Top Management s Attention to Discontinuous Technological ... · Maula,Keil,andZahra: TopManagement’sAttentiontoDiscontinuities Organization Science 24(3), pp. 926–947, ©2013

Zurich Open Repository andArchiveUniversity of ZurichMain LibraryStrickhofstrasse 39CH-8057 Zurichwww.zora.uzh.ch

Year: 2013

Top management’s attention to discontinuous technological change:corporate venture capital as an alert mechanism

Maula, Markku ; Keil, Thomas ; Zahra, Shaker A

Abstract: Technological discontinuities pose serious challenges to top managers’ attention. These dis-continuities, which often occur at the fringes of an industry, are usually driven by innovative and (often)venture capital-backed start-ups creating new products and transforming existing industries in waysthat are difficult for incumbent managers to understand against the backdrop of their existing cogni-tive schemata. However, failing to appreciate and embrace successful technological discontinuities mightendanger incumbents’ very existence. Extending the attention-based view, we explore whether andhow interorganizational relationships guide top managers’ attention either to or away from technologicaldiscontinuities. We propose that homophilous relationships (e.g., alliances with industry peers) shouldexhibit a negative relationship with incumbents’ timely attention to technological discontinuities, whereasheterophilous relationships (e.g., with venture capitalists as a result of coinvestments) should exhibit apositive relationship. Furthermore, we hypothesize that the status of the partners strengthens the effect ofhomophilous and heterophilous relationships with the timely attention of top managers to technologicaldiscontinuities. Based on a longitudinal study of the incumbents in four information and communi-cations technology industry sectors, we find that heterophilous ties through corporate venture capital(CVC), coinvesting with high-status venture capital firms, exhibit a strong positive relationship withtimely attention. CVC, when it connects senior management to high-status venture capitalists throughcoinvestments, has a special role in directing top managers’ attention to technological discontinuities andensuing business opportunities. Implications for the understanding of the role of interorganizational tiesas structural determinants of top managers’ attention are discussed.

DOI: https://doi.org/10.1287/orsc.1120.0775

Posted at the Zurich Open Repository and Archive, University of ZurichZORA URL: https://doi.org/10.5167/uzh-90640Journal ArticlePublished Version

Originally published at:Maula, Markku; Keil, Thomas; Zahra, Shaker A (2013). Top management’s attention to discontinuoustechnological change: corporate venture capital as an alert mechanism. Organization Science, 24(3):923-947.DOI: https://doi.org/10.1287/orsc.1120.0775

This art icle was downloaded by: [130.60.47.58] On: 10 February 2014, At : 08:19

Publ isher: Inst it ut e for Operat ions Research and t he Management Sciences (INFORMS)

INFORMS is locat ed in Maryland, USA

Organization Science

Publicat ion det ails, including inst ruct ions for aut hors and subscript ion informat ion:

ht t p: / / pubsonl ine. informs.org

Top Management ’ s At t ent ion t o Discont inuous

Technological Change: Corporat e Vent ure Capit al as an

Alert Mechanism

Markku V. J. Maula, Thomas Keil , Shaker A. Zahra,

To cite this article:

Markku V. J. Maula, Thomas Keil , Shaker A. Zahra, (2013) Top Management ’ s At t ent ion t o Discont inuous Technological

Change: Corporat e Vent ure Capit al as an Alert Mechanism. Organizat ion Science 24(3):926-947. ht t p: / / dx.doi.org/ 10.1287/

orsc.1120.0775

Full terms and conditions of use: http:/ / pubsonline. informs.org/ page/ terms-and-conditions

This art icle may be used only for the purposes of research, teaching, and/ or pr ivate study. Commercial useor systemat ic downloading (by robots or other autom at ic processes) is prohibited without explicit Publisherapproval. For more informat ion, contact perm [email protected].

The Publisher does not warrant or guarantee the art icle’s accuracy, completeness, m erchantability, fitnessfor a part icular purpose, or non- infr ingement . Descript ions of, or references to, products or publicat ions, orinclusion of an advert isement in this art icle, neither const itutes nor implies a guarantee, endorsement , orsupport of claims made of that product , publicat ion, or service.

Copyright © 2013, I NFORMS

Please scroll down for article—it is on subsequent pages

INFORMS is t he largest professional societ y in t he world for professionals in t he f ields of operat ions research, management

science, and analyt ics.

For more informat ion on INFORMS, it s publ icat ions, membership, or meet ings visit ht t p: / / www. informs.org

OrganizationScienceVol. 24, No. 3, May–June 2013, pp. 926–947

ISSN 1047-7039 (print) � ISSN 1526-5455 (online) http://dx.doi.org/10.1287/orsc.1120.0775

©2013 INFORMS

Top Management’s Attention to DiscontinuousTechnological Change: Corporate Venture Capital

as an Alert Mechanism

Markku V. J. Maula, Thomas KeilDepartment of Industrial Engineering and Management, Institute of Strategy, Aalto University,

FI-00076 Aalto, Finland {[email protected], [email protected]}

Shaker A. ZahraStrategic Management and Organization Department, Carlson School of Management, University of Minnesota, and

Garry S. Holmes Center for Entrepreneurship, Minneapolis, Minnesota 55455, [email protected]

Technological discontinuities pose serious challenges to top managers’ attention. These discontinuities, which oftenoccur at the fringes of an industry, are usually driven by innovative and (often) venture capital-backed start-ups creating

new products and transforming existing industries in ways that are difficult for incumbent managers to understand againstthe backdrop of their existing cognitive schemata. However, failing to appreciate and embrace successful technologicaldiscontinuities might endanger incumbents’ very existence. Extending the attention-based view, we explore whether andhow interorganizational relationships guide top managers’ attention either to or away from technological discontinuities.We propose that homophilous relationships (e.g., alliances with industry peers) should exhibit a negative relationshipwith incumbents’ timely attention to technological discontinuities, whereas heterophilous relationships (e.g., with venturecapitalists as a result of coinvestments) should exhibit a positive relationship. Furthermore, we hypothesize that the status ofthe partners strengthens the effect of homophilous and heterophilous relationships with the timely attention of top managersto technological discontinuities. Based on a longitudinal study of the incumbents in four information and communicationstechnology industry sectors, we find that heterophilous ties through corporate venture capital (CVC), coinvesting withhigh-status venture capital firms, exhibit a strong positive relationship with timely attention. CVC, when it connectssenior management to high-status venture capitalists through coinvestments, has a special role in directing top managers’attention to technological discontinuities and ensuing business opportunities. Implications for the understanding of the roleof interorganizational ties as structural determinants of top managers’ attention are discussed.

Key words : attention; corporate venture capital; homophily; heterophily; status; technological discontinuityHistory : Published online in Articles in Advance September 10, 2012.

IntroductionRecent research highlights the important role of man-agerial attention in developing effective organizationalresponses to technological discontinuities (Barr 1998,Barr et al. 1992, Eggers and Kaplan 2009, Kaplan2008, Kaplan et al. 2003, Kaplan and Tripsas 2008,Tripsas and Gavetti 2000, Virany et al. 1992). Thisresearch suggests that a firm’s response to technologicaldiscontinuities is shaped by how its senior managerspay attention to and interpret technological change (Daftand Weick 1984) and how they translate this cog-nition into effective strategic action (Bourgeois andEisenhardt 1988). This research emphasizes that in timesof discontinuous change, top management’s attentionallocation processes are central to shaping a firm’seffective response to new technological paradigms (Barret al. 1992, Eggers and Kaplan 2009, Kaplan 2008,Ocasio 1997).Previous studies have provided ample evidence that

senior management’s attention is important, but prior

research also suggests that addressing technologicaldiscontinuities poses important cognitive challenges totop managers’ attention systems. Top managers usu-ally focus their attention on information from familiarsources, such as their existing competition (Peteraf andShanley 1997, Porac et al. 1995) or existing alliancepartners. Given their time constraints, executives ofteninterpret information using existing heuristics, cogni-tive frames, and knowledge categories based on pastexperience (Barr 1998, Barr and Huff 1997, Barr et al.1992, Leonard-Barton 1992, Levinthal and March 1993).Although these information processing and sensemak-ing approaches are efficient in times of incrementalchange, when both the existing information sources andthe information that they provide are reliable, they oftenfail in situations of discontinuous change. In these sit-uations, knowledge pertaining to technological discon-tinuities is often fundamentally different from existingknowledge and does not fit neatly into existing cogni-tive frames (Kaplan and Tripsas 2008). Technological

926

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.

Maula, Keil, and Zahra: Top Management’s Attention to Discontinuities

Organization Science 24(3), pp. 926–947, © 2013 INFORMS 927

discontinuities are fundamental shifts from one dominanttechnology to another. These shifts frequently changethe fabric of the industry, the rules of competition,and the identity of the industry’s participants. Thesediscontinuities often occur outside of an industry orat its fringes (Christensen 1997, Henderson and Clark1990), creating new products and business models thatare difficult to understand against the backdrop of exist-ing business logic (Barnett and Pontikes 2008, Tripsasand Gavetti 2000). These new models typically involvea set of players that are not part of the networks inwhich incumbent firms participate (Koka and Prescott2008). However, failing to respond effectively to thesediscontinuities may endanger the very existence of anincumbent.Prior research findings suggest that top management’s

attention allocation is a prerequisite to an effectiveorganizational response to technological discontinuities.However, executives may fail to pay attention to thesechanges because of their information processing systemsand cognitive frames. Thus, we are left with two relatedquestions: What factors might support or hamper topmanagement’s attention allocation to technological dis-continuities? How can organizations construct effectivemechanisms to guide top managers’ attention to techno-logical discontinuities?In this paper, we extend the attention-based view of

the firm (Hoffman and Ocasio 2001; Ocasio 1997, 2011;Ocasio and Joseph 2005) to theorize on the structuraldeterminants of top management’s attention. In partic-ular, we focus on the external mechanisms that firmsemploy to direct their top management’s attention totechnological discontinuities. We propose that varioustypes of interorganizational relationships have differingeffects on top management’s attention and, in particular,that homophilous relationships with industry peers maynegatively affect timely attention to emerging technolog-ical discontinuities, whereas heterophilous relationships(e.g., relationships with prominent venture capitalists(VCs)) may guide top management’s attention towardemerging discontinuities. In line with the principle of thestructural distribution of attention (Ocasio 1997), interor-ganizational relationships allow corporations to create astrategic context in which top managers are exposed totechnological discontinuities. We further argue that theeffects of these external mechanisms are strengthened bythe status of the relationship partner.We test our arguments on the role of homophilous

and heterophilous interorganizational ties and the role ofpartners’ status in these ties in shaping management’sattention to technological discontinuities using a sam-ple of the largest U.S. corporations from four infor-mation and communication technology industry sectorsbetween 1989 and 2000. Incumbents in these industriesfaced two technological discontinuities, the Internet andwireless, that profoundly altered their industry definition

and the rules of competitive rivalry. The emergence ofthe Internet and wireless technologies (e.g., mobile tele-phony and wireless local area networks (WLANs)) alsospawned a large number of new business opportunitiesand subfields that required substantially different skillsand competencies than those that incumbents typicallyhad, sometimes blinding them to the potential gainsfrom competing in this arena. Constructing full allianceand VC syndication networks to examine the effects ofhomophilous and heterophilous interorganizational tiesand partner status in these two different networks, wefound support for a strong positive effect of partnerstatus in heterophilous ties. Homophilous ties throughalliances with industry peers do not have a clear negativeeffect as predicted, but the effect of these relationshipsappears to depend partially on partners’ attention to dis-continuous opportunities.Our study extends the attention-based view from an

examination of the structural determinants within theorganization to include also interorganizational relation-ships. Our results suggest that corporations can directtop management’s attention to technological discontinu-ities by creating an appropriate external structural con-text for a firm. We view interorganizational relationshipsas a mechanism through which incumbents expose theirtop management to important issues in the operatingenvironment, enabling them to overcome potential blindspots in their attention allocation system.We also contribute to research on technological dis-

continuities and change. Recent research has begun todevelop a cognitive perspective on technological changeand firms’ responses to this change (Kaplan and Tripsas2008, Tripsas and Gavetti 2000). Our findings extendprior research that has emphasized the importance ofattention to discontinuities by highlighting the need for adeeper understanding of the circumstances under whichtop managers pay attention to such discontinuities.Our study also adds to the social network litera-

ture. There is a long-standing debate on which net-work structures provide the most benefits to firms (Burt1992, Coleman 1988, Lee 2007), but researchers haveonly recently begun to explore the importance of net-work composition in explaining a firm’s gains as aresult of the multitude of networks to which it is con-nected (Lee 2008). We contribute to this literature bydistinguishing between relationships with homophilousand heterophilous partners and by showing the dif-fering effects of these relationships on top manage-ment’s attention. In addition, we hypothesize and testdiffering effects of partner status in homophilous andheterophilous interorganizational relationships. Althoughthe high status of partners in heterophilous ties exhibitsa positive effect on top management’s timely attentionto discontinuities, the status of partners in homophilousties seems to have a different effect.The results contribute to the emerging literature on

corporate venture capital (CVC), especially its role in

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.


928 Organization Science 24(3), pp. 926–947, © 2013 INFORMS

corporate innovation activities. In CVC investments,large corporations invest in start-ups alongside tradi-tional VC firms (Gompers and Lerner 1998). Theseinvestments are usually made with the intent of gain-ing strategic benefits (e.g., learning about emerging newtechnologies or business models) as well as financialbenefits (Dushnitsky and Lenox 2005a, b; Wadhwa andKotha 2006). CVC investments have frequently beenviewed as a window on technology or a technologyradar (Dushnitsky and Lenox 2006, Keil et al. 2008a,Siegel et al. 1988) that helps to enhance the recogni-tion of emerging technologies and related business mod-els among CVC managers and senior executives. Priorstudies have shown that CVC investments are relatedto increased innovation rates when measured throughpatenting rates (Dushnitsky and Lenox 2005b, Schildtet al. 2005, Wadhwa and Kotha 2006). More recently,research has highlighted the limits of direct knowledgeflows from ventures to the corporate parent (Dushnitskyand Shaver 2009). Recent studies have also begun tocompare learning from CVC investments and otherforms of interorganizational relationships (Dushnitskyand Lavie 2010, Keil et al. 2008b). We add to this streamof research by emphasizing that the contribution of CVCactivities to corporate innovation might be less crucialin transferring patentable knowledge to the corporateparent than in directing top management’s attention toimportant technological changes in the firm’s externalenvironment.

Technological Discontinuities andFirm Attention

Incumbents and Technological Discontinuities

Technological change is often an incremental but cumu-lative process that is punctuated by short revolution-ary periods in the form of discontinuities (Tushman andAnderson 1986). Some of these discontinuities representmajor technological shifts that are so significant that nochange in scale, efficiency, or design can keep exist-ing technologies competitive (Anderson and Tushman1990, Tushman and Anderson 1986). Faced with suchdiscontinuities, some incumbents fail to adapt effectivelyto these changes, leading to erosion of their marketpositions (Henderson 1993, Henderson and Clark 1990,Hill and Rothaermel 2003, Rosenbloom and Christensen1994, Tushman and Anderson 1986).There is agreement that incumbents encounter seri-

ous difficulties in adapting to technological discontinu-ities, but the reasons behind these difficulties remain asubject of debate. These reasons may include the dif-ferent economic incentives of new entrants and incum-bents in pursuing discontinuous innovation (Christensen1997); forces of inertia, such as organizational iden-tity (Tripsas 2009), path dependency, escalation of com-mitment, and irreversible commitments (Henderson and

Clark 1990, Tushman and Anderson 1986); existingorganizational routines and capabilities that lead tosuboptimal responses to radical shifts (Levinthal andMarch 1993, Nelson and Winter 1982); cognitivebarriers in noticing or interpreting new technologies(Danneels 2011, Gilbert 2005, Leonard-Barton 1992);and the embeddedness of incumbents within establishedindustry networks that do not initially value the newtechnology (Rosenbloom and Christensen 1994).Recent explanations of incumbents’ difficulties in

recognizing and responding to technological disconti-nuities underscore senior management’s attention as akey explanatory variable (Barr 1998, Cho and Hambrick2006, Kaplan 2008, Kaplan et al. 2003, Kaplan andTripsas 2008, Ocasio 1997). For instance, in a series ofstudies, Kaplan and her colleagues (Eggers and Kaplan2009, Kaplan 2008, Kaplan et al. 2003) found thatincumbents whose chief executive officers (CEOs) ortop management teams paid timely attention to emergingtechnological discontinuities were more likely to investin new technology and to produce effective responsesto emerging competitive threats. Although these studiesestablish a clear link between senior managers’ atten-tion to discontinuities and their effective firm response,they leave a question unanswered: Why do senior man-agers in some incumbent firms pay attention to thesediscontinuities, whereas others do not? In other words,what are the factors that influence senior managers’attention allocation to emerging technological discon-tinuities? Insights into these issues can be gained byapplying the attention-based view of the firm, developedby Ocasio (1997) and discussed in the next section.

Managerial Attention to Discontinuities

Drawing on the Carnegie School, particularly the workof Herbert Simon (Simon 1947), Ocasio (1997) sug-gests that the structuring of attention in organizationsis one of the central explanations of firms’ behavior.Organizational attention is defined as “the distinct focusof time and effort by the firm on a particular set ofissues, problems, opportunities, and threats and on a par-ticular set of skills, routines, programs, projects, andprocedures” (Ocasio 1997, p. 188). From this perspec-tive, organizations function as systems of procedural andcommunicational channels that receive issues and possi-ble responses to these issues from the environment andfeed this information to decision makers for processing(Ocasio 1997, Vissa et al. 2010). The decision outcomesare therefore not determined solely by the characteristicsof the decision makers but are shaped by the specificorganizational context and the situations of individualdecision makers (Ocasio 1997, p. 189). Ocasio (1997)identifies three principles that underlie the attention-based view: (1) managerial action is usually focused onthe issues to which managers pay attention; (2) attentionis situated, meaning that the context in which decision

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



makers are embedded affects their attention and the out-comes of their decisions; and (3) attention is structurallydistributed—that is, the context in which decision mak-ers find themselves depends on how organizations struc-turally distribute and control the allocation of issues,answers, and decision makers.To understand why senior managers may pay attention

to technological discontinuities, it is particularly impor-tant to understand the structural determinants of atten-tion. Ocasio (1997) proposes that the rules of the game,resources, players, and social positions within a firmcombine to determine the legitimacy, importance, andrelevance of issues and possible organizational responsesto these issues. The rules of the game usually reflectthe formal and informal principles of action, interaction,and interpretation upon which decision makers draw inaccomplishing their tasks (Ocasio 1997). In the faceof technological discontinuities, existing rules of thegame can become counterproductive because they mayfilter information about a given discontinuity, therebylimiting top management’s attention to that discontinu-ity (Leonard-Barton 1992, Levinthal and March 1993).Similarly, existing resources may bias the attention ofdecision makers toward response repertoires that buildupon these existing resources (Leonard-Barton 1992,Levinthal and March 1993).In addition to the rules of the game and resources,

the role of players and structural positions as structuraldeterminants of attention allocation are important inunderstanding firms’ responses to discontinuities. Play-ers refer to individuals and groups within or outsidean organization that influence decision makers. To date,research on the role of different players has focusedlargely on the CEO and other top management teammembers and their role in shaping organizational atten-tion (Cho and Hambrick 2006, Eggers and Kaplan 2009,Hambrick and Mason 1984, Kaplan 2008, Kaplan et al.2003, Yadav et al. 2007) and firms’ responses to emerg-ing issues. For instance, Cho and Hambrick (2006) showthat top management teams’ demographics influencefirms’ attention to industry deregulation and that thischange in attention allocation leads to changes in com-pany strategy. Similarly, Yadav et al. (2007) show thatthe future orientation of a firm’s CEO is a predictor ofthat firm’s speed in responding to technological change.Closely related to the concept of players is the struc-

tural position of decision makers and other players inthe corporation. The distribution of structural positionsshapes attention patterns by providing decision makersand other players with interests, values, and identitiesthat guide their attention to different aspects of the orga-nization’s environment (Ocasio 1997). For instance, inthe context of attention to subsidiary issues, Bouquet andBirkinshaw (2008) find that the structural position of asubsidiary within the network of subsidiaries, measuredthrough the attractiveness of its market and its strengths

compared to other subsidiaries, is an important predictorof the attention a subsidiary receives from the parentmultinational corporation.Taken together, these studies have greatly enriched

our understanding of how the characteristics of players,such as the composition of the top management teamor the CEO’s future orientation, influence their attentionallocation patterns. Yet the focus of these prior studieshas been on players and structural positions that existwithin the boundaries of the firm. However, attentionpatterns in organizations are also influenced by playersoutside the firm’s boundaries (Ocasio 1997); the influ-ence of these players on attention patterns has receivedrelatively little recognition in the literature. Furthermore,structural positions, such as status, may shape attentionpatterns and guide decision makers’ attention to differ-ent aspects of the environment (Ocasio 1997). In thefollowing sections, we will develop hypotheses abouthow players external to a firm influence attention allo-cation patterns within the firm, and we will empiricallytest these hypotheses. In particular, we will focus onthe effects of heterophilous versus homophilous interor-ganizational partners (types of players) and their status(structural position) as structural antecedents of top man-agement’s attention to technological discontinuities.

Interorganizational Relationships andManagerial AttentionInterorganizational relationships constitute an importantcommunication channel through which organizationsand their decision makers interact with different playersin a firm’s external environment. Social network liter-ature points out that through these interorganizationalrelationships, firms receive important information aboutevents in their external environment and their potentialimplications for the firm (e.g., Koka and Prescott 2008,Lee 2007). Interorganizational relationships can help toreduce a firm’s uncertainty about the nature of busi-ness opportunities by giving the firm’s decision makersaccess to different external information sources (Burt1992, Granovetter 1973).We expect interorganizational relationships to play

a particularly important role in guiding attentionrelated to technological discontinuities. As noted, thesediscontinuities often emerge at the fringes of an industry,outside the normal focus of attention of incumbents’senior decision makers (e.g., Tripsas 1997). These dis-continuities often draw upon radically new knowledgeor require operating and organizing principles that arenot easily comprehended within industries’ existingcognitive frameworks. The fact that these discontinu-ities involve a high degree of market and technolog-ical uncertainty makes the evaluation of their likelyimpact on incumbents’ businesses exceedingly challeng-ing. The uncertainty that surrounds these discontinu-ities may even lead some decision makers to ignore

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



them. Interorganizational relationships can reduce theperceived uncertainty (Lee 2007) of these discontinu-ities, drawing attention to them and even shaping seniorexecutives’ attention to these discontinuities. In situa-tions of high uncertainty, decision makers’ perceptionsof the quality of information received through interorga-nizational relationships is often shaped by the status ofthe source.However, research also suggests that network bene-

fits do not occur under all circumstances. Interorganiza-tional networks may even become a liability when thenature of these networks is not aligned with the infor-mation needs of the focal firm (Gargiulo and Benassi2000, Koka and Prescott 2008). For instance, researchon social networks suggests that individuals and organi-zations exhibit a tendency toward homophily (Luo andDeng 2009, McPherson et al. 2001), defined as the ten-dency to associate with similar others and to trust theiradvice (McDonald et al. 2008). Such homophilous net-work contacts may hold similar views (McDonald et al.2008) and may not possess sufficiently dissimilar infor-mation to help steer attention allocation patterns awayfrom routine processing in light of high uncertainty, orthey may even have incentives to direct attention awayfrom emerging discontinuities. In the next section, wedevelop hypotheses for different types of interorgani-zational relationships and their potential impact on thetimely attention to discontinuous business opportunitiesin incumbent firms.

Homophilous Relationships with Industry Peers and

Top Management’s Attention to Discontinuities

One set of interorganizational relationships that mayshape the attention allocation of incumbents’ top man-agers are links that these firms establish with otherincumbent firms in the same industry. Incumbents fre-quently form close alliances with other incumbents toextend supplier relationships and customer relationshipsor to create horizontal cooperation. In their allianceswith industry peers, incumbents cooperate on a broad setof issues that include early stage research, technologystandardization, and manufacturing as well as joint go-to-market strategies or joint marketing. Although theserelationships have diverse purposes, they are character-ized by intense business development activities and fre-quent information exchange (Dyer and Singh 1998). Forinstance, in a research and development (R&D) alliance,firms may jointly develop new technology and, in theprocess, exchange detailed information about technolog-ical trends. Furthermore, in alliances with customers orsuppliers, information about technological developmentsis frequently exchanged to align roadmaps (Heide andMiner 1992). Even in less technology-oriented relation-ships, such as manufacturing or marketing relationships,partners may exchange information about market trends

that influence joint understanding of technological shifts(Hagedoorn 1993).Alliance relationships with industry peers are charac-

terized by a high level of homophily because firms in thesame industry frequently share a large number of simi-larities in their organizational forms, knowledge, experi-ences, and mental frameworks (Porac et al. 1989, Poracet al. 1995, Prahalad and Bettis 1986). Even within anindustry, firms frequently form links with relatively sim-ilar partners (Luo and Deng 2009). Similarity amongalliance partners may make it easy to absorb informationfrom a partner (Lane and Lubatkin 1998) and may pro-vide organizational decision makers with psychologicalcomfort when assessing information from these famil-iar sources (McDonald et al. 2008), but homophilousrelationships may also direct top management’s atten-tion away from technological discontinuities for severalreasons.First, industry peers and the focal firm may have rel-

atively similar knowledge and information. Given thatdiscontinuous technological change often arises outsideof the industry or, at least, at its fringes (Christensen1997, Henderson and Clark 1990) and frequently buildson knowledge that is new to the industry (Kaplan andTripsas 2008), peer incumbents are less likely to be goodsources of knowledge and information about emergingtechnological discontinuities.Second, because firms in the same industry often

show striking similarities in their definitions of com-petitors, operating principles, and technologies (Bettisand Prahalad 1995; Porac et al. 1989, 1995), a peerincumbent’s views on emerging discontinuities are likelyto resemble those held inside the focal firm and aretherefore unlikely to guide incumbent decision makers’attention toward technological discontinuities; instead,inertial forces within the incumbent firm may becemented. In line with this reasoning, Koka and Prescott(2008) argue that dramatic changes in the environmentoften modify the existing bases of competition, and firmsthat are centrally positioned in existing industry net-works are likely to be adversely affected by this shiftbecause the information they gain from their existingties may lose relevance after a change.Third, even if incumbents have information about dis-

continuities and recognize the value of this information,they may not have an incentive to share this informationwith a within-industry peer. Incumbents’ alliance part-ners are likely to have a considerable stake in existingtechnologies and may be equally threatened by a discon-tinuity. Thus, they may use these alliance relationships toinfluence peers against adopting a technological solutionthat could threaten both the focal firm and its partner.Finally, the nature of the relationship with peer

incumbents may bias management’s attention to rou-tine business development issues. Managing ongoingalliance relationships often requires substantial time and

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



managerial attention. Thus, most of these relationshipsare dominated by routine business development issuesthat will be relatively comfortable for managers in thefocal firm given the similarity of the alliance part-ner. Thus, alliances with peers may further direct andeven bias top management’s attention to relatively rou-tine business development issues rather than promotingmindful interpretation of weak signals that are related toemerging technological discontinuities (Salvato 2009).Thus, we hypothesize as follows.

Hypothesis 1 (H1). The larger the number of homo-

philous interorganizational ties to industry peers, the

later the time at which top management will pay atten-

tion to emerging discontinuous technological change.

Status, Homophily, and Top Management’s

Attention to Discontinuities

We expect the negative relationship of homophilous rela-tionships to industry peers and the timely attention oftop management to technological discontinuities to befurther strengthened if these relationships are developedwith high-status partners (Podolny 2005). The status ofa player in an industry can be defined as the posi-tion it occupies in the social hierarchy (Podolny 2005).Contrary to a firm’s reputation, which directly reflectspast behavior (Fombrun and Shanley 1990), status is areflection of the patterns of relations and affiliations inwhich a firm is engaged. High-status firms are incum-bents that have developed a central position in the socialnetwork of an industry over time. We expect status tostrengthen the effects of homophily through two interre-lated mechanisms.First, given how status accrues over time, high-status

industry peers are less likely to be engaged in or holdinformation about technological discontinuities com-pared with lower-status industry peers. Firms gain statusby being selective in their interactions with others andby focusing on relationships with other high-status firmsin an industry at the expense of relationships with moreperipheral firms (Podolny 1993). As we have argued,technological discontinuities often emerge at the fringesor outside of an industry and are frequently created bycompletely new firms that have no status in the indus-try. Therefore, the strategies used by high-status firmsto maintain their status may limit their ability to inter-act with firms that pursue technological discontinuities.Given that high-status industry peers are less likely to beinvolved in inducing technological discontinuities, thesefirms will likely influence top managers’ attention pat-terns toward the incremental developments of existingtechnologies.High-status firms are not only less likely to engage

in and hold information on technological discontinuitiesbut the focal firm’s top management will also attach aparticularly high weight to information from high-status

incumbent peers. In assessing information from differ-ent external sources, top managers must simultaneouslyevaluate both the information and the quality of thesource of information. Research on social networks sug-gests that in such situations, the status of the actor isfrequently used as a signal for the quality of the actorand, by extension, for the quality of their information(Podolny 1993, 1994; Podolny 2001). Because managersface a high level of uncertainty about the ability of apartner to provide high-quality information about emerg-ing discontinuity, they revert to the general assessmentof that partner by industry peers that is reflected in itsindustry status. For instance, members of the top man-agement team are more likely to give greater weight toinformation from IBM, Microsoft, or Hewlett-Packard—all high-status firms in the industries we studied—thanfrom a marginal firm or a newcomer to the industry.In light of emerging technological discontinuities,

high-status firms may utilize the weight that others(peers) attach to their knowledge and information toshape technological trends in their favor by activelyinfluencing attention patterns among suppliers, cus-tomers, and horizontal alliance partners. For instance, inthe early 2000s, incumbent telecommunications infras-tructure companies were highly invested in cellulartechnology for wireless data transmission. When sev-eral alternative wireless data technologies were pro-posed by relative newcomers to the industry, one of theresponses by high-status incumbents was to downplaythe importance of these technologies in their officialcommunications.Taken together, the above observations lead us to pro-

pose that high-status industry peers may be less likelyto hold information on technological discontinuities andmay have a particularly strong influence in drawing focalfirm’s attention toward incremental technological devel-opments. Therefore, we hypothesize as follows.

Hypothesis 2 (H2). The higher a partner’s status in

homophilous interorganizational ties, the later the time

at which top management will pay attention to emerging

discontinuous technological change.

Heterophilous Interorganizational Relationships

with VCs and Attention to Discontinuities

Although homophilous interorganizational relationshipswith industry peers may be a source of inertia inmanagers’ attention allocation to technological discon-tinuities, we expect that relationships with firms thatexhibit heterogeneity from an incumbent would facili-tate top management’s attention to discontinuities. Byforming heterophilous interorganizational relationships,an incumbent may be able to gain access to more diverseinformation and diverging viewpoints that can help toreshape attention patterns within the incumbent. Someheterophilous relationships may further create external

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



structures that are particularly geared to focusing topmanagement’s attention on nonroutine events and thusimproving attention quality (Rerup 2009, Salvato 2009).Our study focuses on one specific set of relation-

ships that incumbents form by coinvesting with VCswhen making CVC investments (Dushnitsky and Lenox2005a) in start-ups. In CVC investments, large corpora-tions invest in start-ups alongside traditional VC firms(Gompers and Lerner 1998), usually aiming to gain bothfinancial and strategic benefits, such as knowledge aboutemerging technologies or business models (Dushnitskyand Lenox 2005a, b; Wadhwa and Kotha 2006). CVCinvestments directly connect the incumbent to start-upsthat focus on exploiting a technological discontinuity.Start-ups are likely to have very different knowledgeand may use very different cognitive frameworks toevaluate emerging technological discontinuity than focalincumbents. For instance, by interacting regularly withstart-ups through board memberships, incumbent man-agers may develop a deeper understanding of emergingdiscontinuity.CVC investments also connect the incumbent to other

VCs that invest in these ventures because most incum-bents coinvest alongside multiple VCs in so-called syn-dicated investments. Syndication enables VCs to sharerisks (Wilson 1968) and exchange information about astart-up’s management and technologies (Lerner 1994).By investing in syndicates with traditional VCs, incum-bents gain access to a community of investors thatspecialize in assessing nascent technologies and spot-ting emerging discontinuities. Given that VCs frequentlyshare information about the ventures they consider dur-ing syndication, access to VCs provides high-qualityinformation about new ventures and discontinuous tech-nologies they may pursue. Access to the VC communitymay be valuable beyond the individual venture beingconsidered because senior managers may gain the oppor-tunity to regularly exchange information with VCs aboutthe broader evolution of technologies and emerging busi-ness subfields. In addition to holding different infor-mation and possessing different cognitive frameworksto evaluate technological discontinuities, VCs are morelikely to draw management’s attention to emerging dis-continuities than their peer incumbents are. VCs haveno (or, at least, less) vested interest in existing technolo-gies. Instead, they may have incentives to actively sup-port emerging discontinuities to support the businessesin which they invest. Therefore, VCs may share infor-mation about and draw incumbent managers’ attentionto potential technological discontinuities.Creating a mechanism to regularly focus on iden-

tifying potential discontinuities is important for thequality of attention to discontinuities. Firms usuallyface a trade-off between strong attentional engagement(Ocasio 2011) to few issues (by focusing their cognitiveresources on a selected set of stimuli over a period oftime and thus improving the quality of their attention to

these stimuli; see Rerup 2009, Weick and Sutcliffe 2006)and emphasizing the deliberate management of nonrou-tine events through mindful exploration of a larger num-ber of issues by the top management team (Ocasio 2011,Salvato 2009). Given the prevalence and short-term pres-sures of ongoing business development, top manage-ment teams may emphasize a smaller number of issues,and ongoing routine business development may crowdout the exploration of emerging discontinuities. Giventhat discussions with VCs typically focus on nonroutineevents and are largely free of the short-term pressuresof routine business development, frequent interactionsbetween VCs and top managers may create a mecha-nism that routinizes the mindful exploration of weaksignals and allow incumbents to overcome the trade-offbetween routine and mindful behavior (Levinthal andRerup 2006, Salvato 2009). This mechanism can expe-dite incumbents’ recognition of technological disconti-nuities. Therefore, we hypothesize as follows.

Hypothesis 3 (H3). The larger the number of het-

erophilous interorganizational ties with VCs, the earlier

top management will pay attention to emerging discon-

tinuous technological change.

Bridging to Networks of High-Status VCs: Status,

Heterophily, and Attention to DiscontinuitiesHeterophilous interorganizational relationships createdthrough CVC investments connect incumbents to emerg-ing communities of firms that focus on commercial-izing technological discontinuities, particularly the VCinvestors who finance these new firms. In this way,incumbent managers may regularly be exposed to diversetypes of information and divergent views about emerg-ing technological discontinuities. However, to ensure alasting impact on top management’s attention patterns, itis important that the information about the technologicaldiscontinuity and its source is credible. In heterophilousrelationships, incumbent managers usually interact withhighly dissimilar partners, which may make it difficult toassess the trustworthiness of information received fromthese sources (Luo and Deng 2009, McDonald et al.2008, McPherson et al. 2001). For instance, top man-agers are unlikely to be easily convinced by the state-ments of a start-up manager or an unknown VC investor.In heterophilous relationships, there is a risk that suchinformation is available but managers may opt to ignoreit because of uncertainty about the quality of informa-tion and the credibility of its source. The social networkliterature (Podolny 2005) suggests that in heterophilousrelationships, the status of the partner plays a critical rolein alleviating this uncertainty and in strengthening theeffect of heterophilous ties.In particular, for CVC investments coinvested with

VCs, we expect that the status of the other VC investorswith whom an incumbent coinvests is critical in guidingtop management’s attention to emerging technologicaldiscontinuities. The new ventures in which an incumbent

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



invests typically have a very limited track record, mak-ing it difficult to form status differentials. However, VCswho coinvest with the incumbent are widely consideredexperts on emerging technological discontinuities andstart-ups, and they usually have well-established trackrecords on which to form status assessments. In fact,the VC industry is one of the most frequently citedexamples of environments where large status differencesexist between VC companies, such as Kleiner PerkinsCaulfield & Byers, Sequoia Capital, and Draper FisherJurvetson, and more peripheral investors (Podolny andCastelluci 1999). Confronted with uncertainty, incum-bent firms’ senior managers are likely to revert to thestatus of the VC firm when evaluating the informationreceived from VC sources. Thus, information obtainedfrom a high-status VC company is more likely to influ-ence senior management’s judgments on the importanceof a technological discontinuity than information fromless centrally placed players.Contrary to peer incumbent networks, in which we

expected centrally positioned high-status peers to haveless information about emerging discontinuities, weexpect the opposite in VC networks. VC firms are deeplyinvolved in commercializing technological discontinu-ities because these discontinuities have high value cre-ation potential (von Burg and Kenney 2000). Theliterature on VC syndication also suggests that these cen-trally positioned VCs usually possess information advan-tages over VCs that are more peripherally positioned(e.g., Freeman 1999, Hochberg et al. 2007, Hochberget al. 2010, Sorenson and Stuart 2001). Knowledge aboutthe quality of the start-up firms in which VCs invest isscarce. To gain confidence in the quality of a start-upand the technology it seeks to commercialize, VCs mustprobe and cross-check several sources of informationabout these technologies and their developers. This duediligence process becomes extremely difficult when theVC holds a peripheral position in the network. In con-trast, VCs that are centrally located in a network cantap into the knowledge of multiple partners, comparetheir information, and make reasonable inferences aboutthe quality of the technology and its future trajectory(Greve 2009, Powell et al. 1996). To protect their sta-tus and trade secrets, centrally positioned VCs typicallysyndicate predominantly with other centrally positionedVCs (Hochberg et al. 2007) or with corporate investorsthat possess complementary resources that compensatefor their lack of status in the VC community (Keil et al.2010). As a result of this network dynamic, high-statusVCs may not only have more influence on top manage-ment’s attention allocation but may also give managerssuperior information about the likely impact of emergingtechnological discontinuities. Therefore, we hypothesizeas follows.

Hypothesis 4 (H4). The higher the status of the part-ners in heterophilous interorganizational ties, the earlier

top management will pay attention to emerging discon-

tinuous technological change.

Methods

Empirical Setting

To test our hypotheses on the role of different typesof interorganizational ties as structural antecedents oftop management’s attention to emerging technologicaldiscontinuities, we collected longitudinal data from thelargest companies (U.S.-based companies that are pub-licly traded in U.S. stock exchanges with revenues above200 million U.S. dollars in 1989) in four informa-tion and communication technologies (ICT) industriescovering the three-digit Standard Industrial Classifica-tion (SIC) codes 357 (computer and office equipment),366 (communications equipment), 367 (electronic com-ponents and accessories), and 737 (computer program-ming, data processing and other computer-related ser-vices) for the period 1989–2000. To prevent survivorbias, we applied the selection criteria at the beginningof the sample period and followed the companies to theend of the sample period or until they were acquiredor ceased operations. This created an unbalanced paneldata set.We focused on the 1989–2000 period, which wit-

nessed the rise of the Internet and wireless technolo-gies (Mowery and Simcoe 2002, Szulanski et al. 2004).At the beginning of the sample period (the beginningof the 1990s), the Internet as a business phenomenonwas still unheard of, and wireless technologies were intheir infancy. However, new start-ups began to explorethe field in the early 1990s. Around 1994, some lead-ing VCs, such as Kleiner Perkins Caufield & Byers,began actively investing in the Internet by funding rev-olutionary start-up companies such as Netscape andAmazon.com. Wireless technologies, such as WLAN,also became important, and VC-backed companiesplayed a significant role in this discontinuity. Overall,there were major technological discontinuities, and wire-less and Internet businesses started emerging in the mid-dle of our sample period.There is significant variance in the timing and the level

of attention paid by the top management of the com-panies we studied to these discontinuities. Some of thefirst incumbents in the sample mentioned the Internet intheir annual reports or 10-K filings in 1993–1995, butmost did so much later, between 1996 and 2000. Forinstance, Microsoft introduced its first Internet browserand mentioned the Internet for the first time in its letterto shareholders in fall 1995, and the Internet was broadlycovered in subsequent letters to shareholders. In fact,in the letter to shareholders in its 1995 annual report,Microsoft declared, “The largest launch in the history ofthe PC business, Windows 95 also launches the personalcomputer in a new role as a platform for the Internet

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



and the world of interactive networks” (Gates 1995).Formal attention to wireless technologies followed asimilar pattern. Our sample companies were generallyactive in different types of interorganizational networksthat included alliances and syndicated VC investments.Approximately 40% of the active sample companies hadcoinvestment ties with VCs.We gathered data on alliances and joint ventures from

Thomson Financial’s SDC Platinum database. Followingthe social network literature (Stuart 1998, Stuart et al.1999) in constructing partner status measures based onalliance network positions, we collected information onall alliances between all companies in SDC Platinumto construct full alliance networks without being con-strained by our focal partners and their immediate part-ners. This process resulted in 250,462 alliance ties in1989–2000. Following the literature on CVC and VCsyndication (Dushnitsky and Lenox 2005a; Dushnitskyand Shaver 2009; Podolny 2001; Sorenson and Stuart2001, 2008), we used the VentureXpert database to iden-tify incumbents’ CVC investments. Again, we wanted touse full syndication networks when calculating partnerstatus measures. We downloaded information on 139,478VC investments covering the 1989–2000 period. For thedependent variables, we collected 10-K filings from theSecurities and Exchange Commission (SEC) and sev-eral alternative sources (e.g., LexisNexis, Mergent) forthose firm-year observations for which the 10-K docu-ment was not available in the SEC EDGAR service. Wealso collected annual reports for all available firm-yearobservations using a large number of sources, includingdatabases of Thomson Financial, LexisNexis, Mergent,and company websites. In many cases, annual reportswere available only as images. In those cases, we scannedthe documents and reproduced the texts of the lettersto shareholders with the aid of optical character recog-nition software, after which the documents were manu-ally inspected and corrected where necessary. Finally, weused Compustat as the source for our financial controlvariables.

Research Design

We employed both event history analysis techniques andfixed effects panel regression models in analyzing theeffects of interorganizational relationships on the tim-ing of top management’s attention to technological dis-continuities. Event history models are well suited forour purpose because our data are right censored for allfirms that did not formally pay attention to technologicaldiscontinuities before they were acquired, before theyceased to exist because of a bankruptcy, or before theend of the analysis period. The hazard models employedin event history analyses, reported later, explicitly con-sidered such censoring.When specifying the hazard model, one can choose

from several parametric models or a semiparametric

model (Hosmer et al. 2008). The semiparametric Coxmodel has the advantage of not making strong assump-tions about the baseline hazards function. This is impor-tant because incorrect parametric assumptions may yieldbiased estimates of the effects of covariates on thehazard rate (Blossfeld and Rohwer 1995). We analyzedthe data using the Cox regression model (Cox 1972),a common choice for analyzing time-to-event data inanalogous settings (e.g., Eggers and Kaplan 2009). Theformal Cox hazards regression model is

h4t5= h04t5 exp8�′X4t591

where h04t5 is the unspecified baseline hazards function,� is a vector of regression coefficients to be estimated,and X is the vector of covariates. The Cox proportionalhazard model does not make any assumptions about thebaseline hazard h04t5. Cox’s partial likelihood estimatorprovides an effective way of estimating � without requir-ing estimates of h04t5. However, the Cox proportionalhazards model assumes that the hazard rates are propor-tional. Thus, h04t5 does not depend on the covariatesincluded in the model. The assumption of proportionalhazards in the Cox model was tested without evidenceof nonproportionality.The data on the timing of top management’s attention

to technological discontinuities were discrete becauseeach corporation was observed annually. This led to tiesamong companies in the timing of their top manage-ment’s attention to technological discontinuities. There-fore, we used the Efron method, which is suitable fordata with a large number of ties, but we found no sig-nificant differences in the results compared with otherchoices (Cleves et al. 2004).1 As additional robustnesstests, we reran the duration analysis using discrete-time survival analysis (a logit model with annual yeardummies and robust standard errors clustered by firm)and piecewise exponential models with robust standarderrors without changes in the main results.The independent variables used in the hazard mod-

els changed over time. To accommodate time-varyingcovariates, we divided the time period during which eachfirm was observed into yearly spells (Tuma and Hannan1984). All time-changing covariates were updated annu-ally. Each annual spell was treated as right censored,except for those spells that terminated in recognizing thetechnological discontinuity. Our models assumed thatinterorganizational relationships in year t−1 would leadto formal attention by top management to discontinuitiesin year t. This one-year lag also helped us to ensurethe causality of the relationship. Furthermore, for inde-pendent variables, we calculated network measures withtwo-year moving windows. Thus, to explain attention inyear t, we used relationships from t− 2 to t− 1, exceptfor two robustness test models, in which we tested sim-ple one-year lagged annual measures t − 1. To account

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



for the potential endogeneity of CVC activity, we esti-mated two-stage models that predicted coinvestmentswith VCs in the first stage and controlled for this in thesecond stage of the model, which estimated the effectsof interorganizational ties on the timing of top manage-ment’s attention to discontinuities.

Measures

Dependent Variables. Our primary dependent vari-able was the timing of the formal attention to majordiscontinuities by an incumbent’s top management. Fol-lowing other studies that measured management cog-nition for a large number of companies over a longperiod of time (e.g., Barr 1998, Eggers and Kaplan2009, Gerdes 2003, Kaplan 2008, Kaplan et al. 2003,Osborne et al. 2001, Schnatterly 2003), we derived thedata for the dependent variable from companies’ offi-cial reports. As a source for our primary measures, weused companies’ annual 10-K filings with the SEC, as inprior research (e.g., Feldman et al. 2010, Gerdes 2003,Schnatterly 2003, Yuthas et al. 2002). Official SEC fil-ings, such as Item 1 (Business) in annual 10-K fil-ings, have several advantages over other types of cor-porate documents. They are comprehensively available,audited, and comparable across firms and over time(Feldman et al. 2010). The comprehensive availability ofthe documents is important because even a small numberof missing documents would make our time-to-attentionvariable indeterminable. Other potential sources of infor-mation, such as letters to shareholders, press releases, orspeeches by senior executives, are not available as con-sistently for all firms in the sample. This factor excludedthe use of letters to shareholders in annual reports in theevent history models. Internal sources, such as minutesfrom board meetings, though desirable, would have beennearly impossible to obtain. Although measuring cog-nitive constructs based on publicly available documentshas limitations (e.g., Fiol 1995, Fiss and Zajac 2006),recent research has shown that these documents reflectjudgments about the importance of issues among seniormanagement. As a result, they offer reasonable prox-ies for senior management’s attention that can be con-structed through text analysis (Cho and Hambrick 2006,Duriau et al. 2007, Eggers and Kaplan 2009, Kaplan2008). In additional analyses that explained annual atten-tion by top management using fixed effects panel regres-sion models, we used letters to shareholders as thesource of data, with similar results. To operationalizethese measures, we used counts of mentions of theInternet (or wireless, in other analyses) divided by thenumber of words in the document (Item 1, Business inthe 10-K form in the main analyses and letter to share-holders in additional robustness tests).As previously stated, we examined two discontinu-

ities that substantially affected ICT companies during

the sample period 1989–2000: the Internet and wirelesstechnologies. Both technological discontinuities pro-foundly affected a large number of industries, redefin-ing how we think about information and communicationtechnologies (Szulanski et al. 2004, Yoffie 1997). Aswith many other discontinuous technologies, Internet-based and wireless-based businesses were first developedby new entrepreneurial ventures (Christensen 1997). Forinstance, whereas none of the sample corporations men-tioned the Internet in their 10-K filings before 1993,dozens of VCs had made investments in companies thatwere active in developing technologies relevant for theInternet since the 1980s, according to VentureXpert.To determine the time when top management paid

attention to these discontinuities, we searched the annual10-K filings of all sample companies over the entiresample period for occurrences of terms commonly usedby top management to refer to focal discontinuitieswithin the “Business” section that describes the businessin the Form 10-K (Item 1). We then recorded the yearof the first mention of the discontinuity. To examine theInternet, we focused on the phrase “the Internet,” whichconsistently referred to the Internet as a discontinuity.To validate our measure, we conducted further con-tent analysis that included key-word-in-context (KWIC)analyses. Although we found that a strict focus on thethe Internet in 10-K files helped to avoid false pos-itives (such as “Internet protocol” or “internetwork-ing,” which were used to refer to technical issues in asmall number of documents much earlier than the Inter-net as a phenomenon was known), we also developedbroader measures based on the KWIC analysis and reranthe main model including additional words (“Internet,”“Web,” “IP,” “on-line,” “online,” “e-commerce,” and “e-business”) with substantially similar results.Thus, our study’s first dependent variable was the time

when top management formally paid attention to theInternet in a firm’s corporate documents. This was mea-sured as the time between January 1989 (alternatively,January 1993 in the robustness tests) and the time ofa company’s first mention of technological discontinu-ity in its 10-K filings. We repeated this approach withwireless technologies (e.g., WLANs) as a focal disconti-nuity. Following the same logic we used when studyingthe Internet, we identified the first mention of variationsof the term “wireless” in the 10-K filings and testedthe hypotheses with this dependent variable. To measuretop management’s attention to wireless technologies, weconsidered other terms, such as “mobile” and “cellu-lar,” but we found that the word “wireless” was the bestchoice for this analysis.2

In addition to the first mentions of the focal discon-tinuities that we used in the event history analyses, weoperationalized another dependent variable to measureannual formal attention by top management to thesediscontinuities as the relative share they received either

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



in the 10-K filings or in letters to shareholders whennormalized by document length. We ran several robust-ness tests that included alternative dependent variablesto ensure the validity and reliability of our measure-ment approach. These tests are presented in the Resultssection.

Independent Variables. The first independent variable(H1) was homophilous interorganizational ties to peerincumbents, measured as the number of alliances or jointventures with peers in the sample. To construct this vari-able, we counted all alliances and joint ventures of afocal firm with all other firms in our sample within thesame three-digit industry while excluding direct com-petitors with a four-digit SIC code match.3 Our choiceof size and industry to capture partner homophily isjustified because size is an important yet easily mea-sured criterion to evaluate a firm’s available resources.For instance, small and medium-sized enterprises in thesame industry face very different competitive conditionsthan large firms do. We further limited the alliances andjoint ventures to relationships within the same indus-try because industry boundaries often shape the men-tal models of a firm and its top management (Poracet al. 1995). To test H3, we created a measure of het-erophilous interorganizational ties by counting coinvest-ment ties with VC firms. The idea behind this measure isthat by establishing a CVC unit that coinvests with pri-vate VC firms, an incumbent firm could connect to thenetworks of start-ups and VCs. These connections wereoperationalized through syndicated investments along-side VCs (i.e., the CVC fund in which an incumbentis the sole limited partner coinvesting with private VCs,thus creating links between the corporation and VCcompanies and start-ups). Syndicated investments fre-quently involve the sharing of information about start-ups, their technologies, and the markets they are likelyto serve (Gompers and Lerner 1999, Sorenson and Stu-art 2001). Information about emerging technologies typ-ically reaches VC networks early (Bygrave et al. 2001,von Burg and Kenney 2000).To test H2 and H4 on the effects of partners’ sta-

tus in homophilous and heterophilous ties, we first cre-ated measures for partner status in homophilous and het-erophilous ties. To measure status in alliance networksand VC syndication networks, respectively, we followedPodolny (1993, 2001, 2005) in using Bonacich’s (1987)centrality measure. The Bonacich centrality ci4�1�5 wasdefined as

ci4�1�5=∑

j

4�+�cj5Rij1

where Rij is an element of the relational matrix R, andeach element of R is the number of companies in whichfirms i and j has invested together (or alliances betweenthe companies, in alliance networks). � is the degree to

which the centrality of i is a function of the centralitiesof other firms; it is the radius of the influence of i. Ifwe did not expect the VC syndication or alliance rela-tionships of j to directly benefit i, � should be small.We followed earlier studies and set � as three-quartersof the reciprocal of the largest eigenvalue of R (Podolny1993, Sorenson and Stuart 2001).For alliance relationships, we calculated the partners’

Bonacich’s centrality measures using the position ofevery alliance partner in the full network formed by allalliances established in the specified moving window.In our analyses, we chose two years, but we also testedother windows without a significant difference in results.We then aggregated the partner measures for focal firmsby averaging across all partners with which a focal firmhad formed alliances or joint ventures in a given year.For the CVC coinvestments with VCs, we calculatedthe centrality of all VCs with which a focal firm hadcoinvested in the given moving window (two years inreported analyses, except for two reported robustnesstests with one-year windows) using the full VC syndica-tion networks. We then averaged these values to arriveat an aggregate value.The results of this measurement approach appear

valid. Examining the data shows that in 1994, forinstance, Microsoft had one of the highest partner sta-tus values in heterophilous ties in the sample based onthis measure because it coinvested with Kleiner PerkinsCaufield & Byers (the most central private VC in ourdata that year) and other top-tier VCs.

Control Variables. The analyses also included sev-eral firm- and industry-related control variables. We con-trolled for top management’s future orientation basedon a text analysis of letters to shareholders or 10-K fil-ings, following the operationalization of Yadav et al.(2007), based on the use of the wording “will” in thosedocuments. Top management’s future orientation couldinfluence both the attention they paid to discontinuitiesand the tools they used to facilitate timely attention(e.g., CVC).We also captured incumbents’ investments in R&D as

a measure of how much weight companies gave to tech-nology and innovation issues. To separate R&D effectsfrom company size effects, we used R&D intensity (i.e.,the firm’s R&D expenditure divided by its annual sales)instead of using R&D expenditure directly.We further controlled for firm size and profitability.

We measured company size as the logarithm of a com-pany’s annual revenues (in millions of U.S. dollars),derived from Compustat. Size may influence the timingof senior executives’ attention to technological disconti-nuities because larger firms may have formal, dedicatedunits to follow such changes. Profitability was measuredas the return on assets (ROA) because past performancemight influence a company’s environmental scanningand CVC investments.

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



In addition to these firm-level controls, we controlledfor industry differences by dummy coding a firm’sindustry by its three-digit SIC classification. Dummycodes were developed for the SIC 366 (communi-cations equipment), SIC 367 (electronic componentsand accessories), and SIC 737 (computer programming,data processing, etc.) sectors. SIC 357 (computer andoffice equipment) served as the reference industry class.Finally, in the duration models, the piecewise con-stant proportional hazards specification of the modelscontrolled for the increasing adoption of the Internetover time by changing annual baseline hazard rates.In discrete-time duration analyses that we ran as robust-ness tests and in fixed effects panel regressions, weapplied year dummies. We lagged time-varying indepen-dent and control variables by one year to reduce the riskof reverse causality.

ResultsTable 1 presents the correlations and descriptive statisticsfor the study’s variables. As the data in Table 1 show,the correlations between the independent and the controlvariables were low to moderate. Variance inflation fac-tor statistics were also low (below 2) in comparison tothe typically used threshold of 10, suggesting that mul-ticollinearity was not a problem in our analysis.We used event history analyses to test the hypothe-

sized relationships between interorganizational relation-ships and the timing of top management’s attention todiscontinuities. Model 1 in Table 2 reports the basemodel using Cox proportional hazards regression toexplain the time-to-attention to the Internet. Only thecommunications equipment industry dummy was weaklysignificant (negative). Other control variables were notsignificant. Model 1, like all other models, included theinverse Mill’s ratio that is used to control for the poten-tial endogeneity of CVC investments syndicated withVCs.4 This ratio was also not significant.Hypotheses 1 and 3 are tested first in Model 2,

with the first mention of the Internet as the dependentvariable. Hypothesis 1 predicted that focal firms’ het-erophilous ties to peer incumbents would have a negativeeffect on top management’s timely attention to techno-logical discontinuities. As noted in Model 2, the coeffi-cient for homophilous ties is negative but insignificant.These results failed to support our hypothesis. Addition-ally, the results concerning H3, which predicted a pos-itive effect on heterophilous interorganizational ties toVCs, failed to provide support for the hypothesis.Model 3 is similar to Model 2 but adds the partner

status measures for H2 and H4. Concerning the pre-dicted negative effect for partner status in homophilousinterorganizational ties in H2, the coefficient is insignif-icant. Hypothesis 4, which predicted a positive relationbetween partner status in heterophilous interorganiza-tional ties and timely attention by top management to Table

1DescriptiveStatisticsandCorrelations

Variable

Mean

S.D

.Min

Max

12

34

56

78

9

1Attentionpaid

totheInternet t

00201

00610

00000

50860

2Attentionpaid

towirelesst

00418

10304

00000

160351

00016

3Homophilo

usinterorganizationaltiest−

2−t−

100654

20461

00000

220000

−00007

−00059

4Statusofpartners

inhomophilo

us

450879

1710444

00000

115360567

−00007

−00049

00582∗

interorganizationaltiest−

2−t−

1

5Heterophilo

usinterorganizationaltiest−

2−t−

170511

450674

00000

8860000

00168∗

00072

00021

00016

6Statusofpartners

inheterophilo

us

830485

2330326

00000

115880176

00276∗

00125∗

00067

00057

00390∗

interorganizationaltiest−

2−t−

1

7To

pmanagement’s

future

orientationt−

110813

10440

00000

90300

00007

−00025

−00028

00007

−00023

−00008

8Firm

sizet−

170068

10521

−40423

110380

00084∗

−00036

00377∗

00252∗

00222∗

00347∗

−00019

9R&D

intensityt−

100076

00060

00000

00453

00116∗

00103∗

00020

−00021

00023

00143∗

00126∗

00091∗

10

Firm

profitability

t−1

00044

00356

−10305

80590

00019

00002

−00008

−00021

00022

00041

−00043

−00031

−00064

Note.Firm-yearobservationsare

usedin

Models

6and12.

∗p<0005.

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.

Maula,Keil,

andZahra:TopManagem

ent’s

Atten

tionto

Disco

ntin

uities

938

Organizatio

nScien

ce24(3),pp.926–947,©2013IN

FORMS

Table 2 Top Management’s Attention to Discontinuous Technological Change

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Model 11 Model 12

Internet Internet Internet Internet Internet Internet Wireless Wireless Wireless Wireless Wireless Wireless

Dependent variable= 1st mention 1st mention 1st mention 1st mention 1st mention mentions 1st mention 1st mention 1st mention 1st mention 1st mention mentions

H1. Homophilous −00095 −00070 00061 00089 −00004 −00075 −00002 −00200 −00206∗ 00000

interorganizational tiest−2−t−1 4000785 4000895 4000495 4000815 4000225 4000865 4000625 4001925 4000965 4000195

H2. Status of partners in 00000 −00001 00001 00000 −00002 00003 00001 00000

homophilous 4000015 4000015 4000015 4000005 4000025 4000045 4000025 4000005

interorganizational tiest−2−t−1

H3. Heterophilous 00004 00002 −00004 −00004 00001+ 00003 00001 00000 −00005 00001

interorganizational tiest−2−t−1 4000045 4000045 4000045 4000075 4000015 4000035 4000035 4000065 4000115 4000015

H4. Status of partners in 00001∗ 00001∗ 00002∗ 00000∗∗ 00001∗∗ 00003∗∗ 00005∗∗ 00000+

heterophilous 4000015 4000015 4000015 4000005 4000015 4000015 4000025 4000005

interorganizational tiest−2−t−1

Top management’s future −00032 −00006 00004 00117 00104 −00033 −00151 −00152 −00152 −00077 −00080 −00014

orientationt−1 4001095 4001085 4001125 4001085 4001105 4000305 4001245 4001285 4001235 4001195 4001235 4000295

Firm sizet−1 00092 00328 00395 −00295 −00327 −00139 00260 00348 00443∗ 00546 00583 −00017

4001915 4002655 4002615 4002975 4003015 4001485 4002275 4002415 4002245 4003555 4003555 4000945

R&D intensityt−1 00911 10974 20330 00259 −00149 −00126 80240∗∗ 80550∗∗ 80827∗∗ 30267 30810 50938

4300175 4303245 4303295 4405915 4408085 4104585 4109375 4200415 4109565 4203165 4205005 4307845

Firm profitabilityt−1 00102 00158 00157 00602 00827 −00068 −00058 −00056 −00131 −00621 −00156 00012

4001475 4001555 4001585 4009965 4100525 4000725 4003575 4004365 4006455 4200415 4200795 4000705

Industry dummy: −10072+ −00986 −00870 −00799 −00542 20475∗∗ 20478∗∗ 20698∗∗ 10210 10127

Communications equipment 4005575 4006095 4006055 4005435 4005625 4005365 4005435 4005755 4007495 4007425

Industry dummy: Electronic −00620 −00700 −00714 −00757+ −00547 10601∗∗ 10560∗∗ 10744∗∗ 00663 00589

components 4004335 4004585 4004575 4004305 4004365 4004695 4004775 4005015 4006115 4006795

Industry dummy: Software 00295 00390 00423 −00057 00084 00836+ 00803+ 00898+ 00413 00324

4003705 4003825 4003855 4004375 4004475 4004785 4004845 4004905 4006215 4006745

Inverse Mill’s ratiot−1 −00418 −00137 00116 −00358 −00263 −00136 00168 00289 00591+ 10630∗∗ 10623∗∗ −00043

4003435 4004945 4004685 4004805 4004775 4000955 4003705 4003915 4003285 4005005 4005395 4001135

Year dummies Incl. Incl.

Constant 10746 00552

4102625 4007405

Observations 557 557 557 276 276 693 524 524 524 195 195 693

Notes. Models 1–5 explain the hazards for corporations to pay attention to the Internet in the 10-K filing Item 1, Business, for the first time during an annual spell. Models 7–11 repeat this

for wireless. In these models, positive coefficients indicate greater hazard, i.e., earlier attention. Models 1–5 and 7–11 are Cox proportion hazards models with time-varying regressors

and robust standard errors. Models 4 and 5 and 10 and 11 limit the sample to companies that eventually pay attention to the discontinuity. Models 5 and 11 use one-year windows (t−15

for hypothesis variables instead of the two-year windows 4t−2− t−15 in all other models. Models 6 and 12 are fixed effect panel regression models explaining normalized counts of the

mentions of discontinuities in the Form 10-K Business sections (number of mentions divided by the number of words in the document×11000). Robust standard errors are in parentheses.∗∗p < 0001; ∗p < 0005; +p < 0010 (two-tailed tests for control variables, one-tailed tests for directional hypotheses).

Downloaded from informs.org by [130.60.47.58] on 10 February 2014, at 08:19 . For personal use only, all rights reserved.



discontinuities, received strong support; the partner sta-tus measure is positive and significant.In Model 4, we sought to ensure that our findings were

not influenced by the varying importance of the Inter-net for our sample companies because some companiesdid not recognize the Internet in their 10-K filings dur-ing the sample period (these companies mostly exitedfrom the sample before the end of the period becausethey had been acquired by or merged with another com-pany or defaulted). Consequently, in Model 4, we lim-ited the sample to incumbent companies that, at somepoint during the time period, recognized the Internet intheir 10-K filings. The main results of Model 3 remainedstrong when the analysis was limited to this subset ofcompanies. In additional unreported robustness analyses,we replicated Model 4 using a discrete-time compet-ing risks model in which we identified firms that hadexited our sample before the right censoring in 2000without recognizing the Internet in their filings (usuallyby being acquired or after ceasing operations because ofpoor performance). The main results remained strong,but homophilous ties increased the hazard of exit fromour sample (i.e., increased the risk of business failure).In unreported robustness tests, we ran a discrete-timeduration model with self-selection correction (equivalentto Model 3). The results were qualitatively similar tothose obtained in Model 3, reaffirming the robustness ofour finding that the status of partners in heterophilousties is important. Furthermore, we ran a piecewise expo-nential model and obtained similar results. In Model 5,we repeated Model 4 using only one-year lagged annualnetwork measures instead of a two-year moving windowto focus the analyses on very recent network ties that arelikely to be particularly relevant for learning about dis-continuities. The results were similar to those reported.To ensure that our results were robust across differ-

ent ways of defining and measuring top management’sattention to discontinuities, we devised another opera-tionalization based on annual weighted counts of men-tions in 10-K filings. This is a common measurementapproach in other empirical research on the attention-based view when attention is used as an independentvariable explaining organizational responses (Eggers andKaplan 2009, Kaplan 2008, Kaplan et al. 2003). How-ever, it should be noted that this continuous measureof the strength of attention is conceptually quite differ-ent from our hypotheses and primary analyses focusingon the time to attention. Therefore, we consider thiscomplementary evidence. Model 6 in Table 2 reportsour results using a fixed effects panel regression thatexplains the relative frequency of mentions of the Inter-net in 10-K filings. This model shows that the status ofpartners in heterophilous ties has positive and signifi-cant coefficients, supporting H4. Other hypotheses arenot supported, although the coefficient for H1 is in thehypothesized direction. These results are consistent with

those obtained in the event history models. Furthermore,increasing our confidence in the main finding about thepositive effect of partner status in heterophilous tieson top management’s attention to discontinuities, weobserved the same result also when operationalizing themeasure using letters to shareholders instead of 10-Kfiles in unreported robustness analyses.As explained in the Methods section, to help gen-

eralize our findings and to ensure that these findingsare not idiosyncratic to one particular discontinuity, weconsidered another important discontinuity in the ICTsector, wireless technologies. The results reported inModels 7–12 were similar to and supportive of our mainfindings based on time-to-attention to the Internet. How-ever, one qualitative difference in the results was that H1received support in Model 11. Although the results werevery similar, we consider the Internet the primary settingfor our hypotheses because wireless technology was aless significant discontinuity than the Internet was duringthis time period. For instance, the adoption of wirelesstechnologies occurred over a significantly longer periodof time, with some companies paying attention to wire-less in 1989–1990 but fewer companies paying attentionto it at the end of the sample period. Thus, wireless tech-nology is a secondary setting for our hypothesis tests.Our analyses suggested that homophilous ties did not

consistently have the expected significant negative effecton timely attention to discontinuities by top manage-ment. Therefore, we decided to further explore this rela-tionship. Because the firms in our sample gradually paidincreasing attention to the Internet and wireless tech-nologies, one possible explanation for the lack of signif-icance might be the changing effect of peer incumbentsover time. In the beginning, when peer companies donot pay attention to emerging discontinuities, they maybe expected to have an inertial negative effect. However,as incumbents increase their attention to the disconti-nuity, they may become advocates of this discontinuity,and their influence on other incumbents that have notpaid formal attention to the discontinuity might change.In fact, once the top management of peer incumbentfirms have paid attention, it is difficult to argue that theywould continue to have an inertial effect. Instead, theireffect could become positive from the perspective of afocal company that has not yet paid attention. Conse-quently, we ran an unreported post hoc analysis of thispotential explanation by dividing annual homophilousties based on whether the top management of the part-ner incumbent paid attention to the discontinuity. In thepanel data analysis in the Internet case, we found thatbefore the partner had paid attention, the effect wasnegative, and after the partner had paid attention, theties had a positive effect, which weakly supports H1.We consider this finding interesting and a logical par-tial explanation for our insignificant results in the pri-mary tests of H1. At the same time, we note that the

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



evidence is quite weak and not replicable in our eventhistory analyses or in the wireless case. A proper testof the moderating effect of peer attention on the effectof homophilous ties would require a different researchsetting that includes controls for other channels (e.g.,competitor intelligence) through which incumbents learnfrom each other.

DiscussionIn this study, we aimed to examine the role of differenttypes of interorganizational ties and the status of part-ners within these ties as structural antecedents of topmanagement’s attention to emerging technological dis-continuities. We tested our hypotheses in a longitudinalstudy of the incumbents in four information and commu-nications technology industry sectors, examining theirties in full alliance and VC networks. We found that thestatus of partners in heterophilous ties formed throughcoinvestments with high-status VC firms was a signifi-cant predictor of timely attention to such discontinuitiesby top management. Homophilous ties did not have aconsistently significant effect, but our ex post analysissuggested that this result might be due to the changingeffects of peers when their top management begins topay attention to a given discontinuity. Our findings con-tribute to several bodies of literature, as discussed next.

Technological Discontinuities and

Top Management’s Attention

A growing body of research on technological disconti-nuities highlights the role of the attention of incumbentfirms’ top managers in responding to discontinuous tech-nological change (e.g., Barr 1998, Kaplan 2008, Kaplanand Tripsas 2008, Tripsas and Gavetti 2000, Viranyet al. 1992). For instance, in work on discontinuities infiber optics and pharmaceuticals, Kaplan and colleagues(Eggers and Kaplan 2009, Kaplan 2008, Kaplan et al.2003) have shown that top management’s attention pre-dicts effective firm responses. By interpreting changesin the external environment and formulating (or at leastorchestrating the formulation of) appropriate and timelyresponses, top managers can shape a firm’s strategy.However, prior studies have not addressed the questionof why top managers in some firms pay attention to thesediscontinuities whereas top managers in other companiesdo not.Prior research has explained the failure to pay atten-

tion to discontinuities largely by invoking cognitiveand incentive factors that work against timely atten-tion by top management to technological discontinuities.These discontinuities are often based on knowledgethat is distant from the incumbents’ existing knowl-edge base (Kaplan and Tripsas 2008) and thereforemay lead to business models that directly contradict theexperience of a firm’s senior executives (Tripsas and

Gavetti 2000). In some cases, technological discontinu-ities may also run counter to deeply held beliefs withinthe firm that form the organization’s identity (Tripsas2009). As a result, environmental cues about these dis-continuities are easily misinterpreted by top management(Barr 1998, Porac et al. 1995, Reger and Palmer 1996,Tripsas and Gavetti 2000). Furthermore, top manage-ment frequently has personal incentives to not embracesuch discontinuities because of the risks of investing inunproven new technologies (Kaplan 2008, Kaplan andHenderson 2005).Our results suggest that some of the interorga-

nizational relationships in which incumbents engagecan play a crucial role in guiding their top manage-ment’s attention to technological discontinuities. Indeed,we find that heterophilous ties with high-status part-ners through coinvestments with high-status VCs pos-itively affect top managers’ timely attention, whereashomophilous interorganizational relationships do nothave a significant impact in this regard. These resultsextend prior arguments in the social network litera-ture that posit that access to diverse sources of infor-mation should be positively related to the ability toadapt to environmental change (Koka and Prescott 2008,Lee 2007). Our results suggest that in addition to thediversity of the information sources, the status of aninformation source plays an important role becauseknowledge and information from these sources carry dif-ferent weights with corporate decision makers. The highstatus of partners in heterophilous ties, such as syndica-tion with high-status VCs, may increase the credibilitythat top managers attach to information from these ties.

Attention-Based View of the Firm

Our results also contribute to the attention-based viewof the firm. In particular, the literature on the attention-based view has noted the importance of structural deter-minants of organizational attention. For instance, Ocasio(1997) argues that the rules of the game, resources, play-ers, and social positions within the firm generate a setof values that order the legitimacy, importance, and rele-vance of issues and possible organizational responses tothese issues. However, prior theorizing in the attention-based view has been largely restricted to structureswithin organizations. Our study extends this researchby emphasizing structures outside of a firm’s bound-aries that might influence top management’s attention.Our core finding—that heterophilous interorganizationalrelationships with high-status partners affect top man-agement’s attention to technological discontinuities—suggests that such structures can play an important rolein shaping attention patterns. As a result, theorizingwithin the attention-based view should be extended toinclude structures beyond the boundaries of the firm.Decision makers in organizations are embedded in multi-ple interorganizational structures that may influence theirattention patterns.

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



Our core finding further suggests a theoretical linkbetween the literature on attention and the literature onstatus, which warrants additional theoretical and empir-ical work. Status and status expectations influence rela-tionship formation across organizations (Chung et al.2000) and are central to intraorganizational group forma-tion processes (Ruef et al. 2003) and interactions duringtask performance (Ridgeway 2001) within these groups.In the context of the present study, status expecta-tions may influence how executives in high-status indus-trial corporations relate to high-status or low-status VCswith whom their firm coinvests and, consequently, howthey evaluate information from these VCs. Althoughour empirical results are limited to the effects of statuswithin heterogeneous interorganizational relationships,status differentials may be important theoretical expla-nations for attention allocation processes in intraorga-nizational contexts. Status expectations may shape thestructures that affect attention allocation and may haveimportant effects on interaction patterns within intraor-ganizational social structures.The link identified in this study between status and

attention may also shed light on bottom-up attention pro-cesses (Ocasio 2011). The decision makers in the firmswe studied are faced with an ecology of informationregarding new technologies that compete for attention.Given the uncertainty regarding these technologies, deci-sion makers sometimes find it difficult to evaluate thisinformation based on its quality (Podolny 1993, 1994;Podolny 2001) and may therefore revert to the status ofthe sources to determine the salience of the information.In this view, status moderates bottom-up processes ofattention allocation.In our empirical analyses, we failed to find consis-

tent support for the hypothesized effects of homophilousrelationships. However, our post hoc analysis, whichdivided homophilous relationships into those with part-ners that paid attention to discontinuity and those thatdid not pay attention to discontinuity, suggests that situa-tional factors (in this case, the fact that a partner had paidattention to the discontinuity) may interact with struc-tures that guide attention allocation. Although tentativegiven the limitations of our data, this finding suggests theneed to further explore how the interaction of structuraland situational factors could affect attention allocationprocesses.Our findings further contribute to recent research on

attention to rare events (Rerup 2009, Weick and Sutcliffe2006). This research has argued that maintaining suffi-cient quality of attention to emerging rare events is chal-lenging, particularly in light of short-term pressures thatarise from ongoing routine activities that may exhaustorganizational attention structures (Levinthal and Rerup2006, Salvato 2009). Regular, focused discussions witha select group of high-status VCs may create a contextin which top managers’ attention focuses on rare events,

freeing discussions from the short-term pressures of rou-tine business development and allowing them to focus onemerging technologies and business models. These reg-ular discussions can increase the quality of attention andfacilitate mindful exploration of potential discontinuitiesby the top management team (Ocasio 2011, Salvato2009). By regularly interacting with high-status VCs, topmanagers may create a mechanism that routinizes themindful exploration of weak signals and allows themto overcome the trade-off between routine and mindfulbehavior (Levinthal and Rerup 2006, Salvato 2009).

Interorganizational Networks

Our analyses offer a fine-grained picture of the rolesof network structure and network composition in socialnetworks. Scholars generally agree that, under most cir-cumstances, social networks formed through interorga-nizational relationships provide benefits to the focal firm(Burt 1992, Coleman 1988, Podolny and Stuart 1995,Powell et al. 1996). However, there is a debate aboutthe type of network that provides the most benefits. Thisdebate has focused on the structures of various networks,emphasizing the different benefits derived from sparseand dense networks (Burt 1992, Coleman 1988, Lee2007) and largely ignoring the differences among thefirms with which a focal firm connects in a network.A considerable amount of prior literature on social

networks has viewed (or, at least, empirically treated)all nodes in a network as equal. However, someresearchers have recently underscored the importanceof network composition in explaining the differenteffects of networks to which a focal firm is connected(Lee 2008). Our finding that heterophilous ties throughcoinvestments with high-status VCs have a consistentlysignificant positive effect on timely attention by topmanagement, whereas homophilous interorganizationalties through alliances to industry peers have no suchunivocal effect, further emphasizes the importance ofnetwork composition. Our analysis of full alliance andVC syndication networks suggests that treating all nodesin a network as equal or focusing solely on networkstructure can lead to inconclusive or even misleadingresults. In particular, the inconclusive finding regardinghomophilous ties is interesting. Our post hoc analysissuggests that homophilous ties seem to have initial iner-tial effects, whereas these ties seem to have positiveeffects when partner firms pay attention to discontinuity.This finding, albeit tentative, suggests that some effectsof social networks are likely to be dependent on dif-ferences across nodes that may not be fully capturedby network structure measures. The results also suggestthat treating all interorganizational relationships as onesocial network in which the firm is embedded may over-simplify reality and mask some of the effects of socialnetworks. Future theorizing would benefit from distin-guishing between the multiple networks to which a firm

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



is connected. Different networks serve different purposesand consist of different types of actors, which may havedifferent effects on the focal firm. Such fine-grained the-orizing will create more realistic models of how socialnetworks influence firm behavior and its outcomes.

Corporate Venture Capital

Our results contribute to the literature on the role ofCVC in the corporate innovation process. Althoughit is somewhat cyclical in volume, CVC investing hasbecome an integral tool of corporate innovation activities(Dushnitsky 2006, Dushnitsky and Lenox 2005a). Priorresearch on CVC investments has examined the potentiallearning benefits of these investments (Dushnitsky andLenox 2005b, Dushnitsky and Shaver 2009, Schildt et al.2005, Wadhwa and Kotha 2006). Early research hasexplored the effect of CVC investments on innovationrates (Dushnitsky and Lenox 2005b) and on exploratoryknowledge creation and learning (Schildt et al. 2005,Wadhwa and Kotha 2006) by linking these investmentsto patenting rates. More recently, studies have pointed tothe limits of knowledge transfer between start-ups andincumbents (Dushnitsky and Shaver 2009), suggestingthat incumbents’ CVC investments might fail in circum-stances where learning might be most valuable for theincumbent. Overall, these findings question the impor-tance of CVC for corporate innovation.Our study contributes to this debate by showing that

the role of CVC activities in corporate innovation mightbe related less to transferring patentable knowledge tothe parent firm and more to directing top management’sattention to major changes in the firm’s environment.CVC can thus be viewed as radar that identifies andhighlights emerging technologies and new businesses(Dushnitsky and Lenox 2006, Keil et al. 2008a, Siegelet al. 1988). As such, CVC investments can significantlyinfluence the cognizance of business opportunities andrelated business models among CVC managers and topmanagers. Even when an incumbent does not transfera specific technology that a start-up can commercial-ize, CVC investments may provide important insightsinto the evolution of a technological field. Informa-tion received from CVC investments may also influ-ence how senior executives think about the likelihoodthat a technological area will become important for theincumbent. These higher-level learning processes are notalways influenced by the individual start-ups in whichthe incumbent invests. Instead, they are shaped by theinformation CVCs receive in the process of screeningventures together with high-status VCs and the portfolioof deal proposals they process for investment purposestogether with syndicating VCs. For instance, althoughMicrosoft did not directly invest in the first Internetstart-ups in 1994, its top-tier coinvestor, Kleiner PerkinsCaufield & Byers, was a lead investor in pioneeringInternet start-ups such as Netscape in early 1994. Among

the incumbent firms in our sample, Microsoft was oneof the early companies to pay attention to the Inter-net, with the Internet clearly identified as a key businessdriver in the letter to shareholders of its 1995 annualreport. Although coinvestment relationships of corpo-rations in CVC investments with high-status VC firmsmay not necessarily create direct knowledge flows aboutthe VCs’ other portfolio companies, emerging businessopportunities and start-ups in which high-status fellowboard members or their VC partnerships are publiclyinvesting are likely to receive more attention by fel-low board members than are other start-ups of similarsize or potential business opportunities of similar uncer-tainty. Among corporate tools for recognizing discontin-uous technological change, a large number of indirectties to promising VC-backed start-ups through coinvest-ments with top-tier VCs is a unique feature of coinvestedCVC investments compared with other interorganiza-tional relationships.Given our results, it is reasonable to ask why CVC

investment, a corporate activity that is often relativelysmall in scale and detached from the core business,plays such an important role in influencing senior man-agement’s cognition. One explanation can be foundin the theory of knowledge brokers (Hargadon andSutton 1997, Hargadon 2002). CVC managers are oftenuniquely positioned to participate in both the corporateenvironment and the start-up and VC communities (Keilet al. 2008a). By operating in both environments simul-taneously, CVC managers are well positioned to act asknowledge brokers who mediate and address cognitiveconflicts that arise between the cognitive frameworksof existing business unit managers and frameworksrelated to technological discontinuity. Overall, our anal-yses show that CVC activities may play a unique roleby providing a window on technological discontinuities.

Managerial Implications

Our results have several implications for managerialpractice. They suggest that to avoid being trapped inthe innovator’s dilemma (Christensen 1997, Rosenbloomand Christensen 1994), incumbents must develop rela-tionships that can direct their top management’s atten-tion to these discontinuities. The results emphasize thatincumbents can create a radar for technological discon-tinuities by actively using syndicated CVC investmentswith top-tier VCs to monitor emerging developmentsin start-up networks. Technologies developed by start-ups are often the forerunners of technological shifts inan industry. Building relationships to such start-ups andthe investors that fund them can direct top managers’attention to technological discontinuities. However, theeffectiveness of this tool depends on an incumbent’sability to coinvest with the highest-status VCs. Conse-quently, incumbents need to structure their investment

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



activities in ways that enable cooperation with high-status VCs. Executives must think strategically abouttheir CVC activities by setting clear investment goals.For instance, executives must identify the types of tech-nological trends their companies should follow and howto best position their companies in VC networks to fol-low these trends.When considering the costs and benefits of CVC

for corporations, it is worth noting that although CVCrequires capital to be committed, each investment in asyndicate with independent VCs should have a positiverisk-adjusted return expectation as a stand-alone invest-ment. With a long-term commitment to CVC investmentactivity, the strategic benefits gained from these invest-ments should be relatively cheaper in comparison to cor-responding benefits from R&D expenditures and otheractivities that create direct expenses.

Limitations

Our study has limitations that should be recognizedwhen interpreting its results. Specifically, our measure-ment approach forces us to abstract from the interpre-tative processes in which top managers engage as theyaddress technological change in their environment (Lantand Shapira 2001). However, the approach we have fol-lowed is in line with similar studies that measure cog-nitive constructs across a large number of organizationsover a long period of time. A word count-based mea-sure might be one of the few ways to capture issuesof managerial cognition in large sample studies (Eggersand Kaplan 2009, Kaplan 2008).Because we construct our attention measure using let-

ters to shareholders and 10-K reports, we are unable tomeasure valence as an opportunity or threat or to mea-sure the specific nature of the understanding. Althoughvalence and specific understanding would be of theoret-ical interest, the methods we use have been found tobe significantly weak in assessing such constructs (Fiol1995). Therefore, we must leave the issue of valence tofuture research.Our sample was drawn from four information and

communications technology industry sectors, providingan interesting setting in which to examine a technolog-ical discontinuity such as the Internet. Companies inthese industries are expected to be well positioned torecognize the strategic importance of the Internet (andwireless technologies in the robustness analysis). Ourindustry choice enabled us to select a sample of firmsthat were similarly affected by the same discontinu-ities. Because industries differ in their knowledge basesas well as in their rate and sources of technologicalchange and because the evolutionary paths of technolo-gies might differ, the results may not apply to otherindustries. Thus, constructive replications of our studyin other industry settings and for other technologies mayyield a richer understanding of how CVC investments

can support incumbents’ early recognition of technolog-ical discontinuities.A final limitation concerns our approach to measur-

ing the networks in which our focal firms were embed-ded. To make the analysis manageable, our analysisof network structures focused exclusively on direct tiesand partners’ status in alliance and VC networks whileignoring information flows through indirect ties. Futureresearch should explore the effects of such indirect tiesin data sets that are smaller in size and therefore moremanageable.

Directions for Future Research

Our results highlight several avenues for future research.Future researchers should study different network com-positions in more detail and should develop a moredetailed picture of the effects of the multiple social net-works in which an incumbent is embedded. Althoughwe were able to distinguish between peer alliances andCVC coinvestments with high-status VCs, other dis-tinctions can be made and should be incorporated intofuture social network studies. Researchers should alsoexplore the time-varying effects of peer alliances that weidentified.Our results suggest that external structural determi-

nants are an important factor in guiding organizationalattention to environmental events. Future research couldanalyze the role of other external determinants of man-agement attention, such as research consortia, tradeassociations, or board interlocks. Future research wouldbenefit also from investigating how internal and externaldeterminants interact to shape organizational attention.This research would not only advance understanding ofattention structures but would also provide importantfindings for the design of organizational issue manage-ment systems.Our findings highlight a need for more fine-grained

research into the processes used by incumbents to cap-ture knowledge and information from CVC investmentsand the deals they screen. The ways that incumbentsgather, analyze, and interpret information about pendingtechnological changes are issues worthy of investigation.The systems and processes that incumbents employ forthese purposes also require further study. It is essen-tial to investigate the processes that incumbents employto make effective use of information in designing theirstrategies. A related issue for future research involvesthe different mechanisms incumbents apply to identifytechnological discontinuities.

AcknowledgmentsThe authors acknowledge the valuable comments we havereceived from senior editor William Ocasio and three anony-mous reviewers. Earlier versions of the manuscript were com-mented on by Vikas Aggarwal, Raffi Amit, Erkko Autio, YvesDoz, Kathleen M. Eisenhardt, Eileen Fischer, Sarah Kaplan,

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



Riitta Katila, Xiaowei Rose Luo, Michelle Rogan, and partic-ipants at the Academy of Management Meetings and semi-nars at Bocconi University, Helsinki University of Technology,INSEAD, Rensselaer Polytechnic Institute (RPI), SchulichSchool of Business at York University, St. Petersburg StateUniversity, University of Southern California, University of St.Gallen, and Zurich University. The superb research assistanceof Yihua Jiang, Juhana Joensuu, Mikko Jääskeläinen, KarriLehtovuori, Olli Paukkeri, Heikki Saukola, Henri Schildt, andVeera Siivonen during different phases of the process is alsoacknowledged with gratitude. The paper has been partiallywritten during M. Maula’s visiting scholarships at INSEADand SCANCOR, Stanford University. The authors acknowl-edge the financial support from the Research Programme forAdvanced Technology Policy (ProACT) of the Ministry ofEmployment and the Economy of Finland and Tekes, theFinnish Funding Agency for Technology and Innovation, aswell as the Academy of Finland [Grant 112129].

Endnotes1We also reran the analyses using as an alternative, theBreslow method, which is the default in Stata (but theoreti-cally less suitable for our situation) and obtained qualitativelysimilar results for the hypothesized relationships.2We found that the word “wireless” was commonly used torefer to the new phenomenon (e.g., DSC Communications ina December 1990 10-K filing: “0 0 0will allow the Companyto provide a new group of products to both the current cel-lular marketplace and the emerging wireless communicationsmarket” (DSC Communications 1990); and Oracle in a May1995 10-K filing: “The networks can range from local andwide area networks to ISDN, Internet, broadband and wire-less” (Oracle 1995)). In contrast, “mobile” was used for manyother purposes, such as referring to mobile users who couldlog in to wired networks in remote locations (e.g., 3Com in aMay 1995 10-K filing: “3Com’s AccessBuilder remote accessservers give these mobile users simplified analog or digital(ISDN) dial-up access to the network”; see 3Com 1995).3We also reran the analyses without this limitation in the sameindustry to test the effects of the industry limitation on theeffects of alliances with incumbents without significant differ-ences in the results.4In the selection model, we estimated the inverse Mill’s ratiobased on a probit model explaining whether the focal com-pany engaged in syndicated CVC investments in a particularyear. The independent variables included top management’sfuture orientation 4t−15, sales (logged, t−1), R&D intensity4t − 15, a dummy for missing R&D data 4t − 15, return onassets (ROA, t − 1), whether the focal company engaged insyndicated CVC investments in the previous period, year dum-mies, and industry dummies. Of these variables, sales, R&Dintensity, ROA, and past CVC activity were all positive andsignificant. Additionally, some of the year and industry dum-mies were significant.

ReferencesAnderson P, Tushman ML (1990) Technological discontinuities and

dominant design: A cyclical model of technological change.Admin. Sci. Quart. 35(4):604–633.

Barnett WP, Pontikes EG (2008) The Red Queen, success bias, andorganizational inertia. Management Sci. 54(7):1237–1251.

Barr PS (1998) Adapting to unfamiliar environmental events: A lookat the evolution of interpretation and its role in strategic change.Organ. Sci. 9(6):644–669.

Barr PS, Huff AS (1997) Seeing isn’t believing: Understanding diver-sity in the timing of strategic response. J. Management Stud.

34(3):337–370.

Barr PS, Stimpert JL, Huff AS (1992) Cognitive change, strategicaction, and organizational renewal. Strategic Management J.

13(S1):15–36.

Bettis RA, Prahalad CK (1995) The dominant logic: Retrospectiveand extension. Strategic Management J. 16(1):5–14.

Blossfeld H-P, Rohwer G (1995) Techniques of Event History Mod-

eling: New Approaches to Causal Analysis (Lawrence ErlbaumAssociates, Mahwah, NJ).

Bonacich P (1987) Power and centrality: A family of measures. Amer.J. Sociol. 92(5):1170–1182.

Bouquet C, Birkinshaw J (2008) Weight versus voice: How foreignsubsidiaries gain attention from corporate headquarters. Acad.Management J. 51(3):577–601.

Bourgeois LJ III, Eisenhardt KM (1988) Strategic decision processesin high velocity environments. Management Sci. 34(7):816–835.

Burt RS (1992) Structural Holes: The Social Structure of Competition(Harvard University Press, Cambridge, MA).

Bygrave WD, Lange JE, Kotha RR, Stock W (2001) Venture capi-tal investments and the growth of revolutionary new industries.Bygrave WD, Autio E, Brush CG, Davidsson P, Green PG,Reynolds PD, Sapienza HJ, eds. Frontiers in Entrepreneurship

Research 2001 (Babson College, Babson Park, MA), 523–535.

Cho TS, Hambrick DC (2006) Attention as the mediator between topmanagement team characteristics and strategic change: The caseof airline deregulation. Organ. Sci. 17(4):453–469.

Christensen CM (1997) Innovator’s Dilemma: When New Technolo-

gies Cause Great Firms to Fail (Harvard Business School Press,Boston).

Chung S, Singh H, Lee K (2000) Complementarity, status similar-ity and social capital as drivers of alliance formation. StrategicManagement J. 21(1):1–22.

Cleves MA, Gould WW, Gutierrez RG (2004) An Introduction to Sur-

vival Analysis Using Stata: Revised Edition (Stata Press, CollegeStation, TX).

Coleman JS (1988) Social capital in the creation of human capital.Amer. J. Sociol. 94(Supplement):S95–S120.

Cox DR (1972) Regression models and life-tables. J. Roy. Statist. Soc.Ser. B 34(2):187–220.

Daft RL, Weick KE (1984) Toward a model of organizations as inter-pretation systems. Acad. Management Rev. 9(2):284–295.

Danneels E (2011) Trying to become a different type of company:Dynamic capability at Smith Corona. Strategic Management J.

32(1):1–31.

DSC Communications (1990) DSC Communications CorporateAnnual Report, Form 10-K of SEC filing. http://www.dsccommunications.com.

Duriau VJ, Reger RK, Pfarrer MD (2007) A content analysis ofthe content analysis literature in organization studies: Researchthemes, data sources, and methodological refinements. Organ.Res. Methods 10(1):5–34.

Dushnitsky G (2006) Corporate venture capital: Past evidence andfuture directions. Casson M, Yeung B, Basu A, Wadeson N, eds.The Oxford Handbook of Entrepreneurship (Oxford UniversityPress, Oxford, UK), 387–431.

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



Dushnitsky G, Lavie D (2010) How alliance formation shapescorporate venture capital investment in the software indus-try: A resource-based perspective. Strategic Entrepreneurship J.

4(1):22–48.

Dushnitsky G, Lenox MJ (2005a) When do firms undertakeR&D by investing in new ventures? Strategic Management J.

26(10):947–965.

Dushnitsky G, Lenox MJ (2005b) When do incumbents learn fromentrepreneurial ventures? Corporate venture capital and invest-ing firm innovation rates. Res. Policy 34(5):615–639.

Dushnitsky G, Lenox MJ (2006) When does corporate venture capitalinvestment create firm value? J. Bus. Venturing 21(6):753–772.

Dushnitsky G, Shaver JM (2009) Limitations to interorganizationalknowledge acquisitions: The paradox of corporate venture cap-ital. Strategic Management J. 30(10):1045–1064.

Dyer JH, Singh H (1998) The relational view: Cooperative strategyand sources of interorganizational competitive advantage. Acad.Management Rev. 23(4):660–679.

Eggers JP, Kaplan S (2009) Cognition and renewal: Comparing CEOand organizational effects on incumbent adaptation to technicalchange. Organ. Sci. 20(2):461–477.

Feldman R, Govindaraj S, Livnat J, Segal B (2010) Management’stone change, post earnings announcement drift and accruals. Rev.Accounting Stud. 15(4):915–953.

Fiol CM (1995) Corporate communications: Comparing execu-tives’ private and public statements. Acad. Management J.

38(2):522–536.

Fiss PC, Zajac EJ (2006) The symbolic management of strategicchange: Sensegiving via framing and decoupling. Acad. Man-

agement J. 49(6):1173–1193.

Fombrun C, Shanley M (1990) What’s in a name? Reputation buildingand corporate strategy. Acad. Management J. 33(2):233–258.

Freeman J (1999) Venture capital as an economy of time. LeendersRTAJ, Gabbay S, eds. Corporate Social Capital and Liability

(Kluwer, Norwell, MA), 460–482.

Gargiulo M, Benassi M (2000) Trapped in your own net? Networkcohesion structural holes, and the adaptation of social capital.Organ. Sci. 11(2):183–196.

Gates WH (1995) Letter to shareholders. Microsoft, Redmond, WA.http://www.microsoft.com/investor/AnnualReports/default.aspx.

Gerdes J Jr (2003) EDGAR-Analyzer: Automating the analysis of cor-porate data contained in the SEC’s EDGAR database. DecisionSupport Systems 35(1):7–29.

Gilbert CG (2005) Unbundling the structure of inertia: Resource vs.routine rigidity. Acad. Management J. 48(5):741–763.

Gompers PA, Lerner J (1998) The determinants of corporate venturecapital successes: Organizational structure, incentives, and com-plementarities. NBER Working Paper 6725, National Bureau ofEconomic Research, Cambridge, MA.

Gompers PA, Lerner J (1999) The Venture Capital Cycle (MIT Press,Cambridge, MA).

Granovetter MS (1973) The strength of weak ties. Amer. J. Sociol.78(6):1360–1380.

Greve HR (2009) Bigger and safer: The diffusion of competitiveadvantage. Strategic Management J. 30(1):1–23.

Hagedoorn J (1993) Understanding the rationale of strategic technol-ogy partnering: Inter-organizational modes of cooperation andsectoral differences. Strategic Management J. 14(5):371–385.

Hambrick DC, Mason PA (1984) Upper echelons: The organizationas a reflection of its top managers. Acad. Management Rev.

9(2):193–206.

Hargadon AB (2002) Brokering knowledge: Linking learning andinnovation. Staw BM, Kramer RM, eds. Research in Organiza-

tional Behavior, Vol. 24 (Elsevier, Oxford, UK), 41–85.

Hargadon A, Sutton RI (1997) Technology brokering and inno-vation in a product development firm. Admin. Sci. Quart.

42(4):716–749.

Heide JB, Miner AS (1992) The shadow of the future—Effects ofanticipated interaction and frequency of contact on buyer-sellercooperation. Acad. Management J. 35(2):265–291.

Henderson RM (1993) Underinvestment and incompetence asresponses to radical innovation: Evidence from the photolitho-graphic alignment equipment industry. RAND J. Econom.

24(2):248–270.

Henderson RM, Clark KB (1990) Architectural innovation: Thereconfiguration of existing product technologies and the failureof established firms. Admin. Sci. Quart. 35(1):9–20.

Hill CWL, Rothaermel FT (2003) The performance of incumbentfirms in the face of radical technological innovation. Acad. Man-

agement Rev. 28(2):257–274.

Hochberg YV, Ljungqvist A, Lu Y (2007) Whom you know mat-ters: Venture capital networks and investment performance.J. Finance 62(1):251–301.

Hochberg YV, Ljungqvist A, Lu Y (2010) Networking as a barrier toentry and the competitive supply of venture capital. J. Finance65(3):829–859.

Hoffman AJ, Ocasio W (2001) Not all events are attended equally:Toward a middle-range theory of industry attention to externalevents. Organ. Sci. 12(4):414–434.

Hosmer DW, Lemeshow S, May S (2008) Applied Survival Analy-

sis: Regression Modeling of Time-to-Event Data (John Wiley &Sons, Hoboken, NJ).

Kaplan S (2008) Cognition, capabilities, and incentives: Assessingfirm response to the fiber-optic revolution. Acad. Management J.

51(4):672–695.

Kaplan S, Henderson R (2005) Inertia and incentives: Bridging orga-nizational economics and organizational theory. Organ. Sci.

16(5):509–521.

Kaplan S, Tripsas M (2008) Thinking about technology: Applying acognitive lens to technical change. Res. Policy 37(5):790–805.

Kaplan S, Murray F, Henderson R (2003) Discontinuities and seniormanagement: Assessing the role of recognition in pharmaceu-tical firm response to biotechnology. Indust. Corporate Change

12(2):203–233.

Keil T, Autio E, George G (2008a) Corporate venture capital, disem-bodied experimentation and capability development. J. Manage-

ment. Stud. 45(8):1475–1505.

Keil T, Maula MV, Wilson C (2010) Unique resources of corpo-rate venture capitalists as a key to entry into rigid venturecapital syndication networks. Entrepreneurship Theory Practice

34(1):83–103.

Keil T, Maula MVJ, Schildt H, Zahra SA (2008b) The effect of gov-ernance mode and relatedness of external business developmentactivities on innovative performance. Strategic Management J.

29(8):895–907.

Koka BR, Prescott JE (2008) Designing alliance networks: The influ-ence of network position, environmental change, and strategy onfirm performance. Strategic Management J. 29(6):639–661.

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



Lane PJ, Lubatkin M (1998) Relative absorptive capacity and interor-ganizational learning. Strategic Management J. 19(5):461–477.

Lant TK, Shapira Z (2001) Organizational Cognition: Computationand Interpretation (Lawrence Erlbaum Associates, Mahwah, NJ).

Lee GK (2007) The significance of network resources in the race toenter emerging product markets: The convergence of telephonycommunications and computer networking, 1989–2001. Strate-gic Management J. 28(1):17–37.

Lee GK (2008) Relevance of organizational capabilities and itsdynamics: What to learn from entrants’ product portfolios aboutthe determinants of entry timing. Strategic Management J.

29(12):1257–1280.

Leonard-Barton D (1992) Core capabilities and core rigidities:A paradox in managing new product development. Strategic

Management J. 13(S1):111–125.

Lerner J (1994) The syndication of venture capital investments. Finan-cial Management 23(3):16–27.

Levinthal D, Rerup C (2006) Crossing an apparent chasm: Bridgingmindful and less-mindful perspectives on organizational learn-ing. Organ. Sci. 17(4):502–513.

Levinthal DA, March JG (1993) The myopia of learning. StrategicManagement J. 14(S2):95–112.

Luo XR, Deng L (2009) Do birds of a feather flock higher?The effects of partner similarity on innovation in strategicalliances in knowledge-intensive industries. J. Management

Stud. 46(6):1005–1030.

McDonald ML, Khanna P, Westphal JD (2008) Getting them tothink outside the circle: Corporate governance, CEOs’ externaladvice networks, and firm performance. Academy Management

J. 51(3):453–475.

McPherson M, Smith-Lovin L, Cook JM (2001) Birds of a feather:Homophily in social networks. Annual Rev. Sociol. 27:415–444.

Mowery DC, Simcoe T (2002) Is the Internet a US invention? An eco-nomic and technological history of computer networking. Res.Policy 31(8–9):1369–1387.

Nelson RR, Winter SG (1982) An Evolutionary Theory of Eco-

nomic Change (Belknap Press of Harvard University Press,Cambridge, MA).

Ocasio W (1997) Towards an attention-based view of the firm. Strate-gic Management J. 18(S1):187–206.

Ocasio W (2011) Attention to attention. Organ. Sci. 22(5):1286–1296.

Ocasio W, Joseph J (2005) An attention-based theory of strategyformulation: Linking micro- and macroperspectives in strategyprocesses. Szutanski G, Porac J, Doz Y, eds. Strategy Pro-

cess (Advances in Strategic Management, Vol. 22) (Emerald,Bingley, UK), 39–61.

Oracle (1995) Oracle Corporation Annual Report, Form 10-Kof SEC filing. http://www.sec.gov/Archives/edgar/data/777676/0000891618-95-000501.txt.

Osborne JD, Stubbart CI, Ramaprasad A (2001) Strategic groupsand competitive enactment: A study of dynamic relationshipsbetween mental models and performance. Strategic Manage-

ment J. 22(5):435–454.

Peteraf M, Shanley M (1997) Getting to know you: A theory of strate-gic group identity. Strategic Management J. 18(S1):165–186.

Podolny JM (1993) A status-based model of market competition.Amer. J. Sociol. 98(4):829–872.

Podolny JM (1994) Market uncertainty and the social character ofeconomic exchange. Admin. Sci. Quart. 39(3):458–483.

Podolny JM (2001) Networks as pipes and prisms of the market.Amer. J. Sociol. 107(1):33–60.

Podolny JM (2005) Status Signals (Princeton University Press,Princeton, NJ).

Podolny JM, Castelluci F (1999) Choosing ties from the inside of aprism: Egocentric uncertainty and status in venture capital mar-kets. Leenders RTAJ, Gabbay SM, eds. Corporate Social Capitaland Liability (Kluwer, Norwell, MA), 431–445.

Podolny JM, Stuart TE (1995) A role-based ecology of technological-change. Amer. J. Sociol. 100(5):1224–1260.

Porac JF, Thomas H, Baden-Fuller C (1989) Competitive groups ascognitive communities The case of Scottish knitwear manufac-turers. J. Management Stud. 26(4):397–316.

Porac JF, Thomas H, Wilson F, Paton D, Kanfer A (1995) Rivalryand the industry model of Scottish knitwear producers. Admin.Sci. Quart. 40(2):203–227.

Powell WW, Koput KW, Smith-Doerr L (1996) Interorganizationalcollaboration and the locus of innovation: Networks of learningin biotechnology. Admin. Sci. Quart. 41(1):116–145.

Prahalad CK, Bettis RA (1986) The dominant logic: A new linkagebetween diversity and performance. Strategic Management J.

7(6):485–501.

Reger RK, Palmer TB (1996) Managerial categorization of competi-tors: Using old maps to navigate new environments. Organ. Sci.7(1):22–39.

Rerup C (2009) Attentional triangulation: Learning from unexpectedrare crises. Organ. Sci. 20(5):876–893.

Ridgeway CL (2001) Gender, status, and leadership. J. Soc. Issues57(4):637–655.

Rosenbloom RS, Christensen CM (1994) Technological disconti-nuities, organizational capabilities, and strategic commitments.Indust. Corporate Change 3(3):655–685.

Ruef M, Aldrich HE, Carter NM (2003) The structure of found-ing teams: Homophily, strong ties, and isolation among USentrepreneur. Amer. Sociol. Rev. 68(2):195–222.

Salvato C (2009) Capabilities unveiled: The role of ordinary activitiesin the evolution of product development processes. Organ. Sci.20(2):384–409.

Schildt HA, Maula MVJ, Keil T (2005) Explorative and exploita-tive learning from external corporate ventures. EntrepreneurshipTheory Practice 29(4):493–515.

Schnatterly K (2003) Increasing firm value through detection andprevention of white-collar crime. Strategic Management J.

24(7):587–614.

Siegel R, Siegel E, MacMillan IC (1988) Corporate venture capital-ists: Autonomy, obstacles and performance. J. Bus. Venturing3(3):233–247.

Simon HA (1947) Administrative Behavior (Macmillan, New York).

Sorenson O, Stuart TE (2001) Syndication networks and the spa-tial distribution of venture capital investments. Amer. J. Sociol.106(6):1546–1588.

Sorenson O, Stuart TE (2008) Bringing the context back in: Settingsand the search for syndicate partners in venture capital invest-ment networks. Admin. Sci. Quart. 53(2):266–294.

Stuart TE (1998) Network positions and propensities to collabo-rate: An investigation of strategic alliance formation in a high-technology industry. Admin. Sci. Quart. 43(3):668–698.

Stuart TE, Hoang H, Hybels RC (1999) Interorganizational endorse-ments and the performance of entrepreneurial ventures. Admin.Sci. Quart. 44(2):315–349.

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.



Szulanski G, Doz Y, Ovetzky Y (2004) Incumbents’ framing: Threeestablished companies respond to the Internet. Baum JAC,McGahan AM, eds. Business Strategy Over the Industry Life-

cycle (Advances in Strategic Management, Vol. 21) (Emerald,Bingley, UK).

3Com (1995) 3Com Corporation Annual Report, Form 10-Kof SEC filing. http://www.sec.gov/Archives/edgar/data/738076/0000738076-95-000011.txt.

Tripsas M (1997) Unraveling the process of creative destruction:Complementary assets and incumbent survival in the typesetterindustry. Strategic Management J. 18(Special issue):119–142.

Tripsas M (2009) Technology, identity, and inertia through thelens of “The Digital Photography Company.” Organ. Sci.

20(2):441–460.

Tripsas M, Gavetti G (2000) Capabilities, cognition, and iner-tia: Evidence from digital imaging. Strategic Management J.

21(10–11):1147–1161.

Tuma NB, Hannan MT (1984) Social Dynamics: Models and Methods

(Academic Press, Orlando, FL).

Tushman ML, Anderson P (1986) Technological discontinuitiesand organizational environments. Admin. Sci. Quart. 31(3):439–465.

Virany B, Tushman ML, Romanelli E (1992) Executive successionand organization outcomes in turbulent environments: An orga-nization learning approach. Organ. Sci. 3(1):72–91.

Vissa B, Greve HR, Chen W-R (2010) Business group affiliationand firm search behavior in India: Responsiveness and focus ofattention. Organ. Sci. 21(3):696–712.

von Burg U, Kenney M (2000) Venture capital and the birth of thelocal area networking industry. Res. Policy 29(9):1135–1155.

Wadhwa A, Kotha S (2006) Knowledge creation through external ven-turing: Evidence from the telecommunications equipment man-ufacturing industry. Academy Management J. 49(4):819–835.

Weick KE, Sutcliffe KA (2006) Mindfulness and the quality of orga-nizational attention. Organ. Sci. 17(4):514–524.

Wilson R (1968) The theory of syndicates. Econometrica 36(1):119–132.

Yadav MS, Prabhu JC, Chandy RK (2007) Managing the future: CEOattention and innovation outcomes. J. Marketing 71(4):84–101.

Yoffie DB (1997) Introduction: CHESS and competing in the age ofdigital convergence. Yoffie DB, ed. Competing in the Age of

Digital Convergence (Harvard Business School Press, Boston),1–36.

Yuthas K, Rogers R, Dillard JF (2002) Communicative action andcorporate annual reports. J. Bus. Ethics 41(1):141–157.

Markku V. J. Maula is a professor of venture capital,head of the Institute of Strategy, and the founding director ofthe Aalto Ventures Program at Aalto University, Finland. Hisresearch centers on the intersection of strategy, entrepreneur-ship, innovation, and finance with a particular focus on venturecapital and private equity, corporate venturing, mergers andacquisitions, and innovation.

Thomas Keil is a professor of strategic management atAalto University in Finland. He holds a D.Sc. (Tech) degreefrom the Helsinki University of Technology. His researchfocuses on corporate entrepreneurship; strategic renewal;mergers and acquisitions; and more generally, the intersectionof strategic management, entrepreneurship, and innovation inhigh-tech industries.

Shaker A. Zahra is a department chair, the Robert E.Buuck Chair in Entrepreneurship, and a professor of strategyand organization at the Carlson School of Management, Uni-versity of Minnesota. His research centers on entrepreneurshipand innovation in technology and science-based global indus-tries and international entrepreneurship.

Dow

nlo

aded

fro

m i

nfo

rms.

org

by [

130.6

0.4

7.5

8]

on 1

0 F

ebru

ary 2

014, at

08:1

9 . F

or

per

sonal

use

only

, al

l ri

ghts

res

erved

.

Documents

Top Management s Attention to Discontinuous Technological ... · Maula,Keil,andZahra: TopManagement’sAttentiontoDiscontinuities Organization Science 24(3), pp. 926–947, ©2013