Master Thesis 2015 – 2016

(1)

Master Thesis 2015 – 2016

MSc Business Administration (Strategic Innovations Management)

by

Neli Argirova Marovska

Student number: s2654792

University of Groningen

Faculty of Economics and Business

January 2016

Topic: The impact of information flexibility on the project – level

innovation performance: a meta – analysis

Supervisor: Dr. Hans van der Bij

Co – assessor: Dr. Wim Biemans

(2)

The impact of information flexibility on the project – level

innovation performance: a meta – analysis

Abstract

The dynamic and uncertain business environment in which projects operate has required reactive, adaptable and flexible product and project characteristics and actions for the sake of overcoming the internal and external threats. The ability of projects to change, modify and define their products during the course of the development stages based on the internal and external conditions would significantly impact the performance of not only the product itself, but the way innovative projects perform. Knowing how to ameliorate the performance of projects and the manner in which information flexibility affects them effectively and efficiently is a vital ability in today’s uncertain business environment. Bearing this in mind, a growing number of studies have attempted to shed more light on the relationship between information flexibility and the project – level innovative performance. Therefore, this paper undertakes a meta – analysis in order to provide a more detailed picture on the correlating relationships between the different aspects of information flexibility and the project – level performance based on the existing literature. The results from this study suggest that the most promising moderators that reached homogeneity belong to the “Iterations” and the “Feedback” metafactors. The third metafactor “Frequency” did not provide any homogeneous results.

Introduction

(3)

effective way so as to positively influence the end – performance of their innovative initiatives (e.g. product/service/process innovation practices).

In the existing literature so far, the concept of flexibility has been a research topic, the idea of which is to accommodate activities in the innovation projects under the conditions of dynamic, fast – changing and uncertain environment conditions. The central role of projects’ performance alongside with the impact of flexibility is understandable since the development and the success of the innovative initiatives of a business are heavily dependent on the introduced link and under the mentioned conditions. Thus, the project level could be regarded as the basis for the business activities of a company. Therefore, by focusing on this level, a better and a more concentrated perspective of the functioning of the business is achieved.

According to Biazzo’s article (2009), based on a number of seminal studies carried out by other researches, the definition of flexibility in innovation and new product development is “the ability to embrace environmental turbulence rapidly adapting to new technological or market information that emerges over the course of a project” (Biazzo, 2009). This definition of flexibility is further developed by support practices, which allow flexibility to exist in new product development projects (Biazzo, 2009). According to the article, one of the practices, required by flexibility is the ability to create mechanisms that sense and analyze the market environment in a systematic way, thus, allowing companies to constantly receive market – specific information (Biazzo, 2009). The market information itself would then shed light on the evolution of the needs of customers, thus, providing the designers with insights from the received feedback. This will enable them to see how to change their initial prototypes to better fit the desires of the consumers and the changes in the market / technological environment (Biazzo, 2009). Bearing in mind the article of Biazzo (2009) alongside with other pre and post published studies provides an interesting opportunity for a new research that aims to dwell deeper and shed more light on the impact of flexibility on the project – level innovative performance.

Research objective

(4)

will use the collected articles from the existing literature and through the means of a carried quantitative meta- analysis of the gathered papers will try to fill in the gap in the theory related to information flexibility, i.e. the manner in which the innovation is defined and developed specifically and how and to what extent it affects the project – level innovative performance. Ergo, the main objective of this Master thesis would be to contribute to the existing theoretical field of research on the role of information flexibility and how it impacts the project – level innovative performance. By analyzing the consistency and the nature of the results from the meta – analysis on the basis of the collected literature, more light on the topic would be shed. Through the means of a meta – analysis, the gathered articles are analysed and checked for conclusive results that might occur and help close the topic. Also, by undertaking this research approach, any level of heterogeneity that needs to be further researched till homogeneity is reached is clearly defined. The title proposed for this Master thesis is “The impact of information flexibility on

the project – level innovation performance: a meta – analysis”. This study will question the

level of strength or weakness of the mentioned impact in the relationship between information flexibility and project – level innovation performance. The received results from the calculated metafactors and their potential moderators will allow us to draw conclusive statements, when possible, thus, suggesting closure on some of the theoretical issues and debates in regards to this topic. Moreover, through this meta – analysis the inconclusive results would be clarified and discussed. This will lead to uncovering any potential future theoretical research objectives that need attention in order for all of the information flexibility and project – level performance aspects to be fully covered and debated.

Literature Review

Project performance and evaluation

(5)

On the other hand, Lim and Mohamed (1999) present two views on project performance success factors: the macro and the micro viewpoints. The first viewpoint is concerned with macro factors that influence the success or failure of a project (e.g. economy, supervision, etc.) (Lim & Mohamed, 1999). The micro viewpoint, however, regards the set of micro completion criteria (e.g. time, cost, quality, performance, safety) as the completion criteria, which determines the project success (Lim & Mohamed, 1999). The completion criteria of the micro viewpoint are influenced by a list of factors, such as: technological, commercial, finance, human, environment, risk, etc. (Lim & Mohamed, 1999).

For the purposes of this meta – analysis, the dependent variables for the project – level innovation performance would be defined and extracted from the gathered literature, based on their correlating link to the independent variables, so that the locus of the study is preserved.

Flexibility in terms of efficiency

Flexibility, defined as “the capacity to adapt” or the ability “to accommodate change with minimal degradation of performance”, has become more and more important (Golden & Powell, 2000). In the paper of Golden and Powell (2000) it is argued that one of the most important measures of flexibility is efficiency or “the ability to maintain efficiency while accommodating or adapting to change”, which in the case of internal resources would relate to the extent to which they are efficiently improved by flexibility and thus, can positively influence the performance of projects (Golden & Powell, 2000).

Flexibility in NPD projects and the need to change during the course of the project

(6)

throughout the development process, the uncertainty gradually decreases because of problem solving and knowledge acquisition, the time spent and the cost of unforeseen reworks, either external or internal, increases in the later phases of the project (Biazzo, 2009). Thus, the idea of design flexibility presented by Thomke (1997) can be defined as the “function of the incremental cost and time of modifying design choices: The higher the cost and time of modifying a design, the lower the design flexibility”, ergo, high design flexibility in projects leads to a better performance as it reduces time and cost (Thomke, 1997). Moreover, the “high flexibility derived from the use of flexible technologies” (Thomke, 1997) and the adoption of flexible technologies is one of the major strategies for increasing the flexibility in new product development projects so as to reduce the time and cost to modify the project design (Thomke and Reinertsen, 1998). Furthermore, Iansiti (1995) argued that development models, which possess flexibility, would try to delay their product definitions. In addition, the early freezing of their product concept and the stage overlap in NPD projects in an attempt to surpass the concurrent engineering (Iansiti, 1995). To reduce the effect of the costly iterations due to some unpredictable modifications, an attempt to analyse and foresee arising opportunities and constraints and to “freeze” the product or process definition at the beginning of the project is a way to deal with the issue of the high spending on multiple iterations (Souder & Moenaert, 1992). Another approach, close to the first one, would require a sharp and early product definition strategy (Kalyanaram & Krishnan, 1997). However, Verganti (1999) argues that in terms of NPD, reaction is required. Through an efficient and rapid manner of introducing modifications during the course of the later stages of the project, the definition freeze milestone will be placed as close to the market launch phase as possible (Verganti, 1999). Under these conditions, flexibility would be allowed to occur (Verganti, 1999). Bearing this in mind, design flexibility could be regarded as a function of incremental time and cost, which modifies the design choices. Hence, it is argued that higher levels of time and cost, needed to modify a design would present lower design flexibility and vice – versa (Thomke, 1997). However, design flexibility as defined by Thomke (1997) is concentrated more on the potential aspect of the information flexibility, which the projects possess. Instead, for the purposes of this study the focus will be on the flexibility, which arises and is necessary upon receiving and obtaining new information as discussed in Biazzo (2009).

(7)

which represent slightly different, yet complementary, aspects of information flexibility in regards to the project – level innovative performance (see Appendix 2).

Iterations

(8)

The article of Terwiesch and Loch (1999) builds upon the description of multiple iterations provided by Eisenhardt and Tabrizi (1995). According to the latter article, for the reduction of time – to – market in mature and stable mainframe microcomputer industries, shortening and overlapping of the activities and rewarded developers, is the answer to the desired decrease in the time – to – market (Terwiesch & Loch, 1999). However, in the case of rapidly changing dynamic markets, the compression strategy would fail to provide significant accelerations (Terwiesch & Loch, 1999). Hence, in such uncertain and vibrant markets, Terwiesch and Loch (1999) agree with the experiential strategy proposed by Eisenhardt and Tabrizi (1195), consisting of multiple design iterations and shorter times between milestones (Terwiesch & Loch, 1999). Therefore, it could be argued that the clear distinction between the types of markets and industries is essential in order to undertake an opportunity – creating strategy and not one that might potentially lead to negative performance results and compromise the commercialization of the product. Terwiesch and Loch (1999) also categorize redesign iteration as at least 10% product components modifications, prototyping being a classical example of iteration and debugging not considered as such (Terwiesch & Loch, 1999).

(9)

freedom to change and readjust some of the product’s features for less time and money. Moreover, the idea underlines the benefit of being proactive during the course of the project towards the market and technological turbulence through flexible designs as a means of effectively coping with the environment, neutralizing potential threats and keeping up the level of a competitive new product development.

(10)

are not effectively handled, the project would fail and the company will underperform in regards to its competition in the market. Thus, the unpredictable changes that come up should be treated carefully and accordingly to the early freezed product definition as soon as possible so as not to lose the time – to – market advantage.

Feedback

(11)

the project is developed. Therefore, it is argued that projects, in which the developing team has received earlier new information and feedback on the product performance and features impacts positively the end results of these projects. Ergo, influencing strongly the continuing development activities, would outperform projects, which lack feedback during their development stage and are therefore, more restrained in terms of implementing timely and appropriate readjustments, modifications and changes to their product concepts (MacCormack, Verganti, & Iansiti, 2001).

Lynn et al. (2000) discuss the role of information acquisition in new product development projects as a collection set of primary and secondary data received from a variety of sources, such as customers, competitors, suppliers, markets, etc. (Lynn, Reilly, & Akgun, 2000). The new information acquired enables project teams to form a clearer picture of emerging market needs, competitor’s position and market movements before undertaking necessary, timely and systematic design changes, which influence the end – performance and the success of the new product development projects (Lynn, Reilly, & Akgun, 2000). However, in regards to this meta – factor – i.e. feedback, the information implementation variable will be included in this analysis. This comes as a result from the information acquisition activities, since it represents the ability of teams to implement the information acquired in order to undertake market – strategy – related actions, thus, performing better and faster, reaching new – product timeliness (Lynn, Reilly, & Akgun, 2000). Therefore, in this paper, again the path of the received feedback moves from the external environment into the internal environment of the project, thus, impacting its performance.

(12)

gathering and tracking of customer feedback aiming at using it for continuous process improvement that would be beneficial for the project itself by reducing the chance of late product corrections. In this article, the directionality of the received feedback is from the external environment to the internal one in which the project is developed. Hence, this variable is used in this “Feedback” meta - factor (Bardhan, Krishnan, & Lin, 2013).

Pinto and Prescott (1990) discuss the impact of monitoring and feedback as a tactical dimension, which influences the project success during its lifecycle and its end performance (Pinto & Prescott, 1990). During the course of the project, all of its important aspects are monitored so as to facilitate the following of the project progress in terms of budget and schedule (Pinto & Prescott, 1990). The progress, which would be actually made, would then be regularly compared with the schedule and the results from the project reviews would be shared among all team members. This facilitates and enables the monitoring of the impact of the feedback on the schedule and budget (Pinto & Prescott, 1990). Through regular observations of the project progress, the received feedback could help the development team to improve its practices and readjust them accordingly to the budget and the schedule and to implement modifications when needed in order to enhance the development of the project (Pinto & Prescott, 1990). Here, the benefits from information flexibility in regards to the performance of the project are not so much concentrated on the advantages of a richer choice selection of multiple iterations before choosing the one to be commercialized. The advantages of receiving information during the course of the project would facilitate the improvement of the internal practices and the planning and financial aspect of the project itself, thus, strongly influencing its performance. Therefore, in this study the directionality of the received feedback and its benefits moves from the internal environment to the external one, i.e. the end performance of the commercialized product.

(13)

definition allows the project team to undertake any new changes that occur from the received feedback, thus, allowing for flexibility to take place during the project development stages.

Frequency

Similarly to the characteristics of multiple iterations given by Eisenhardt and Tabrizi (1995), the paper of Callahan and Moretton (2001) presents building frequency as a valuable independent variable. It impacts the project performance, distinguishing the amount of multiple iterations with their frequency, therefore, posing the opportunity to include and analyse it as separate meta - factor. According to the article, the product development process is accelerated through frequent designs, which facilitate the improvement and the ability of the development team “to recognize the requirement for a shift in direction” (Callahan & Moretton, 2001). In the software industry researched by the authors, the spiral model of development was found to be increasingly influential, since it is an iterative approach in which “the design process cycles though the stages of specification development, high level design, detailed design and testing” (Callahan & Moretton, 2001). In particular, load builds facilitate the finding of errors thought the testing of the product. Hence, in the sense of a radically new product development project frequent load builds enable the development team to follow the product’s progress though the feedback received and act accordingly (Callahan & Moretton, 2001). Each load build uncovers new information about the product, thus, giving a better understanding of its features. Therefore, frequent builds would depend on the specific needs of the project and the amount of time necessary to complete the build in a successful manner (Callahan & Moretton, 2001). To further elaborate on the findings of this article, it could be stated that in this case, it is not the specific characteristics of the external environment that would require frequent builds so as to reach success, but the characteristics of the project itself that would demand for more or less frequent builds. Ergo, unlike the articles in the “iterations” meta – factor, which mainly focus on the external aspect, this paper approaches the topic by analyzing the internal characteristics of the project that lead to build frequency as a necessity.

(14)

Sherman, & Davies - Cooper, 1998). The negative impact on performance that design change frequency has in the article is the result from the negative correlation that it has with cycle time, which is important for the successful completion of the project. Hence, it represents a contradictory view in relation to the build frequency variable by Callahan and Moretton (2001), which is positively related to the project performance. Here, the delay in freezing the product definition is regarded as a weakness that compromises the successful performance of the project and is therefore regarded as a negative trait of the project.

Methodology

The meta – analysis is conducted by using the method of quantitative review of the existing academic literature, as we found inconsistent findings in literature (Song, Podoynitsyna, van der Bij, & Halman, 2008). The relationship between information flexibility and the project performance is the focus of the Master Thesis and hence, all gathered articles are therefore, concentrated at analyzing this specific relationship.

(15)

and how these factors affect the project – level performance in relation to the dependent variables.

Data collection

The final paper aimed to collect a range of 8 to 10 articles based on the aforementioned relationship. This is done so that conflicting options are taken into consideration for a broader scope on the relationship between the independent variables and the dependent variables. In this study we found a number of 10 articles (see Appendix 2), divided into 3 meta – factor categories as previously mentioned. The academic papers were targeted in such a manner so as to look for comparability between the articles, so that the meta – analysis could be conducted. Therefore, articles that were researching information flexibility dimensions in relation to the project – level innovation performance were the focus of the article search and collection. The final set of 10 articles proved that they are comparable and therefore, the meta – analysis can be conducted by using them.

(16)

the ten articles (see Appendix 2) that possessed correlation matrixes and Cronbach alphas for both its independent and dependent variables formed the meta – factors in this Master Thesis.

Variables

Ultimately, the Master Thesis provided independent variables, which were related to information flexibility and were placed and discussed under iteration, feedback and frequency categories. Under the “iteration” metafactor, definitions of independent variables such as iterations (Eisenhardt & Tabrizi, 1995) and (Terwiesch & Loch, 1999), flexible project specifications (Candi, van den Ende, & Gemser, 2013), prototype development proficiency (Souder, Sherman, & Davies - Cooper, 1998) and freezing the product design early (Zirger & Harthley, 1996) were gathered. The second metafactor “feedback” consists of the following independent variables - market/technological feedback (MacCormack, Verganti, & Iansiti, 2001), customer feedback as part of process maturity (Bardhan, Krishnan, & Lin, 2013), monitoring and feedback (Pinto & Prescott, 1990) and information implementation (Lynn, Reilly, & Akgun, 2000). The final third metafactor “frequency” includes independent variables such as build frequency (Callahan & Moretton, 2001) and design change frequency (Souder, Sherman, & Davies - Cooper, 1998).

In addition, variables such as development time; cycle time; new product success; project duration; speed – to – market and architecture design effort were extracted from the literature as representatives for the dependent variables. They were included in the three metafactors, i.e. “iteration”, “feedback” and “frequency”, as the role of the dependent variables on the project – level innovative performance was confirmed alongside with their correlating relationship with the independent variables stated in the paragraph above.

Analysis

(17)

the explained variance can be calculated in Excel. Corrections are made for sample size differences and measurement error. The aim of the calculation of the results in Excel would be to determine whether the results are interpretable (i.e. homogeneous) or heterogeneous, in which case the meta – analysis has to be redone on a sub - sample level till homogeneous results are accomplished. The criterion for ‘homogeneity’ for a meta-factor is an explained variance >=75%. Therefore, if an explained variance for a meta – factor is less than the criterion for “homogeneity”, a moderator analysis would be undertaken in search for homogeneous results.

Discussion of the Results

The first metafactor, which was calculated, was named “Iterations” (see Appendix 2). Through the use of the equations, proposed by Hunter and Schmidt (1990) and used in Song et al. (2008), the necessary formulas for the variables in this metafactor were applied. The initial results for the “iterations” metafactor only included the articles in Appendix 2. After the correction for the sample size and measurement size to total variance, variance artifact, new sample size and real variance, the result for the percentage explained is 17,555, which suggests heterogeneity (see Table 1). Due to the fact that the level of heterogeneity was too high and it is difficult to draw overall conclusions, a moderator analysis on a sub – group level for this metafactor was undertaken as stated in the Analysis part of the Methodology section. Moreover, due to the heterogeneous results, it could be argued that the importance of the factor would depend on a situation, hence, the need of a subgroup analysis.

Table 1 Correlation r Sample size Cronbach alpha α Cronbach alpha β

Eisenhardt and Tabrizi (1995) -0,37 72 1 1 Candi et al. (2013) -0,18 132 1 0,8 Terwiesch and Loch (1998) 0,26 140 1 1 Zirger and Harthley (1996) -0,03 44 1 1 Correction for sample size ṝ -0,039484536

Correction for measurement error -0,04055491 Total variance 0,05936159 Variance artifact 4,46185E-06 New sampling error 0,010416616 Real variance 0,048940512

Percentage explained 17,55525397 Heterogenious

Standard deviation 0,221225026 Cronbach alpha √α Cronbach alpha √β Real var. - Stand. Deviation -0,172284514 1 1 Real var. + Stand. Deviation 0,270165538 1 0,894427191

1 1

1 1 VARINDEP VARDEP 1 0,973606798 0,973606798 0 0,002786405 MEANINDEP MEANDEP

(18)

The moderator analysis was conducted based on the directionality of the received information, which influenced the implemented changes. In some of the articles in the “Iterations” metafactor, the view on information flexibility was first discuss as an internal matter, which is necessary for better external adjustments and performance. The only articles, which argued that it is the external environment, which impacts the existence of information flexibility in a particular project, is the article of Terwiesch and Loch (1999) (see Table 2). The moderator analysis, based on information directionality, provide homogeneous results (see Table 2). The results proved that the directionality of the information is an important moderator that needs to be taken into consideration (see Table 2).

In addition, another approach in search for another moderator was undertaken. The next sub – group analysis for the “Iterations” metafactor, was focused on the dependent variable time (see Table 3). In this moderator analysis, the articles were selected based on a common dependent variable – “time” and the paper, which did not enter the analysis was the one of Candi et al. (2013), since its dependent variable was performance. After the moderator analysis was calculated, the results suggested once again a heterogeneous relationship (see Table 3). Which means that the dependent variable time is insignificant moderator for the “Iterations” metafactor, since it failed to reach the >= 75% level to yield homogeneity.

Table 2 Correlation r Sample size n Cronbach alpha α Cronbach apha β

Eisenhardt and Tabrizi (1995) -0,37 72 1 1

Candi et al. (2013) -0,18 132 1 0,8

Zirger and Harthley (1996) -0,03 44 1 1

Correction for sample size ṝ -0,208548387 Correction for measurement error -0,216155087 Total varience 0,01365757 Varience artifact 0,000167477 New sampling error 0,012198617 Real variance 0,001291476

Percentage explained 90,54388062 Homogeneous

Cronbach alpha √α Cronbach alpha √β

1 1

1 0,894427191

1 1 VARINDEP VARDEP

1 0,964809064 0,964809064 0 0,003715206

MEANINDEP MEANDEP

VAR/MEAN VAR/MEAN SUMVAR/MEAN

(19)

However, the first moderator analysis concentrated on the directionality of the received information flow provided high homogeneous results of 90%, thus, confirming the moderator as vital for this metafactor. Ergo, the overall results for the “Iterations” metafactor provide conclusive results in terms of directionality as a moderator and inconclusive results in terms of the dependent variable time.

The second metafactor, called “Feedback” also showed results of heterogeneous correlations (see Table 4).

Therefore, similarly to the first metafactor, it was subjected to sub – group analysis in search for the moderator. The required moderator analysis was conducted on the basis of common dependent variables, i.e. “time” (see Table 5). However, the moderator “time” failed to

Terwiesch and Loch (1998) 0,26 140 1 1

Correction for sample size ṝ 0,03296875 Correction for measurement error 0,03296875 Total varience 0,074539624

Varience artifact 0

New sampling error 0,011857679 Real variance 0,062681945

Standard deviation 0,250363625 Cronbach alpha √α Cronbach alpha √β

Real var. - Stand. deviation -0,18768168 1 1

Real var. + Stand. Deviation 0,313045569 1 1

1 1 VARINDEP VARDEP

1 1 1 0 0

MEANINDEP MEANDEP Moderator analysis of the "Iterations" metafactor.

Performed based on the dependent variable "time". VAR/MEAN VAR/MEAN SUMVAR/MEAN

The article of Candi et al. is excluded, since it's dependent 0 0 0

variable is performance.

Table 4 Correlation r Sample size n Cronbach alpha α Cronbach alpha β

Lynn et al. (2000) 0,42 308 0,9 0,81 MacCormac et al. (2001) 0,248 29 1 1 Bardhan et al. (2013) 0,248 637 1 1 Pinto and Prescott (1990) 0,48 408 0,9 0,87 Correction for sample size ṝ 0,354824891

Correction for measurement error 0,380061327 Total variance 0,011071846 Variance artifact 0,000443229 New sampling error 0,002217851

Real variance 0,008410766 Cronbach alpha √α Cronbach alpha √β Percentage explained 24,03465525 Heterogenious 0,948683298 0,9

1 1

Standard deviation 0,09171023 1 1

Real var. - Stand. Deviation -0,083299463 0,948683298 0,932737905 VARINDEP VARDEP Real var. + Stand. Deviation 0,100120996 0,974341649 0,958184476 0,933599043 0,000877801 0,002510012

MEANINDEP MEANDEP

(20)

achieve the >= 75% homogeneity level, but succeeded to raise the percentage explained number, suggesting that there is a slight moderating effect, which should be further investigated.

The next moderator analysis for this metafactor included the article of Pinto and Prescott (1990) to the ones of Lynn et al. (2000) and Bardhan et al. (2013). The reason for this was that while the dependent variable for Pinto and Prescott (1990) was not precisely “time”, but success, it was used in the meaning of “speed – to – market” and “time” (see Table 6). However, this moderator analysis also failed to reach to homogeneous results and even performed worse that the initial moderator analysis from Table 5.

A third moderator analysis was performed, following the logic of the directionality of the received feedback. The articles selected for this sub – group analysis were gathered on the principle of externally received feedback, which then impacted the creation of multiple iterations, design modifications and later product design definition (see Table 7). The article of Pinto and

Lynn et al. (2000) 0,42 308 0,9 0,81 Bardhan et al. (2013) 0,248 637 1 1 Correction for sample size ṝ 0,304059259

Correction for measurement error 0,328490913 Total variance 0,006499552 Variance artifact 0,000611527 New sampling error 0,001746858 Real variance 0,004141167

Cronabch alpha √α Cronbach alpha √β 0,948683298 0,9

Only the time related dependent variables from the 1 1 VARINDEP VARDEP Feedback meta - factor are included here. 0,974341649 0,95 0,925624567 0,001316702 0,005

MEANINDEP MEANDEP

VAR/MEAN VAR/MEAN SUMVAR/MEAN 0,001351376 0,005263158 0,006614534

Lynn et al. (2000) 0,42 308 0,9 0,81

Bardhan et al. (2013) 0,248 637 1 1

Pinto and Prescott (1990) 0,48 408 0,9 0,87

Correction for sample size ṝ 0,35711456 Correction for measurement error 0,391597783 Total variance 0,011059322 Variance artifact 0,00046698 New sampling error 0,001691561

Real variance 0,008900781

Cronbach alpha √α Cronnbach alpha √β

Standard deviation 0,094343953 0,948683298 0,9

Real var. - Stand. Deviation -0,085443172 1 1

Real var. + Stand. Deviation 0,103244735 0,948683298 0,932737905 VARRINDEP VARDEP

0,965788865 0,944245968 0,911942242 0,000877801 0,002599327 Moderator analysis of the feedback articles,

selected based on the dependent variable (time). MEANINDEP MEANDEP MacCormack is left behind, because the dependent

variable is (quality). VAR/MEAN VAR/MEAN SUMVAR/MEAN

Pinto and Prescott's dependent variable is success, but 0,000908896 0,002752807 0,003661702 as a speed - to - market. Therefore, here included with

(21)

Prescott (1990) was left behind since the directionality of the received feedback was not from the external environment, but from the internal setting in which the project existed. The results of this moderator analysis are promising, since they are the closest to the >= 75% level of homogeneity, which means that the directionality of the received feedback has more significance to the “Feedback” metafactor, than the other analyses of moderators.

Unfortunately, the result from the last moderator analysis is still low to conclude homogeneous relationship and is therefore, inconclusive, but more promising than the preceding analyses. However, a final attempt on this metafactor was conducted, which gathered on a sub – group level the articles of Lynn et al. (2000) and Pinto and Prescott (1990) (see Table 8). The moderator that is defined here and on the basis of which these two articles were combined on a sub – group level is the role of feedback as part of the market – related strategy (Lynn et al., 2000) and the tactics (Pinto and Prescott, 1990) of projects. Both articles define information implementation and monitoring and feedback as parts of the strategy and tactics of projects. The way products are modified, during the development process, based on the received information / feedback impacts the strategy / tactics of the project and therefore, the performance of the project itself.

Lynn et al. (2000) 0,42 308 0,9 0,81

Bardhan et al. (2013) 0,248 637 1 1

MacCormac et al. (2001) 0,248 29 1 1

Correction for sample size ṝ 0,302390144 Correction for measurement error 0,287310336 Total varience 0,006396817 Varience artifact 0,000396973 New sampling error 0,003038149 Real variance 0,002961696

Cronbach alpha √α Cronbach alpha √β 0,948683298 0,9

1 1

1 1 VARINDEP VARDEP

0,982894433 0,966666667 0,950131285 0,000877801 0,003333333 MEANINDEP MEANDEP

(22)

This moderator analysis proved to be successful, as it reached homogeneity, due to the fact that the real variance was a negative number and since it is impossible to be a negative number it was put as a zero. Therefore, the percentage explained was 100%, suggesting that the result is significant (see Table 8). This significance can be also explained by the fact that the confidence interval includes only one point and this point is the corrected correlation coefficient (see Table 8). Hence, it could be concluded that the role of feedback and information implementation as part of the tactics and strategy of projects is a definitive moderator and it impacts the performance.

“Frequency” is the last third metafactor included in this study. The results from the initial metafactor are presented in Table 9. “Frequency” as a metafactor also did not reach the >=75% level so as to conclude homogeneity (see Table 9).

Due to the fact that “Frequency” consists only of 2 articles, a moderator analysis looks impossible. Moreover, the two articles that form the metafactor are researching the same

Lynn et al. (2000) 0,42 308 0,9 0,81

Pinto and Prescott (1990) 0,48 408 0,9 0,87

Correction for sample size ṝ 0,454189944 Correction for measurement error 0,522451395 Total varience 0,000882444 Varience artifact 0,000120636 New sampling error 0,00256779

Real variance -0,001805981 0

Percentage explained - 100 Homogeneity

Standard deviation 0 Cronbach alpha √α Cronbach alpha √β

Confidence interval 0 0,948683298 0,9

0,948683298 0,932737905 VARINDEP VARDEP

0,948683298 0,916368953 0,86934392 0 0,000535885

MEANINDEP MEANDEP

0 0,000584792 0,000584792

Callahan and Moretton (2001) -0,13 44 1 1 Souder et al. (1998) -0,56 101 0,66 1 Correction for sample size ṝ -0,429517241

Correction for measurement error -0,473975206 Total variance 0,039081836 Variance artifact 0,003582235 New sampling error 0,009301603 Real variance 0,026197998

Percentage explained 32,96630796 Heterogenious Cronbach alpha √α Cronbach alpha √β

1 1

0,81240384 1 VARINDEP VARDEP 0,90620192 1 0,90620192 0,01759616 0 MEANINDEP MEANDEP

(23)

dependent variable “time” and they both focus on the internal characteristics of projects that lead to build / design frequency. Hence, a subgroup division cannot be performed.

An attempt to combine the “Frequency” metafactor with the “Iterations” one was attempted. The logic behind this action was the fact that the difference between multiple iterations / prototypes and build / design frequency could be the matter of a wrong interpretation of multiple in terms of numbers and frequent in terms of regular. However, if multiple is similar to frequent, then the combining of the two metafactors would be reasonably backed – up and the semantics of the two words would not play diversifying role. After calculating the united articles from “Iterations” with the ones from “Frequency”, the results still showed a very high level of heterogeneity (see Table 10). Despite the fact that the variance artifact is low, the percentage explained shows that the articles are not combined in the most efficient manner.

To make sure that there were no neglected moderators, this new metafactor was subjected to two additional sub – group analyses. The first moderator analysis was based on the common dependent variable “time” (see Table 11). However, the results from this moderator analysis proved to be inconclusive, since they failed to prove homogeneity.

Callahan and Moretton (2001) -0,13 44 1 1

Souder et al. (1998) -0,56 101 0,66 1

Candi et al. (2013) -0,18 132 1 0,8

Correction for sample size ṝ -0,145590994 Correction for measurement error -0,138557296 Total varience 0,083970992 Varience artifact 0,00016842 New sampling error 0,011374971 Real variance 0,072427601

Cronbach alpha √α Cronbach alpha √β

Real var. - Stand. Deviation -0,196696164 0,81240384 1

1 0,894427191

1 1

1 1 VARINDEP VARDEP

0,968733973 0,982404532 0,951688646 0,005865387 0,001857603 MEANINDEP MEANDEP

(24)

The second sub – group analysis was conducted on the basis of the common direction in which the implementation of the changes to the product definition occurred. In the case of this analysis, the directionality of the implemented changes during the course of the project happened from the internal environment, in which the project was developing and resulted in the performance of the product in the external environment (see Table 12).

Unfortunately, this attempt to find a moderator that leads to homogeneous results was also unsuccessful. Ergo, even with the combination of papers from the two original metafactors, no significant conclusions could be extracted from the results due to the lack of homogeneity.

Callahan and Moretton (2001) -0,13 44 1 1

Souder et al. (1998) -0,56 101 0,66 1

Correction for sample size ṝ -0,134264339 Correction for measurement error -0,139498205 Total variance 0,111094284 Variance artifact 0,000164785 New sampling error 0,012175141

Real variance 0,098754359

Cronabch alpha √α Cronbach alpha √β

Real var. - Stand. Deviation -0,215497704 0,81240384 1

1 1

1 1 VARINDEP VARDEP

Moderator analysis. Without the Candi et al. (2013) article. 0,962480768 1 0,962480768 0,00879808 0 The dependent variables for these articles is (time),

while for Candi was performance. MEANINDEP MEANDEP

0,009141045 0 0,009141045

Callahan and Moretton (2001) -0,13 44 1 1 Souder et al. (1998) -0,56 101 0,66 1 Eisenhardt and Tabrizi (1995) -0,37 72 1 1 Candi et al. (2013) -0,18 132 1 0,8 Zirger and Harthley (1996) -0,03 44 1 1 Correction for sample size ṝ -0,290076336

Correction for measurement error -0,307884888 Total varience 0,034406355 Varience artifact 0,000806947 New sampling error 0,012704763 Real variance 0,020894645

Cronbach alpha √α Cronbach aplha √β

1 1

0,81240384 1

1 1

Iteration and Frequency moderator analysis, based on 1 0,894427191

the directionality of the implemented changes, which 1 1 VARINDEP VARDEP influences the multiple and frequent prototypes. 0,962480768 0,978885438 0,942158408 0,007038464 0,002229124 The directionality is from the internal environment of

the project to the external market/technological environment. MEANINDEP MEANDEP The article of Terwiesch from Iterations is excluded as

(25)

Future Research Directions

Given the end results so far, ideas about future research directions have emerged and could be regarded as an interesting theoretical venture. This paragraph will discuss some of the issues that came up unresolved after all the meta calculation were done.

Throughout the study, the articles that were used in the calculations were all based on product – oriented innovative projects. This is understandable since the majority of the literature, related to projects, performance, flexibility, etc. is concentrated on new product development. Therefore, it would be interesting to analyse how and whether information flexibility impacts service projects in the same manner as product projects and if yes, what are the possible similarities and differences between them. This will give a better and broader perspective on the project performance aspect of the research and place some focus on these projects that deal with intangible activities and how they could be modified by new information and feedback in the course of their formation. Also the nature of the project could serve as a basis for an additional moderator, since it will provide a different setting in which the projects exits and a different manner in which they are organised and managed. Moreover, the differentiation in the type of the project could possible decrease the level of heterogeneity of some of the results of this study and propose interesting insights to be taken into consideration.

(26)

related to the dependent variables that might shed more light on the studied relationship and provide opportunity for better results is “quality”. As stated at the beginning of the Literature Review, time alongside with cost and quality are among the main aspects by which we analyse the project performance. Ergo, it is only logical to suggest that more attention is invested in the last two factors. They will provide a broader and more in - depth aspect of the nature of project innovative performance. By complementing and diversifying the so far gathered information on the analysed relationship, these two additional dependent variables will hopefully have a direct impact on the homogeneity of results, thus, allowing for more conclusive statements to be drawn.

Another promising future research objective that could be of value for this meta – analysis would be to study the impact of information flexibility on the project – level innovative performance in different types of industries. By the latter statement, it is meant that the majority of the articles that research these topics are based on projects that are part of high – tech, computer, and electronics based companies, i.e. high – velocity industries. While, this is understandable given the fact that these companies are usually the most proactive and active in terms of developing innovative projects, it would be interesting to see whether information flexibility has the same impact on the project performance if the project itself is developed in a more stable, moderately competitive and less dynamic business environment.

A final suggestion for a future research that might complement the results of this study and provide an opportunity to reach higher homogeneity among the defined metafactors would be a clear differentiation among radical vs. incremental innovative projects. Again through clearing out the different nature of innovation, a more detailed look at project performance could be reached and whether there is a difference in the impact of all of the aspects of information flexibility on the performance of projects depending on the nature of the innovation. Moreover, this will allow for further attempts on this topic to benefit from more choices to assemble their moderator analysis sub – groups, thus, impacting the end results of the study.

Limitations of the study

(27)

(28)

found by October), the focus of the study shifted to another flexibility dimension – the one of information flexibility.

Conclusion

The topic of information flexibility and how it impacts the project – level innovative performance discusses important aspects of the product development and how it is shaped by different variables, situations, external and internal influences, uncertainty, etc. This study conducted a meta – analysis of ten gathered articles, which were concentrated on the correlation between the different aspects of information flexibility (e.g. iterations, multiple prototyping, feedback, early freezing of the product definition, etc.) and how they impact the performance of projects. While the topic is gaining maturity in terms of the theoretical side of the matter, the results from this meta – study showed that there are no conclusive statements for all of the three metafactors that could be discussed presently. The reasons for the inconclusive results could be due to the relatively small sample as mentioned in the Limitations section or to the relatively similar approach towards and view on the topic in all of the articles. However, one of the most promising metafactors was “Feedback” and its accompanying moderator analysis concentrated on the directionality of the received feedback. The only two homogeneous results in the study were provided by one of the moderating analysis from the “Iterations” metafactor and by one from the “Feedback” metafactor. The reached homogeneity proved that the directionality of the changes and the information is vital when it comes to multiple iterations and design specifications. The second reached homogeneity proved that when feedback and information implementation take part in the strategy and tactics of projects, by leading to product definition modifications and product changes, that would impact the project performance and moderate it. Therefore, future research could place greater focus on the in - depth analysis of these two metafactors and any other undetected moderators and artifacts in order to provide more conclusive results regarding the researched relationship.

(29)

(30)

Bibliography

Bardhan, I., Krishnan, V., & Lin, S. (2013). Team Dispersion, Information Technology, and Project Performance. Production and Operations Management, 22(6), 1478-1493.

Barklay, C., & Osei - Bryson, K. M. (2010). Project Performance Development Framework: An Approach for Developing Performance Criteria & Measures for (IS) Projects. International Journal of Production Economics, 124, 272-292.

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

Callahan, J., & Moretton, B. (2001). Reducing Software Development Time. International Journal of Project Management, 19, 59-70.

Candi, M., van den Ende, J., & Gemser, G. (2013). Organising Innovation Projects under Technological Turbolence. Technovation, 33, 133-141.

Eisenhardt, K., & Tabrizi, B. (1995). Accelerating Adaptive Processes: Product Innovation in the Global Computer Industry. Administrative Science Quaterly, 40, 84-110.

Golden, W., & Powell, P. (2000). Towards a Definition of Flexibility: In Search of the Holy Grail. The International Journal of Management Science, 28, 373-384.

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

Kalyanaram, G., & Krishnan, V. (1997). Deliberate Product Definition: Customizing the Product Definition Process. Journal of Marketing Research, 34(2), 276-285.

Lim, C. S., & Mohamed, M. Z. (1999). Criteria of Project Success: An Explanatory Re - Examination. International Journal of Project Management, 17(4), 243-248.

Lynn, G., Reilly, R., & Akgun, A. (2000). Knowledge Management in New Product Teams: Practices and Outcomes. IEEE Transactions on Engineering Management, 47(2), 221-232.

MacCormack, A., Verganti, R., & Iansiti, M. (2001). Developing Products on "Internet Time": The Anatomy of a Flexible Development Process. Management Science, 47(1), 133-150. Pinto, J., & Prescott, J. (1990). Planning and Tactical Factors in the Project Implementation

Process. Journal of Management Studies, 27(3), 305-327.

Song, M., Podoynitsyna, K., van der Bij, H., & Halman, J. (2008). A Meta - Analysis on the Performance of High - Tech Start - Ups. Journal of Product Innovation Management, 25(1), 7-28.

Souder, W. E., & Moenaert, R. K. (1992). Integrating Markeeting and R&D Project Personel within Innovation Projects: An Information Uncertainty Model. Journal of Management Studies, 29(4), 485-512.

Souder, W., Sherman, D., & Davies - Cooper, R. (1998). Environmental Uncertainty, Organisational Integrations, and New Product Development Effectiveness: A Test of Contingency Theory. Journal of Product Innovation Management, 15, 520-533.

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

Thomke, S. H. (1997). The Role of Flexibility in the Development of New Products: An Empirical Study. Research Policy, 26(1), 105-119.

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

(31)

Wild, J. J., & Wild, K. L. (2012). Globalisation. In J. J. Wild, & K. L. Wild, International Business: The Chanllenges of Globalisation (6th ed., pp. 24-53). Boston: Pierson.

(32)

Appendices

Appendix 1: Provisional Table of Content

Master Thesis 2015 – 2016