BIG DATA WITHIN INNOVATION PROCESSES

(1)

BIG DATA WITHIN INNOVATION PROCESSES

____________________________________________________________

A Quantitative Study of Fortune 200 Global Companies

SOFIE E. A. PERSSON

Supervisors: Prof. Johan Brink & Prof. Richard Tee Innovation- & Industrial Management

Masters Degree Project Graduate School

(2)

BIG DATA WITHIN INNOVATION PROCESSES By Sofie Persson

© Sofie Persson LUISS Guido Carli, Viale Romania 32, 00197 Roma RM, Lazio, Italy All rights reserved. No part of this thesis may be reproduced without the written permission by the authors Contact: sofie.persson.mail@gmail.com

(3)

Abstract

Essay/Thesis: Master’s Degree Thesis / 30 ECTS

Program: MSc Innovation- & Industrial Management

Level: Second Cycle

Semester/year: Spring/2018

Supervisors: Prof. Johan Brink & Prof. Richard Tee

Keyword: Big data, big data analytics, market orientation, innovation processes, pre-development activities, innovative productivity

Background: The importance of innovation simply continues to stagger. Albeit, many firms continue to struggle in the launch of new technologies that meet the objectives of successful innovation. As a result, the life cycles of innovations are shortened and inadequate returns endure. Current research suggests modern technologies of big data can facilitate businesses as they strive to generate value-generating innovations. However, few explain what firms actually aim to find and how the extracted information supports activities within innovation.

Purpose: The purpose of this study is to investigate firms’ usage of big data to facilitate market orientation related to the innovative process, while assess the impact such efforts have on firms’ innovative output.

Theory: Shaping the research are the fields of: proactive- and responsive market orientation (Narver & Slater, 1990), pre-development theory inspired by the stage-gate framework and innovative productivity, as described by Kim and Mauborgne (2004). The latter enacts the comparative measure of innovative performance within this report, since it represents both the speed of innovative processes whilst simultaneously represents value-generating output of innovation.

Method: This study is conducted through a deductive approach to research. Four hypotheses were constructed for testing against the gathered findings. With a cross-sectional design, an online questionnaire was distributed via LinkedIn to the Fortune 200 Global Companies of 2017 to collect quantifiable data. To test the hypotheses five main concepts were constructed on the basis of previous theory and the operationalization within adjacent research.

Result: Given the four originally stated hypotheses due for assessment, three were accepted. Consequently, the findings suggest: (1) big data adoption; big data analysis within proactive market-oriented activities; and big data derived customer insight used within pre-development activities all demonstrate positive and significant correlations with innovative productivity. (2) The studied population preferably pursues market-oriented activities to extract and assess the latent needs of customers vis-à-vis expressed needs.

(4)

Thank You

Given this report, the process has not always been as straightforward as one may have wished. With that said, I would personally like to thank my supervisors, Prof. Johan Brink and Prof. Richard Tee who supported me with constructive feedback along the way. I would also like to extend my gratitude to all the respondents who made this report possible by taking the time to conduct the online questionnaire. But most of all, I would like to give a special thank you to my friends who enacted sounding boards from start to finish, I owe this to you.

(5)

List of Tables & Figures

FIGURE 1: FORTUNE 1000 REPORT BIG DATA INITIATIVES & SUCCESS RATE ... 6

FIGURE 2: TYPOLOGY OF INNOVATIVENESS ... 10

FIGURE 3: SEQUENTIAL VERSUS PARTIALLY PARALLEL INNOVATION PROCESSES ... 15

FIGURE 4: AN OVERVIEW OF THE STAGE-GATE SYSTEM ... 16

FIGURE 5: CONCEPTUAL MODEL ... 29

FIGURE 6: QUESTION 1 – INDUSTRY REPRESENTATION ... 43

FIGURE 7: QUESTION 2 – MAIN REVENUE STREAM OF BUSINESS ... 44

FIGURE 8: QUESTION 3 – DEPARTMENT OF RESPONDENT ... 44

FIGURE 9: QUESTION 4 – PROFESSIONAL LEVEL OF RESPONDENT ... 45

TABLE 1: QUESTION 5 – RESPONDENT'S TIME WITH EMPLOYER ... 45

TABLE 2: MEASURE OF CONCEPTS & UNDERLYING INDICATORS ... 46

TABLE 3: CORRELATION MATRIX OF INPROD- & BDA INDICATORS ... 48

TABLE 4: CORRELATION MATRIX OF INPROD- & REMO INDICATORS ... 50

TABLE 5: CORRELATION MATRIX OF INPROD- & PROMO INDICATORS ... 50

TABLE 6: CORRELATION MATRIX OF INPROD- & PREDBD INDICATORS ... 51

TABLE 7: CORRELATION MATRIX OF CONCEPTS ... 52

(8)

1. Introduction

This section of the study will start off by introducing the debated relationship between big data and modern innovation – both in the sense of its confirmed- and expected outcomes.

Furthermore, the necessity of the study is implied with regards to the shortage of research within the relevant domain, which limits further comprehension the details of the relationship. Finally, the section includes the purpose, research question and delimitations of this study.

1.1 Background

The effective use of big data is believed to change the innovation game and imply vivid cost benefits to modern businesses (Davenport, 2013). For decades innovation in itself has been deemed a strategic weapon for competitive advantage, but in modern times, it represents a prerequisite not to be evaporated by creative destruction (Anthony et al., 2018). The reason behind such change lies in the ever-shorter product life cycles, which pressure businesses to innovate on a frequent basis, not to become obsolete.

To cope with the intensified global competition, research underlines companies’ strong belief in the launch of new products to generate future growth and sustained leadership in their respective markets (LaValle et al., 2011). While considered a necessity, a frequent launch of new products has been proven inadequate by itself to race. The failure rate of new technology product exceeds 50 percent, whilst approximately 75 percent of retail products do not even surpass the USD 7.5 million-mark in first-year annual revenues (Schneider & Hall, 2011).

Firms that exhibit success in relation to new technology are commonly concluded to simultaneously meet two objectives: maximizing the fit with customer needs while minimizing time-to-market (Shilling & Hill, 1998). An activity often expressed to ease both of these efforts is market orientation (Cooper, 1990; Narver et. al., 2004). Meaning, to actively seek to understand one’s customers and to comprehend their demands. Firstly, by knowing what customers want, their needs are also further feasible to incorporate into new technology and thereto maximize fit. Secondly, having information readily available already in the early stages of innovation, allows for the confidence to make more speedy decisions, which result in a quicker time-to-market (Zhan et al., 2017; Cooper, 1990). Consequentially, enabling success and above average returns (Sánchez & Pérez, 2003; Afonso et al., 2008).

(9)

Whilst surveys and customer focus groups have long led the way for gaining customer insight and feedback, a third ‘wave’ has swept advanced economies due to modern digitalization – namely big data (Hamm, 2012). Many businesses have struggled with the sets of data they have gathered from traditional surveys and focus groups, much due to the rules of statistics.

Errors that inevitably have caused derived pathways to be equally incorrect. (Uncles, 2011) Some of those challenges constitute: sample representation, trustworthiness of participants, causation appearance by merged data sets, the positive illusion of managers and errors in performance measurements. Moreover, obtaining customer insight and doing your

‘homework’ may be costly and time consuming (Cooper, 1990). However, as Davenport (2013) points out, the modern digital trail deems to offer a more economical and accurate pathway for future innovations.

Our frequent online presence results in the possibility for organizations to now access information about everything from how we vote, what sneakers we like, where we are located to our nightly pulse (Morabito, 2015; Chen et al., 2012). What’s more, with the right means businesses can now obtain this information in real-time and at considerable scale, consequently lower the probability of sampling error and response trustworthiness. Likewise, the internet has changed the reach of customers. Businesses necessarily do not have to wait for surveys, facing the dilemma of targeting the ‘right people’ in addition to gaining general knowledge about their existing- or desired customer base. Hence, one may expect these extensive hypothetical customer samples and their respective insight potential; to shift businesses inclination toward market orientation and future innovation processes (Morabito, 2015).

1.2 Problem Discussion

According to NewVantage - Big Data Executive Survey (2017), major corporations have already jumped on the big data wave. As outlined in Figure 1, we can conclude that big data contributes to value in several aspects of businesses’ activity already, including the areas previously touched upon: innovation, product launch and efficiency. Yet, the authors of the same survey do not express how big data contributes to value within innovative processes more explicitly, which is equally important in order to generate understanding of modern businesses. Consequently, to enhance this comprehension equates the main purpose of this report.

(10)

Figure 1: Fortune 1000 Report Big Data Initiatives & Success Rate

Source: NewVantage (2017), p. 8

Although the big data phenomenon has grown in popularity, both within academia and business, research that discusses and empirically suggests the practical use and affects of big data within innovation processes, still remain few (Mishra et al., 2017). Nevertheless, companies that have made heavy investments and already introduced new technologies, as an outcome of those efforts are American Express, General Electric and Capital One, just to mention a few. Still, the same group of companies see themselves as simply having taken a small bite of the endless opportunities that are to be expected along the data-driven transformation journey.

Henceforth, the question remains, what influence does big data truly have on innovative activities and innovative output? Is it safe to assume that firms that now utilize their chance to understand their customers better, also more frequently meet the objectives related to successful innovative output? In other words, manage to commercialize products that maximize the fit with customer needs, while simultaneously have processes that allow for minimal time-to-market.

1.3 Purpose

The purpose of this study is to investigate firm usage of big data to facilitate market orientation related to the innovative process, while assess the impact such efforts have on innovative output.

1.4 Research Question

• How does big data facilitate modern innovation processes and affect innovative output?

(11)

1.5 Delimitations

The element of big data is argued to have many applications (Morabito, 2015), but in this research the focus will merely regard its facilitation within innovative activities related to market orientation and pre-development. With that said, the term ‘market orientation’ will therefor imply insights provided by big data analytics. Given the above stated purpose, the utilization of findings related to these activities will simply be measured in relation to innovative productivity and its underlying components; time-to-market and value-generating output. Consequently, the focus of big data facilitation won’t be directed toward stand-alone marketing efforts, market expansion analysis or product reviews.

As previously touched upon the main emphasis of this paper involves the pre-development stages of the innovation process, meaning the activities that are specifically executed within those stages. With that said, the probable support big data can contribute to within the stages of development and post-development exceed the scope of this research. The reason for neglecting these stages comes down to a matter of complexity and time constraint. I.e. the element of big data tends to be differently applied to these stages (Chen et al., 2012), especially within activates related to modern manufacturing and would thereto require a much broader approach of this research.

As for the applicability of the findings, they are limited to the companies constituting the Fortune 200 Global of 2017, as they are the ones representing the empirical findings within this study. With that said, no further specific constraints have been made to the scope of research, geographical or otherwise.

(12)

2. Literature Review & Hypotheses

This section includes the theoretical framework that lays the foundation for the formulated hypotheses of this research. Firstly, the line of theory is divided into four main rubrics;

innovation, the innovation process, market orientation and big data. How each theory relates or has been applied to this research can be found following the various topics. Subsequently, the reasoning behind the hypotheses are expressed, that in turn aim to support the main research question of this study. Finally, the section is concluded by the display of a conceptual model, which also explicitly summarizes how the hypotheses and stated theory are to coincide.

2.1 Innovation

As a term, innovation has been interpreted in various ways over the years, which has caused the lack of an authoritarian and clear definition of the concept (Baregheh et al., 2009).

Although the construct was denoted more frequently following the mid 20^th century, the processes related to innovation joint by technological- and economical change were known to be important decades before (Veblen, 1899; Lorenzi et al., 1912; Schumpeter, 1934).

Similarly, Damanpour (1991) also posits innovation to be commonly associated with change, as organizations innovate to influence their current environment or as a response to a changing environment – internal or external. Hence, its not surprising that ‘innovation’ has taken a widespread of shapes and forms besides exerted by various department and disciplines differently. Thereto, explaining the challenges of creating a single defining term.

When applied, studied and exerted by various disciplines with countless perspectives, the range of definitions will be equally broad as a result (Damanpour & Schneider, 2006). The problem that arises following this vague solidarity is the undermining of what actually constitutes as ‘innovation’ and its nature (Zairi, 1994; Cooper 1998; Adams et al., 2006).

For instance, one of the earlier definitions was stated accordingly: “Innovation is the generation, acceptance and implementation of new ideas, processes, products or services”

(Thompson, 1965, p. 2). Although, this view of implementation is still upheld as it does not differ radically from more recent interpretations (e.g. West & Anderson, 1996; Wong et al., 2009). Then there are researchers like Kimberly (1981) that suggest the notion ought to include the different types of innovation to more adequately imply its meaning. Moreover, another common approach primarily emphasizes the degree of newness to best imply

(13)

innovation, as it translates the element of change (Van du Ven et al., 1986; Damanpour, 1996). In accordance, Zaltman et al. (1973) defines innovation as: “the process of developing a new item, the new item itself and the process of adopting the new item” (p. 10).

Nevertheless, within this research, the disciplinary perspective of knowledge management will represent the adopted view of innovation. The reason is the current debate that uplifts the importance of information within the process of innovation (Herkema, 2003; Gloet &

Terziovski, 2004; Darroch & McNaugthon, 2002; Parlby & Taylor, 2000; Cavusgil et al., 2003; Plessis, 2007). In essence, the dialogue proclaims that present- and future innovative performance is highly dependent upon knowledge, much due to the rise of creative industries.

Modern firms are stated to face further complex environments, where both the preference of customers and technology change rapidly, whereto information could facilitate necessary comprehension. In accordance, the following definition applies to this research: “Innovation as the creation of new knowledge and ideas to facilitate new business outcomes, aimed at improving internal business processes and structures and to create market driven products and services. Innovation encompasses both radical and incremental innovation.” (Plessis, 2007, p. 21)

2.1.1 Categorization of Innovation

Once the concept of innovation is clearly defined, diverse innovative ideas and outputs are equally relevant to denote to offer precision within applicable contexts. For decades, economists and researchers have distinguished between various categories of innovation. The first divide categorizes the type of innovations under two main rubrics: product- or process innovations (Utterback & Abernathy, 1975). The former indicates the act of introducing something to the market, which improves the quality and range of products offered. The latter implies new pathways of delivering or producing a good or service. (Gloet & Terziovski, 2004; Greenhalgh & Rogers, 2010) Theoretically the terms are easily separated, whereas practically they are fairly tangled. That is, more than often one business’s product innovation gives birth to another company’s process innovation and vice versa (Plessis, 2007).

Confusion arises similar to the chicken and the egg.

The second longstanding categorization of innovation implicates the level of innovativeness;

often represented by a divide between radical- and incremental technological change (Duchesneau et al., 1979; Hage 1980; Daft & Becker 1978). The latter is commonly referred to enhance the performance of existing technologies, within the dimensions that customers in

(14)

main markets already value (Greenhalgh & Rogers, 2010). On the contrary, radical technologies offer the customers something completely new and give rise to a whole new industry of innovative products (Plessis, 2007). The internet or the light bulb, are both common examples to describe such technology. However, not all researchers find the division to be all that cut and dry as they describe innovativeness i.e. level of newness of a product or service (Garcia & Calantone, 2001). In fact, much similar to issues related to the definition of innovation, sub-categories of the concept face as well (especially product innovations). For example, some researchers compare product innovativeness as it relates to the firm (e.g. Cooper, 1979; Green et al., 1995) whilst others use technological attributes as benchmark (e.g. Atuahene-Gima, 1995; Veryzer, 1998) or other parameters. In addition, according Garcia and Calantone (2001), researchers use their own terminology for various types of innovation, which substantiate confusion even further. As a result, the same authors suggest contributions and findings are difficult to compare within relevant domains, which therefor ought to be modified or better rationalized. Their suggestion is displayed in Figure 2, where they organize the innovativeness of products as a second-order factor.

Figure 2: Typology of Innovativeness

Source: Garcia & Calantone (2001), p. 124

In essence, the research duo argues that the construct of innovativeness is in fact comprised of two individual components: newness to firm (micro level) and newness to industry (macro level). On that basis, and contrary to previous research, they instead suggest three main types of product innovation: radical-, incremental- and really new innovation. (Ibid.) Similar to previous theory related to radical innovation (Hage 1980; Daft & Becker 1978), such technological change is represented by discontinuation at both macro- and micro level visible in Figure 2 (Garcia & Calantone, 2001). However, incremental innovations are simply

(15)

represented by discontinuations at micro level and therefor have no relation to its overall industry. Finally, the new construct ‘really new innovation’ is recognized by a stand-alone change at macro level (market- or technology newness) joint by any shift in micro-factors (market- or technology know-how) (Ibid.). Their definition of the new concept was derived from previous suggestions related to moderately innovative products: “consisting of lines to the firm, but where the products were not as innovative (that is not new to the market) and new items in existing product lines for the firm” (Kleinschmidt & Cooper, 1991, p. 243, cited in Garcia & Calantone, 2001).

Not to forget, an increasing share of innovation is represented by services today (Miles, 2006). In fact, Barras (1986; 1990) was early in noting that the IT-revolution would have the same implications for the service sector, as the industrial revolution had on the manufacturing sector. Moreover, similarly to product innovations, services too range between incremental- and radical innovations. (Paré, 2004; Skålén et al., 2015). Contrary to production innovation though, service innovation is commonly stated to be further based on the merits of customization and thus traditionally more dependent on the producer-customer relationship (Barras, 1986). As a consequence, alternative processes have enacted blueprints to explain the processes of service innovation, e.g. the reverse product cycle (Ibid.). However, modern technology is suggested to alter these attributes that cause division between product-, process- and service innovation. With abundant volumes of information generated by customers, businesses no longer have to wait for customers to express their requests for new services – their generated data will do it for them (Lehrer et al., 2018). I.e. readily available information can now be applied as an inbound resource to innovation, instead of be dependent upon relationships. Much like any other resource related to product innovation. As a consequence, to distinguish between product and service, may deem excessive and has therefor been exempted from the analysis of this report.

Within this study, a richer typology of product innovativeness and a joint view of modern service innovation are thereto exhibited. Additionally, the research scope of innovation can take various levels, here ranked according to range: project-, firm-, regional- and sectorial level (Verhees & Meulenberg, 2004). Whereas, in the case of this research the firm level scope demonstrates the later findings and analysis to more easily compare individual actors, regardless of innovation or firm type.

(16)

2.1.2 Measures of Innovation

When striving to map innovative activity or compare firms ‘innovativeness’, the rightful measures or adequate parameters to apply are fairly difficult to decide upon. The root of this dilemma can be explained by various innovative activities as businesses represent different industries, business sizes and countries (Garcia & Calantone, 2001). On firm level, traditional proxies of innovative input include R&D expenditure (Grilishes & Mairesse, 1990) and share of skilled employees (specifically engineers and scientists) that perform R&D. However, expenditure analysis underlines the majority of funds to go towards development (D) and smaller volumes toward research (R), since new technologies require efforts within business production activities. (Greenhalgh & Rogers, 2010) In other words, the measure of R&D expenditure may very well simply highlight new product development activities alone, which do not equal innovative activities per se (Cohen & Levinthal, 1989). Moreover, inputs in R&D do not necessarily equate successful output (Pavitt, 1985). In some industries the results are not instantly visible either. Subsequently, the measure undoubtedly has its limitations, when not used to compare the levels of input between similar industries or businesses.

(Greenhalgh & Rogers, 2010)

To continue, innovative output is commonly measured via IPR-generation. The proxy is fairly popular amongst economists, where patent statistics are performed (Pavitt, 1982; Archibugi, 1992). However, similar attention is not given to trademarks, copyrights and design, which make firms in unrelated industries hard to compare (Greenhalgh & Rogers, 2010).

Furthermore, not all patents indicate profitable or actually produced innovations. Albeit, the notion of patents imply tangible, technological novelty and underline the company’s belief in the financial value of the invention, which explain the proxy’s lasting popularity (Archibugi, 1992)

Due to the complexity of measuring innovative activity, inputs and outputs should not be measured separately, as the process of innovation is not linear (Von Hippel, 2009). In an effort to merge innovative activity into a single number, the emergence of the innovation index has grown in popularity (Greenhalgh & Rogers, 2010). Meaning, straight out ask firms about their innovative efforts. They can take a shape to suit firm-, industry- or sectorial level, however no such index is yet to be standardized on a firm level basis. The method is also criticized as findings may vary depending on how the index is orchestrated, how different indicators are measured and weighed in comparison to the rest. A well-known example of

(17)

this is The Community Innovation Survey (CIS). Initiated in 1991 by the European Commission and sent out every other year to European firms with more than ten employees (European Commission, 2012). Questions are industry and firm neutral and seek to understand firm’s introduction of new innovations, where they distinguish between new to the firm and new to the industry. In fact, similar to the separation made by Garcia &

Calantone (2001). This specific survey will get further attention in the later stages of this report. Another common approach to an index is through the use of a literature-based innovation output. Where one monitors the news about innovations sent to technical journals etc. (Greenhalgh & Rogers, 2010), first attempted by Gort and Klepper (1982). The benefit to this approach is that all firm-sizes are recognized and thus allow for otherwise overviewed insights.

Based on the latter debate, this research has thereto applied an index-based approach for the collection of empirical data. With the research question in mind that aims to conclude innovative performance on firm level, in a group of various industries, an index approach hence deem the most suitable if its measures are equally neutral. Especially since innovative inputs or outputs are explicitly difficult use as unifying benchmarks in unrelated industries (Cohen & Levinthal, 1989; Pavitt, 1985; Greenhalgh & Rogers, 2010). The specifics of the method are not being outlined here nonetheless, but are kept for later stages of this report.

2.1.3 Innovative Productivity

Innovation is often stated to be single most important factor for long-term growth (Greenhalgh & Rogers, 2010; Shilling & Hill, 1998; Cooper & Kleinschmidt, 1991; Anthony et al., 2018).Yet, innovative efforts often fail (Schneider & Hall, 2011), which inevitably limits firms that continue to see inadequate yields from new technologies. In addition, according to Hamel and Getz (2004), budgets dedicated to R&D are expressed to have decreased and henceforth pressure innovation teams to ‘do more with less’. As a consequence, funds have to be spent efficiently. But that alone will not promise successful innovation. A firm cannot simply outgrow its rivals unless it can out-innovate them, which requires an enhanced consistency of adequate returns (Ibid.). Simply put, future innovations do not merely depend on the ability to generate new technology efficiently, but productively.

(18)

The notion of innovative productivity is perhaps not the most common concept within the field of innovation. In previous research the concept is expressed as an above-average level of innovativeness that takes place within a given time period and input of resources (Cooper

& Edgett, 2005). Hamel and Getz (2004), denote it as “more bang for every innovation buck”

(p. 27). However, in this research the concept will take the perhaps more concrete interpretation used by Kim and Mauborgne (2004), followed by Cooper and Edgett (2005); a firm that exhibits innovative productivity both launches value-generating innovations, while simultaneously exhibits a short time-to-market. In fact, the term suggests the exact objectives that were previously discussed in relation to successful innovation (Shilling & Hill, 1998).

Namely, the maximization of fit with customer demands and the minimization of time-to- market.

Within this study, innovative productivity will play a central role as the concept represents triumph where most firms struggle in relation to innovative efforts. Meaning, they fail to launch new technologies that customers genuinely want (Schneider & Hall, 2011) while at the same time struggle with diminishing budgets (Hamel & Getz, 2004). In addition, with the knowledge of difficulties related to comparing innovative outputs and inputs amongst actors (Cohen & Levinthal, 1989; Pavitt, 1985; Greenhalgh & Roger, 2010), this research has instead attended to the benchmark of innovative productivity. As a result, the purpose of outlining how big data facilitates innovative processes, has not merely been benchmarked against a more concrete measure; but also benchmarked against a notion that implies

‘rightfully’ executed innovative efforts.

2.2 The Innovation Processes

By thinking of innovation as a process, the otherwise diffuse art of creation can more easily be managed and supervised. More often than not, large corporations start off with a broader range of innovations, to then through a set of checkpoints to winnow out the less promising ideas from the promising ones (Christensen et al., 2008; Kahn, 2018). An organized- and systematic approach removes variances in pathways by which innovation efforts are handled (Shilling & Hill, 1998; Harmancioglu et al., 2007). With that said, not all firms follow the same line of method from idea to finalized products. In fact, with today’s increasing volumes of outsourcing, agile processes and levels of open innovation, the path from idea to commercialization allows for numerous compositions of the innovation process (Grönlund et al., 2010; Cooper, 2009, Zhan et al., 2017). Albeit, if we instead assess the order of

(19)

undertakings the innovation goes through, the underlying construct amongst models deem fairly similar. In Figure 3, we see two common and simple versions explaining the process of innovation. At the top, we can find one of the more conventional pathways of the innovative process (Rothwell, 1994), where development proceeds sequentially between the functional groups (Schilling & Hill, 1998). Similar models came about as a result of competitive intensification within markets (e.g. Clark, 1980), where larger companies fought to enhance efficiency and thus gain additional market share. At the bottom, we find a further overlapping process approach (e.g. Graves, 1987), which is the result of external pressure for speedier time-to-market and quality (Dumaine, 1989; Rothwell, 1994). The parallel method accordingly allowed for internal cooperation at an earlier stage, which enabled assessments of idea feasibility sooner as well. Essentially, the innovation could now be killed or be reworked sooner if needed, on the basis of later stages, henceforth increase the overall speed of processes. (Schilling & Hill, 1998; Rothwell & Zegveld, 1985)

Figure 3: Sequential Versus Partially Parallel Innovation Processes

Source: Schilling & Hill (1998), p. 73

In essence, regardless of model, the flow of activities is present throughout. In addition, although there have been a range of alternatives to the ones outlined in Figure 3, the fundamental functions and activities within seldom seem to change (e.g. Clark, 1980; Graves, 1987; Cooper, 2009; Christensen et al., 2008; Khan, 2018). With that said, as this research aims to assess how modern assets facilitate the activities within the process, irrespective of applied model, further attention has thereby been given to better understand the functions.

Consequently, the frequently cited, yet detailed theory of stage-gate has aided as framework.

(20)

2.2.1 The Stage-Gate Model

The model was first developed by Cooper and Kleinschmidt (1986); where they describe the prototypical innovation process as a set of activities that they refer to as stages, divided by gates (see Figure 4). Stages include needed activities to be able to move forward to the next gate, whereas gates constitute checkpoints along the way, to determine whether innovations live up to certain measures (Cooper, 2009). The reason for dividing specified activities into a sequence of tasks originates from the purpose of minimizing risk and uncertainties that follow innovation (Cooper & Kleinschmidt, 1991; Griffin, 1997).

The discussed project-specific activities are normally performed by cross-functional teams, which represent various business units and thus require cooperation and enhanced communication (Engwall et al., 2005) As the invention moves closer to the final stage, financial investments grow progressively, whilst the project-specific risk declines (McDermott & O’Connor, 2002). A project is normally evaluated based on pre-established criteria at the various gates and then labeled kill/go/hold depending on how well it fulfills those criteria (Bers et al., 2014). Moving along the innovation process from idea to commercialization, the requirements steepen as well as the importance of the activities within each stage. Although, not all businesses undertake all activities of Figure 4 (for reasons that will be discussed later) most processes include stages of feasibility testing, development and launch (Grönlund et al., 2010). As a consequence, the overall sequence of activities has here been divided similarly into three to simplify: pre-development-, development- and post- development stages. They are to be described individually below, with regard to their various undertakings by the support from the selected framework of stage-gate.

Figure 4: An Overview of The Stage-Gate System

Source: Cooper (1990), p. 46

(21)

2.2.2 The Pre-Development Stages

The first part of the innovation process is initiated by the presentation of an original idea, which then is to be forwarded to initial screening at the first gate. Assuming the innovative idea is accepted, it will go through a series of assessment before it moves forward to development (Kahn, 2018). The technical merits start off fairly vague, with only a handful of requirements at Stage 1, to then steepen as the idea approaches development. Technically, assessments are related to development- and manufacturing feasibilities, costs and time-to- market estimations. Once the project has reached Gate 3, the same criteria needs to be defined and the product’s specification outlined to the smallest detail (Cooper, 1990).

Feasibility testing also involves the estimation of market potential, to gather information related to possible market size, competitive advantage and expected market demand.

Moreover, as these pursuits exhibit the understanding of customers, i.e. who they are and what they want, market-oriented activities are here sought to facilitate (Narver et al., 2004;

Zhan et al., 2017). Adequate findings through such an approach are intended to support the firm with information directing e.g. what attributes to consider from a technical perspective and what customers the product will speak to once commercialized. Henceforth, the same insight can be found useful during the establishment of financial estimations. Numbers derived from potential market insights, also allow for more concrete financial assessments of projects, especially as costs, payback period and revenues appear more precise in the later stages (Cooper, 2009). In turn, projects that deem unprofitable or unattractive can either be put in ‘hold’ for further adjustments to sail through to development, or simply be killed and thus terminated. (Christensen et al., 2008) According to Cooper and Kleinschmidt (1991) these are the stages that separate winners from losers, as they take more ‘justified’ decisions from the start. However, not all researchers tend to agree with the same notion, as they suggest the overall approach to be too time consuming (Uncles, 2011; Mann & Jones, 2002;

Sethi et al., 2001).

2.2.3 The Development Stage

If the project has been given a ‘go’ at Gate 3, it has thereto been viewed as a promising opportunity since heavy funding normally is required for development (Cooper, 1990). At this stage, the idea is to be developed according to the specifications established in the previous stages. Furthermore, operations- and marketing plans are to be thoroughly examined and made ready for implementation. (Rothwell, 1994; Kahn, 2018) Likewise, at this stage the

(22)

financial analyses are normally updated as costs and time-to-market are more clearly outlined, whilst potential issues and applications regarding IPR are to be undertaken (Cooper, 2009; Greenhalgh & Rogers, 2010).

2.2.4 The Post-Development Stages

Following the development of a prototype or close-to-finished product, the innovative idea is transferred to a new set of stages. Post-development, the product or service is commonly reassessed on the basis of its attractiveness and how well it relates to the targeted customer group. Prior to commercialization, the details of operational- and marketing plans are to be finalized, joint by the occurrence of a number of checks to validate a future launch between Gate 4-5. (Cooper, 1990) A number of analyses are often conducted to facilitate these checks, which in turn may include: monetary feasibility, customer perception of overall offer, in- house testing to determine user quality, manufacturing and production optimization through trial and error. (Grönlund et al., 2010) If the idea still deems lucrative, a ‘go’ at the final gate allows project to sail through to the stage of commercialization. There, pre-established strategies are to be set in motion as the product is launched. (Cooper & Kleinschmidt, 1991) Finally, after the product has been launched, a comparison of projections and the final outcome allow the organization to determine the strengths and weaknesses of the idea (Grönlund et al., 2010). The appraisal can take the shape of financial- or product performance measures and may generate knowledge for projects ahead. Post-launch, the team involved in the project-specific innovation process is dissolved after the review activity and the innovation continues as an established member of the product portfolio, after which it is finally terminated. (Cooper, 1990)

2.2.5 Theory Meets Practice

As businesses tend to have different organizational design, their innovative activities vary too. The same applies to products and services, not all innovations require equally long routes from idea to commercialization (Cooper 1990). E.g. if a corporation is already familiar with the required technology and manufacturing procedures for the intended project, it will be handled less thoroughly vis-à-vis an innovation that represents completely unknown territory.

Hence, the more uncertainty involved, financially or in terms of novelty, the more cautious and overall rigid innovation process will be performed in order to minimize the project- specific risk (McDermott & O’Connor, 2002).

(23)

To continue, even though alternative innovation process models are outside the scope of this research, empirical findings show that the overall current approach of firms is surprisingly similar to the stage-gate model (Grönlund et al., 2010). Even in the cases of best-practice examples, according to Griffin (1997). So even though some alterations have been made to the original stage-gate model to incorporate industry-specific elements, the initial framework is to the highest degree suitable to describe today’s innovative activities (Grönlund et al., 2010; Cooper, 2009; Bers et al., 2014). Conclusively, in order to compare businesses in unrelated industries, a theoretical framework that’s applicable to the majority deemed the most useful to support the research question at hand.

2.3 Market Orientation

The theory of market orientation has commonly been referred to within the field of innovation research (e.g. Cooper, 2009; Uncles, 2011; Zhan et al., 2017). Yet, it first appeared in relation to the domain of marketing (Kotler, 1984) and still do. Thereto, it does not deem odd that the two most prominent conceptualizers of term represent the field of marketing and strategy, although the two have slightly different interpretation of the concept.

The first pair, Kohli and Jaworski (1990) view market orientation as a market intelligence perspective: “The organization-wide generation of market intelligence, dissemination of the intelligence across departments and organization-wide responsiveness to it” (p. 6). The second pair on the other hand, rather sees the concept as an organizational culture or firm approach (Narver & Slater, 1990). In addition, they were also the first to recognize the positive relationship between market orientation and business profitability, which follows this stated definition: “the organization culture that most effectively and efficiently creates the necessary behaviors for the creation of superior value for buyers and, thus, continuous superior performance for the business" (ibid., p. 21). Since the couple’s original publication, several researchers have made similar conclusions, which is thereto why the latter definition has been applied within this research (e.g. Atuahene-Gima, 1995; Deshpande et al., 1993;

Han et al., 1998; Jaworski & Kohli, 1993; Li & Calantone, 1998; Pelham & Wilson, 1996).

Nevertheless, not all researchers are equally convinced about the debated benefits. Most commonly, their arguments suggest that market-oriented activities diminish the chances for

‘true’ innovative output (Berthon et al., 1999) and unbiased efforts within R&D (Frosch, 1996). Others express their concerns related to the confusion market-oriented efforts may cause within business processes (Macdonald, 1995), whilst some claim firms who pursue

(24)

such an approach instead increase their chances of loosing their industry leadership. Whereto, they give customers too much of a voice (Christensen & Bower, 1996). According to Uncles (2011) the method is not always as simple or successful as researchers tend to suggest, which inevitably results in unsuccessful attempts. More often than not, the reasons for failure carry traditional issues related to sample representation and the rules of statistics that may occur within e.g. surveys or focus groups. As a consequence, one may assume that not all businesses may be equally inclined to incorporate the approach of market orientation, which will therefor be equally respected along this study.

To continue, in accordance with Narver and Slater (1990), market orientation is a construct that rests on three behavioral pillars, namely customer-, competitor orientation and inter- function coordination. Although as this research focuses primarily on the two former, since the purpose of the study relates to the effects of less biased inbound information, by the use of big data. Simply put, not the internal spread of inbound information.

For those firm who do exhibit the approach, market orientation has been considered a singular method, namely through market driven- (Jakowski et al., 2000) or responsive market orientation (Narver et al., 2004). However, later research has found that attempts too can deem proactive, where latent customer needs construct the focus of orientation (ibid., 2004).

In order to grasp the variance between the methods, they are individually described below, to then outline their relationship to innovation and the stage-gate model.

2.3.1 Responsive Market Orientation

Empirical analyses have over a long period mainly focused on responsive market orientation.

Where organizations aim to comprehend and approach customer verbalized needs and explicitly expressed intelligence. In other words, expressed needs or solutions are those desires the buyers can articulate, as they are aware of them (Narver et al. 2004). E.g. if a customer expresses its need for ‘transportation’, the pronounced solution could be a car.

However, the approach is not always easy to successfully complete or to gain a competitive advantage from. On the one hand, the main obstacle related to the method implies believed causality due to the composition of dependent- and independent variables, which results in enhanced correlations (Uncles, 2011). In addition, the perspective of the customer or employee, who’s voice is intended to generate valuable intelligence, could generate equally incorrect conclusions if the sample is not adequate in size for the intended purpose. On the

(25)

other hand, if successfully handled, the expressed needs are obtainable for all competitors within an industry. That is, they ought to generate similarity in product and services offered.

In turn this may entail low levels of differentiation in addition to price wars or intense marketing competition (Narver et al., 2004). Complications like this have resulted in the loss of the approach’s initial allure. (Uncles, 2011)

2.3.2 Proactive Market Orientation

This orientation has received less attention in comparison to the previous mentioned.

Businesses who take such an approach intend to comprehend and discover latent needs.

(Narver et al., 2004) The notion of these needs and the expressed were introduced simultaneously (Kohli & Jaworski, 1990; Narver & Slater, 1990; Slater & Narver 1995), albeit a separate measure to describe them is newly introduced (Narver et al, 2004). What sets the orientation aside is the fact that organizations attempt to understand the customer’s needs, before the customer itself is made aware of them. Subsequently, it is a more challenging task that requires the business to again and again surpass the expectations of the customer (Rust &

Oliver, 2000). Contrary to responsive market orientation, the businesses here ‘lead’ the customer when they satisfy unexpressed needs, which in turn denote the proactivity of the approach (Narver et al., 2004). Firms that declare to hold this approach are assumed to be motivated to discover something new (Garcia et al., 2003). Although, firms ability to do so successfully, rely upon adequate skills to diverge from this knowhow and thereto capitalize on found unexplored market opportunities (Koza & Lewin, 1998). In order to pinpoint these needs, extensive customer observations have previously been performed. E.g. Procter &

Gamble shadowed 80 households by the use of camera teams in order to obtain insight into different customer routines around the globe (Nelson, 2001). Nevertheless, these efforts are simplified today by the adoption of modern technology and software (Narver et al. 2004;

Hofacker et al., 2016).

2.3.3 Market Orientation & Innovation

The relation between innovation success and market orientation has been confirmed on multiple occasions (Zhan et al., 2017; Cooper, 1986; Mann & Jones, 2002; Sethi et al., 2005).

Nonetheless, recent data underlines that businesses execute inadequate efforts to comprehend its customers and environment, before moving ahead with commercialization. The stated outcomes include products that lack a defined market, with no specific target group where marketing efforts are equally diffuse, undifferentiated offers and products with undesired or

(26)

an abundance of features. In fact, 75 percent of all consumer goods fail to generate satisfactory returns (Schneider & Hall, 2011)

Cooper (1990) was early to underline the importance of market orientation along the stage- gate process. He argued that businesses believe that the ‘homework’ activity implies longer processes, which in turn causes them to miss out on valuable market opportunities. Even if the process was somewhat prolonged, the alternative was higher probability of product failure due to weak pre-development data gathering. In fact, successful products have been proven to occupy more time and furthermore require additional investments vis-à-vis those who failed.

In addition, the importance of market orientation becomes evident, as the steps that separate the winners from losers, are argued to be the ones that lead up to development stage (stage 3 in Figure 4). (Cooper & Kleinschmidt, 1991) I.e. the ‘homework’ phase of the innovation process. In addition, other researchers seem to have arrived at the same conclusion, as they too underline the significant role that market orientation has within innovative processes (e.g.

Drazin & Schoonhoven, 1996; Zhen et al., 2017; Hofacker et al., 2016; Hamm, 2011;

Morabito, 2015).

2.4 Big Data

Although the importance of market-orientation is nothing new, related challenges have decreased the approach’s initial appeal (Uncles, 2011). However, based on recent discussions, the new capabilities of big data are suggested to ease those exact efforts (Chen et al., 2012; Hamm, 2012; Zhang et al., 2015). In fact, some even argue an adoption of such capabilities is critical for future novel and successful products (Chuang et al., 2015). So, what is in fact big data? McKinsey (2011) describe the concept as a dataset with an incredible dimension, that in turn requires advanced software tools in order to be handled, stored and analyzed. Even though the data element has opened up for new methods to extract intelligence that result in new business advantages, it is not just a ‘new’ way of conducting analytics. It differs in three main aspects, namely in volume, velocity and variance. (McAfee

& Brynjolfsson, 2012) Yet, the overall concept may still seem fairly vague.

Thereto, further attention is given to these three attributes, as both relevant definitions and its significance deem easier to comprehend as a result. Nevertheless, the applied definition within this research follows: “Big data is data that exceeds the processing capacity of conventional database systems. The data is too big, moves too fast, or doesn’t fit the

(27)

structures of your databases architectures. To gain value from this data, you must choose an alternative way to process it” (Dumbill, 2013, p. 1)

2.4.1 Volume

We may have left a digital trail before when browsing the internet, but the amounts of data have completely staggered the last few years, with no signs of slowing down (Marshall et al., 2015). From every update on social media, the GPS-activity when you drive your car, to the clicks you make on a webpage and more, are now collected and stored on massive hard drives, allowing for real-time data gathering (McKinsey, 2011; Zhen et al., 2017). In 2017, we produced more than 20 zettabytes (1 ZB = 10¹²GB) per annum, which equate almost 20- fold the volume of 2010. The increase is the result of the embedded devices we now use and the increasing number of online users, which in turn make the average person’s interaction level with data further frequent (Morabito, 2015; Hofacker et al., 2016). In 2010 we averaged 85 interactions daily per capita and in 2015 it increased to 218. Yet, according to estimations it will grow from 601 to a whopping 4785 times per day between 2020-2025, contributing to the estimated 160 ZB per year by the end of the same period. (Reinsel et al., 2017)

Henceforth, with more data crossing the internet every second than constituted the whole internet 20 years ago, businesses find themselves with petabytes of full data sets and therefor also a completely different insight than previously expected (McAfee & Brynjolfsson, 2012).

2.4.2 Velocity

As previously mentioned, it is not just the size of the data that sets aside big data and big data analytics from traditional business analytics. The speed of the data generation is in some cases even increasingly essential. Real-time intelligence or close to real-time information allow for more agile businesses than before. (McAfee & Brynjolfsson, 2012) Nonetheless, the insights are merely accessible for companies who have equipped themselves with adequate technology to process the produced information (Johnson et al., 2017; Chen et al., 2012). Whereto human capacity is insufficient to successfully manage it (Morabito, 2015).

The latter becomes especially relevant for businesses in sectors where decisions of customers frequently fluctuate. Subsequently, products also tend to be shorter lived if customers’

decisions change often, thereto rapid alignment with new needs and intelligence equate prerequisites not to become obsolete (Johnson et al., 2017; Anthony et al., 2018).

(28)

2.4.3 Variance

The currently produced data is generated by various sources, which moreover are fairly new to the market (McAfee & Brynjolfsson, 2012). E.g. everything from geographical locations of smartphone users, customer pulse charts from smart watches to views on Netflix (Hamm, 2012). In turn, this also allows for different compositions of data sets. According to McKinsey Quarterly (2010) the majority of data is now originated by ‘things’. As data mainly represents a by-product of modern electronics, much due to the fact that electronics now have a higher degree of connectivity than ever before. The latter implies that we can expect a new type of data, which will have a different face than the one we have started to grow accustomed to (DBMS-based and web-based). Data derived from an internet-connected machine used for manufacturing, will not equate the shape of web-based intelligence for instance. Conclusively, in order to deal with new variety, broader spectrums of algorithms and software continue to play more significant roles for value creation. (Chen et al., 2012) 2.4.4 Big Data Analytics

So how do we make sense of all this data? The notion of intelligence has been around since the middle of the 20^th century, whilst business intelligence has become wildly popular since the 1990’s and with it came business analytics (BI) just a few years later (Chen et al., 2012;

Davenport, 2006). Today, the world of BI has been extended by the element of big data, which requires business to possess specialized process technologies, visualization tools, storage capacity and management capabilities to enable such activity (Chaudhuri et al., 2011). In the report “Business Intelligence and Analytics – From Big Data to Big Impact”, Chen et al. (2012) divide intelligence into three separate categories, two of which involve big data:

Web-Based Intelligence (2.0) - Unstructured Content: This relates to the analytical processes with the ability to sort data characterized by social media, and websites. It basically comes down to the understanding of digitally present text- and content data (Doan et al., 2011). An abundance of information can be extracted from this kind of content and for various purposes, such as to understand customers, products, businesses and industries. E.g.

Google Analytics make it possible to map a customer’s online behavior; what they purchase, how they search online and consequently establish patterns. The advantage of having such insight in excess allows for businesses to optimize their content to align with customers’

online behavior, expressed comments on social media, social games, blogs etc. (O’Reilly,

(29)

2005). Lusch et al. (2010) have described web-based intelligence as a mean to initiate a virtual ‘conversation’ between businesses and the market, where text- and web-mining techniques enable companies to actually understand what the customer want without interaction.

Mobile- & Sensor-Based Content (3.0): As electronic devices, cars appliances etc. in our lives become further connected, new versions of data sets arise. They in turn require equally innovative technologies to be sorted and handled. The analytics differ in the sense that they need to comprehend and process geographical specific-, individual- and contextual data that is generated by modern technology equipped with sensors (Chen et al., 2012). In order to reap the full benefits of such processes, businesses have to secure storage capacities and be able to process real-time or near-time data, i.e. conduct stream analytics (Tönjes et al., 2014). E.g.

estimate flight arrival can be much more accurately predicted by the use of available weather data, GPS-locations and airport flight schedules, in comparison to the pilot’s own estimate. A major U.S. airline managed to adopt the idea, as they collected and combined massive amounts of sensor-based information from various sources every 4.6^th second to allow for enhanced predictions, through patterns in the data and automated operations (McAfee &

Brynjolfsson, 2012). Consequently, these types of analytics open up the opportunity for organizations to incorporate timely decision-making processes, automated machine learning and more, if properly incorporated (Qiu et al., 2016; Etzion, 2015).

Within this research, the impact of big data plays a central role. It is thereto important to comprehend its alternative occurrences and specific attributes in order to assess its impact within innovative productivity (quick to market with value-generating innovations). By understanding the range of possibilities related to big data, an equally supported approach to data collection has therefor been feasible. That is, the approached firms possible capabilities were contained and respected within the scope of data collection due for analysis. With that said, the hypotheses that are to enable a conclusive answer to the main research question therefor constitute the next line of focus.

(30)

2.5 Hypotheses Development

At this point, the various theories are to mend together to support the overall purpose of this research: “…to investigate firm usage of big data to facilitate market orientation related to the innovative process, while assess the impact such efforts have on innovative output”. To begin with, the first point of action is thereto set to assess the overall effect and relationship that can be found between the two concepts that outskirt the purpose – namely big data and innovative productivity.

Successful innovation is as previously debated reliant upon satisfying two main objectives simultaneously: maximizing the fit with customer needs while minimizing time-to-market (Shilling & Hill, 1998, Mann & Jones, 2002; Sethi et al., 2001; Barczak, 1995; Griffin, 1997;

Quesada & Gazo, 2007; Liao & Barnes, 2015; Afonso et al., 2008). When a firm manages to deliver upon these objectives, later research suggests the firm to exhibit productive innovative activities or so called innovative productivity (Cooper & Edgett, 2005; Hamel &

Getz 2004, Kim & Mauborgne 2004). According to longstanding debates, the underlying factors that allow for these objectives to be fulfilled are: (1) pre-development research to enhance the confidence and speed within decisions (Cooper, 1980; Lovelace et al., 2001;

Cooper & Kleinschmidt, 2011; Balbontin et al., 1999; Barczak, 1995; Calantone et al., 1997;

Dwyer & Mellor 1991; Griffin 1997) and (2) the understanding of customers’ needs to be able produce innovations that customers truly want (Cooper & Kleinschmidt, 2011; Souder et al., 1997; Evanschitzky et al., 2012; Williamson & Yin 2014; Lin et al., 2010; Rese & Baier, 2011; Narver et al., 2004).

However, these views are as previously discussed not free from criticism. Some researchers suggest the innovativeness of new technology would suffer if the customer’s voice weighs in too much (Bower, 1996; Frosch, 1996). Others are suggesting difficulties within pre- development research to be directing businesses onto wrongful paths, to be complicated or to be overly time-consuming (Uncles, 2011; Mann & Jones, 2002; Sethi et al., 2005). Both sides of the aisle have valid points, but as this research aims to assess; what role does modern technology have in this debate? In turn, big data is suggested to facilitate speedier processes, more detailed customer insight and enhancement in the understanding of markets and products (McKinsey, 2011; Wong, 2012; Zhan et al., 2017) – i.e. the above stated success factors for successful innovation. A recent study conducted by IBM claims big data to support overall firm productivity when applied to the innovation process. In turn, they argue

BIG DATA WITHIN INNOVATION PROCESSES