Radicalness -0.36 -0.40† -0.36† Employees (Log) 0.45† 0.46 0.45 Revenue (Log) -0.27 -0.30 -0.27 Main Effects Centralisation -0.27 -0.26 Delay of Freezing 0.31† 0.32 Gate Conditionality -0.31† -0.31† Moderator

Industry Type (mc) x Centralisation (mc)

0.25

Industry Type(mc) x Freezing (mc) -0.25

Industry Type (mc) x Gate Conditionality (mc)

0.50†

R² 0.28* 0.45† 0.55

F 2.76* 2.94* 2.31*

(28)

28

5. Discussion

This study investigated the effect of different flexibility dimensions, classified by Biazzo, during NPD processes, on the performance of that NPD project. The results discussed in the previous chapter, enable us to answer the research question this study proposed: What is the relationship between flexibility of NPD processes and the NPD project’s performance in different industries?

The empirical results show that a delay in freezing the product definition to a later moment in time has a positive effect on the performance of the product innovation. This provides evidence that flexibility on the informational dimension of NPD processes is beneficial for the outcomes of the project. This is in contingence with the information processing theory and the newer models of Stage-Gate product development processes. The delay of freezing the product definition enables the project to adapt to new information and/ or changes in customer preference. Moreover, it allows for iterations to be made to the project, in accordance with the newly available information.

Furthermore, openness to interpret criteria and the status quo, and allowing criteria to be met in later stages have been proven to negatively influence the performance of the NPD project. Thus, temporal flexibility decreases the performance of product innovations. This is contrary to the expectations, which are based on Sethi and Iqbal’s (2008) study. They propose after all that rigorous gate controls – where no permission is granted to projects to continue to the next stage – mitigate learning effects. Those effects are seen as critical for effective new product development. When their study is combined with the findings of this paper, there could be advocated for an underlying cause. A reason could be that gate conditionality actually shows a curvilinear (inverted U-shape) effect in relation to the performance of NPD projects, although this has not been tested in this research. Too little rigidity among criteria for milestones may mitigate the initial benefit of this flexibility dimension, such as adaptation to the environment. More gate conditionality than necessary could result in biased decision making, since criteria are not taken strict anymore. The organisation could be flogging a dead horse; the firm may keep investing in a project that will never be feasible.

(29)

29 interpreted with caution, as only centralisation is measured as construct for this dimension, and not the formalisation or the amount of structuration of the development process, for instance. In other words, centralisation does not seem to have an influence on the performance of product developments. One of the explanations for this finding could come from knowledge-based theory: centralisation means the decision making is centred in one individual or one organisational level. If this individual or unit has excellent managerial capacities, centralisation does not have to be a bad thing. This manager may be very well suited to make calls regarding important decisions, even though it may impair the flexibility of the project. These managing skills could lead to equally good performance outcomes as a flexible structured project where the decision-making is decentralised.

All in all, the flexibility of NPD processes have been proven to have mixed effects on the performance of the NPD project. Whereas flexibility in the informational dimension has a positive influence on the performance, the flexibility in the temporal dimension has a negative influence on the performance. The structural dimension does not seem to have an effect on the performance at all. Hypotheses are only partly supported.

When looking at the influence of the industry type on the previously described relations, only one moderating effect seems to be present. The negative relationship between gate conditionality and performance is stronger in technology developers. This means that being more flexible in the evaluation of criteria at gates between stages, results in even less performance when the firm is defined as a technology developing firm.

5.1 Theoretical implications

(30)

30 as well. Technological developers as classified by Von Nordenflycht perform lesser when they use more flexibility in the evaluation of gate criteria, in other words, use more gate conditionality.

5.2 Managerial implications

Managers should consider the uncertainty regarding their NPD projects and the needed flexibility. As explained before, a centralised structure may be beneficial in coherence with very capable managers. However, when this coherence is not there, a decentralised structure with authority in knowledgeable employees could be better suited. This study also shows that a delay of freezing the product definition to a later moment in time, could be beneficial for the project’s performance. Blindly following a strategy where the product definition is only defined at the end, as a consequence of this study, would be foolish. Managers should consider the amount of information available at the start and based on the knowledge flows at a later stage in the project, for the defining moment of the product specifications. A product that is not defined at the latest stage may be difficult to promote, as buyers could not fully know what they will buy. Therefore, caution is needed in interpreting the result and implementing such a strategy in the NPD process. Managers should also consider the amount of gate conditionality they use. Perhaps, some projects would be better off when they would be stopped, in relation to the amount of resources they require. Stricter enforcement of criteria and milestones could aid in this. However, as will be explained in the next section, gate conditionality should be further investigated to fully understand how this construct relates to the performance.

(31)

31

5.2 Limitations and future research

This study has been conducted to best efforts. However, it knows several limitations. One major flaw is the relatively small sample size of less than 50 organisations. The research team has found great difficulty in convincing organisations that are involved in NPD projects, to participate in this study. Therefore, the sample is perhaps not homogeneous and big enough, which makes it hard to generalise the results to practice. Secondly, the dimensions of flexibility have only been measured on one construct each. For example, centralisation is used to measure structural flexibility. However, formalisation could for example also be part of this dimension, but is not examined in this study. The scope of this study is limited due to the amount of time available. Therefore, it is uncertain that the informational dimension, which is measured by the freezing point, will always be positively related to the performance. Factors as the number of iterations and the number of changes according to feedback from customers could also be seen as flexibility of the informational dimension, but at this point it is unclear how these would relate to performance. Thirdly, the professional service industries have now been represented by organisations that are positioned in the healthcare industry. Therefore, generalisation towards any other professional service industry should be met with caution, as it may be different in for example the accountancy industry. Furthermore, the lack of significant moderating effects of industry type could be related to the heterogeneous firms sample. The organisations are classified according to healthcare industry or non-healthcare industry. But even within the healthcare industry-typed sample, there is a lot of variation in type of firm. Medtech conglomerates, healthcare purchasers, academic hospital spin-offs and small medtech start-ups are all within the healthcare class. The independent variables may differ too much within this group already in order to gain clear insights in the moderating effects of industry type. This heterogeneity would not matter in an enormous sample size. This class, however, only contained 11 firms.

(32)

(33)

33

6. 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. Sage.

Aldrich, H. E. (2008). Organizations and environments. Stanford, CA: Stanford Business Books.

Ahuja, G., & Katila, R. (2001). Technological acquisitions and the innovation performance of acquiring firms: A longitudinal study. Strategic management journal, 22(3), 197-220. Amit, R., & Zott, C. (2012). Creating Value through business model innovation. MITsloan Management Review, 53(3), 40-50.

Bain, J. S. (1968). Industrial organization. New York: Wiley.

Bessant, J. and Francis, D. (2004). Developing Parallel Routines for Radical Product Innovation. Paper presented at the Proceedings of the 11th International Product Development Management Conference, Dublin, Ireland.

Bessant, J., Lamming, R., Noke, H., & Phillips, W. (2005). Managing innovation beyond the steady state. Technovation, 25(12), 1366-1376.

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

Carrillo, J. E. (2005). Industry Clockspeed and the Pace of New Product Development. Production and Operations Management, 14(2), 125-141.

(34)

34 Cooper, R.G. (1990). Stage-Gate Systems: A New Tool for Managing New Products. Business Horizons 33(3):44–54.

Cooper, R. G. (2008). Perspective: The stage‐gate® idea‐to‐launch process—update, what's new, and nexgen systems. Journal of product innovation management, 25(3), 213-232.

Cooper, R. G. (2014). What's Next?: After Stage-Gate. Research-Technology Management, 57(1), 20-31.

Cooper, R.G. (2017) Idea-to-Launch Gating Systems: Better, Faster, and More Agile, Research-Technology Management, 60:1, 48-52,

Cooper, R. G., & Sommer, A. F. (2018). Agile–Stage-Gate for Manufacturers: Changing the Way New Products Are Developed Integrating Agile project management methods into a Stage-Gate system offers both opportunities and challenges. Research-Technology Management, 61(2), 17-26.

Covin, J. G., & Slevin, D. P. (1989). Strategic management of small firms in hostile and benign environments. Strategic management journal, 10(1), 75-87.

Davis, M.S. (1992) Health Care: Innovation, Impact and Challenge. School of Policy Studies-School of Public Administration, Queens University.

DeLong, T., & Nanda, A. (2003). Professional services: Text and cases. New York: McGraw-Hill/Irwin.

Downey, H. K., Hellriegel, D., & Slocum, J. W. (1975). Environmental Uncertainty: The Construct and Its Application. Administrative Science Quarterly, 20(4),

Duignan, J.quantitative research methods. In A Dictionary of Business Research Methods. : Oxford University Press.

Galbraith, J. R. (1974). Organization design: An information processing view. Interfaces, 4(3), 28-36.

(35)

35 Graham, N. (1995). Quality in health care: theory, application, and evolution. Gaithersburg: Aspen

Greenwood, R., & Empson, L. (2003). The professional partnership: Relic or exemplary form of governance?. Organization studies, 24(6), 909-933.

Hagedoorn, J. (1996). Innovation and Entrepreneurship: Schumpeter Revisited. Industrial and Corporate Change, 5(3), 883-896.

Han, J. K., Kim, N., & Srivastava, R. K. (1998). Market Orientation and Organizational Performance: Is Innovation a Missing Link? Journal of Marketing, 62(4), 30.

Hitt, M. A., & Ireland, R. D. (1985). Corporate distinctive competence, strategy, industry and performance. Strategic Management Journal, 6(3), 273-293.

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

Izetbegović, S., Stojkanović, G., Švrakić, S., & Mehmedbašić, E. (2013). Innovating in Health Care – Modern Challenges. Journal of Health Sciences, 3(3)

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.

Jaworski, B., & Kohli, A. (1993). Market orientation: Antecedents and consequences. Journal of Marketing, 57(3), 53-53.

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.

(36)

36 Laursen, K., & Foss, N. J. (2003). New human resource management practices, complementarities and the impact on innovation performance. Cambridge Journal of economics, 27(2), 243-263.

Laursen, K., & Salter, A. (2005). Open for innovation: the role of openness in explaining innovation performance among U.K. manufacturing firms. Strategic Management Journal, 27(2), 131-150.

Loch, C. (2000). Tailoring Product Development to Strategy: Case of a European Technology Manufacturer. European Management Review 18(3):246–258.

Lorsch, J., & Tierney, T. J. (2002). Aligning the stars: How to succeed when professionals drive results. Harvard Business Review Press.

March, J. G., & Simon, H. A. (1993). Organizations. Cambridge, Mass: Blackwell.

Martin, J.L., Norris B.J., Murphy E., Crowe, J.A. (2008) Medical Device Development: The Challenge for Ergonomics, Applied Ergonomics 39 (3), 271-283

Mintzberg, Henry. “Structure in 5s: A Synthesis of the Research on Organization Design.” Management Science, vol. 26, no. 3, 1980, pp. 322–341.

Nordenflycht, A. V. (2010). What Is A Professional Service Firm? Toward A Theory And Taxonomy Of Knowledge-Intensive Firms. Academy of Management Review, 35(1), 155-174.

Phillips, W., Noke, H., Bessant, J., and Lamming, R. (2004). Beyond the Steady State: Managing Discontinuous Product and Process Innovation. AIM Research Working Paper no. 9, Advanced Institute of Management Research (AIM).

Pisano, G. P. (2015). You need an innovation strategy. Harvard Business Review, 93(6), 44-54.

Porter, M. E. (1979). How competitive forces shape strategy.

(37)

37 Radas, Sonja & Božić, Ljiljana. (2009). The Antecedents of SME Innovativeness in an Emerging Transition Economy. Technovation. 29.

Rochford, L., & Rudelius, W. (1997). New product development process: Stages and successes in the medical products industry. Industrial Marketing Management, 26(1) Rottier, P. A., & Rodrigues, V. (2008). Agile development in a medical device company. In Agile, 2008. AGILE'08. Conference (pp. 218-223)..

Sawhney, M., Wolcott, R., & Arroniz, I. (2007). The 12 different ways for companies to innovate. IEEE Engineering Management Review, 35(1), 45-45.

Schleimer, S. C., Faems, D. (2016). Connecting Interﬁrm and Intraﬁrm Collaboration in NPD Projects: Does Innovation Context Matter? Journal of product innovation fmanagement, 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.

Span, K. (2011). Het belang en de reikwijdte van innovaties in de gezondheidszorg. Tijdschrift voor gezondheidswetenschappen, 89(2), 91-92.

Sperry, R and A Jetter (2009). Theoretical Framework for Managing the Front End of Innovation under Uncertainty, PICMET 2009 Proceedings, Portland, USA.

Starbuck, W. H. (1992). Learning by knowledge‐intensive firms. Journal of management Studies, 29(6), 713-740.

Stock, R. M., Six, B., & Zacharias, N. A. (2012). Linking multiple layers of innovation-oriented corporate culture, product program innovativeness, and business performance: a contingency approach. Journal of the Academy of Marketing Science, 41(3), 283-299. Takiah M. Iskandar, (1996) "Industry type: a factor in materiality judgements and risk assessments", Managerial Auditing Journal, Vol. 11 Issue: 3, pp.4-10

(38)

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

Thompson, J. D. (1967). Organizations in action: Social science bases of administrative theory. New York: McGraw-Hill.

Tushman, M. L., & Nadler, D. A. (1978). Information processing as an integrating concept in organizational design. Academy of management review, 3(3), 613-624.

Winch, G., & Schneider, E. (1993). Managing the knowledge‐based organization: the case of architectural practice. Journal of management studies, 30(6), 923-937.

(39)

39

Appendix A – Survey Questions

Construct Item

number Question (in Dutch) Answer

Performance Als u de uitkomsten van het project vergelijkt met de verwachtingen bij aanvang van het project, hoe heeft het project dan gepresteerd op:

1 Productontwikkelingskosten Likert scale 1-7

2 Productkwaliteit “”

3 Technische prestatie m.b.t. productspecificaties “”

4 Tijd tot marktintroductie “”

5 Marktaandeel “”

6 Winstgevendheid van het product “”

7 Commercieel succes van het product “”

Centralisation De volgende vijf stellingen hebben te maken met de centralisatie van het project.

1 Men kon weinig doen tijdens het project zonder dat een

leidinggevende ergens over moest beslissen. Likert scale 1-7 2 Als iemand zelfstandig een beslissing wilde nemen

tijdens het project, werd dit ontmoedigd. “” 3 Zelfs kleine beslissingen moesten worden voorgelegd

aan een leidinggevende. “”

4 Projectleden moesten voor vrijwel alles vooraf

toestemming vragen aan een leidinggevende. “” 5 De meeste beslissingen van de projectleden moesten

worden goedgekeurd door hun leidinggevende. “” freezing the

product specification

1 Welk gedeelte van het project was al gereed toen de

productdefinitie vastgesteld werd (in %)? numeric value Gate

Conditionality De volgende vier stellingen hebben betrekking op de beslissingscriteria die gebruikt zijn bij de diverse mijlpalen.

1 Bij beoordelingen bij de mijlpalen in het project was het geoorloofd dat aan bepaalde criteria nog niet voldaan was, onder de voorwaarde dat hier later wel aan zou worden voldaan.

Likert scale 1-7

2 De beoordelingscriteria lieten toe dat het project een andere volgorde van fasen doorliep, waardoor het mogelijk was door te gaan met het project terwijl nog niet aan alle criteria werd voldaan.

“”

3 De beoordelingscriteria lieten toe dat verschillende ontwikkelingsactiviteiten goedgekeurd werden op een ander moment dan gepland.

“” 4 Bij elke mijlpaal moest aan alle projectcriteria voldaan

worden voordat het project door mocht naar de volgende fase.

“”

Industry Type 1 In welke bedrijfstak(ken) bevindt het bedrijf zich? text

Revenue 1 Wat is de grootte van uw bedrijf, gemeten in omzet? (in

euros) numeric value

Firm Size 1 Wat is de grootte van uw bedrijf, gemeten in het aantal

werknemers? numeric value

Technological

Radicality 1 Tijdens het project werd gebruik gemaakt van technologie die volledig nieuw is voor het bedrijf. likert scale 1-7

B2B/B2C ratio 1 De verhouding tussen B2B en B2C omzet in het gehele

bedrijf was: 5 categories ordinal

(40)

40

Appendix B – Factor analysis, internal consistency and computed

variables

Factor analysis – component matrix:

Loading Cronbach’s Alpha of all included items

Chronbach’s Alpha if deleted M_Perf1* M_Perf2 0.58 0.73 M_Perf3 0.52 0.74 M_Perf4 0.45 0.79** M_Perf5 0.66 0.70 M_Perf6 0.66 0.74 M_Perf7 0.76 0.77 0.71 Central1 0.81 0.89 Central2 0.52 0.91** Central3 0.91 0.84 Central4 0.85 0.86 Central5 0.82 0.89 0.87 Gatecon1 0.53 0.66 Gatecon2 0.54 0.38 Gatecon3 0.48 0.68** 0.66 Gatecon4* -0.59

*Excluded from next analyses

The effect of Flexibility on Performance during New Product Development Processes