Вы находитесь на странице: 1из 21
Forecasting & Social Change 101 (2015) 62 – 82 Contents lists available at ScienceDirect Technological

Contents lists available at ScienceDirect

Technological Forecasting & Social Change

Technological Forecasting & Social Change Exploring the cognitive value of technology foresight: The

Exploring the cognitive value of technology foresight: The case of the Cisco Technology Radar

Siri Boe-Lillegraven a ,c , , Stephan Monterde b

Boe-Lillegraven a , c , ⁎ , Stephan Monterde b a Aarhus University, School of Business

a Aarhus University, School of Business and Social Sciences, Department of Business Administration, Bartholins Allé 10, 8000 Aarhus C, Denmark

b Cisco Systems, Inc., Corporate Development, Business Development, Avenue des Uttins 5, 1180 Rolle, Switzerland

c The Sino-Danish Center for Education and Research, Beijing, China

article info

Article history:

Received 1 July 2013 Received in revised form 27 March 2014 Accepted 29 July 2014 Available online 13 October 2014

Keywords:

Organizational future orientation Technology foresight Technology radar Cognition Decision-making Dual process theory Mental models

abstract

There is much literature on how corporate foresight should be practiced, but fewer explanations of through which mechanisms it creates value. We seek to deepen this understanding by applying a cognitive lens. We mainly explore how and why foresight can influence ways of processing information, first conceptually and then empirically through the case of the Cisco Technology Radar. Through consulting multiple individuals, opposed to focusing solely on statements from top managers or observable outputs on the firm level, we uncover a broad set of hidden valuesfrom this technology foresight initiative. We conclude that a critical mechanism of a system such as the radar is its probing of analytical thinking, as well as its way of making people connect and exchange views across functions and departments. We also observe that deeper and more frequent involvement corresponds with greater perceived value of the system. Our findings have implications for research seeking to further understand through which mechanisms foresight creates value, and for the practice of designing, implementing and motivating for participation in technology foresight. © 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

1. Introduction

By the end of 2012 the Cisco Technology Radar a foresight system for identification, selection and verification of emerging technologies (Rohrbeck et al., 2006) was widely considered a success. This was not only due to the several hundred scouted technologies. More importantly, the technology reports were increasingly spotted on the tables of key decision-makers within the company. The radar scouting process also arose considerable interest from external clients. Still, as one of the authors can personally testify, people kept asking for the ultimate proofthat the radar created value. So what, they asked. Where was the story of the technology that thanks to the foresight system had been turned into a successful business?

Corresponding author at: Bartholins allé 10, 8000 Århus C, Denmark. Tel.:

+ 45 24247457 (cell), + 45 84165028 (office). E-mail addresses: Siri@badm.au.dk (S. Boe-Lillegraven), smonterd@cisco.com (S. Monterde).

Where was the profitable innovation that would never have happened had the radar project not existed? Exactly how or why did the technology foresight system make a difference? Generally, foresight is known to bring value through different roles; exploring new business fields (strategist role), increasing the number of innovation concepts and ideas (initiator role) and increasing the quality of the output of innovation projects (opponent role) (Rohrbeck and Gemünden, 2011). Research within the scenario foresight literature also specifies the usefulness of foresight for changing mindsets and decision- making styles, which is important for enabling managers to realize the need for change and take action. However, direct links between the role of foresight and specific outcomes can be hard to establish. This was also the situation with regard to the Cisco Technology Radar. True, some of the technologies that have been sourced by Cisco could be tracked back to the scouted technology reports. Proving that it was the radar initiative that made the difference was still hard. In a large multi-national company there

0040-1625 © 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

63

are thousands of internal and external factors impacting the organization. Even if all those variables could be controlled for, the effort required to measure the impact of the radar would be too demanding and costly. In Cisco, the people responsible for the radar were more distracted than truly disturbed by the So what?question. They considered other factors as proof enough that the foresight system was creating corporate value for example that more and more people asked to be involved and that certain departments requested tailored versions of the process. It was still hard to describe this kind of value more accurately. They really like it!is hardly a satisfactory answer to the So what?-question. Or is it? In the broader literature on strategic management there has been a growing interest in the role of cognition or managerial mindsets for enabling change. In this paper we make an effort to further specify how and why the technology radar makes a difference by looking into cognitive aspects of foresight. Instead of searching for examples of technologies, products or services that were sourced and developed thanks to the foresight initiative, we are in this paper focusing on the potential role of technology foresight for changing mindsets and ultimately, how this relates to value creation. We do this theoretically in Section 2 and empirically in Section 3 by asking the following guiding research question:

How and why can engagement in foresight impact ways of thinking and processing information?

In other words, we contribute to the understanding of value creation from foresight by highlighting benefits not likely to be directly measureable through looking at products or other end results, but rather by keeping an eye on the process. This makes it necessary to look deeper than the corporate leveldescriptions of value stemming from foresight, and consider also how individuals might perceive the benefits from participating. Through our framework and analysis we show how systematic probing for analytic thinking through foresight can be a key catalyst for changing mindsets, which will ultimately influence strategy as well as decision-making. We also highlight how the organizational context and the design of the foresight process can catalyze mental model updates. The paper is structured as follows: First, we establish a theoretical framework from which to approach the case (Section 2). To help us specify our focus we derive in total 8 propositions 5 regarding individual and 3 regarding collective aspects of the cognitive role of foresight. After summarizing the propositions (Section 2.5) we present our case (Section 3), more specifically the setting and methods (Section 3.1), the background of the Cisco Technology Radar (Section 3.2) and our results (Section 3.3). Finally, we discuss our results (Section 4), highlight key implications and limita- tions as well as suggest several avenues for further research before we conclude (Section 4.3).

2. Literature

In the following we theoretically consider not only the final outcomes of foresight, but also its potential for affecting habits and mindsets of individuals and thus creating value by enabling organizational renewal and change. We find that there is a

research gap particularly with regard to three aspects. I) Most current research considering the value of foresight does it on the organizational level meaning that deeperparts of the foresight process concerning the value perceived by individual stakeholders are less accounted for, particularly from a cognitive perspective. II) It is somewhat open whether the link between foresight and mental model change can be established also for activities that do not include scenarios, for instance technology foresight. III) While there are some arguments for why foresight can change mindsets, such as due to the memorability of scenarios, there is lack of a solid theoretical foundation from which to explore the cognitive value of foresight. A better understanding of the collective cognitive process of foresight what happens when individuals come together as participants in a foresight system is also needed. Whetten (1989) refers to theory-development authorities such as Dubin (1978) when listing four essential elements a complete theory must contain; what, how, why and who/ where/when (see Fig. 1). Our aim, in addition to exploring the how, is to contribute with further developments especially for the whyrelated to the cognitive dimensions of foresight. We are attempting first to build and then to verify the theoretical framework. In our efforts to identify literature describing the value creation of foresight, and more specifically cognitive factors, we have encountered a problem described by Rohrbeck and Bade (2012) in a recent review: The literature explores similar phenomena, but uses different terminologies. We identified papers describing the value creation of corporate foresight, of strategic foresight, of technology foresight and of specific tools and methods such as scenario planning and horizon scanning. Our empirical study is focusing on a specific case of technology foresight, but we are concerned with processes and mecha- nisms that can also be found in other approaches to foresight. We therefore felt it would do more harm than good only to consider literature on specific technology foresight tools. Instead, we have chosen to consider a broader literature on value creation from foresight, as long as the foresight dealt with in the papers involves systematic efforts to interpret and look out for the future. We follow the example of Rohrbeck and Bade and consider the larger field we position ourselves within as organizational future orientation. While our empirical study certainly makes our results more relevant for other radar projects or technology foresight initiatives on a more general level, we would argue that our theoretical framework can also serve as input for future research on any organizational foresight effort that takes the form of a widely implemented process, close to the best practices described in the literature (Rohrbeck and Gemünden, 2008).

2.1. Current understanding of foresight and cognition

Literature on the value creation of foresight is often concerned with its role for managing uncertainty. For example, foresight systems have been highlighted as a means for identifying drivers of change to handle the effect and response uncertainty through bettingon discontinuities, or more often creation of strategic options (Vecchiato and Roveda, 2010). Corporate foresight is also seen as valuable for perceiving, interpreting and responding to change as well as influencing other actors and enabling learning (Rohrbeck and Schwarz,

64

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

Forecasting & Social Change 101 (2015) 62 – 82 Fig. 1. The focus of this paper.

Fig. 1. The focus of this paper. We aim to answer how and why engagement in foresight can impact cognition, thus contributing mostly to the howand whyelements of foresight theory. Source: Own design, solely based on Whetten (1989).

2013). Related more to the future studies literature, it has been proposed that a key value of foresight is to enable organizations to learn the future faster, from problem exploration to problem resolution and action (Burt and van der Heijden, 2008; Burt, 2010). These roles can be linked to quantifiable value contributions such as the number of new products and services, the number of reported trends and technologies or the number of decisions made using foresight. But should they only be reduced to such measures? Particularly in the literature on scenario foresight, there is emphasis on the value for changing decision-making styles and mindsets of managers (Chermack and Nimon, 2008; Chermack, 2004; Glick et al., 2012; Meissner and Wulf, 2012), which by the strategic management literature has been identified as crucial for change and long-term survival (Huff et al., 1992; Reger and Palmer, 1996; Gavetti and Rivkin, 2007). A well- known example is Polaroid that, due to a prevailing mental model, failed to adapt to the growth of digital imaging (Tripsas

and Gavetti, 2000). In a critical survey on strategic foresight and organizational learning, arguments for the cognitive value of foresight in a wider context than scenarios are put forward by Bootz (2010). He talks about the cognitive virtues of the foresight attitude which facilitate the questioning of individual representations(p 1591) and suggests that beyond the traditional role of decision-making support, some foresight studies () could be regarded as knowledge management and tools to drive change(p 1593). Literature on horizon scanning (Miles and Saritas, 2012; Palomino et al., 2012; Brown et al., 2005) also emphasizes the role of this foresight approach for directly assisting and improving decision-making through its structured process of identifying emerging trends and techno- logical developments. Particularly with reference to public sector foresight programs, there have still been calls and attempts for a closer assessment of the impact of foresight for policy- and decision-making (Calof and Smith, 2012). Experts have concluded that while design of the foresight process as

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

65

well as organizational design is important for foresight to have impact, the attitudes and commitment of participants also play

a key role (Calof et al., 2012). However, the underlying

mechanisms through which foresight gains impact could still

be further elaborated on.

Bootz et al. (Bootz, 2010; Bootz and Monti, 2008; Bootz et al., 2009) also provided a matrix typology which helps clarify that there are both direct and indirect strategic implications of foresight. Strategic orientation and change drive have most direct implication for strategy, while decision-making support and mobilization have more indirect strategic impact. There are however still many unanswered questions with regard to the specific cognitive role, such as how the ability of foresight to change mindsets can be explained.

2.2. Specifying the role of mental models

Literature describing the cognitive aspects of foresight, and more widely the cognitive aspects of organizational and strategic management, frequently refers to mental models. Mental models are a comprehension of a specific domain in the mind

of an individual and are therefore important for organizations

and management to understand in order to enhance learning (Miller, 2003; Senge, 2006; Starkey et al., 2004). Within the scenario literature, Chermack (2004) describes mental models as

the lenses through which we see the world(p 301), including our beliefs, values and biases. Shared mental modelsthat can

be achieved by strategic conversation are also mentioned, in line

with van der Heijden (1997). Burt and Chermack (2008) refer to

the aha experience that moment when new insights have been gleaned and a mental model has shifted(p 290). Rook (2013) proposes a synthesized and specific definition of a concentrated, personally constructed, internal conception, of external phenomena (historical, existing or projected), or experience, that affects how a person acts.We build on this definition, and thus view mental models as individually held. Exactly what happens when mental models are shifted with foresight? Memorability is a central argument for why scenarios can help alter mental models (Chermack, 2004; Bood and Postma, 1997). In a review, Bood and Postma (1997)

specifically indicated that foresight in the form of scenarios can hinder bottlenecks to strategic decision-making because they a) decrease managers' inherent cognitive-inertia (the tendency

to stick with world views based on outdated information even

when new information is available), b) shorten the feedback lag between strategic decisions and the results and c) solve inefficiency and bias due to either too much or too little disagreement between members of a management team. However, from a cognitive perspective, deeper explanation of how foresight contributes to update mental models is needed. One might also speculate that other foresight processes such as scouting, scanning and assessment of technologies bring expe- riences that are less compelling and less likely to stick. Does this mean that they are also less likely to update mental models?

2.3. The dual view of doing and thinking

By drawing on literature on the main ways in which we process information, additional arguments for how and why foresight contributes to updating mental models can be put forward. Within cognitive and social psychology, dual process

theories have been around for 30 years or more (Evans, 2008) and have sustained their ground despite a paradigm shift in the psychology of reasoning (Elqayam and Over, 2012). Following the dual process account of Evans (2006), most activities involving conscious thinking outside the scope of everyday routines seem to contribute to updating mental models of individuals. According to this account, humans consider one hypothesis of the future, or mental model, at a time. Central for generating and considering these models are two main kinds of reasoning processes heuristic and analytical that alternate in controlling behavior. These processes are most popularly known as System 1 and System 2, as used for instance by Kahneman (2011) in his best-selling book on reasoning and decision biases, Thinking, Fast and Slow. 1 It will be outside the scope of this paper to consider the ongoing debate concerning the details of different dual process accounts. We will focus on the main aspects of the heuristicanalytic theory of reasoning (Evans, 2008; Evans, 2006; Evans and Over, 1996) and how it relates to updating mental models and affecting decision-making of individuals participating in foresight. The theory allows for an assumption that System 1 is more dominant in activities involving routines, while System 2 is more dominant in explorative activities. Dominance of System 2 can be seen as critically important for acquiring new knowledge outside of the boxand the success of foresight.

2.3.1. Mental exploration and exploitation When we talk about innovation, especially on the company level, March's (1991) terminology of exploration and exploita- tion in organizations is widely used. The ways to design dual structures for innovation (Duncan, 1976) by balancing explo- ration and exploitation in firms are discussed in the large and still expanding research on ambidexterity. Foresight has been linked to organizational ambidexterity through its identifica- tion of discontinuous change and development of radical innovations (Rohrbeck, 2010a) clearly related to exploration. One should be careful of drawing direct links between the frameworks, but we find it useful to refer to exploration and exploitation when using dual process theory to investigate the cognitive role of foresight. We link System 1 processes to exploitation and System 2 processes to exploration. As with exploration and exploitation on the individual or firm level, System 1 and System 2 are seen to represent two distinctive kinds of activities, in this case not in terms of observable actions, but rather processes in the brain. Most authors agree on differentiating between processes that are unconscious, rapid and automatic, and those that are conscious, slow, and deliberative (Evans, 2008). System 1 is generally perceived as fast and automatic, while System 2 is slow and controlled. Other words associated with the systems are heuristic, intuitive, holistic, and impulsive (System 1) and analytic, reflective, rational, and systematic (System 2) (Evans, 2008). As with exploitation activities in organizations, the System 1 processes are associated with routines. To save capacity, the brain applies routines to our reasoning, operated by System 1. This system generates mental models for consideration of the analytic system, by default to represent the most probable states of affairs given the current

1 Kahneman specied that the system labels should be interpreted as ctitious characters.

66

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

understanding. However, like in organizations, too many routines or too much System 1 thinkingcan result in a blind spot for taking into account important changes in the environ- ment. This is because the heuristic system contextualizes problems with reference to prior knowledge, the so-called relevance principle (Evans, 2008). Related to the example of Polaroid, it is likely that decision-makers rarely analytically questioned the very basis for their business model and to a large extent continued exploitation via existing routines mentally as well as practically. With intervention from System 2, implying more analytic processing and hypothetical thinking, the likelihood of updating the current understanding increases. However, these analytical processes are also prone to biases, due to our tendency to satisfice; to hold on to representations of reality that are good enough and thus save effort (Evans, 2006). In addition to cognitive capacity, factors such as personal motivation and task instructions are central for determining how much we will let the analytic processes intervene in decision-making. Time is also a key factor. Due to the capacity demands for System 2 processing, the analytic mode is less likely to dominate if decisions are made under time pressure. One important clarification is needed. We are probably not talking about two completely different physical systems in the brain (Evans, 2008), rather two main types of processes (Evans, 2006). As with activities of exploration and exploitation, one way of processing does not directly rule out the other. For application to foresight it is sufficient to assume that before a decision or judgment is made and thus action takes place, both types of processes will have been activated in the brain of an individual. The relevant question is which process is the more dominating; to which degree the analytic processes intervene and thus to which degree the mental model is updated.

2.3.2. Implications for foresight Fig. 2 summarizes the revised heuristicanalytic theory of reasoning as proposed by Evans in his 2006 extension (Evans, 2006) applied to foresight. In sum, the theory provides a good explanation for how employees and managers judge, decide and act, for instance related to perceived future business opportunities: They consider one hypothesis about the future at a time (singularity principle) and confine themselves to it until they find good reason to give it up (satisficing principle). More active intervention of analytic processing may result in modification or replacement of the mental model about the future. Cognitive ability (general intelligence), task instructions (for example strongly probing for deductive reasoning), personal motivation (for example that the problem seems personally relevant for the solver), personality or cultural context (resulting in a tendency to employ analytical thinking more often) and time are key factors that determine the degree of intervention. Evans extension largely builds on an idea proposed by Evans and Over (1996) that the analytic system is (more) involved when hypothetical thinking imagining possibilities that go beyond the factual knowledge of the world is required. A deeper explanation for why foresight is useful for updating mental models is thereby offered: When decision-makers engage in thinking about possible scenarios or trend outcomes for the future, analytic processing is employed to a larger extent. In a situation with no foresight process to stimulate analytic

processing, the heuristic system is more likely to bias actors to expect things in line with past experiences and not consider things outside the box. Thus, the possibility for updating the current mental model would be present every time a person engages in hypothetical thinking about the future. Even if the analytic processing would fail to intervene at some final decision-making-point, the likelihood that some updating has happened before will be increased compared to a situation with no prior engagement in foresight all other things being equal. We can therefore propose that

1a) The revision of current mental models is helped by engagement in foresight

When putting forward such a proposition, it also becomes relevant to discuss the role of the frequency of engagement. Intuitively, it is tempting to infer that more foresight engagement would, if proposition 1a holds, be better than less engagement. Considering the heuristicanalytic theory of reasoning, such a view can also be backed up by referring to mental processes. Due to the satisficing principle, there is a risk that individuals participating in foresight might still be biased to settle for good enoughand therefore sometimes fail to take important information into account. To overcome the satisficing principle and maximize the update of mental models, it would thus help to increase both the delivery of content for analytic processing provided by the heuristic system, and the analytical consider- ation of the hypothesis. All this is likely to be helped by regular participation in foresight, since this will bring more hypotheses about the future up for consideration as well as probe for more frequent analytical thinking, than only occasional engagement all other things equal. We therefore propose that

1b) Regular engagement in foresight increases the individual cognitive payoff from foresight, compared to only occa- sional engagement

Critically examining the latter proposition, however, we are led to question what is meant by regular engagement. To really get to the bottom of the cognitive mechanisms, we should also specify what kind of engagement we are referring to. For instance, participants in a scenario workshop are subject to other instructions and perform different actions than the readers and viewers of the final results from that workshop. Since both task instructions and personal motivation are relevant for determining the intervention of analytic process- ing, it is likely that individuals who are actively participating will update their understanding to a larger extent than passive readers, because they are probed more through task instruc- tions and in greater need of mobilizing motivation to act accordingly. We therefore propose that

1c) Participative (active) engagement in foresight increases the individual cognitive payoff compared to engagement only as a reader of end results (passive)

However, as suggested also in the foresight literature (Peck, 2009; Bezold, 2010), there may be personality differences when it comes to motivation for engagement in foresight, where motivation and perceived payoff are likely to be contingent on a match with personal preferences in thinking and doing. Such claims are in line with the heuristicanalytic theory of reasoning, which also suggests that individual differences may make certain individuals engage in analytical

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

67

Forecasting & Social Change 101 (2015) 62 – 82 67 Fig. 2. Dual processing and foresight.

Fig. 2. Dual processing and foresight. The top line conceptualizes how a business problem is tackled in a situation without engagement in foresight (driven b y heuristic processing, no attention to potential disruptions). The bottom line conceptualizes how participation in foresight stimulates modification of current beliefs through generation of scenarios that foster analytic thinking. Source: Own development, building on the dual process account of J.S.B.T. Evans; The heuristicanalytic theory of reasoning: extension and evaluation, Psy-chon. Bull. Rev. 13 (2006) 378395.

thinking more often and more naturally than others. The literature on scenario planning has already found evidence for links between scenario planning and an increased reliance on intuitive decision-making style (Chermack and Nimon, 2008). Technology foresight is however rather different from scenario planning, in that it relies more on fixed processes and systematic analysis, likely to be more related to analytic thinking. We therefore propose that

1d) Individuals with a high preference for analytic thinking are more likely to see larger cognitive benefits from technol- ogy foresight than individuals with a lower preference.

Moreover, due to the capacity demands of analytic processing, it is less likely that individuals will be able to update mental models under time pressure, where heuristics are more employed. Updating the mental models in due time before a critical decision-making-point is therefore likely to be important to make an informed high quality decision under time pressure. We thus propose that

1e) Preparation through foresight activities before a critical decision-making-point will I) shorten the time needed to take an informed decision at a critical decision-making point and II) increase the quality of a decision taken under time pressure compared to a situation with no preparation through foresight.

2.4. Collective aspects

Thus far we have concentrated on how cognition is relevant for foresight mainly on the individual level. However, collective aspects must also be explored in order to determine the final impact of such mechanisms on the value and success of foresight. Since foresight systems and processes often involve many individuals; what is the role of cognitive aspects for the value creation of foresight when individuals come together? Knowledge about value outcomes on the collective level has often been fueled by searches for best practices in how to organize foresight activities. For instance, Rohrbeck and

68

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

Gemünden (2008) offer a framework for benchmarking fore- sight, built from case studies. They point out the dimensions information usage, method sophistication, people and networksand organization and cultureas central for determining the value of foresight and further subdivide them into 19 characteristics. Battistella (2014) also provides an organizing framework based on case studies, and specifies the different effects of structure, coordination, decision processes and control systems on the efficiency (related to return on investment) and effectiveness (related to goal achievement and satisfaction) of foresight. A third recent paper provides a complex framework to judge the contribution of foresight towards more participatory knowledge societies(Amanatidou, 2014), and sees foresight as a system of elements and interactionsthat are both affecting and being affected by internal as well as and external factors. These examples show that a) numerous aspects play a role in determining the value of foresight when analyzing it on the collective level, and b) there has been a need for systematizing these aspects and sorting them into smaller categories in order to make the frameworks relevant for practice and much progress has been made by recent contributions. There is however still room for more clarification on how collective aspects influence the value of foresight related to cognition. In the following, we make an effort to explain how and why individual engagement in foresight, stimulated by analytic processing, can be transformed into a collective payoff in the first place. We also use insights from social cognition to specify how shared and conflicting mindsets matter for the value of foresight, and how control mechanisms govern cognitive payoffs through influencing motivations.

2.4.1. Collaboration and cognitive bonus Based on the recent contributions to the foresight literature (Rohrbeck and Gemünden, 2008; Battistella, 2014; Amanatidou, 2014), it can be summarized that there are certain general ways of organizing for foresight that will yield better payoffs than others, and that several factors influence these payoffs. The direct relations with cognitive processes are still to be further specified. Still, the insights regarding best practices can be built on to understand the rel ation between individual cognitive payoffs and how such payoffs can be enforced through interaction. Firstly, using the terminology of Rohrbeck and Gemünden, people and networks will likely moderate how much informa- tion is brought in and cued for consideration by analytic processes. Secondly, organizational aspects such as time available and perceived risks and rewards will likely moderate engagements in foresight thinking, thus also affect the amount of new perspectives considered. Thirdly, the general openness and willingness to share different views will mediate the effects of the former two points. If debating the status quo is very common and accepted in the organization, individual mental models can be updated and/or further enhanced by exchange with other updated individuals. If there is not a culture for challenging current points of views, the mental models of a foresighted minoritymay be negatively affected through pressure from a larger group of lagging individuals who are rejecting input. We can therefore assume, rooted in the literature on how foresight ideally should be established and organized, that

collective processes can both moderate and mediate the outcomes of foresight, from a cognitive perspective. As such, we suggest that

2a) When foresight is at its best practice, individuals (and the firm) receive a collective cognitive bonus from participation.

The explanation for this bonuscan again be found in the basic understanding of how the mind works, and is related to capacity constraints. Each individual only has a certain capacity for analytic processing. However, when exchanging views or mental models with other individuals engaging in similar hypothetical thinking about the future, they are offered a shortcut compared to processing everything themselves from scratch. It follows from this that there are potential efficiency gains from involving many rather than few people in foresight activities at least if ambitions for engaging in hypothetical thinking are high. Outsourcingsome parts of the analytic processes to other individuals and receiving back more finishedmental models for heuristic cuing saves time and effort, for both the individual and for the group. This latter explanation is in line with the idea that assumptions of mental representations at the individual level should not be lifted to understand organizational level outcomes without taking interaction into account (Allard-Poesi, 1998).

2.4.2. Culture, coincident meanings and control mechanisms When a foresight system is introduced into an organization, it is normally not done from a blank sheet perspective. Therefore, it is unsurprising that achieving a fit between the organizational culture and foresight is emphasized. Organiza- tional culture can already be directly linked to cognitive aspects, in that it is defined as a pattern of beliefs and expectations shared by the organization's members (Schwartz and Davis, 1981). A related view for instance, is of culture as normative order, with the implication that culture can act as a social control system by shaping the behaviors of individuals and groups. Norms can vary in two dimensions: the degree of consensus regarding the norm, and the intensity of the approval/disapproval. Only when there is both intensity and consensus can a strong culture exist (O'Reilly, 1989). The individual perspective the dual process perspective also suggests that culture influences individuals' tendency to engage in analytical thinking (Evans, 2006). However, there are good arguments for questioning a) if "culture" is the right label to use when considering organiza- tional aspects of foresight, and b) the sometimes seemingly implicit assumption that shared mindsets are what create strong value outcomes from foresight. For instance, based on the view of culture as normative order, we are at first glance dealing with somewhat of a contradiction in regards to the cognitive value of foresight. For foresight to be effective, mindsets should constantly be opened up, updated and negotiated across functions and departments, indicating the importance of conflicting beliefs. However, some kind of consensus or shared mindset must also be present, with regard to the perceived importance and usefulness of foresight, for people to actually participate in foresight activities. To resolve how the mental states of alignment and complementarity can coexist, it must first be acknowledged

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

69

that representations at the collective level are something beyond simple aggregations of the mental models of individuals. Second, that such representations are not stable. The view of culture as normative order hints at this by explicating how norms can vary in both intensity and degree of consensus. However, by drawing further on the perspective of Gray et al. (1985), and their metaphors of construction and destruction of coincident meanings in organizations, we get closer to understanding the underlying mechanisms in play. Based on their account, we can assume that at any-given time, some organizational meanings will be coincident and others will not. Moreover, it is similar valuing that fuels organizational members' meanings and indicate the highest degrees of coincidence. Such values are often not expressed in daily conversation, but remain tacit influences on choices and behavior. For instance, full coincidence of meaning is possible if all members of an academic department perceive the new department head will work for each of their interests(p 87). If individuals have different perceptions of how their personal welfare is influenced by certain experiences, the coincidence will be lower. Applied to foresight, this first of all explains how shared representations (regarding the usefulness of foresight, perhaps due to an understanding that it will ultimately lead to personal gains for the participant) and disparate views (with regard to the interpretation and relevance of the issues put under focus by foresight activities) are not contradictory. It also provides the insight that, in order for individuals willingly to engage in foresight, they do not necessarily need to share the same views on why such an engagement would be useful, as long as they all share the understanding that it is somehow a good idea to take part. It is however suggested by Gray et al. that greater coincidence among the tacit value systems produces more socially binding meanings. A foresight system that responds to a broad set of values reflecting many individual interests could possibly still be more effective than a system that is geared towards one main rationale for participation. The reason for this is that the balance between seeking to control the behavior of individuals (in effect, secure their motivated participation) and avoiding a streamliningof meanings (which can be harmful to the quality of the foresight process) will be easier to manage. This idea represents an important difference from the assumption that a good fit between foresight and the organizational culture means that individual foresight mindsetsmust necessarily be similar in all aspects and, consequently, actively controlled by management. Nevertheless, social cognition theory tells us that the need for socially shared cognition to believe that one's ideas correspond to the ideas of the group is a fundamental influence on the processing that occurs in the human brain, and is related to an overall need to belong. Put simply, people do not do well unless they are motivated to get along with at least a few other people(p 4849) (Fiske and Taylor, 2013). However, this functions as yet another reminder that "shared" does not necessarily mean identical, and that motivations do indeed play a role in influencing the cognitive modes of individuals. What can this tell us about value creation from foresight at a collective level, related to cognition? Most of all, it nuances the prescriptions regarding how to organize for foresight. While the culture-as-normative-order-view would be likely to suggest managerial focus on cultural control mechanisms in order to secure increased approval for foresight

through creation of strong norms, the constructiondecon- struction rather inspires governing for diverse beliefs and values. In line with these two approaches to understanding the creation of meaning and change in organizations, we will in the following put forward two further propositions. First, when it comes to control mechanisms, four basic mechanisms have been linked to creating shared views: I) systems that provide for participation; II) clear and visible actions from management that support the cultural values; III) consistent messages from coworkers regarding how to behave and interpret key events; and IV) comprehensive reward systems that focus on recogni- tion and approval rather than (just) money. Based on this line of thought, we propose that

2b) Cultural control mechanisms will be important for influencing motivations to engage in foresight and thus enforce the participation in, and value creation from organization-wide foresight systems.

However, taking into account the insights from the constructiondeconstruction view, and remembering that too much consensus could influence analytical thinking negatively, we also propose that

2c) A foresight system that corresponds to a broad set of values reflecting many individual interests rather than one main shared rationale for participation is important for securing maximum motivation to participate in foresight, while at the same time avoiding limitation of diversity in thinking within the foresight activities.

A useful perspective to help understand the link between propositions 2b and 2c is to take into account a more nuanced view of organizational culture. The competing values frame- work (Cameron and Quinn, 1999) acknowledges that multiple values are present in most organizations, but distinguishes between four main organizational culture types. Central in the competing values analysis is the organizational glue, which unites people within a culture. The clan culture organization is held together by loyalty or tradition, while the adhocracy culture is held together by commitment to experimentation and innovation. For the market culture, the glue is an emphasis on winning, while formal rules and policies hold the hierarchy culture together (Cameron and Quinn, 1999). This framework can be seen as a useful practical approach to balancing the implications of propositions 2b and 2c. An important driver for utilizing the cognitive value of foresight could still be to implement some control mechanisms, but at the same time making sure that the system set-up can function as gluefor the different culture types. Ensuring people unite and talk across organizational cultures, even though they do not necessarily value participation in the same way, could be particularly important in the early stages of introducing the system. This is because, at this point it would be more difficult to deconstruct existing meanings and rationales in the organization, so foresight efforts would have a larger need for fitwith the existing culture(s) in order for individuals to grasp and buy into its purpose. However, in the longer run, the foresight system could potentially also contribute to change culture through influencing ways of thinking and constructing new meanings. This is more likely to happen after a critical mass of users have joined and seen the benefits. As reflected in the propositions above, the challenge for managers would then

70

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

be to manage the deconstruction and construction of meanings in relation to foresight, without negatively influencing the cognitive diversity among the individual foresighters.

2.5. Summary of the propositions

Guided by our overall research question we have reviewed relevant literature and suggested how revision of mental models is helped by engagement in foresight (proposition 1a) and how frequent and active engagement increases cognitive payoffs (proposition 1b and 1c) compared to passive and less frequent participation. Furthermore, we have proposed that technology foresight might be particularly beneficial for analytic thinkers (proposition 1d), and that foresight can shorten the time needed to take an informed decision at a critical point in time as well as increase decision quality (proposition 1e). We have also proposed that there is a potential cognitive bonus when foresight is at its best practice (proposition 2a) and that cultural control mechanisms are important for motivating for foresight thinking (proposition 2b). However, a foresight system that corresponds to a broader set of values reflecting many individual interests rather than one main shared rationale for participation is also important for avoiding limitation of diversity in thinking within the foresight activities (proposition 2c). After the introduction of our methods (Section 3.1) and case (Section 3.2), we will, in the next chapter, present our empirical findings (Section 3.3), and discuss how they validate the propositions and our overall research questions (Section 4).

3. Case

In the following we retain the cognitive lens and apply it to studying the case of the Cisco Technology Radar. A case study was considered useful in order to extract deep knowledge, since it is an ideal methodology for a holistic, in-depth investigation (Feagin et al., 1991). The single case approach allowed for exploring the phenomenon in its context without sacrificing the richness of the studied incident and its context (Eisenhardt and Graebner, 2007), meaning that we acquired useful business-relevant insights deep enough to also inform our specific theoretical propositions. While it is correct that statistical generalization from a case is not possible, analytic generalization can still be made when previously developed theory is used as a template against which to compare the empirical results of the case study(Yin, 1984). In the terminology of Stake (1995), this can thus be seen as an instrumental case study, meaning that the case is used to understand more than what is obvious to the observer. The unit of analysis is the case organization, with use of data also from the community within the organization that has been, or is, engaged in foresight. The setting and methods of the case study will be described below, followed by an initial description of the company and the relevant events and stakeholders. The results of the study will then be presented and discussed based on the key dimensions of our theoretical framework.

3.1. Setting and methods

Cisco Systems (Nasdaq CSCO) is a multi-national technology giant founded in 1984. Evolving from a sole focus on routers and

switches, Cisco has greatly shaped the infrastructure underlying the Internet and eventually transformed itself into an end-to-end network provider. Our interest in the company is related to the Cisco Technology Radar project. One of the authors has been involved with this project throughout its existence, while the other author has been familiar with it since autumn 2012, when the case study began. Even though this paper is specifically concerned with the aspects of foresight related to cognition, we chose to include broader spanning questions in the first part of our investigation to secure a good basic understanding and avoid missing any important perspectives. We capitalized on the outsider point of view of one of the authors by conducting a general report of the case, citing 35 sources, most of them open-access, as a first step. These insights were then validated and complemented through personal interviews with the second author, who should be considered an internal expert, as well as studies of internal documents. The insights provided by the expert authorwere further validated through talks with key members of the organization at different levels. We also greatly benefited from one of the authors many experiences as an observer in radar process meetings and as a facilitator of scouting activities. We were, however, aware that even though the insider perspective was an advantage in terms of having good access to details and in-depth knowledge about the case, it could also lead to bias if this knowledge was not challenged and old truthswere taken for granted. We therefore consistently engaged in critical discussions, challenging one another's assumptions. We also applied relevant frameworks, such as the best practice dimensions of Rohrbeck and Gemünden (2008) and the OCAI framework of Cameron and Quinn (1999) in order to achieve a solid theoretical basis for discussing the case. Each author scored the Cisco foresight system using the frameworks, before having a joint discussion to reach the final scores. Finally, we chose to gather additional data using approaches where the data analysis was less dependent on our interpretations. Central to the more qualitative analysis, were 25 written testimonials from different stakeholders. These were originally collected as part of the preparation for an internal innovation competition at Cisco, where the stakeholders were asked to provide their opinion about the radar. To analyze the statements we chose the method known as cognitive mapping, or more accurately in our case; causal mapping, as it can enable a more detailed examination of subjective beliefs and at the same time offers a meaningful way to portray the data (Swan, 1997). In order to locate the central beliefs regarding why the radar was perceived to be valuable, statements were coded and reorganized, and maps representing relations between smaller statements were combined into a larger map, along with recommendations from the literature (Huff, 1990). Specifically, we looked for the underlying rationales of the endorsements, often following key words such as because, and, or it is. The main advantage of this method was that underlying beliefs and rationales could be identified; tacit information such as this can be harder to get hold of through more direct questions in interviews and questionnaires, since it is not always in the conscious mind of the respondent. Through the analysis of statements that had been freely formulated by the informants, we were able to map out data-rooted rationales less vulnerable to our own phrasing and language.

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

71

However, we later triangulated the causal mapping results through informal talks with key stakeholders within Cisco, as well as an anonymous online questionnaire. The choice of the questionnaire was based on the rationale of complementing the insights that one of the authors in particular gained through deep involvement in the radar programme, with a measure- ment providing harderand more specific data. The question- naire was built to further check the relation between (degree of) engagement in foresight, thinking styles and perceived value outcomes, necessary to more accurately answer some of the propositions. We used a standard Likert-scale to measure the degree of agreement to pre-defined statements. The statements were developed based on themes identified through the document studies and the mapping of the 25 testimonials. For instance, we asked participants to indicate to what degree they agreed with statements such as: The radar helps save time,or the radar slows down the decision- making process. The respondents were reminded that there were no right or wrong answers, and that their honesty was important. We also collected information on the respondents' degree of involvement in the foresight project and measured their decision-making style. The latter was achieved through use of the General Decision-Making Style (GDMS) question- naire of Scott and Bruce (1994), a validated measure also previously used in foresight research (Chermack and Nimon, 2008). We built in questions to be able control for degree of familiarity with the radar and to check for differences between different roles (such as scout, reader, panelist) and positions (such as director, manager, VP). In order to not bias the respondents by making them reflect on their decision-making style before answering the other questions, we chose to ask the decision-making style questions towards the end of the questionnaire. These questions were followed only by a question re-inviting them to rate the value of the radar as a decision-making tool after having become more aware of their own decision-making habits, and an open comment-boxinviting any additional comments. The questionnaire was distributed via a link sent through e- mail, targeting some 200 key stakeholders active in the technology radar. This yielded around 60 replies, among which were 39 fully completed questionnaires mainly from manager- and director level decision-makers. Although the sample is small, the questionnaire was seen as valuable particularly because the responses were anonymous. More- over, it was an efficient way of getting in touch with respondents that otherwise could have proved difficult to reach, given the size of the company and the challenge of coordinating with many people at different locations. For selected measures the questionnaire data were also found to be acceptable for simple statistical analysis. However, it is our belief that the more robust insights from the case are the ones that do not rely solely on the questionnaire, but rather are based on a mix of our qualitative and quantitative data.

3.2. Background and description of the Cisco Technology Radar

The formal history of the Cisco Technology Radar started in 2011, when one of the authors secured the early funding for the project after some intensive months of planning and prepara- tion. Cisco, despite its roughly 66,000 employees in 2011, still had a somewhat entrepreneurial culture. This meant that there

was no structured and consistent way of identifying, selecting, assessing and disseminating company-wide technology intel- ligence. The idea was to establish a network of technology scouts to provide early identification of novel technologies and trends, to enable informed strategic decision-making and to help stimulate innovation. After assessment the key trends would be visualized along a so-called technology radar, a format greatly inspired by research publications regarding similar projects (Rohrbeck, 2010b; Rohrbeck et al., 2006). The radar reports would serve as input for strategic decision- making as well as to stimulate innovation in Cisco through dissemination of knowledge and best practices. The recruit- ment of scouts, assessment panelists and other key stake- holders would span across many departments, functions and geographies. During its first year the radar was successfully expanded from its first test version involving only employees working in corporate development, to become a company-wide technol- ogy foresight system. For example, the 9th volume of the radar published in June 2012 was based on 233 novel technologies submitted by 64 technology scouts from multiple organizations across North and South America, Europe, Japan and Australia. The technologies had been selected and assessed by technical and business leaders including Fellows, Distinguished Engi- neers, Directors and VPs. The Cisco Technology Radar website served more than 250 unique visitors, and had more than 1,900 page views per month. Moreover, printed copies of the Technology Radar were distributed quarterly to key decision- makers, including Cisco Fellows, the Operating Committee and selected executives. The Cisco radar can be described as both an outcome and a process. Regarding its outcome, selected technologies are summarized in a figure resembling that of a radar screen. The plotted technologies are displayed horizontally along the radar according to their relation to different technology fields or organizational units. They are also organized vertically based on their maturity level, and plotted using symbols indicating their relevance for the company (see a visual example of the radar in Fig. 3). The idea of the condensed format is to provide easy access to a quick summary of which technologies Cisco should be keeping a particular eye on, with the possibility of reading into the details in supplementing documents. The visual radar, is in other words the main element and is regularly disseminated to key decision-makers and experts within the company. However, much of its value is also believed to be created during the process leading up to the visual plot, as well as what happens when the technology plotis once again taken up for discussion and re-prioritized in the process leading up to the next quarterly radar report. Effort has therefore been put into optimizing the radar process, which in addition to the dissemination phase consists of technology identification, selection and assessment (see Fig. 3 for overview). In the first phase, the focus is on bringing in information. In principle, everyone in Cisco can assist with this by becoming a technology scout a person engaged in identifying new and relevant technologies for Cisco. An interac- tive website has been set up where employees can both view and contribute to the radar by submitting scouting reports. However, in line with recommendations from other radar projects, a community of more specialized and regularly engaged scouts secures a satisfactory level of expertise and continuity in the

72

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

Identification: An interna onal network of technology scouts submits novel technologies. All Cisco employees can
Identification: An interna onal network of technology
scouts submits novel technologies. All Cisco employees
can become technology scouts. Inputs are also
increasingly being gathered through joint efforts with
academia, venture capitalists, customers and partners.
Selection: The radar team and a panel of senior technical
experts review all technologies on a quarterly basis. The
selec on criteria are built upon the concepts of
Technology, Impact and Novelty.
Assessment: The selected technologies undergo a deeper
analysis led by the scout. A one-page technology profile is
wri en for an execu ve audience, pu ng the technology
in a business context and limi ng the content to unbiased
facts and findings. Senior technical and business leaders
assess the relevance of each selected technology for the
company. The assessment criteria are based on two axes:
market opportunity and implementa on risk.

Dissemination: The Cisco Technology Radar is published every quarter. The radar screen displays all technologies along with their maturity, their posi on in the IT/Telecom value chain, and their relevance as rated by the assessment panel. The Technology Radar is distributed throughout Cisco using a website, a PDF version and a newsle er.

Graphs and descriptions: Cisco 2013

Fig. 3. Overview of technology radar process graphs and description: CISCO 2013.

scouting process. These scouts are recruited through numerous channels, for instance through the HR department or by consulting internal information brokers people who possess detailed information on who would fulfill the scouting criteria of being both highly knowledgeable within a field and having a vivid social network. Efforts are also increasingly made to recruit

external scouts from academia and other collaboration partners. While the scouts work mainly individually preparing initial reports about emerging technologies, there is much more interaction in the next phases of selection and assessment. Every quarter, a panel of editors agrees on a maximum of 20 technologies to be presented to a panel of engineer experts, often

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

73

by the scouts themselves. The experts discuss the relevance of the technologies according to criteria such as novelty, changes in complementary technologies and degree of awareness about the technology and its application. After the panel has agreed on which technologies they see as the most relevant, scouts can be asked to prepare more detailed reports. Finally, the selected technologies go through a reality checkwhere they are rated by an assessment committee, who discuss their market impact and realization potential. The members of this committee are typically technical leaders and directors, ensuring that key stakeholders within Cisco take part in the radar process and feel ownership with regard to the results. Finally, based on the ratings of the assessment committee, the technologies go into the visual radar screen, which is disseminated to the wider organization. It is natural to question about the difference between the radar process and other common foresight methods, such as scenario planning and technology forecasting. Similar to scenario planning, there can be said to be an emphasis on generating images about the future, particularly in that scouts are asked to think about which technologies will be relevant for Cisco from a long-term perspective. However, the explicit focus can be said to be more on identifying and assessing all relevant technologies to look out for, rather than using them to construct more or less likely scenarios describing what can or will actually happen. The most visual outcome of the radar process is the radar illustration, which has the potential for diving deeper into quite technical descriptions of the different technologies, depending on the interest of the reader. In this way, the idea of the radar is perhaps more about having a constant working tool, to synchronize efforts across functions and departments, and to secure alignment with the company's strategy through the processes of expert selection and assessment. Similar to other technology forecasting methods, however, there is also an emphasis on rational and explicit methodology, with the aim of improving decision-making. For instance, it was learned through trial and error how critical it was to separate scouting (information input) from the discussion of the technologies (information prioritizing/rat- ing). Anonymous voting is practised to minimize for instance, social biases in the decision-making when the technologies are selected and assessed. The criteria that the selection is based on are also designed to minimize what could otherwise be quite emotional influences on the rating, due to the fact that deeply engaged experts from different business areas within Cisco are put in a setting where they need to converge. Another foresight method that can be said to have some similarities with the radar process is horizon scanning, which is being increasingly used to support governments and private organizations in their policy- and decision-making (Palomino et al., 2012), perhaps particularly in the medical and health industry (Miles and Saritas, 2012; Brown et al., 2005). Horizon scanning, sometimes described as a subset of environmental scanning, according to most definitions involves a systematic approach to bringing and assessing information about future trends. The technology radar can certainly be included in such a broad definition, however compared to typical horizon scan- ning methodologies, the information gathering phase in the radar must be said to be less based on systematic research and literature reviews. Also, the assessment information gathered could be said to be more dependent on people interacting and

discussing, rather than a more document processing based assessment. Finally, as has also been called for with regard to horizon scanning (Miles and Saritas, 2012), there is perhaps more emphasis on the format for dissemination of the scanning results in the radar process. Still, for those very familiar with the horizon scanning approaches, it might make sense to describe the radar as a deep and technology focused form of network based long-term horizon scanning, where assessment of technologies by key decision-makers is built into the process before the final report and results are more widely disseminated. A key element of the Cisco Technology Radar initiative has also been the structured approach to implementation and roll- out. Collaboration with the HR-department was set up at an early stage to improve the motivation and recruitment of scouts. Scouting was for instance incorporated in the courses offered for internal training, and completion of scouting made the participant eligible for points that could be used to buyother rewards. E-mail signature badges for bronze, silver and gold scouts were also developed as part of a more intrinsic reward system. The bar for offering tailored versions of the radar for different departments was, at the same time, held low, in order to maximize use and buy-in. The service department, who liked the approach but felt they needed a version adjusted to their focus on products and services rather than technolo- gies, were among the first to make use of a custom-made version. To secure early funding and further buy-in from key decision-makers, a stakeholder management framework was actively used by the radar implementation group. Despite having gained substantial momentum during the first couple of years, the So what?"-question kept re-appearing from time to time, asking for ultimate evidence of a scouted technology leading to a major business opportunity for Cisco. Still, engaged participants both inside Cisco and increased interest from various external partners, kept assuring the core radar team that they were creating value. In late 2013, Cisco sent out a media alert saying that they would host a public technology radar webcast, broadcasting an expert panel discussion on which technology trends would define the next year and beyond. In February 2014, press reports about the details of Cisco's first public Technology Radar Studybegan to emerge. As for the original, internal radar project, it was as per early 2014 still under continuous development.

3.3. Results

In the theoretical part of this paper we focused on how the literature could inform our main research question about how and why engagement in foresight can impact ways of thinking and processing information. The question guided our selection and discussion of literature, from which we generated 8 propositions. However, in order to identify potential value contributions related to cognition, we started by broadly mapping out the perceived values of the Cisco Technology Radar. In the following we will first present the general findings and then move on to discuss specific cognitive aspects. In the discussion ( Section 4 ) the findings will be linked more directly to the overall research question we started out with, followed by a discussion of implications, limitations and conclusion.

74

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

Forecasting & Social Change 101 (2015) 62 – 82 Fig. 4. Rationales for endorsing the technology

Fig. 4. Rationales for endorsing the technology radar. Grouping of different reasons given for endorsing the technology radar. Collected from 25 written testimonials. Rationales in bold frames represent the 5 main groups. Supplementary explanations (if any) are shown in the connected frames. Illustration made using the knowledge modeling program CmapTools.

3.3.1. Findings related to the value contribution of the radar Through our studies of corporate documents we clarified that the main goals of the radar are described in the following way by Cisco: I) Early identification of novel technologies and trends, II) enabling of informed strategic decision-making and III) stimulation of innovation. Cisco's emphasis on these roles of foresight can be explained through looking at the company's history as well as recent developments. Historically, Cisco has been renowned for its acquisition strategy. Taking over smaller, innovative firms has been the preferred way of getting access to expertise and technology, making expensive and perhaps unproductive internal research less called for. Cisco iPrize, the company's global innovation contest, and its iZone internal employee ideation site are however examples of Cisco's commitment to fostering open and crowd-sourced innovation in the later years. Thus, the Technology Radar initiative in Cisco can be seen as part of this movement towards internally fostered innovation as well as innovation in collaboration with key partners and customers:

Novel technologies are at the core of product and service innovation. They provide new routes to differentiation, cost reduction, and lead to new business opportunities. By disseminating information about novel technologies devel- oped both inside and outside of Cisco, and by accelerating external engagements with our ecosystem through the sharing of best practices and the development of joint technology visions, the Cisco Technology Radar paves the way for open innovation.reads a recent introduction document presenting the radar project.

Acquisitions are however still an important part of the growth strategy in Cisco. Because the radar locates, assesses and filters information about threats or opportunities impacting Cisco's future business, it is also seen as valuable for assisting decision-making about acquisitions. Additionally, values related to the efficiency of the radar are highlighted in corporate documents and presentations. Implicitly, decision-makers have limited time for details, but the radar consolidates information

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

75

and provides them with the intelligence needed to stay on top of the technology landscape. Individual stakeholders within Cisco, however, value slight-

ly different aspects and use different wording than in the more

formal corporate descriptions of the radar. The main rationales for key stakeholders' endorsements are summarized in Fig. 4 and can be grouped in five main themes: Stakeholders appraise

the radar because they genuinely like the process, they believe

it is good for business, it brings new perspectives, helps them

carry out their tasks and helps communication. As visible from the figure, some highlighted features particularly relate to their own personal benefit, for instance that they are given contact time with the top management level through participation in the radar process. Others relate it more directly to corporate goals, or to outcomes such as better communication and collaboration. Some also say it helps them carry out foresight related tasks, perhaps indicating that some of the radar activities are among their formal responsibilities. Many, however, simply highlight that they really like the radar process, be it because of the structured way of working, because they are impressed with the team behind it, or that the radar as a tool simply is the best. In the anonymous survey we had the opportunity to ask respondents to further rate the values. Here, we saw that stakeholders familiar with the radar, rated values related to collaboration and communication higher than some of the values initially expressed at the organizational level. The radar helps understand what other peers are doing, the radar helps people communicateand the radar helps people connect and socializewere agreed more upon than corporate languagestatements such as the radar is a key input into strategic activitiesand the radar is a key pillar of Cisco's Innovation Engine. Table 1 shows the descriptive statistics from the survey. In summary, the value contribution of the technology radar in Cisco seems to vary more than corporate documents describing its' purpose would initially indicate. Even though individuals agree that the radar is valuable, the underlying rationales for why they believe so differ between 5 main hidden values” — some overlapping with and some compli- mentary to the values as communicated on the corporate level.

3.3.2. Findings related to the radars inuence on individual cognition

3.3.2.1. Proposition 1a. The radar as an outcome is reported to be

a tool for decision-makers. Moreover, the radar process itself

seems to foster analytical thinking, in line with our expecta- tions (proposition 1a). For instance, technologies that have been on the radar for more than one year are systematically reviewed to decrease the threat of aged information or outdated mental models. Moreover, in the selection and assessment process participants are forced to analytically process and reconsider their current understandings through considering mandatory questions and decision criteria. During the selection of technologies, engineers are helped to realize how the scouted technology has relevance to Cisco's strategy, as one participant put it when endorsing the radar. In the selection panel the panelists first discuss and then anony- mously vote on which technologies to select, meaning that points of views are converted and stakeholders upgrade their

understanding of relevant technologies during these meetings. In the assessment panel there is, however, more focus on extractionof mental models of the panelists, rather than revision. This is because they rate the technologies based on their current understandings, and there is less debate. Still, through being exposed to the technologies that have already gone through the selection phase, assessment panelists' views on the technology landscape are revised. As for the scouts, they also seem to experience revision of current knowledge after having submitted their initial reports. One engineer stated that the radar helped him connect with leaders and business units he would not otherwise interface with. Through this, I learn their business requirements and their technical objections, and adjust the proposal to accom- modate those requirements and objections or engage with the business unit leader to explain how we can leapfrog competition with the new technologyhe wrote as part of his endorsement of the radar. One statement from a key Cisco customer also shows that participation in the radar yielded some "Aha" experiences or update of current mental models:

The ideas that we discussed have some significant legs to take us some places we would have not previously thought of but seem obvious now, he wrote. In the questionnaire (see Table 1 for details) we further tested the perceived relevance of the radar for influencing analytical thinking and changing mindsets. Most of the respondents familiar with the radar (N = 32) agreed somewhat or strongly that the radar helps provide an overview of the most important technologies (81%), that it helps assess technologies (78%) and that it allows for faster (72%) and better (75%) information processing. The statement that the radar changes mindsets received more ambivalent answers, however with most answers still on the agree-side (47% somewhat or strongly agree, 38% uncertain, 15% somewhat or strongly disagree). Most respondents (75%) agreed or strongly agreed that the Radar improves my current understanding, indicat- ing revision of mental models through engagement in the foresight process.

3.3.2.2. Proposition 1b. We expected to find that people who engaged more regularly in radar activities would report greater benefits (proposition 1b). From the questionnaire data we saw a consistent tendency that respondents who reported high familiarity and frequent exposure to the radar assessed it more positively. For example, all respondents exposed to the radar on a monthly or weekly basis (N = 16) agreed or somewhat agreed that the radar allows for better information processing, while people who were quarterly or even more rarely exposed (N = 23) disagreed or were uncertain to a larger extent. The same tendency was observed for all questions regarding the cognitive payoff from the radar, with the largest differences regarding the statements the radar helps save time, the radar facilitates decision-makingand the radar changes mindsets. For the last one we found a strong positive relationship with level of exposure through statistical testing (Pearson's r of 0.42, p = b 0.01). Moderate positive relation- ships between self-reported familiarity and the total assess- ment scores of the radar and between the self-reported exposure to the radar and the total assessment scores (Pearson's r of 0.3 for both) were also found. The p-values were however slightly above the acceptable level (p = 0.06)

76

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

Table 1 Answers from key radar community members. Answers from anonymous online questionnaire developed to further check the relation between engagement in foresight, thinking styles and perceived value outcomes. Decision-making styles were measured based on the measure of Scott and Bruce (1994).

Item a

N b

Mean

Std Dev

Min

Max

Exposure to radar Familiarity with radar The radar helps provide an overview of the most important techn. The radar helps assess technologies The radar allows for faster information processing The radar allows for better information processing The radar helps understand what other peers are doing The radar helps people communicate The radar helps save time The radar helps people connect and socialize The radar is a key input into strategic activities The radar facilitates decision-making The radar helps in uence decision-makers The radar changes mindsets The radar increases the quality of work The radar helps make more objective and less biased decisions The radar helps people do what they should have done anyway The radar helps save money The radar helps work faster The radar is a key pillar of Cisco's Innovation Engine The radar doesn't impact the quality of discussions The radar helps people carry out their daily tasks The radar changes the organizational culture The radar is irrelevant for business The radar slows down the decision-making process Sum radar assessment The radar improves my current understanding The radar helps me process information faster The radar helps me be more objective and less biased The radar helps me reduce riskiness of decisions The radar helps me convince others of decisions Sum decision assessment Rational decision-making style Intuitive decision-making style Dependent decision-making style Avoidant decision-making style Spontaneous decision-making style

39

2.36

0.93

1

4

39

3.74

1.43

1

5

32

4.09

0.86

2

5

32

4.06

1.01

2

5

32

3.97

0.97

2

5

32

3.91

0.89

2

5

32

3.88

0.83

2

5

32

3.56

0.84

2

5

32

3.56

0.80

2

5

32

3.41

0.76

2

5

32

3.41

1.13

1

5

32

3.41

0.98

2

5

32

3.38

0.87

1

5

32

3.34

1.00

1

5

32

3.22

0.83

1

5

32

3.22

0.87

2

5

32

3.19

0.69

2

5

32

3.19

0.64

2

5

32

3.13

0.79

1

5

32

3.09

1.12

1

5

32

3.03

1.00

1

5

32

2.78

0.87

1

5

32

2.69

0.90

1

5

32

2.47

1.22

1

5

32

2.25

0.84

1

4

32

63.91

13.53

30

91

32

4.0625

1.0453

2

5

32

3.6875

1.0298

1

5

32

3.5625

1.1341

1

5

32

3.4063

0.9456

1

5

32

3.375

0.907

1

5

32

18.094

4.4603

8

24

39

4.3654

0.4793

3.5

5

39

3.559

0.5215

2.4

5

39

3.5692

0.6736

1.8

5

39

1.841

0.9075

1

5

39

2.3179

0.8344

1

5

a A Likert scale ranging (strongly disagree, somewhat disagree, uncertain, somewhat agree, strongly agree) was used for all items, except for exposure(daily, weekly, monthly, quarterly, other) and familiarity (very unfamiliar, somewhat unfamiliar, somewhat familiar, familiar, very familiar). Answers were coded by assigning scores 15. For the assessment questions, means above 3 represent more agreement than disagreement.

b Respondents that answered they were somewhat unfamiliar and very unfamiliar with the radar were excluded from some comparisons, as they were found to lack the basic knowledge required to make a qualied assessment.

and the sample size (N = 39) indicate that these results should be interpreted with great care.

3.3.2.3. Proposition 1c. We also expected that individuals who were more actively and deeply engaged in radar activities would report higher cognitively related benefits (proposition 1c). Using the questionnaire data, we checked this by comparing the total assessment scores of readers, occasional scouts, regular scouts, selection panelists and assessment panelists. If one respondent reported belonging to more than one category, the category indicating most frequent involve- ment was selected. The categories of assessment panelist and selection panelist were grouped together, as many double- categorized these roles. Judging from the mean total assess- ment scores of the different groups familiar with the radar, selection and assessment panelists rated the payoff from the radar much higher (69.5) than readers (61.5), occasional scouts (61.33) and regular scouts (60.66). When accepting double- categorization, Spearman's correlation tests returned a signif- icant moderate positive relationship (0.38, p = 0.03) between assessment score and being a selection panelist, indicating that

this finding was likely not by happenstance. The other categories had non-significant very weak relationships, mean- ing that these results could easily be due to happenstance. The finding that selection panelists perceive the value of the radar as higher, fits with observations that selection panelists negotiatemental models and convert more than other groups. It is likely that they have a higher cognitive payoff than scouts who to a larger degree feed inexpert knowledge and assessors that feed outtheir ratings of the technologies.

3.3.2.4. Proposition 1d. Based on the theory, we also expected that analytic thinkers would see larger cognitive benefits from the radar than individuals with lower preferences for analytical thinking (proposition 1d). From talks with key stakeholders and analysis of the rationales for endorsing the radar, we have already observed that a key perceived value of the radar was its structured process. Unfortunately, we do not have information about the relation of these value assessments to individual preferences of ways of thinking. Based on the questionnaire data we could however examine the relationship between scores given to the radar's value for decision-making and the

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

77

scores returned particularly for the rational decision-making style from the GDMS questionnaire (refer to Table 1 for descriptive statistics). First of all, we found that the rational decision-making style (preference for logical evaluation of alternatives, thorough search) was the dominant one for near all respondents. This made it less relevant to group the respondents into sub-groups based on their dominant decision-making style. Instead we checked if a higher score on rational style (mean 4.4) had a relation to giving a high assessment score of the radar. We found positive relationships between high rational decision-making scores and agreement that the radar helps make more objective and less biased decisions. In an attempt to find counter-evidence we also checked the relationship between high scores on intuition and the agreement that the radar decreases bias, but this was negative. The same was the case for the assessment of the radar's ability to reduce the riskiness of decisions; high scores on rational decision-making style were positively related to high scores when assessing this statement, while high scores on intuitive decision-making were not or weakly negatively related. It should be stressed that these results were not statistically significant and more data would be needed to draw conclusions solely based on these findings. There are however indications that the radar foresight system is simply more naturalfor rational decision-makers. Individuals that scored high on rational decision-making style to a larger degree saw the radar process as a natural part of daily work, evident by a strong positive relationship with the statements the radar helps people do what they should have done anywayand the radar helps people carry out their daily tasks. Here, statistical testing indicated that the results represent a clear tendency and are likely not due to happenstance (Spearman's rho correla- tions of 0.43 and 0.39, and p-values of 0.013 and 0.026, respectively).

3.3.2.5. Proposition 1e. Our final expectation related to individ- ual cognitive payoffs from foresight, was to find evidence that the radar had shortened the time needed to take an informed decision at a critical point in time, and had increased the quality of decisions in Cisco (related to proposition 1e). While the value for decision-making was spelled out in several corporate documents describing the purpose of the radar, relatively few stakeholders highlighted the value for decision-making directly when they freely assessed it as evident from Fig. 4. Still, many did so indirectly by indicating that the radar helped them in their daily tasks (which involved decision-making) and that the radar opened up new perspectives. The brief and consistent format was highlighted in the endorsements, as well as the rigorous process for selecting and assessing trends. As one internal user reported: This high octaneresult is a tremen- dous time saver and mind opener for executives who must stay on top of the technology landscape with an eye to (the) impact on their business. Without the Technology Radar, the alterna- tives of either being oblivious to the less immediate technology scene or wading through dozens of industry publications with scattered or irrelevant information content, are either foolish or prohibitively time-consuming, respectively. Through the open for comments-boxin the questionnaire one respondent requested that the radar should be even more of an integral part of decision-making. In my organization there is a huge desire to increase the use of technology

intelligence and the radar in strategic decision-making, he wrote. However, another wrote that very few executives mention Technology Radar as having any influence on their decision-making processand asked for proof of the radar's effect on business decisions, if such existed. Still, we saw that respondents familiar with the radar (N = 32) rather agreed than disagreed with statements that the radar facilitates decision-making, that it helps influence decision-makers and that it helps make more objective and less biased decisions (see Table 1 for descriptive statistics). The same was the case for statements that the radar leads to better and faster information processing and that it helps save time.

Through the more qualitative data collection we also became aware of one actual event where the decision-making time was likely reduced due to the foresight system. The example in mind is the decision to acquire the company Intucell, a provider of advanced self-optimizing network (SON) software, in early 2013. This technology had been spotted, selected, assessed and put into the radar ahead of the acquisition. It had therefore already been presented to and processed by several key decision-makers in due time before the possibility to acquire Intucell arose. However, acquiring a company is about much more than just assessing its technology, and it is not possible to know whether Intucell would have been acquired anyway if the technology had not been rst put on the radar. Still, it would likely have been difcult to shortcutthe technology assessment necessary for the acquisition without the systematic foresight process. Since time to make the decision to acquire Intucell was limited, a lower quality decision could have been the outcome if no foresight had taken place, since less information about the technology would have been processed.

3.3.3. Findings related to collective aspects inuencing cognitive value

3.3.3.1. Proposition 2a. In order to understand how collective aspects influence the value creation from the radar in Cisco, we first needed a good understanding of the corporate environ- ment and how advanced the Cisco radar initiative was compared to foresight best practices. Based on the framework of Rohrbeck and Gemünden (2008) we judged the radar as close to best practice or as best practice in most respects, with some notable weaknesses (for instance limited scope with regard to information usage, limited external networks). Especially, the deliberate openness for anyone to participate in scouting still gave reason to expect results indicating a so- called collective bonus from foresight participation (proposi- tion 2a). One of the stakeholders assessing the radar indeed stated that the radar was especially effective as a productivity tool for Cisco's CTOs as it gives them a broader view of the technology landscape () without having to invest additional resources to pursue individual topics” — indicating an efficiency gain for CTOs when they with low cognitive effort take overinformation that has been more systematically processed by other individuals. We also already observed that participants in the selection panel negotiatedhypothetical futures through discussing the technologies before voting anonymously on their importance for Cisco. In other words, they profited from

78

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

other individuals who engaged in similar hypothetical thinking. Several others stress in their endorsements that the radar is a tremendous time saver, saves time thanks to a well- structured processand that it works in the direction of improved productivity at operational and executive level. Although these statements were given in a situation where there was a bias towards positive assessments, the anonymous answers in the questionnaire confirm the trend of valuing the radar for its time-saving abilities, potentially indicating a cognitive bonus (see Table 1 for overview of answers).

3.3.3.2. Proposition 2b. We also expected to find that the

organizational context influenced the motivation for, and therefore also the degree, of foresight thinking. Through our more qualitative data collection we looked into which control mechanisms had been at play when setting up and managing the radar (related to proposition 2b). Evident from internal project reports, the radar initiative in Cisco was deliberately not kept to a small corporate elite, but was from early on made open for anyone through the scouting website and through adjusting the process to fit wishes from different departments. The mainly intrinsic reward system put together in cooperation with the HR department can also be seen as a mechanism for controlling participation, along with substantial efforts to communicate the value of the radar internally as well as externally. However, through our investigations we noted that collaboration opportunities and visibility to other departments were mentioned much more frequently than monetary rewards or recognition (through, for instance badges), as reasons for participating in the radar process. The high agreement in the questionnaire that the radar helps under- stand what other peers are doing, and that it helps communi- cation (see Table 1 for details), supports an interpretation that social rewardsmay have been most important in the case of the radar. However, at the beginning of the implementation, e- mail badges and recognition from the boss might still have been important for recruiting scouts, since the social benefits of the radar were yet to be proved. However, in one of the early meetings to secure funding for the radar, one piece of feedback from a senior engineer was along the lines of Is this not what we are already meant to be doing?. One way of interpreting this statement is that many of the mechanisms for stimulating peoples' motivation for participating in the radar were already in place at the point of implementation. The extra incentives and communication efforts put into place when rolling out the radar might have strengthened the intensity of agreement that foresight thinking was important, but the organizational context could already have been foresight friendly.

3.3.3.3. Proposition 2c. As reflected in the findings about the

different rationales for endorsing the radar, individuals in Cisco do not necessarily see exactly the same value in the radar tool and process. In line with proposition 2c, we, in fact, expected to find that the radars' ability to correspond to a broad rather than

narrow set of interests would be crucial for the quality of the radar outcomes. Indeed, the 25 testimonials we analyzed show that people highlight quite different advantages of the radar although they agree that it is overall valuable (see Fig. 4 for details), indicating different tacit value systems. We applied the

competing values framework (Cameron and Quinn, 1999) to

further bring light to the key motivational drivers in Cisco, from

a collective point of view. The results indicate that we can

hardly talk about one unified foresight mindsetor foresight culturewithin Cisco, as a traditional approach to managing organizational culture would perhaps prescribe. True, the culture in Cisco has been described as entrepreneurial, but after closer assessment we considered there to be a blend of different cultures, depending on the business unit. For instance, the organization can be said to be held together by its commitment to innovation, which might be strongly related to foresight thinking. However, more operationally related drivers such as achievement and goal accomplishment are also present. Success can be defined by winning in the marketplace, but also by efficiency and increasingly by having the most unique products, in accordance with strategic goals. This value assessment by consulting individuals as well as performing the organizational level analysis helps clarify that the Cisco Technology Radar indeed seems to correspond to a broad rather than a narrow set of motivations. However, it is more difficult to assess how this ability directly influences the quality of foresight thinking. An argument stating that the radar's ability to correspond with different interests will keep people's mindsets more independent (and thus increase the quality of the process through preventing group thinking) may make intuitive sense, but cannot be directly supported or rejected based on our data.

4. Discussion

Table 2 summarizes our propositions and the degree of support found in our data. The most convincing results are related to proposition 1a, 1b, 1c and 2b, while we consider the total support for the remaining propositions to be weak or at

best medium. Overall, we still find all the propositions helpful as input for further answering our guiding research question about how and why engagement in foresight can impact ways

of thinking and processing information.

The case provided evidence that engagement in the technology radar seemed to stimulate analytical thinking and that it helped the revision of mental models (proposition 1a). The expected trend that regular engagement in the radar process (proposition 1b) as well as active rather than passive participation (proposition 1c) increased the payoff was also supported by the questionnaire data. However, the findings are based on a small sample and with only partial statistical significance of the results. As for 1c, however, through the case study we were able to further specify the meaning of activein the case of the Cisco Technology Radar; it was in fact the group that engaged in most discussion and negotiation of viewsthat tended to rate the radar value highest. This can be taken as further indication that for foresight to be able to change ones way of thinking, significant effort in terms of not only acquiring information, but also processing and discussing it, is a key factor. However, an alternative interpretation of the question- naire data backing up this conclusion could be that the individuals most deeply engaged in the discussions (selection panelists) feel more ownership of the radar process and thus are biased to rate the payoffs higher. Still, the relation between deeper engagement in the radar process and the perceived value of it remains.

79S.

Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

Table 2 Overview and evaluation of results.

Proposition

Types of data, evidence/counterevidence (e/c)

Degree of

Robustness

support

 

(weak,

medium,

strong)

1a) The revision of current mental models is helped by engagement in foresight

Observations and interviews (e), radar assessments (e), answers from questionnaire (e) Answers from questionnaire (e)

Medium to

Evidence can be considered mostly anecdotal and/or based on retrospective judgments. More experimental design needed to rmly establish causality.

strong

1b) Regular engagement in foresight increases the individual cognitive payoff from foresight, compared to only occasional engagement 1c) Participative (active) engagement in foresight increases the individual cognitive payoff compared to engagement only as a reader of end results (passive) 1d) Individuals with a high preference for analytic thinking are more likely to see larger cognitive benets from technology foresight than individuals with a lower preference.

Medium to

Larger sample is recommended for future research.

 

strong

Answers from questionnaire (e), observations and interviews (e)

Strong

Questionnaire sample is very small, but more qualitative data indicate the same trend. A larger-sample study particularly focused on this question could be interesting in the future. The total ndings are more supportive than counter-evidential, but the re- sults based on the questionnaire were not statistically signi cant. However, statistically signi cant results indicated that the radar might be more natural for rational decision-makers. Data support that several decision-makers perceive the radar as helpful for making better and faster decisions. Mainly anecdotal evidence.

Radar assessments, answers from questionnaire (e)

Weak to

medium

 

Weak to

1e) Preparation through foresight activities before a critical decision- making-point will I) shorten the time needed to take an informed decision at a critical decision-making point and II) increase the quality of a decision taken under time pressure compared to a situation with no preparation through foresight. 2a) When foresight is at its best practice, individuals (and the rm) receive a collective cognitive bonus from participation.

Document studies (e), observations and interviews (e), questionnaire (e/c)

medium

Weak to

Case analysis through best practice framework, radar assessments (e), observations and interviews (e), questionnaire (e) Observations and interviews (e), document studies (e)

medium

Evidence from qualitative data is stronger than questionnaire evidence. Robustness is vulnerable to the interpretations of the researchers. Future research could test the relation in a more experimental design.

2b) Cultural control mechanisms will be important for in uencing motivations to engage in foresight and thus enforce the participation in, and value creation from organization-wide foresight systems.

Medium

Findings are ultimately based on the perceptions of individuals and interpretations of the researchers. There is indication that many mechanisms for motivating for foresight were already in place before the implementation of the radar, and that additional motivation efforts mainly strengthened the intensity of agreement. Results indicate that we can hardly talk about one uni ed foresight mindset within Cisco, and that individuals indeed seem to have different tacit value systems related to participation. The relation between appealing to a broad set of values and the quality of foresight thinking cannot be further speci ed by the data.

2c) A foresight system that corresponds to a broad set of values reecting many individual interests rather than one main shared rationale for par- ticipation is important for securing maximum motivation to participate in foresight, while at the same time avoiding limitation of diversity in think- ing within the foresight activities.

Case analysis through OCAI framework (e), radar assessments (e), document studies (e)

Weak

80

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

It is also worth reflecting on the result which shows that the radar process seems to feel more natural for rational decision- makers. This indicates that the radar foresight process could have a better fit with people who already have an analytically dominated thinking style, perhaps in contrast to foresight approaches more concerned with creating and presenting scenarios. However, we cannot know to what extent the match between thinking style and design of the radar process affected the value creation from it, since we could not establish statistically significant relations between perceived payoff and thinking style (proposition 1d). It is also hard to assess to what extent participation in the radar could have switched peoples ways of making decisions, even though our case analysis indicates that most of the process is designed to probe analytical thinking and more systemized decision-making. The questionnaire data supports the expectation that the radar process allows for faster and better information process- ing (related to proposition 1e). It was, however, only on the general level that we could establish evidence that indicated a specific event where the radar enabled a faster and more informed decision. Still, this finding adds to the literature about the different roles of foresight, suggesting that the influence on thinking and decision-making might be considered a measur- able result in itself. Our qualitative analysis indicated some kind of collective payoff related to the saving of processing time and capacity (proposition 2a). The presence of such a cognitive bonusseems contingent on a critical mass of individuals participating as well as a certain level of quality in the process (in our analysis assessed through the best practice framework). As suggested also by previous research, control mechanisms do seem to have had a role in motivating for participation in the foresight process (proposition 2b). While the culture in Cisco seems to have been radar friendlyalready upon introduction, particularly intrinsic rewards put into place with help from the HR department could have helped intensify the agreement regarding the importance of the radar. Still, evident from the efforts to validate proposition 2c, the radar indeed seems to appeal to a broad set of values, indicating rather a lower coincidence of meaning given to participation in the radar. Even though a direct relation to the quality of the process was hard to establish in our empirical analysis, the way Cisco seems to have managed very diverse motivations and sub corporate culturesis notable. The technology radar seems to function as some kind of organizational gluein itself, uniting people from different departments and with different tacit value systems. Even more than the literature gave reason to expect, a success criteria and important role of the radar seems to be the way it makes people think and talk. Indeed, part of the value creation from the radar can be linked to the way it influences thinking, and the interaction enabled by the radar seems to catalyze the changes in mindsets on the individual level.

4.1. Implications

Our findings have implications for research seeking to further understand through which mechanisms foresight creates value, and for the practice of designing, implementing and motivating for participation in technology foresight. First, we show that cognitive aspects have a role in determining the value creation from foresight. Awareness

about how certain activities can influence the thinking styles of individuals (for instance analytic versus heuristic processing of information) and that individuals have different habits and preferences with regard to information processing could help the success of future systematic processes for identifying and assessing trends and technologies. The study of the Cisco Technology Radar case also indicates that the effect of changed mindsets can be seen and measured as a value in its own right. Tracking the degree of changed mindsets for example through cognitive mapping (Huff, 1990) could in the future be one additional way to assess the value creation from foresight research and practice. Second, a more deeply rooted explanation for how foresight contributes to update mental models is provided through our theoretical framework. In future research, formal hypothesis based on this explanation could be tested in a more experimental design, in order to strengthen the practical arguments for why foresight works. For instance, pre- and post-measures where mental models are mapped with regard to specific understand- ings of the future, possibly supplemented with measures of thinking styles, could be obtained in connection with different foresight projects. Comparison between active and more passive involvement in foresight and the effects on mental models and thinking styles could also be explored in such a design. It is worth mentioning that the finding related to analytic processing could be seen as contradictory to existing research within scenario foresight, which focuses more on scenarios' memorability and their ability to shift decision-makers over to use more intuition (Chermack and Nimon, 2008). However, intuition is per definition involving decision-making based on existing knowledge, which may lead to biases. This paper argues, based on theory as well as empirical findings, that stimulating analytic thinking is a key success criterion of technology foresight. It thus also compliments the research on foresight and decision biases and adds weight to the claims that foresight indeed can reduce such bias (Meissner and Wulf, 2012). Building on previous findings from the scenario literature and using the terminology introduced in this paper, links could perhaps be made between heuristic processing and making the mental models stick, complimenting the analytic processing needed to update them. The roles and the balance between analytic and intuitive thinking could be further researched in relation to foresight. Third, the observation that while cultural control mecha- nisms seem important for the success of corporate foresight, the technology radar in fact corresponds to a broad rather than narrow set of values, may have implications for further research as well as the practical design and implementation of foresight systems. Securing a glue effectby building control mechanisms that work in different corporate sub-cultures may increase the success probability of technology foresight systems. Focusing on efforts that intensify agreement about the foresight being useful, rather than aspiring to streamline value perceptions and motivations seems fruitful, based on our case evidence. In sum, research as well as the design and implementation of foresight systems can be well served by focusing not only on the requested output of foresight in terms of new products or new technologies scouted, but also its value creation through the systematized information processing and interaction between decision-makers. In the Cisco case we saw that individuals participating in the foresight highlighted values

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

81

supplementary to the values communicated from the corporate level. Many of these hidden valueswere related to social benefits from foresight. In this perspective, the fact that people like itmight indeed be a relevant aspect to consider when asked to provide evidence for the usefulness of a foresight system.

4.2. Limitations and clarications

This paper contributes with a theoretical framework that was tested and further explored through a practical case. Thus, there are automatically a number of limitations present. First of all, the empirical findings were linked to a specific version of technology foresight and, for instance, not scenario planning, and cannot uncritically be generalized to a wide foresight context. However, as detailed in the case description, the radar initiative does have similarities with other technology foresight methods and it can therefore be expected that particularly the implications are relevant also for methodologies such as horizon scanning and technological forecasting. The empirical case served to partly validate our theoretical framework, which hopefully will continue to be challenged in further research. Perhaps additional theory could have shed different light on our data and analysis, for instance perspec- tives on single and double loop learning (Argyris and Schön, 1978). The importance of questioning in the radar process could have been examined more closely, as this is seen as a driver of double-loop learning and therefore related to change. Explicitly mapping out the connection between analytic processing, heuristic processing and single or double loop learning could be a task for future research. Argyris' own work on theories in use(Argyris, 2002) and work comparing organizational learning theory and theories about different cognitive styles (Korth, 2000) could serve as inspiration. The evidence found in the case can rightfully be criticized for being anecdotal, based on a critically small sample and/or vulnerable to the perceptions of respondents as well as interpretations of the researchers. Even though it can be seen as an advantage that our study used multiple methods in order to compensate for and reduce some of these risks, future studies should perhaps aspire to limit the scope by concentrat- ing on fewer questions and instead increase the pool of same- source data. For instance, a more formal hypothesis could be generated as a basis for surveying a larger sample, or the claims related to analytical processing could be tested through a more experimental design. With regard to the finding that the foresight system seems to be seen as more valuable, or at least natural, for rational decision-makers, there are many unanswered questions. We do not know if the participation in the radar process had any effect on the decision-making styles of the participants. Could it be that it stimulates people to switch to more rational decision- making? Or is this type of foresight system more beneficial for people with rational styles to begin with, while scenario foresight is more beneficial for intuitive decision-makers? Future research could explore this further. Due to concerns about length and focus this paper refrains from discussing the issue of enough. Even though there can be established links between analytic thinking and the successful update of mental models of the future, it does not mean that everyone always should maximize this updating.We

will in this relation end with the well-known point made by March (1991) that processes like learning, analysis and technological change involve a delicate trade-off between exploration and exploitation(p 85). One could imagine that if all individuals are constantly challenging and updating all organizational truths and routines, it would hurt the current business too much and tip the efficiency curve downwards to a general state of inefficiency. Routines and decisions based on oldknowledge can be extremely beneficial as long as the environment does not change too much so that old ways do not apply anymore. The question about which level of mindset updateis enough is a big one, which hopefully will be revisited and addressed in future research.

4.3. Conclusion

In this paper we had the ambition to advancing the understanding of how and why foresight can influence ways of thinking. We did this first by generating propositions rooted in relevant literature, which were further tested and validated through an empirical case study of the technology radar initiative in Cisco. We found that the radar process indeed seems to influence ways of thinking through a process probing for analytical information processing and exchange of world- views, resulting in more frequent update of mental models. We also found indications of a relationship between more frequent and active involvement in the radar activities and the payoff in terms of perceived value. The design of the radar process, enabling interaction and communication across departments and functions, seems important for transforming individual benefits into payoffs observable on the collective level. The ability of the foresight system to correspond to a broad set of values and motivate people across departments and functions its glue effect has likely been a critical success factor in the case of Cisco. In summary, a central value of this technology foresight tool seems to be its' systematic probing of analytical thinking, which helps update mental models and thus influences decision-making and enables change. While the analytic probing is helped by specific criteria for reporting and discussing information, it is the interaction made possible by the foresight process that seem to determine the ultimate payoff for individuals as well as on the corporate level.

References

82

S. Boe-Lillegraven, S. Monterde / Technological Forecasting & Social Change 101 (2015) 6282

Bootz, J.P., Monti, R., Bootz, J.P., Monti, R., 2009. Proposition d'une typologie des démarches de prospective participative pour les entreprises, Trois cas illustratifs: EDF R&D, AXA France et BASF Agro, Management et avenir, juin. Brown, I.T., Smale, A., Verma, A., Momandwall, S., 2005. Medical technology horizon scanning. Australas. Phys. Eng. Sci. Med. 28, 200203. Burt, G., 2010. Learning the future fasterthe role of Vickers appreciative system. Futures 42, 265270. Burt, G., Chermack, T.J., 2008. Learning with scenarios: summary and critical issues. Adv. Dev. Hum. Resour. 10, 285295. Burt, G., van der Heijden, K., 2008. Towards a framework to understand purpose in futures studies: the role of Vickers' appreciative system. Technol. Forecast. Soc. Chang. 75, 11091127. Calof, J., Smith, J.E., 2012. Foresight impacts from around the world: a special issue. Foresight 14, 514. Calof, J., Miller, R., Jackson, M., 2012. Towards impactful foresight: viewpoints from foresight consultants and academics. Foresight 14, 8297. Cameron, K.S., Quinn, R.E., 1999. Diagnosing and Changing Organizational Culture. Addison-Wesley, Reading. Chermack, T.J., 2004. Improving decision-making with scenario planning. Futures 36, 295309. Chermack, T.J., Nimon, K., 2008. The effects of scenario planning on participant decision-making style. Hum. Resour. Dev. Q. 19, 351372. Dubin, R., 1978. Theory Development. Free Press, New York. Duncan, R., 1976. The ambidextrous organization: designing dual structures for innovation. In: Killman, R.H., Pondy, L.R., Sleven, D. (Eds.), The Management of Organization. North Holland, New York, pp. 167188. Eisenhardt, K., Graebner, M., 2007. Theory building from cases: opportunities and challenges. Acad. Manag. J. 50, 2532. Elqayam, S., Over, D., 2012. Probabilities, beliefs, and dual processing: the paradigm shift in the psychology of reasoning. Mind & Society 11, 2740. Evans, J.S.B.T., 2006. The heuristicanalytic theory of reasoning: extension and evaluation. Psychon. Bull. Rev. 13, 378395. Evans, J.S.B.T., 2008. Dual-processing accounts of reasoning, judgment, and social cognition. Annu. Rev. Psychol. 59, 255278. Evans, J.S.B.T., Over, D.E., 1996. Rationality and Reasoning. Psychology Press, Hove, UK. Feagin, J., Orum, A., Sjoberg, G.E., 1991. A Case for Case Study. University of North Carolina Press, Chapel Hill, NC. Fiske, S.T., Taylor, S.E., 2013. Social Cognition: From Brains to Culture. Sage Publications, London. Gavetti, G., Rivkin, J.W., 2007. On the origin of strategy: action and cognition over time. Organ. Sci. 18, 420439. Glick, M.B., Chermack, T.J., Luckel, H., Gauck, B.Q., 2012. Effects of scenario planning on participant mental models. Eur. J. Train. Dev. 36, 488507. Gray, B., Bougon, M.G., Donnellon, A., 1985. Organizations as constructions and deconstructions of meaning. J. Manag. 11. Huff, A., 1990. Mapping Strategic Thought. Wiley, Chichester. Huff, J.O., Huff, A.S., Thomas, H., 1992. Strategic renewal and the interaction of cumulative stress and inertia. Strateg. Manag. J. 13, 5575. Kahneman, D., 2011. Thinking, Fast and Slow. Farrar, Straus and Giroux, New York. Korth, S.J., 2000. Single and double-loop learning: exploring potential influence of cognitive style. Organ. Dev. J. 18, 8798. March, J.G., 1991. Exploration and exploitation in organizational learning. Organ. Sci. 2, 7187. Meissner, P., Wulf, T., 2012. Cognitive benefits of scenario planning: its impact on biases and decision quality. Technol. Forecast. Soc. Chang. 80, 801814. Miles, I., Saritas, O., 2012. The depth of the horizon: searching, scanning and widening horizons. Futures 14, 530545. Miller, G., 2003. The cognitive revolution: a historical perspective. Trends Cogn. Sci. 7, 141144. O'Reilly, C., 1989. Corporations, culture, and commitment: motivation and social control in organizations. Calif. Manag. Rev. 314, 925. Palomino, M.A., Bardsley, S., Bown, K., Lurio, J.D., Ellwood, P., Holland-Smith, D., Huggins, B., Vincenti, A., Woodroof, H., Owen, R., 2012. Web-based horizon scanning: concepts and practice. Foresight 14. Peck, J., 2009. Some theories of social change for futures practitioners. J. Futur. Stud. 14, 107122. Reger, R.K., Palmer, T.B., 1996. Managerial categorization of competitors: using old maps to navigate new environments. Organ. Sci. 7, 2239.

Stephan Monterde is the Director of the Technology Intelligence and Strategy group in Cisco. He holds a Bachelor of Electronics and an MSc in Communication Systems from the Swiss Federal Institute of Technology (EPFL) in Lausanne, Switzerland. He also holds an MBA from IMD with honors in leadership. He is leading Cisco's Technology Intelligence efforts, aiming at capturing and delivering information on technology trends, opportunities and threats. Prior to joining Cisco in 2007, he was Head of Messaging Services at Orange Group, FT Global Operations.

Siri Boe-Lillegraven is a PhD researcher at the Department of Business Administration at Aarhus University, Denmark. She is also afliated with the Sino-Danish Center for Education and Research, Beijing. Her educational background spans from business (MBA), innovation and entrepreneurship (Msc level), journalism and sociology (Ba). She has 10 years of industry experience, mainly from the newspaper industry and from a science park/ incubator. Her PhD project is concerned with investigating micro-drivers of organizational change.