Embed Size (px)
Citation preview
UNBUNDLING THE STRUCTURE OF INERTIA:RESOURCE VERSUS ROUTINE RIGIDITY
CLARK G. GILBERTHarvard University
I work to unbundle the structure of inertia into two distinct categories: resourcerigidity (failure to change resource investment patterns) and routine rigidity (failure tochange organizational processes that use those resources). Given discontinuouschange, a researcher’s failure to recognize these distinctions can generate conflictingfindings regarding effects of threat perception on inertia. Using field data on theresponse of newspaper organizations to the rise of digital media, I show that a strongperception of threat helps overcome resource rigidity but simultaneously amplifiesroutine rigidity. I develop an interpretive model exploring mechanisms for overcomingthese divergent behaviors.
The inability of incumbent firms to overcomeorganizational inertia when threatened with dis-continuous technological change has been a topicof repeated scholarly inquiry (Hannan & Freeman,1977; Henderson & Clark, 1990; Levinthal, 1992;Tushman & O’Reilly, 1996). One of the reasons thistopic receives so much attention is that incumbentfailure is extremely prevalent (Christensen &Rosenbloom, 1995; Tushman & Anderson, 1986).Perhaps even more intriguing is that failure oftenoccurs even when managers are aware of the needto change (Johnson, 1988). I propose that some ofthe difficulty observed (both in management prac-tice and scholarly research) occurs because of afailure to differentiate between two very distinctforms of inertia. Previous definitions of incumbentinertia describe it as the inability to enact internalchange in the face of significant external change(Miller & Friesen, 1980; Tushman & Romanelli,1985). I divide this general phenomenon into two
distinct categories (1) failure to change resourceinvestment patterns (Christensen & Bower, 1996;Henderson, 1993) and (2) failure to change the or-ganizational processes that use those resource in-vestments (Leonard-Barton, 1992; Nelson & Winter,1982). The first category is referred to here as re-source rigidity. The second is referred to here asroutine rigidity.
This article will show that the failure to differen-tiate the aspects of inertia is an oversight in theliterature, partly because it leaves the underlyingphenomenon inadequately described. But the lackof specificity can also lead to conflicting, even con-tradictory, findings. I will show how differentiatingbetween resource and routine rigidity can be par-ticularly important when exploring the role ofthreat perception under conditions of discontinu-ous change. Whereas some scholars suggest thatthreat perception enables response, others arguethat it constrains response. For example, studiesshow that threat can unlock inertia by motivatingchange (Huff, Huff, & Thomas, 1992; Lant, Mil-liken, & Batra, 1992). And yet, threat perception hasalso been shown to increase inertia by narrowingalternatives and focusing response on previouslylearned routines (Dutton & Jackson, 1987; Staw,Sandelands, & Dutton, 1981). I show here how re-conceptualizing inertia as having two distinctforms can help explain this seeming contradiction.Using longitudinal field data, I compare percep-tions in eight newspaper organizations of the emer-gence of digital publishing. Findings show that theway managers perceive the threat of discontinuouschange creates paradoxical links between resourceand routine rigidity. This article also examinesmechanisms for overcoming both types of rigiditysimultaneously.
I would like to thank my dissertation committee, Jo-seph Bower, Clayton Christensen, and Tom Eisenmann,as well as acknowledge the generous financial support ofthe Doctoral Programs of the Harvard Business School.Mike Tushman also provided tremendous support in de-veloping these ideas and preparing them for publication.Other useful comments have come from Jane Dutton,Kathy Eisenhardt, Ranjay Gulati, Nicolaj Siggelkow,Mary Tripsas, and three anonymous AMJ reviewers. I amlikewise grateful to the participants of the Wharton Tech-nology Mini-Conference and the BYU Winter StrategyConference for their suggestions. An earlier version ofthis paper was the winner of the Robert J. Litschert BestDoctoral Student Paper in Business Policy at the Acad-emy of Management annual meeting in 2001 and a dis-sertation finalist in 2002.
� Academy of Management Journal2005, Vol. 48, No. 5, 741–763.
741
THEORETICAL BACKGROUND
The Problem of Incumbent Inertia
In this section, I first specify how resource androutine rigidity represent different types of inertia.I then explore how interpretation might have adifferent effect on each type, looking specifically atconditions of discontinuous change. For this arti-cle, discontinuous change refers to externalchanges that require internal adaptation along apath that is nonlinear relative to a firm’s traditionalinnovation trajectory (Christensen & Bower, 1996;Tushman, Smith, Wood, Westerman, & O’Reilly,2003).
Resource rigidity. Of the many reasons thatfirms might underinvest in discontinuous change,resource dependency and incumbent reinvestmentincentives are two important ones. The theory ofresource dependency originated with the work ofPfeffer and Salancik (1978), which showed that afirm’s external resource providers shape and con-strain its internal strategic choices. Resource pro-viders can include both capital markets and cus-tomer markets. For example, Noda and Bower(1996) demonstrated how public equity marketsplace performance requirements that constrainchanges in business models and product architec-tures to conform to positions that originated at thetime of their initial funding. Christensen andBower (1996) demonstrated how customers canpenetrate a firm’s internal resource allocation sys-tems (Bower, 1970). “When the initial price/perfor-mance characteristics of emerging technologiesrender them competitive only in emerging marketsegments, and not with current customers, resourceallocation mechanisms typically deny resources tosuch technologies” (Christensen & Bower, 1996:198). In these studies, resources were both financialand attention-based (Ocasio, 1997).
The second general argument concerning re-source rigidity relates to market power. A debatepersists in the economics literature as to whetherincumbent firms, strongly positioned with a giventechnology, will invest sufficient resources in dis-continuous change (Arrow, 1962). Gilbert andNewberry (1982) found that if entry to a new tech-nology is blocked, incumbent firms have strongincentives to reinvest in their current market posi-tions and not in the new technology. Reinganum(1983) argued that if incumbent investment in-creases the probability of market adoption in a waythat alters a firm’s otherwise dominant position,the firm has strategic incentives not to invest. Thus,whether constraints stem from a desire to preservemarket power or from blinders created by resourcedependence, they represent powerful inertial
forces blocking incumbent investment in discon-tinuous change.
Routine rigidity. Even when incumbent firmsinvest, there is a second inertial problem. The lit-erature has repeatedly revealed the persistence andinflexibility of firm routines. Routines are definedhere as repeated patterns of response involving in-terdependent activities that become reinforcedthrough structural embeddedness and repeated use(Feldman & Pentland, 2003; Nelson & Winter,1982). Part of the explanation for routine rigidity isthat organizational processes that are tightlyaligned with one environment can be difficult tochange, because they are self-reinforcing and arenot built to adapt to discontinuities (Miller &Friesen, 1980; Siggelkow, 2001; Teece, Pisano, &Shuen, 1997; Tushman & Anderson, 1986). Further,exploitation processes can drive out explorationprocesses, making it difficult to develop new capa-bilities (Benner & Tushman, 2001; Burgelman,1994; March, 1991). Another part of the explana-tion is that the original motivation for designing anorganizational routine can be separated from thepeople executing the routine (Nelson & Winter,1982). The underlying logic pervades the thinkingof the organization, often manifesting as deeplyingrained cognition (Prahalad & Bettis, 1986; Trip-sas & Gavetti, 2000). The challenge is increasedbecause this cognition also becomes tacit (Schein,1985), making it even more difficult to recognizethe sources within the routines that are creating thedifficulties. Thus, managers often rely on a learnedpattern of response that is structurally and cogni-tively reinforced, instead of employing new searchefforts (March & Simon, 1958).
The Effect of Threat Perception on Inertia
Although the literature yields considerable in-sight into the sources of both resource and routinerigidity, less attention has been paid to their differ-ences and possible interactions (Henderson, 1993).Awareness of these differences is particularly im-portant when exploring potential contradictions re-garding the role of threat perception in discontin-uous change—specifically, whether the perceptionof a threat increases or decreases inertia. Threatperception is defined as a deep sense of vulnerabil-ity that is assumed to be negative, likely to result inloss, and largely out of one’s control (Dutton &Jackson, 1987; Jackson & Dutton, 1988). A well-developed literature on strategic change suggeststhreat perception is a response catalyst. Huff andhis colleagues (1992) showed how threat-relatedstress could overcome inertia. Lant and her col-leagues (1992) found that companies threatened
742 OctoberAcademy of Management Journal
with performance decline were more willing tocommit to strategic change. Barr and Huff (1997)showed that managers must view external changesas having a negative impact on performance beforeinternal changes are possible.
But although the strategic management literaturesuggests that threat is a catalyst in reducing inertia,other research shows that threat can actually in-crease inertia. Management scholars who apply so-cial-psychology ideas to threat-motivated responsediscuss three intermediate behaviors that can in-crease inertia. These are contraction of authority,reduced experimentation, and focus on existing re-sources. For example, Staw, Sandelands, and Dut-ton found threat led to “increased centralization ofauthority, more extensive formalization, and stan-dardization of procedures” (1981: 513). Hermann(1963) noted a contraction of authority in crisissituations. Other research shows that threat nar-rows the number of alternatives firm leaders arewilling to consider and reduces the level of exper-imentation in firm response (Ross & Staw, 1993;Staw et al., 1981). Finally, because threat-inducedbehavior is concerned with averting loss, it is di-rected toward preserving current resources ratherthan toward creating new opportunities (Dutton,1992; Hartman & Nelson, 1996; Mittal & Ross,1998).
In an effort to understand the conflicting conclu-sions regarding the impact of threat on inertia, Ipropose two related research questions: (1) Howdoes threat perception affect incumbent inertia inthe face of discontinuous change? and (2) Is theeffect of threat perception different for resourcerigidity and routine rigidity? After describing thisstudy’s research methods, I articulate a set of for-mal observations (propositions) based on compar-ative data from eight field sites. These observationsyield an interpretive model of response to dis-continuous change that addresses the varying im-
pact of threat perception on resource and routinerigidity.
METHODS
The present research might best be described astheory elaboration (Lee, 1999; Lee, Mitchell, & Sa-blynski, 1999) in that it elaborates theoretical linksnot previously addressed in the literature. For ex-ample, previous studies on the role of interpreta-tion have not distinguished between resource androutine rigidity, resulting in the apparent contra-dictions described in the previous section. Thus, Iattempt here to “simplify, reconnect, and redirecttheory” (Lee et al., 1999: 166) on the role of inter-pretation in response to discontinuous change in away that differentiates between these two sourcesof inertia.
Research Setting
The research employed a multicase design thatsupports a “replication logic,” whereby a set ofcases is treated as a series of experiments, eachserving to confirm or disconfirm a set of observa-tions (Yin, 1994). The primary unit of analysis wasan online venture of a newspaper organization re-sponding to digital media; the embedded unitswere the sponsoring newspapers and their corpo-rate managements. Table 1 describes the four com-panies and eight newspapers studied. Each com-pany owned two of the sampled newspapers,though each newspaper operated with significantautonomy. Print circulation across the sampleranged from 200,000 to more than 500,000 averagedaily readers. All of the newspapers but one werelocally distributed only. The competitive dynamicsacross markets were similar; each paper was thelargest competitor in a low-rivalry market. Internetpenetration across each market varied, but not by
TABLE 1Description of the Eight Newspapers Studied
Newspaper Parent OrganizationDaily PrintCirculation
CirculationRange
Online LaunchDate
Number of OnlineEmployees in
2000
Beacon A The Beacon Company 250,000 Local 1994 45Beacon B The Beacon Company 200,000 Local 1995 20Press A The Press Company �500,000 Local/regional 1994 �100Press B The Press Company 400,000 Local/regional 1995 60Expositor A The Expositor Company �500,000 Local/regional 1995 �100Expositor B The Expositor Company 200,000 Local 1996 32Morning News A The Morning News Company �500,000 National 1994 �100Morning News B The Morning News Company 300,000 Local 1996 41
2005 743Gilbert
more than 10 percentage points from highest tolowest. Each of the newspapers launched a Website between 1994 and 1996. Because of the sensi-tivity of the data, the names of the organizationsand the newspapers are disguised.
I chose to examine the response of newspaperfirms to digital publishing for two reasons. First,the effect of electronic publishing on the newspa-per industry generally matched the research ques-tions. Online publishing was a discontinuouschange in that it presented external changes thatrequired nonlinear internal adaptations. For exam-ple, the features of the Internet that early onlineusers valued—access to breaking news, searchabledatabases, live weather and traffic—differed con-siderably from the features that were available inprint. Similarly, the online business model wasalso considerably different, because it was drivenby an altered cost structure, new categories of rev-enue, and different customer requirements. Andyet, this discontinuity threatened to supplant bothprint readership and advertising over time, despitethe lack of initial customer overlap. The secondreason for selecting the newspaper industry is thatsingling out one industry helps control for extrane-ous variation (Eisenhardt, 1989a). The selection ofcase sites was based on theoretical sampling (Glaser
& Strauss, 1967) along a series of polar types that werelikely to extend the emerging theory (Eisenhardt,1989a). Thus, I targeted four innovative ventures andfour rigid ventures for examination.
Data Sources
Data were collected from three main sources:open-ended interviews, archival documents, anddirect observations. Table 2 summarizes thesesources. I also collected over 150 public docu-ments, including press releases, annual reports, an-alyst reports, and industry articles.
Interviews. Of a total 62 interviews, 51 werein-depth, one- to two-hour in-person interviewswith the senior executives at the corporate, news-paper, and online venture levels of the sampledbusinesses. I used semistructured interview tem-plates concerning what motivated a manager tocommit to digital media, how that commitmentevolved over time, the relation between print andonline efforts, and so forth. Additionally, 11 30-minute follow-on telephone interviews were usedto expand on the specific question of why a corpo-rate management chose to separate its digital ven-ture from its print business, or chose to keep theventure integrated with the print newspaper. I at-
TABLE 2Sources of Data
Newspaper
Interviewsa Archival Documents Direct Observations
Corporate Newspapers Ventures Total Number Examples Number Examples
Beacon A 4 (2) 2 3 9 7 Business plan, customer list,internal memo, strategicplan, sales collateral
5 Planning meeting, salescalls
Beacon B 2 (1) 2 2 6 5 Business plan, internal memo n.a. n.a.Press A 3 (1) 2 3 8 5 Business plan, customer list,
internal memo, salescollateral
5 Content developmentmeetings, budgetingmeetings
Press B 4 (2) 2 3 9 6 Business plan, customer list,strategic plan, salescollateral
4 Sales calls, newsroomplanning meetings
Expositor A 3 (1) 2 3 8 2 Business plan, internal memo n.a. Budgeting meetings,planning meetings
Expositor B 3 (2) 2 3 8 6 Business plan, customer lists,internal memo
5 n.a.
Morning News A 3 (1) 2 3 8 2 Business plan, customer lists,internal memo
5 Budgeting meetings,planning meetings,sales calls
Morning News B 2 (1) 2 2 6 n.a. n.a. n.a. n.a.
Total 24 (11) 16 22 62 (11) 33 n.a. 24 n.a.
a Data collection included 51 in-depth one- to two-hour open-ended, in-person interviews with personnel from the indicated areas and11 telephone interviews that were 30-minute follow-up conversations with the corporate executives on their decisions to separate theironline ventures from the parent newspapers or keep them integrated. The total figure includes both the extended in-person interviews andthe shorter telephone interviews. The telephone interviews are also shown in parentheses.
744 OctoberAcademy of Management Journal
tempted to triangulate by using multiple infor-mants and cross-checking information against ar-chival and public documents to avoid retrospectivebias in the interviews. Strict case study protocolwas followed. More than 90 percent of the inter-views were transcribed and entered into a casestudy database.
Archival documents. The 33 internal archivaldocuments that I collected at seven of the eightsites included online business plans, strategy pro-posals, internal memos, annual strategic planningdocuments, customer lists, and historical sales col-lateral material. These documents constituted avaluable primary source of data and offered a wayto cross-check the interviews and to control forretrospective bias. I focused on documents that dis-cussed whether or not to fund an online business.These included business plans starting as early as1990.
Direct observations. Over a year and a half, from2000 to 2001, I recorded meetings among onlineventure, newspaper, and corporate executives, at-tended planning meetings for the online ventures,observed story creation for the newspapers and theWeb sites, monitored sales calls for both the news-papers and the online products, and visited plan-ning meetings. In all, I observed 24 field eventsfrom five of the research sites and recorded theminto the case study database.
Research Process
My investigation was informed by three charac-teristics that have been observed as being associ-ated with threat perception: negative focus, empha-sis on loss, and sense of a lack of control (Dutton &Jackson, 1987; Jackson & Dutton, 1988). I groupedinterview and archival data into time-sequencedarrays to see how threat perception evolved overtime. I also tried to identify instances in which fieldobservations did or did not fit these categoriza-tions. For example, I coded the motivation to com-mit to the Internet as either a threat or an opportu-nity. When a business plan stated that “onlinerevenue would cannibalize print revenue” (nega-tive, loss) or “we can slow it down, but we can’tstop it” (lack of control), the motivation of themanager who had presented the plan was coded asa threat perception. In contrast, if a manager de-scribed online revenue as largely “additive to printrevenue” (positive, gain) or as susceptible to “in-fluence through involvement” (in control), I codedthe manager’s motivation as opportunity percep-tion. Although sample size and the varying charac-teristics of data sources at the different sites did notallow statistical comparison, a second reader, who
was blind to the original coding and purpose of thestudy, cross-checked my coding efforts for each siteover all periods (Miles & Huberman, 1984). In noinstance was there a conflict between the secondreader’s coding and the original coding.
I then developed a set of formally stated obser-vations that were based on early case analysis of aset of matched-pair, polar cases—one innovativeand the other rigid. Below, these formal observa-tions are presented as research propositions. I ana-lyzed the case data and “enfolded” a set of relevantliteratures, following methods for inductive theorydevelopment (Eisenhardt, 1989a; Glaser & Strauss,1967). I then used analytical replication to deter-mine whether the emerging relationships were con-firmed or disconfirmed in the rest of the sample(Eisenhardt, 1989; Yin, 1994). To confirm or dis-confirm the relationships identified in each obser-vation, I arrayed the data following techniques forcross-case pattern sequencing (Eisenhardt, 1989b)and tabular displays (Miles & Huberman, 1984). Aswith deductive hypothesis testing, the formal ob-servations fit a consistent pattern, though they didnot always conform perfectly (Eisenhardt, 1989a;Sutton & Callahan, 1987). I used the individualobservations to construct an interpretive model ofresponse to discontinuous change that differenti-ated inertia into resource and routine rigidity.
ANALYSIS OF DATA
Response in the Absence of Threat: ResourceRigidity
The field analysis confirmed that without a per-ception of threat, there was considerable resourcerigidity around discontinuous change. This wasmanifested in the data by the failure of establishednewspaper firms to invest financial and attention-based resources in digital publishing. I exploredthis occurrence of rigidity by examining the forcesof resource dependence and position reinvestmentincentives.
Resource dependency. Much of the initial re-source rigidity stemmed from resource dependen-cies related to the demands of the established printnewspaper customers—both advertisers and read-ers. These demands were difficult to reconcile withthe requirements of an emerging set of online cus-tomers. The data in Table 3 show that businessproposals for online often stalled for more than twoyears in review in these newspaper organizations.Even when money was provided, operating atten-tion could be equally difficult to secure. An onlinesales representative at the Beacon A recalled this:“Print reps could sell the online product, but with
2005 745Gilbert
varying degrees of success. Their margins werehigher on other products that were easier for themto sell. Online was really just a novelty to them.” “Ioccasionally sell a bundled print and online pack-age,” explained a print rep at the Press A. “There isno standard package, and it is hard to really knowwhat the print advertisers would want.” I gathereddata on the overlap between print and online ad-vertisers at five of the research sites, asking manag-ers to estimate how many of their top 25 onlineadvertisers were represented among the top 25newspaper advertisers. Customer overlap was only7 of a possible 125. The travel category at the PressA was a good example. Of the top ten bookingagents online, only four advertised in print. In sum-mary, the lack of customer overlap, a different sell-ing process, and the lower relative gross marginscombined to limit the amount of time and attentionprint sales reps were willing to invest in online.
Position reinvestment incentives. The resourcerigidity that I observed in the initial response of
these newspaper sites fell almost entirely into thecategory of resource dependence. But even as theseincumbent organizations came to realize the risksincurred by depending so heavily on their tradi-tional customers, they did not necessarily over-come their resource rigidity. Market position incen-tives to continue reinvesting in the core newspaperbusinesses remained strong. Nearly every researchsite conducted huge studies to estimate the canni-balization effect of providing their news informa-tion online. This concern was particularly acute atthe Expositor A and the Press A. The director of thePress Company’s Internet group reflected, “I thinkthe notion that people would start reading theirnewspapers on the screen was quite prevalent.”Managers at both of these sites wanted to stall ordelay any investment in a technology that mightcannibalize the core newspaper business. One se-nior manager described his strategy as “wait andsee,” preferring not to take his paper online beforeadoption became inevitable.
TABLE 3Resource Rigidity
Newspaper EvidenceYears inProposal
InitialCommitment
Date ofLaunch
Examples: Financial and OperationalConstraints
Beacon A Interviews, archival documents,public documents
2 Publishersponsorship
1994 Sales: “Print reps could sell the online product,but with varying degrees of success. Theirmargins were higher on other products thatwere easier for them to sell. Online wasreally just a novelty to them.” (salesrepresentative)
Beacon B Interviews, archival documents 2 Forecastprofitability
1995 Sales: Staff provided training, but then droppedthe program multiple times owing to lack ofclient interest.
Press A Interviews, archival documents,public documents
2 Forecastprofitability
1994 Sales: “I occasionally sell a bundled print andonline package . . . it is hard to really knowwhat print advertisers would want.” (salesrepresentative)
Press B Interviews, archival documents,public documents
2 CEOsponsorship
1995 Newsroom: Editors call online staff “low-browcontent” for using radio feeds for breakingnews, personals, and unedited user-postedcontent and refuse to work together.(newspaper editor)
Expositor A Interviews, archival documents,public documents
3 CEOsponsorship
1995 Budgeting: “And in the end, the only real valueis cash and cash creation. You can’t build abusiness just on potential or hope.” (CEO)
Expositor B Interviews, archival documents,public documents
2 Forecastprofitability
1996 Budgeting: “Look, when we roll these up intoour budgets we miss our targets.” (vicepresident, product development)
Morning News A Interviews, archival documents,public documents
1 Publishersponsorship
1994 Newsroom: “I will be goddamned if someonline reporter is going to call my sourcesand say they are from our paper.” (publisher)
Morning News B Interviews, public documents 2 Forecastprofitability
1996 Sales: “We bundled print and online, but therewere clients who wanted online only. Thesewere less interesting to sales reps and the or-ganization wasn’t ready to deal with thatreality.” (vice president, marketing)
746 OctoberAcademy of Management Journal
Threat Perception and Resource Rigidity
I next examine how a perception of threat af-fected the resource rigidity at each field site, firstpresenting the field data and then summarizing theobserved behavior more formally. The data indicatethat a strong perception of threat was associatedwith increased financial and organizational com-mitment to an online venture. At the various sites,the perception of threat built gradually and atslightly different paces. But by 1997–98, mostnewspapers’ managements had become genuinelyconcerned about the imminence of online newspa-per publishing, even though the advertisers in theirmainstream markets still did not view digital mediaas a valuable advertising outlet. The director ofmarketing at the Morning News B described theevolution of their thinking:
You felt like Chicken Little screaming “The sky isfalling,” but after a while people started listeningwhen they saw what the other competitors weredoing. We made watch lists for TV, radio, verticalstart-ups, telephonecompanies, andCitysearch.City-search was poaching people. . . . The publisher wasunlike some in that he saw the threat.
Whereas in earlier funding discussions, propo-nents tended to describe online publishing as agrowth opportunity meriting investment (oftenwith little success), in 1997–98 these discussionsturned to the perceived threat posed by the Inter-net. Data in Table 4 show that threat perceptiondeveloped across the sample. Threat perceptionwas evidenced as negative statements about thelack of control managers had in preventing losses tothe Internet. For example, the CEO of the PressCompany’s Internet group explained, “There werepeople who thought we would lose half of ourcirculation.” Others worried about losing classifiedadvertising products, including employment, realestate, and auto listings. These products (represent-ing as much as 60 percent of the profit of mostmetropolitan newspapers) appeared to be very por-table to the searchable database format availableonline. One executive recalled his organization’sperception as follows: “McKinsey had come in andhad done a rather startling analysis of the classifiedbusiness. They predicted that 20–30 percent of ourclassified revenue would disappear by 1998. Thatraised enormous alarm bells in some people.” Fi-nally, managers were concerned that the impact ofdigital media was largely out of the newspapercompanies’ control and that it would run its coursewith or without their participation. “What if we doevery damn thing we can think of and executeflawlessly and we still don’t make it?” lamented the
publisher of the Beacon A. “We can slow it down,but we can’t stop it.”
In seven of the cases, this mounting concernabout the Internet threat led to expanded financialand organizational commitment. The archival dataalso showed evidence that threat perception gener-ated resource commitment. For example, a 1997business plan for the Beacon B estimated that 15 to20 percent of its print classified revenues might belost to online. One business plan explained, “If wedon’t cannibalize ourselves, someone else will.”While earlier proposals had emphasized financialreturns and new market opportunities, argumentsfor increased funding now stressed the growingthreat to print revenue from online competition.Despite mounting losses in the online business,frequently exceeding 100 percent of online reve-nues, expenditures grew by as much as 400 percentduring the years that threat perceptions were build-ing (see Table 4). The number of employees allo-cated to the online ventures also increased. Forexample, dedicated online staff at one site ex-panded from 5 to 40 people during an eight-monthperiod in 1998. Staffing levels reached or exceeded100 individuals at many sites (see Table 4).
The expanding resource commitment was notlimited to financial expenditures and headcount; itextended to operating commitments such as man-agement time and effort. The threat posed by digitalmedia overcame customer dependencies that mightotherwise have pulled investment away from thenew technology. “Look, it didn’t make any sensefor us to try to sell this stuff, but we began to feelthat if we didn’t work on it, it might come back tohaunt us,” said a sales manager at the Beacon A.Reporters were asked to summarize articles andstories before they were published in print, andmany were encouraged to write follow up on sto-ries just for the Web.
Indeed, threat perception was observed to have apowerful catalytic effect on both types of resourcerigidity, a situation that can be summarized in myfirst two propositions:
Proposition 1a. The perception of an imminentthreat in the face of discontinuous change en-ables managers to overcome sources of re-source rigidity that stem from resourcedependence.
Proposition 1b. The perception of an imminentthreat in the face of discontinuous change en-ables managers to overcome sources of re-source rigidity that stem from incumbent posi-tion reinvestment incentives.
2005 747Gilbert
TA
BL
E4
Per
cep
tion
ofT
hre
atH
elp
sO
verc
ome
Res
ourc
eR
igid
ity
New
spap
erSt
ated
Sour
ceof
Thr
eat
Yea
rT
hrea
tEm
erge
dEv
iden
ceEx
ampl
esof
Thr
eat
Patt
ern
ofB
ehav
iora
Expe
ndit
ure
Expa
nsio
nbEm
ploy
eeEx
pans
ion
Bea
con
AR
ead
ersh
ip,d
isp
lay
adve
rtis
ing,
clas
sifi
edad
vert
isin
g
1996
Inte
rvie
ws,
arch
ival
doc
um
ents
,p
ubl
icd
ocu
men
tsLa
ckof
con
trol
:“W
hat
ifw
ed
oev
ery
dam
nth
ing
we
can
thin
kof
and
exec
ute
flaw
less
lyan
dw
est
ill
don
’tm
ake
it?
We
can
slow
itd
own
,bu
tw
eca
n’t
stop
it.”
(pu
blis
her
)
Lit
eral
rep
lica
tion
1996
–98:
400%
1996
–98:
153
40
Bea
con
BC
lass
ifie
dad
vert
isin
g19
97In
terv
iew
s,ar
chiv
ald
ocu
men
tsFo
cus
onlo
ss:1
997
inte
rnal
esti
mat
esof
15–2
0%lo
ssof
clas
sifi
edsh
are
by20
01(1
997
stra
tegi
cp
lan
)
Lit
eral
rep
lica
tion
1997
–99:
300%
1997
–99:
5318
Pre
ssA
Rea
der
ship
,dis
pla
yan
dcl
assi
fied
adve
rtis
ing
1997
Inte
rvie
ws,
arch
ival
doc
um
ents
,p
ubl
icd
ocu
men
tsFo
cus
onlo
ss:“
Th
ere
wer
ep
eop
lew
ho
thou
ght
we
wou
ldlo
seh
alf
ofou
rci
rcu
lati
on.”
(CE
O,I
nte
rnet
grou
p)
Lit
eral
rep
lica
tion
1997
–199
9:30
0%19
97–9
9:503
100�
Pre
ssB
Rea
der
ship
,dis
pla
yan
dcl
assi
fied
adve
rtis
ing
1997
Inte
rvie
ws,
arch
ival
doc
um
ents
,p
ubl
icd
ocu
men
tsN
egat
ive
effe
ct:“
Peo
ple
wer
ete
llin
gu
sth
atn
ewsp
aper
sw
ere
hea
din
gto
the
grav
eyar
dan
dw
ew
ere
begi
nn
ing
tobe
liev
eth
em.”
(pu
blis
her
)
Lit
eral
rep
lica
tion
1997
–99:
250%
1997
–99:
303
50
Exp
osit
orA
Sto
ckp
rice
,cl
assi
fied
adve
rtis
ing
1997
Inte
rvie
ws,
arch
ival
doc
um
ents
,p
ubl
icd
ocu
men
tsFo
cus
onlo
ss:“
We
wer
ew
orri
edab
out
clas
sifi
eds
pri
mar
ily.
Th
ese
new
firm
sw
ere
set
toco
me
inan
dta
keou
rm
ost
pro
fita
ble
pie
ceof
busi
nes
s.”
(ear
lyon
lin
ep
resi
den
t)
Lit
eral
rep
lica
tion
1997
–99:
300%
1997
–99:
503
100�
Exp
osit
orB
Rea
der
ship
,dis
pla
yan
dcl
assi
fied
adve
rtis
ing
1998
Inte
rvie
ws,
arch
ival
doc
um
ents
,p
ubl
icd
ocu
men
tsN
egat
ive
effe
ct:“
We
[mid
dle
man
ager
s]h
adbe
enco
nce
rned
for
aw
hil
e,bu
t...
the
reas
onw
efi
nal
lygo
tin
toth
em
arke
tw
asth
atou
rC
EO
was
taki
ng
hea
tfr
omW
all
Str
eet.”
(vic
ep
resi
den
t,p
rod
uct
dev
elop
men
t)
Lit
eral
rep
lica
tion
1998
–200
0:25
0%19
98–2
000:
153
31
Mor
nin
gN
ews
AP
rim
ary
mot
ivat
ion
dis
trib
uti
on/p
rod
uct
ion
savi
ngs
,not
thre
at
n.a
.In
terv
iew
s,ar
chiv
ald
ocu
men
ts,
pu
blic
doc
um
ents
Not
thre
at:“
Eig
hty
per
cen
tof
my
cost
sar
ep
rod
uct
ion
and
dis
trib
uti
on.N
owal
lof
asu
dd
enI
hav
ea
solu
tion
.It
isn
ota
con
ten
tp
lay,
but
am
ajor
cost
red
uce
ran
dp
rod
uct
exp
and
er.”
(pu
blis
her
)
Th
eore
tica
lre
pli
cati
on19
97–9
9:20
0%19
97–9
9:603
100�
Mor
nin
gN
ews
BC
lass
ifie
dad
vert
isin
g19
98In
terv
iew
s,p
ubl
icd
ocu
men
tsN
egat
ive
effe
ct:“
You
felt
like
Ch
icke
nL
ittl
esc
ream
ing
“th
esk
yis
fall
ing,
’bu
taf
ter
aw
hil
ep
eop
lest
arte
dli
sten
ing.
”(v
ice
pre
sid
ent,
mar
keti
ng)
Lit
eral
rep
lica
tion
1998
–200
0:40
0%19
98–2
000:
5340
aT
hes
eca
tego
ries
are
base
don
Yin
’s(1
994)
anal
ytic
alte
chn
iqu
eof
rep
lica
tion
.In
theo
reti
cal
rep
lica
tion
,th
ebe
hav
ior
was
not
obse
rved
,bu
tit
sab
sen
cew
asco
nsi
sten
tw
ith
the
theo
ryar
gued
abov
e.T
he
cost
stru
ctu
reat
the
Mor
nin
gN
ews
Are
veal
edan
opp
ortu
nit
ym
otiv
atio
nex
iste
d,b
ut
for
reas
ons
that
wer
eth
eore
tica
lly
con
sist
ent
wit
hth
ege
ner
alp
atte
rnob
serv
ed.
bP
erio
dba
sed
onre
leva
nt
exp
ansi
onof
thre
atp
erce
pti
onat
asi
te.
Nu
mbe
rsre
pre
sen
tco
mp
any
esti
mat
esfr
omth
est
art
ofye
ar1
ofth
est
ud
yp
erio
dto
the
end
ofye
ar3.
Observations at every primary research site con-firmed the negative relationship between threatperception and resource rigidity (see Table 4).
In Table 4, which summarizes these observa-tions, I have noted whether the data observed fol-lowed Yin’s pattern of replication (1994). When anobserved site replicated the pattern, I called thecase a literal replication. When a site did not followthe pattern, but for reasons that were not inconsis-tent with the proposition, I called the case a theo-retical replication. If sites had not followed thepattern for other reasons, the exceptions would alsohave been noted. Notice in Table 4 that seven of theeight sites showed behavior that was a literal rep-lication of the pattern: threat helped overcome re-source rigidity. The exception to this pattern wasthe Morning News A. At this national newspaper,the Internet actually helped to solve an existingorganizational problem. Thus, it was not discontin-uous to its operating requirements in the same waythat it was for other locally distributed newspaperorganizations. Whereas the other papers had hugeeconomies of local production and distribution, theMorning News A did not. According to the presi-dent and publisher:
This was a wonderful opportunity from the start. Ifyou are a national newspaper with a 3 percent pen-etration, all of a sudden you have an opportunity forvirtually no cost to distribute the product. . . . TheInternet creates huge opportunities to deliver prod-uct in areas that were uneconomical before. . . .Eighty percent of my costs are production and dis-tribution. Now all of a sudden I have a solution.
Also, because it had a limited classified product,this newspaper did not share the fear of cannibal-ization with many of the other sites in the sample.The Internet matched a set of sustaining needs forthe print newspaper business and was acceptedaccordingly. The key distinction was that the op-portunity was not discontinuous for the MorningNews A, as it was for the other research sites. Thus,the newspaper does not follow the pattern of rep-lication observed in the other sites, but for theoret-ically consistent reasons (Yin, 1994).
Threat Perception and Routine Rigidity
Although threat perception reduced resource ri-gidity, it increased routine rigidity. In keeping withprevious research, my findings show that threatperception produces three intermediate behaviorsthat amplify routine rigidity: contraction of author-ity, reduced experimentation, and focus on existingresources (Staw et al., 1981).
Contraction of authority. Data confirmed thatcorporate management asserted its control over de-cision making, withdrawing considerable authorityfrom operating divisions, in each newspaper adop-tion of online publication studied here. This con-traction of authority involved transferring morecontrol to corporate officers, such as the head ofbusiness development, the CEO, or a newly ap-pointed online director. At nearly every site I stud-ied, these individuals or groups assumed control ofstrategy. Table 5 gives examples of the observedcontraction of authority. Sometimes this shift wasaccomplished by imposing business plan templateson local site managers. According to the vice pres-ident of technology and operations at the Beacon A,
It was very centralized in the beginning, which wasvery uncharacteristic, because the culture is verymuch to let these guys run their own businesses. Wehad a basic business model for every site. We gavethem money. We told them they could hire people,but we told them exactly how to do it.
The contraction of authority also took the form ofcentralized decision approval. The Morning NewsCompany’s CEO personally directed online strategyat the flagship paper, and the head of businessdevelopment made strategic, financial, and hiringdecisions for all the other newspapers. Similarly,the Press A required the newspaper CEO to approveall senior hires. Sites had to follow detailed budgetsand adhere to strict marketing plans. The ventureswere allowed little autonomy in local planning.This restriction increased reliance on existing rou-tines because it limited the alternatives considered.Further, operating managers were less likely tochange corporate-imposed routines. As one operat-ing manager explained, “We felt like even thoughwe were all very focused on the online business,the corporate folks kind of had a plan and stuckwith it, even though we could see the failures oc-curring in our local market.” The effect of thisbehavior can be formally stated:
Proposition 2a. Perception of an imminentthreat leads to a contraction of authority thatamplifies routine rigidity.
Reduced experimentation. The contraction ofauthority had a feedback effect on the level of ex-perimentation in online strategy. For example, oneof the papers at the Beacon Company wanted toexperiment with other forms of revenue generation,but it felt constrained by the corporate strategy thatimposed a site template for sales strategy, businessmodels, and product plans. This reduced experi-mentation was not solely a function of corporatecontrol; the data show that the aggressive pace of
2005 749Gilbert
TA
BL
E5
Evi
den
ceof
Th
reat
-Mot
ivat
edB
ehav
iors
Lin
ked
toR
outi
ne
Rig
idit
y,19
96–9
8a
New
spap
er
Con
trac
tion
ofA
utho
rity
Red
uced
Expe
rim
enta
tion
Focu
son
Exis
ting
Res
ourc
es
Beh
avio
rEx
ampl
esB
ehav
ior
Exam
ples
Beh
avio
rEx
ampl
es
Bea
con
AL
iter
alre
pli
cati
onB
reak
ing
from
anea
rly
pol
icy
oflo
cal
dev
elop
men
t,co
rpor
ate-
dev
elop
edbu
sin
ess
pla
nte
mp
late
sfo
ral
lsi
tes,
incl
ud
ing
bud
gets
,an
dor
gan
izat
ion
char
ts
Lit
eral
rep
lica
tion
Loc
alm
arke
tex
per
imen
tsan
dva
riat
ion
dis
con
tin
ued
beca
use
each
site
was
forc
edto
foll
owco
rpor
ate
pla
n
Lit
eral
rep
lica
tion
In19
96,c
orp
orat
ebe
ings
trac
kin
gp
rin
tre
ader
ship
and
onli
ne
aud
ien
ceov
erla
ptr
acke
dm
onth
ly
Bea
con
BL
iter
alre
pli
cati
onP
rod
uct
tem
pla
te,i
ncl
ud
ing
con
ten
tca
tego
ries
,d
eter
min
edby
corp
orat
e
Lit
eral
rep
lica
tion
Non
stag
edd
eplo
ymen
tof
cap
ital
crea
ted
lock
-in
onin
itia
lco
rpor
ate
pla
n
Lit
eral
rep
lica
tion
Pro
du
ctd
evel
opm
ent
base
don
valu
eto
the
pri
nt
pro
du
ct
Pre
ssA
Lit
eral
rep
lica
tion
Wh
erea
sea
rlie
rin
volv
emen
tw
asm
inim
al,t
he
CE
Obe
com
esin
volv
edin
all
onli
ne
bud
get
app
rova
ld
ecis
ion
s
Lit
eral
rep
lica
tion
Des
pit
ein
crea
sin
glo
sses
,re
sou
rce
com
mit
men
tex
pan
ds
wit
hou
tp
ush
ing
onth
ed
rive
rsof
mis
sed
targ
ets
Lit
eral
rep
lica
tion
Tre
men
dou
sco
nce
rns
abou
tca
nn
ibal
izat
ion
,rea
der
ship
stu
die
sd
one
aski
ng
wh
eth
erp
eop
lew
ould
stop
read
ing
the
pap
er
Pre
ssB
Lit
eral
rep
lica
tion
Cor
por
ate
staf
fse
lect
init
ial
onli
ne
dir
ecto
ran
dea
rly
stra
tegy
Th
eore
tica
lre
pli
cati
onS
ome
loca
lex
per
imen
tati
onoc
curs
afte
rou
tsid
eon
lin
ed
irec
tor
ish
ired
;dif
fere
nt
con
ten
tst
rate
gies
,rev
enu
est
ream
test
edan
dch
ange
dba
sed
onm
arke
tte
sts
Th
eore
tica
lre
pli
cati
onF
ocu
sm
oves
away
from
the
new
spap
erto
the
busi
nes
sm
odel
and
cust
omer
requ
irem
ents
that
are
un
iqu
eto
onli
ne
med
ia
Exp
osit
orA
Lit
eral
rep
lica
tion
CE
Od
rive
sen
try
into
onli
ne,
taki
ng
over
pre
viou
sro
leof
loca
lp
ubl
ish
ers
Lit
eral
rep
lica
tion
On
lin
egi
ven
dir
ecti
ven
otto
lose
mon
ey,a
nd
fun
din
gfo
rex
per
imen
tsve
ryd
iffi
cult
Lit
eral
rep
lica
tion
Ad
vert
isin
gp
rod
uct
sbu
ilt
arou
nd
the
nee
ds
ofp
rin
tad
vert
iser
s
Exp
osit
orB
Lit
eral
rep
lica
tion
All
dec
isio
ns
rega
rdin
gcl
assi
fied
stra
tegy
gove
rned
byco
rpor
ate
Lit
eral
rep
lica
tion
Cor
por
ate
dri
ven
clas
sifi
edp
lan
enfo
rced
over
pre
viou
slo
cal
exp
erim
ents
Lit
eral
rep
lica
tion
Can
nib
aliz
atio
nsu
rvey
sco
nd
uct
ed
Mor
nin
gN
ews
AT
heo
reti
cal
rep
lica
tion
CE
Oru
ns
the
onli
ne
busi
nes
sat
flag
ship
pap
erfo
rex
ten
ded
per
iod
Th
eore
tica
lR
epli
cati
onE
xper
imen
tati
onoc
curr
edw
ith
inth
ecu
rren
tbu
sin
ess
mod
elan
dp
rin
tad
vert
iser
cust
omer
requ
irem
ents
Lit
eral
rep
lica
tion
All
pro
du
ctd
evel
opm
ent
exp
end
itu
res
buil
tar
oun
dim
pac
tst
ud
ies
onp
rin
tad
vert
iser
s
Mor
nin
gN
ews
BL
iter
alre
pli
cati
onC
orp
orat
ebu
sin
ess
dev
elop
men
td
irec
tor
crea
tes
site
stra
tegy
and
per
son
ally
app
rove
sal
lfu
nd
ing
Lit
eral
rep
lica
tion
Loc
alm
arke
tex
per
imen
tsan
dva
riat
ion
dis
con
tin
ued
,fo
rced
tofo
llow
corp
orat
e
Lit
eral
rep
lica
tion
Sel
lin
gst
rate
gybu
ilt
arou
nd
“up
-se
llin
g”to
exis
tin
gad
vert
iser
s
aT
hes
eca
tego
ries
are
base
don
Yin
’s(1
994)
tech
niq
ue
ofre
pli
cati
on.
Inth
eore
tica
lre
pli
cati
on,
the
beh
avio
rw
asn
otob
serv
ed,
but
its
abse
nce
was
con
sist
ent
wit
hth
eth
eory
argu
edab
ove.
Th
eM
orn
ing
New
sA
was
opp
ortu
nit
y,n
otth
reat
,m
otiv
ated
.H
owev
er,
the
opp
ortu
nit
yth
en
ewsp
aper
saw
was
rela
ted
toco
stsa
vin
gsin
the
core
busi
nes
s,n
otto
crea
tin
ga
new
busi
nes
s(s
eeT
able
4).T
her
efor
e,th
eex
per
imen
tati
onth
atd
idoc
cur
was
focu
sed
onth
eco
rebu
sin
ess.
Th
us,
the
case
vari
esfr
omth
ep
atte
rn,b
ut
for
theo
reti
call
yco
nsi
sten
tre
ason
s.T
he
Pre
ssB
did
not
foll
owth
eob
serv
edp
atte
rnfo
rre
ason
sd
iscu
ssed
subs
equ
entl
yin
the
text
.
resource commitment also made it more difficult tostep back and change behavior. The sample firmsexpanded resource commitments by more than 100percent per year during the period of acceleratedthreat perception. Recall that employment at theMorning News A went from 5 to 40 people in lessthan eight months. Expenses at the Beacon B dou-bled, then nearly doubled again in 1997 and 1998,contradicting original forecasts of profitability by1996 (see Table 5). Whereas initial disappoint-ments might have prompted managers to regroup,the perception of threat caused them to press aheadon the same misdirected paths. To paraphrase San-tayana’s (1905) definition of a fanatic, it was as ifthe organizations were doubling their speed uponlosing their direction. If an initial response iswrong, then expanding resources may only solidifythose initial tendencies. The Beacon A, for exam-ple, hired more than 40 people to implement anexpansion strategy that closely resembled that ofthe newspaper business model. Because the expan-sion occurred so quickly, the resources investedreinforced rather than reshaped established rou-tines of the parent. The effect of this behavior canbe formally stated:
Proposition 2b. Perception of an imminentthreat leads to a reduced level of experimenta-tion that amplifies routine rigidity.
Focus on existing resources. The fear of canni-balizing the core newspaper business promptedmanagers to focus on their existing resources,rather than consider new options presented by thenew technology (see Table 5). One manager at thePress A expressed fear that the online publicationwould drain revenues from the print publication:“Cannibalization was a huge concern for everyoneinitially. . . . We asked questions about readershipoverlap and whether they would stop reading thepaper.” Functional managers often retained over-sight for the online products. For example, thepresident of the Expositor A Internet site explainedthat “because the classified organization was soworried about defending the print classifieds busi-ness, that group held onto the online business.”Because they were focused on the existing busi-ness, they responded with routines that workedwell in that business. Development of the new tech-nology was often driven by the effect it would haveon the newspaper, despite the parallel growth of aseparate online ad market. The head of sales for thePress Company’s Internet group stated, “On the onehand I should go do whatever we need to do, but onthe other hand there is concern about the paper.”The effect of this behavior can be formally stated:
Proposition 2c. Perception of an imminentthreat leads to a focus on existing resourcesthat amplifies routine rigidity.
Table 6 summarizes evidence on the effects ofeach of these threat-motivated behaviors on routinerigidity. In almost every instance, threat-motivatedresponse led to aggressive replication of the news-paper product and business model. Seven of theeight research sites turned out a product that wasmerely an extension of the newspaper. Indeed,many of the sites republished more than 85 percentof their Web site content directly from the localnewspaper product. The publisher of the Beacon Aobserved this: “We learned [from early involve-ment with new media] that there wasn’t very muchappetite for an ‘electronic’ newspaper. . . . But thatis exactly what we did with the Internet.” Commonfeatures such as discussion boards, site-searchingtools, breaking news from third-party sources, com-munity forums, and other content features com-monly provided by the many new entrants werelargely absent from newspaper sites. Ironically, thetechnologies to develop these products werelargely available and relatively easy to deploy. Thechairman and CEO of the Beacon Companyreflected:
Where I think we missed the boat is that we saw it asan extension of the newspaper. In other words,something richer and deeper than the newspa-per. . . . Our Internet operations were really run bypeople who came out of the newsroom, so they wereeditors who tended to look at this more as anewspaper.
Most sites simply reproduced the newspaper.The online director at one site remarked, “Remem-ber that I had said to the CEO at the time that itmade absolutely no sense to replicate the newspa-per on the Internet. Then I saw the product and itwas just that.”
The rigidity was expressed not only in the prod-uct but also in the business model. I compared theincome statements of the eight research sitesagainst a panel set of five competing online entrantfirms. This analysis identified six categories of rev-enue associated with digital media that differedfrom those built around a print newspaper (e.g.,e-mail and interactive advertising). Whereas a com-parative set of entrants averaged more than fivenew revenue categories per site, most of the news-paper sites had only one (see Table 6). These newcategories accounted for more than 40 percent ofmany of the entrant firm revenue streams. As theCEO of the Expositor Company explained, “Wecouldn’t see any models that we were familiar with,nor any we knew how to make money with.”
2005 751Gilbert
The Notable Exception—Outside Influence andSeparate Structure at the Press B
The most notable exception to the pattern of rou-tine rigidity was the Press B. And although this sitewas similar to the others in most contextual factors,it was the only site that launched a structurallydifferentiated venture from the outset—a decisionthat was largely influenced by the CEO’s externalpersonal network. Like the rest of the sample, thePress B struggled with resource rigidity early on,and threat perception ultimately became the meansto overcome that constraint. Recall that in initialproposals managers had argued, “If we don’t can-nibalize ourselves, someone else will.” Similarly,the Press B management’s early thinking about theform of its Internet product did not significantlydiffer from the assumptions of the managers at theother newspapers in the sample. The Press B news-paper managers assumed that the product would bea “newspaper in electronic form.” But unlike theother incumbents, the Press B did not follow these
initial impulses to replicate the newspaper. Thisshift originated with the suggestion by the CEO thatsomeone from outside of the newspaper industrylook at the online business. The CEO’s view hadbeen shaped from the recommendation of a long-time friend and personal advisor who was based inSilicon Valley and had been observing some of thechanges created by the Internet. The Press B seniormanagement then launched a search for outsideadvice on strategy. A Silicon Valley business exec-utive with a background in new media was hired towrite the original business plan. That original plancalled for an organizational design with significantautonomy from the newspaper. The new venturewas subsequently set up as a wholly owned sub-sidiary and hired outside managers with new me-dia experience. The management team then re-cruited a separate sales force to price and sellonline ads. They also developed a separate brand tosignal that the product, although owned by theparent, would be distinct from the newspaper. Fi-
TABLE 6Routine Rigidity Associated with Threat-Motivated Behavior
NewspaperEarly
ProductPercent Print
ContentaNew Revenue
Categoriesb Examples
Beacon A Extension ofnewspaper
�75% 1 “We learned . . . that there wasn’t very much of an appetite foran ‘electronic’ newspaper . . . But that’s exactly what we didwith the Internet.” (publisher)
Beacon B Extension ofnewspaper
�85% 1 “Where I think we missed the boat is that we saw it as anextension of the newspaper, in other words, something richerand deeper than the newspaper.” (CEO)
Press A Extension ofnewspaper
�85% 3 “Remember that I had said to the CEO at the time that it madeabsolutely no sense to replicate the newspaper on theInternet. Then I saw the product and it was just that.” (CEO,Internet group)
Press B Multisourcedinteractivemedia
�50% 3 “We are really becoming a separate company from thenewspaper. I came from there. I love the paper, but we arenow a different group with a very different way of working.”(online editor)
Expositor A Extension ofnewspaper
�85% 1 “We couldn’t see any models that we were familiar with, norany we knew how to make money with.” (CEO)
Expositor B Extension ofnewspaper
�85% 1 “We failed to recognize the importance of tools such as search,but rather presented this in the layout of a printednewspaper.” (vice president, product development)
Morning News A Extension ofnewspaper
�90% 1 “Where we made our mistake was we missed the next wave ofopportunity. We could have said we want to be a nationalclassified source. We could have become different contentverticals. But we have done very little on content verticals.”(publisher)
Morning News B Extension ofnewspaper
�80% 1 “I don’t see this as that different than what the newspaperscurrently do; it is just another channel.” (vice president,marketing)
a Based on internal estimates at each site in 1998.b Based on 1998 comparisons of print newspaper income statement analysis and new entrant income statement analysis. The entrants
were selected from interviews with new media experts. The comparison sites included Citysearch, Monster.com, Yahoo, i-village, andCNET. Six categories of revenue were identified as being new for print newspapers: fee-based archival access, e-mail marketing, e-mail listrental, fee-based data analysis, behavioral targeting, and demographic targeting.
752 OctoberAcademy of Management Journal
nally, the Internet group was located in officesmore than a mile away from the print offices. Somenewspaper staff were allocated to the venture, andimportant links to the newsroom were preserved.But the locus of the new venture, both physicallyand operationally, was distinct.
Early on, the site evolved into a regional sourceof news and information that was distinctly differ-ent from the newspaper. More than 50 percent ofsite content originated from sources other than thenewspaper (see Table 6). Lead stories were differ-ent, and they rotated throughout the day; sectionsthat did not exist in print were added; and userswere provided with a host of tools that enabledthem to take advantage of digital media, includingtraffic Web cameras, searchable event databases,interactive discussion forums, and new forms ofcontent. Just how different the new product becamewas described by one online editor:
Page views from the newspaper are now barely morethan one-third of the available pages on our site. Weare really becoming a separate company from thenewspaper. I came from there. I love the paper, butwe are now a different group with a very differentway of working. They are one source of informa-tion—an important source. But we buy our contentfrom them like we buy it from anywhere else.
The site captured new and different categories ofrevenue, and it developed its own business model.Though its site was not as innovative as those ofcompeting nonnewspaper entrants, the Press B ev-idenced considerably less routine rigidity than thesites of other online newspapers in the sample (seeTable 6). Whereas, for example, most of the otherscaptured only one new category of revenue, thePress B captured three: fee-based archival access,e-mail marketing, and fee-based usage data analy-sis. Similarly, less than 50 percent of the Press B’ssite content was reused newspaper content.
Outsiders, Structural Differentiation, and ThreatPerception
Though all the research sites debated whether tostructurally differentiate their online ventures, adesire to leverage the assets of the print businessmotivated many to stay integrated with their parentnewspaper organizations. In the Beacon A’s origi-nal 1990 online business plan, the publisher wrote:“The power of the newspaper to provide thrust forthe new services can be harnessed only if itachieves deep levels of integration with the news-paper. Structuring the experiment as an enterpriseseparate from the newspaper would be crippling ifnot fatal.” The vice president of business develop-
ment at the Morning News A stated: “Our basicstrategy is an integrated strategy. . . . In the localinformation market, the newspaper has an advan-tage. To separate the online unit from the newspa-per is to give away a lot of that advantage.”
Nevertheless, four of the research sites in thesample did eventually differentiate the organiza-tional structure of their online ventures. The un-derlying motivation in each case was related tooutside influence on decision making. Two of thesites separated in response to the adamant de-mands of online employees hired from outside thenewspaper. In the case of the Press A, an influentialboard member from the technology industry helpedintroduce the company to a high-profile new mediaexpert who was eventually offered a job as theonline media director. His condition for joining theventure was that it be separated from the parentorganization. Thus, though the venture waslaunched from within the newspaper, it was sepa-rated from it in the summer of 1997. At the BeaconCompany the process was more evolutionary anddriven by interactions with outside partners. Forexample, an outside partner that marketed onlinedirectory services through the newspaper’s Website had asked to be allowed to build its own agencyto sell the online ads. Within a year, the salesagency had tripled the performance of the printsales reps, despite an identical headcount. Thenewspaper eventually purchased the agency, butkept it separate at the demand of the sales director.The agency kept adding online advertising catego-ries beyond the directory ads through 2000, bywhich time it was selling close to two-thirds of theonline ads for the Beacon A site.
Each site that separated its online venture fromthe parent organization cited the influence of out-side parties in the decision process. The process ofdifferentiation often occurred over several years,owing to the reticence of functional units in thenewspaper organization. As mentioned earlier, thePress A separated from the newspaper organizationin the summer of 1997. The Beacon A separated inthe spring of 1999, and the Beacon B followed inthe summer of 1999. To determine whether a ven-ture was structurally differentiated from its parentorganization, I performed analyses that includedunits’ own reported structural classifications aswell as other metrics, such as reporting lines andphysical locations (see Table 7). Note that in thefour sites that separated, corporate managers men-tioned outside influence, while managers whochose to remain integrated did not. This pattern ofobservation leads to a formal proposition:
2005 753Gilbert
TABLE 7Outside Influence and Structural Variation, 2000
NewspaperOutside Influence on
Structural ChoiceTiming ofChanges
DivisionalUnita, b
ReportingLinesc Newsd Salese
PhysicalLocationf Description
Beacon A Outside partner1997; outsideonline CEO, 1999
Spring 1999 D D H H D Set up as a differentiated unitin a different building fromthe parent with primaryreporting through the onlinegeneral manager.
Beacon B Outside advisor,1998
Summer 1999 D D H H H Set up as a differentiated unit,but shared responsibilitieswith some parent functionalstaff. Primary reportingresponsibility still throughonline general manager.
Press A Outside boardmember, 1997;outside onlineCEO, 1997
Summer 1997 D D H D D Set up as a differentiated unitin a different building fromthe parent with primaryreporting through the onlinegeneral manager. Allfunctional staff hiredseparately for online.
Press B CEO’s personaladvisor, 1994;outside businessplan consultant,1995; online CEO,1995
Spring 1995 D D H D D Set up as a differentiated unitin a different building fromthe parent with primaryreporting through the onlinegeneral manager. Allfunctional staff hiredseparately for online.
Expositor A No key outsideinfluencementioned
Remainedintegrated
I I I I H Kept integrated with the parent.Online general manager andsmall support staff hired andlocated in a separatebuilding, but all functionalstaff kept integrated withparent.
Expositor B No key outsideinfluencementioned
Remainedintegrated
I I I I I Kept integrated with the parentorganization. Online generalmanager hired, but allfunctional responsibilitiescoordinated through parent.
Morning NewsA
No key outsideinfluencementioned
Remainedintegrated
H I H I I Structure was hybrid.Newsroom was clearlyhybrid, with some onlinecontent going back to thenewspaper, but mostfunctional reporting wentthrough the parent.
Morning NewsB
No key outsideinfluencementioned
Remainedintegrated
I I H H I Kept integrated with the parentorganization. Online generalmanager hired, but allfunctional responsibilitiescoordinated through parent.
a “I” represents an integrated structure; “H,” a hybrid structure; and “D,” a differentiated structure.b Based on management’s self-description.c Based on primary reporting responsibility of functional staff, e.g., online sales manager reporting to print sales manager (I), online GM
(D), or both (H).d Based on primary responsibility for content development: print newsroom (I), separate online newsroom (D), hybrid (H).e Based on primary responsibility for selling online ads: print staff (I), independent sales reps (D), hybrid (H).f Based on the location of the online venture: within the parent organization (I), physically separated from the parent (D), or a
hybrid (H).
754 OctoberAcademy of Management Journal
Proposition 3. Involving outside influencewhen deciding how to respond to discontinu-ous change will increase the likelihood thatmanagers will structurally differentiate a newventure from its parent organization.
Even before sites differentiated structurally,some managers had begun to perceive that the In-ternet, while perhaps still representing a threat,could become a source of new opportunity for theirnewspapers. The president of the Press Companyexplained:
We were worried about the Web in that it wouldalter the way in which people would get informa-tion, but it was not purely defensive. We hadlaunched into entertainment years ago as a defen-sive move. It eventually became a new source ofgrowth for us. Many of the threats eventually be-come opportunities. The Internet may be the sameway.
This dualistic view was, however, hard to main-tain in an environment in which operating respon-sibilities for the newspaper predominated. Thus,the data reveal that one reason for creating a differ-entiated structure is to decouple the motivation atthe parent from the motivation at the venture. Asthe former head of new media at the Beacon Com-pany reflected, “I didn’t focus people on the threat,especially those managing the new business.Where I did emphasize the threat was in workingwith the print folks to get them off their butts and inarguing for resources.” Data from the other sitesconfirm that structure affects how managers per-ceive their motivation.
The data show that the differentiated structureshelped to create environments where motivationcould be built entirely around the separate oppor-tunity that existed for the online model. “When wesimply changed our name from the newspapername to ‘the city.com’ . . . it changed people’s ex-pectations of what would be on the site. This, inturn, changed how people in our online organiza-tion viewed who they were and what they wereproducing.” The new president of the Beacon Com-pany’s new media group observed, “Now that weare separate, we own the opportunity in a way wenever did when we were still with the newspaper.”Even as the concept of the opportunity was chang-ing in the differentiated units, the parent organiza-tions remained focused on the threat to the corebusiness; discussions there centered on cannibal-ization and the inevitable path of digital media.“This,” cautioned one CEO, “could be the death ofour entire franchise.” That threat perception re-mained high in the newspaper organizations was acritical factor in overcoming resource rigidity, but
opportunity perception was given a chance to de-velop simultaneously in the ventures. Statedformally:
Proposition 4. Structural differentiation canhelp decouple threat perception in a parentfrom an opportunity perception in a newventure.
In the four sites that did not differentiate theirventures from the parent organizations, managerscontinued to be preoccupied with the threat to theirorganizations. The vice president of the Expositor Binsisted, “We continue to see this as a way to pro-tect classifieds, and that if we don’t do it someoneelse will.” Comparing the sites that separated withthose that remained integrated reveals that oppor-tunity perception emerged only where there wasstructural differentiation. Table 8 summarizes com-parative data for Proposition 4.
Not only did outside influence and structuraldifferentiation help to decouple the cognitive per-ceptions in the newspaper organizations from thosein the online venture, but also all three of thesevariables were correlated with relaxed routine ri-gidities. This relaxation was driven largely by theeffect of each of these variables on the three inter-mediate behaviors that increase routine rigidity.For example, structural autonomy lowers the ten-dency for a parent organization to assert authorityover a new venture. Structural autonomy expandedthe ability of venture management to run local ex-periments that would not have been possible in aworld of business model and product templates.Outside influence also helped expand the alterna-tives considered in the new ventures. One manifes-tation of this was the previously noted impact out-siders had on the choices of structure. But outsideinfluence also expanded the product and businessmodel ideas in these ventures. The external partnerat the Beacon A proposed pricing and product cat-egories that the newspaper had not thought of. As aWeb site editor who came from the newspapercommented about the online director who wasbrought in from outside: “He is constantly seeingdigital media in different ways than I am used to orappreciate. At first, it bothered me, but now that Isee it working, I increasingly endorse the input.” Inaddition, structural autonomy and a renewed op-portunity mind-set freed venture managers from anewspaper focus, because their responsibilities inthe parent newspapers were no longer immediate.
The four sites that differentiated structurally alsosubstantially increased their innovation. Whereasthe integrated sites continued to derive as much as90 percent of their Web site content from theirsponsoring newspapers, all of the differentiated
2005 755Gilbert
TA
BL
E8
Dec
reas
ing
Rou
tin
eR
igid
ity,
2000
New
spap
erO
uts
ide
Infl
uen
ceS
tru
ctu
reM
anag
eria
lF
ram
ing
Exa
mp
les
Con
ten
tfr
omP
rin
ta
Loc
alM
ark
etP
enet
rati
onS
core
b
New
Rev
enu
eC
ateg
orie
sc
Bea
con
AO
uts
ide
par
tner
,19
97;
outs
ide
onli
ne
CE
O,
1999
Dif
fere
nti
ated
Op
por
tun
ity
emer
ges
“Now
that
we
are
sep
arat
e,w
eow
nth
eop
por
tun
ity
ina
way
we
nev
erd
idw
hen
we
wer
est
ill
wit
hth
en
ewsp
aper
.”(n
ewh
ead
ofn
ewm
edia
)
45%
1.9
4
Bea
con
BO
uts
ide
advi
sor,
1998
Dif
fere
nti
ated
Op
por
tun
ity
emer
ges
“We
wer
eal
lse
tto
let
peo
ple
buy
ads
onli
ne.
...T
he
pap
ers
did
n’t
wan
tto
buy
into
it.
Th
isis
one
area
wh
ere
we
wil
ld
obe
tter
asa
sep
arat
eco
mp
any.
”(v
ice
pre
sid
ent,
tech
nol
ogy
and
oper
atio
ns)
50%
1.4
4
Pre
ssA
Ou
tsid
ebo
ard
mem
ber,
1997
;ou
tsid
eon
lin
eC
EO
,19
97
Dif
fere
nti
ated
Op
por
tun
ity
emer
ges
“Th
eyju
stsi
tth
ere
and
mak
eu
sfi
gure
itou
t.T
hey
don
’tm
ake
the
dec
isio
ns
for
us.
”(v
ice
pre
sid
ent,
sale
s)
50%
c2.
16
Pre
ssB
CE
O’s
per
son
alad
viso
r,19
94;
outs
ide
busi
nes
sp
lan
con
sult
ant,
1995
;on
lin
eC
EO
,19
95
Dif
fere
nti
ated
Op
por
tun
ity
emer
ges
“Pag
evi
ews
from
the
new
spap
erar
en
owba
rely
mor
eth
anon
e-th
ird
ofth
eav
aila
ble
pag
eson
our
site
....
Th
en
ewsp
aper
ison
eso
urc
eof
info
rmat
ion
,an
imp
orta
nt
sou
rce.
Bu
tw
ebu
you
rco
nte
nt
from
them
like
we
buy
itfr
oman
ywh
ere
else
.”(o
nli
ne
edit
or)
35%
1.7
5
Exp
osit
orA
No
key
outs
ide
infl
uen
cem
enti
oned
Inte
grat
edT
hre
atp
ersi
sts
“Fu
nct
ion
alre
por
tin
gre
lati
onsh
ips
are
extr
emel
yti
me-
con
sum
ing.
It’s
not
just
that
the
grou
ps
thin
kli
keth
en
ewsp
aper
.It
take
sa
lot
lon
ger
tom
ake
coll
ecti
ved
ecis
ion
s.”
(ear
lyon
lin
ep
resi
den
t)
75%
1.1
2
Exp
osit
orB
No
key
outs
ide
infl
uen
cem
enti
oned
Inte
grat
edT
hre
atp
ersi
sts
“We
con
tin
ue
tose
eth
isas
aw
ayto
pro
tect
clas
sifi
eds,
and
that
ifw
ed
on’t
do
itso
meo
ne
else
wil
l.”
(vic
ep
resi
den
t,p
rod
uct
dev
elop
men
t)
70%
0.9
3
Mor
nin
gN
ews
AN
oke
you
tsid
ein
flu
ence
men
tion
edIn
tegr
ated
Th
reat
per
sist
s“O
ur
basi
cst
rate
gyis
anin
tegr
ated
stra
tegy
...
Inth
elo
cal
info
rmat
ion
mar
ket
the
new
spap
erh
asan
adva
nta
ge.
To
sep
arat
eth
eon
lin
eu
nit
from
the
new
spap
eris
togi
veaw
aya
lot
ofth
atad
van
tage
.”(v
ice
pre
sid
ent,
busi
nes
sd
evel
opm
ent)
90%
1.6
2
Mor
nin
gN
ews
BN
oke
you
tsid
ein
flu
ence
men
tion
edIn
tegr
ated
Th
reat
per
sist
sD
iscu
ssio
ns
wit
hm
anag
emen
tre
mai
nce
nte
red
ond
efen
din
gcl
assi
fied
pro
du
cts.
75%
0.5
2
aB
ased
onin
tern
ales
tim
ates
and
inco
me
stat
emen
tan
alys
isat
each
site
.b
Rat
ioof
mon
thly
Web
site
use
rsto
dai
lyn
ewsp
aper
read
ers.
Dat
aco
llec
ted
from
new
spap
ersi
tes
and
chec
ked
agai
nst
Med
iaM
etri
xan
dN
iels
onN
etR
atin
gsas
wel
las
circ
ula
tion
dat
afr
omth
eA
ud
ited
Bu
reau
ofC
ircu
lati
on.
cB
ased
on19
98co
mp
aris
ons
toen
tran
tin
com
est
atem
ents
.E
ntr
ants
wer
ese
lect
edby
new
med
iaex
per
tsan
din
clu
ded
Cit
ysea
rch
,M
onst
er.c
om,
Yah
oo,
i-vi
llag
e,an
dC
NE
T.
Six
cate
gori
esof
reve
nu
ew
ere
iden
tifi
edas
new
toth
en
ewsp
aper
s:(1
)fe
e-ba
sed
arch
ival
acce
ss,(
2)e-
mai
lm
arke
tin
g,(3
)e-
mai
lli
stre
nta
l,(4
)fe
e-ba
sed
dat
aan
alys
is,(
5)be
hav
iora
lta
rget
ing,
and
(6)
dem
ogra
ph
icta
rget
ing.
sites borrowed only 50 percent or less of their con-tent from their sponsoring newspapers (see Table8). This evolution had a positive effect on marketadoption of the new products; sites that had sepa-rated and shifted their emphasis to the emergingopportunity enjoyed much higher local marketpenetration than the other sites in the sample. Icreated a local market penetration score on thebasis of public data, calculating the ratio ofmonthly Web site users to daily newspaper readers;A score of 1.0 implied that a Web site’s monthlyusers were equal to the newspaper’s daily readers;anything above that amount implied more users permonth relative to daily readership. Whereas thedifferentiated sites averaged a score of nearly 1.8,the integrated sites averaged barely 1.0 (see Table8). The separated sites also introduced more inno-vation to their underlying business models, averag-ing close to five new categories of revenue, com-pared with just more than two new categories in theintegrated sites (see Table 8). Both the autonomy tofocus on the separate business and the perceptionof an independent opportunity seemed to facilitategreater divergence from the traditional routines ofthe core business. Thus, a final proposition can bestated:
Proposition 5. Outside influence, structuraldifferentiation, and opportunity framing com-bine to relax routine rigidity in a new venture.
Data confirming this observation were also con-sistent across research sites (see Table 8).
DISCUSSION
Differentiating the Structure of Inertia
In the introduction to this article, I noted theapparent contradiction in the literature regardingthe impact of threat perception on organizationalinertia. Whereas a number of scholars have ob-served threat to be a catalyst that enables organiza-tions to overcome inertia (Barr & Huff, 1997; Cyert& March, 1991; Lant et al., 1992), others have foundevidence that threat actually increases inertia (Dut-ton & Jackson, 1987; Staw, Sandelands, & Dutton,1981). Acknowledging conflicting observations inrelated research on risk, Sitkin and Pablo (1992)hinted that the phenomenon itself might be under-specified. They suggested that part of the confusionstems from viewing outcomes along a single deter-minant of behavior and argued for a model basedon a more complex set of determinants. Scholarshave also observed the need for a more theoreti-cally complex view of inertia in research on re-source commitment and organizational change. For
example, Johnson (1988) showed that inertia iscomprised of both a motivational determinant anda procedural determinant. Using data from the re-sponse of a clothing retailer who was threatenedwith industry change, Johnson showed investmentcan be highly motivated, but also deployed rigidlythrough traditional business routines. Faced withthe performance decline of its core business in“down-market niche” men’s wear, the retailer ini-tially sought to diversify into new markets. But itthen reapplied its existing business model, ratherthan adapting or repositioning its main business.Threat motivated resource commitment, but rou-tines remained locked on the traditional businessmodel.
The current study identifies these unique deter-minants of inertia as resource and routine rigidity.Both constrain adaptation, but they have very dif-ferent underlying causal mechanisms. Data showthat resource rigidity stems from an unwillingnessto invest, while routine rigidity stems from an in-ability to change the patterns and logic that under-lie those investments. The first relates to the moti-vation to respond, the second to the structure ofthat response. Recall that the issues regarding re-source rigidity in the data dealt with allocatingfinancial resources or management time to projectsthat supported the venture. These inertial forceswere very different from those related to the rou-tines and logics traditionally used to develop newscontent and run the newspaper business. Both thecurrent study and the previous research by Johnson(1988) reveal that not only are these determinantsof inertia different, but also that they can moveindependently. By analogy, resource rigidity isconcerned with movement along a line, while rou-tine rigidity deals with the trajectory of the line. Amanager could invest aggressively and still fail toadapt underlying routines. Sull (1999) called sucha pattern of behavior “active inertia.”
Ironically, a closer examination of the previousliterature reveals that earlier measures of inertiaactually align with the observed categorizations ofresource and routine rigidity, even if the definitionslacked the underlying specificity. Specifically, re-examination of previous research shows that threatdecreases resource rigidity but increases routinerigidity in a predictable, repeated pattern across theprevious studies. For example, the literature thatviews threat as a catalyst to response typically mea-sures behavior as a willingness to commit re-sources. Kahneman and Tversky’s (1984) study ofrisk showed that individuals are more often willingto commit financial resources when they perceivethe issues of concern as being in the domain of lossrather than in the domain of gain. Mittal and Ross
2005 757Gilbert
(1998) observed that individuals who had beengiven threat scenarios exhibited a significantlygreater willingness to spend than participants whohad been given opportunity scenarios. Similarly,the literature on strategic change suggests thatthreat-driven response unlocks resources for in-vesting in new strategic initiatives (Lant et al.,1992).
This is very different from the view that threat isa constraint, as is seen in research that measureschanges in organizational process and operationallogic. For example, Herman (1963) measured com-munication and information patterns within an or-ganization. Staw, Sandelands, and Dutton (1981)and Nutt (1984) examined changes in the underly-ing business logic of organizations. Thus, variablesthat measure a willingness to invest financial orattention-based resources (Lant et al., 1992) havebeen confused with variables that measure changein “dominant logic” (Prahalad & Bettis, 1986) orchange in operating routines (Staw et al., 1981).The reason these subtypes of inertia move in dif-ferent directions is that their underlying causalmechanisms differ, a factor that ironically showsup in these previous measurement efforts in therelevant research.
Toward an Interpretive Model of Response toDiscontinuous Change
Failure to recognize the difference between re-source rigidity and routine rigidity risks more thanjust contradictory results. When researchers do notrecognize this distinction, they may also fail toobserve important interactions between the twosubtypes of inertia. For example, research on over-coming inertia should focus on response motiva-tion and resource commitment. And yet, whilevariance perceptions and performance declines of-ten decrease motivational constraints (Barr & Huff,1997; Cyert & March, 1963), they can simulta-neously increase constraints on the underlyinglogic of an organization’s operating routines (John-son, 1988; Sull, 1999). By differentiating types ofinertia, one can focus on their unique implicationsfor organizational response. I used my observationsfrom the study, formalized above as Propositions1a–4, to develop a longitudinal model that mapshow the perception of discontinuous change im-pacts overall inertia. Figure 1 contains this model.
Propositions 1a and 1b support the notion thatthe perception of an imminent threat can unlockresource rigidity (Cyert & March, 1963; Lant et al.,1992). Note that threat perception leads to behaviorthat is different from the behavior observed byChristensen and Bower (1996). Their findings sug-
gest that when discontinuities are led by noncorecustomers, established firms do not allocate re-sources to a new business or technology. And yetthe current study shows that threat perception canlead to intense resource commitment, even in theabsence of core customer demand—recall the rapidgrowth in expenditures observed in the presentsample, which occurred despite the virtual lack ofoverlap between print and online customers. Asone sales representative noted at the Beacon A,“Look, it didn’t make any sense for us to try to sellthis stuff, but we began to feel that if we didn’twork on it, it might come back to haunt us.” Threatwas the catalyst to overcoming resource rigidity.
The data also reveal that the ability to overcomeone type of inertia appears to increase problemswith the other. The observed increase in routinerigidity stems from three intermediate behaviorsthat arise from threat-induced response—contrac-tion of authority, reduced experimentation, and fo-cus on existing resources (Propositions 2a, 2b, 2c).These behaviors were shown to be self-reinforcing.For example, threat perception led to a rigid focuson the existing business. This focus on existingresources was hardened by the aggressive pace ofcommitment, which created lock-in effects and re-duced the ability to experiment. The aggressivedeployment of resources required increasing corpo-rate oversight. And corporate leaders’ contractionof authority further reinforced the focus on theestablished business at the expense of the new op-portunity (see Figure 1). These self-reinforcing be-haviors led to intense routine rigidity, causingmanagers to adhere more closely to familiar rou-tines and behavioral patterns.
Observations on Structural Differentiation:Sources and Implications
Previous research has repeatedly demonstrated alink between structural autonomy and innovation(Christensen, 1997; Tushman & Anderson, 1986;Tushman & O’Reilly, 1996). The current studyhelps expand understanding of the forces that leadto the decision to structurally separate a new ven-ture. The data show that outside influence shapesthe choice to structurally differentiate (Proposition3) and that structural differentiation cultivates anenvironment in which managers are more likely toturn their attention to the independent opportunityassociated with a discontinuity (Proposition 4).And while outside influence does appear to belinked to the decision to separate a venture from itsparent, there are still questions as to why some ofthe newspaper companies studied here incorpo-rated such external influence, while others did not.
758 OctoberAcademy of Management Journal
FIG
UR
E1
An
Inte
rpre
tive
Mod
elof
Iner
tia
inR
esp
onse
toD
isco
nti
nu
ous
Ch
ange
a,
b
aP
rim
ary
ques
tion
vari
able
sar
esh
aded
gray
.
Note that the data do not seem to indicate that thesites that failed to harness outside influence hadaccess to outsiders and then ignored them. In theinterviews with senior managers regarding thechoice of organizational structure, only those whostructurally separated their ventures mentionedoutside influence. I asked the managers at the foursites that remained integrated, “What outsiders didyou involve in your decision?” None of them men-tioned outside individuals or organizations in theirresponses. Thus, the first relevant question appearsto be how outsiders entered firm and individualnetworks. On this point the data are not entirelyconclusive, but all of the companies that eventuallydetached their online ventures seemed to drawtheir outside influence from external networks: anoutside friend of the CEO (the Press B); a boardmember and new hire (the Press A); an apparentlyserendipitous set of comparative successes withoutside partners (the Beacon A); and an outsideadvisor (the Beacon B). That the reach of formaland informal networks would influence the firms’internal decision making is consistent with net-work theories and team composition studies (Gelet-kanycz & Hambrick, 1997; Podolny, 2001; Stuart &Podolny, 1996;). This observation leads to ques-tions regarding how such nodes in a firm’s externalnetwork develop and how they influence the firm.Influence may be a function of status (as indicatedby the role of one CEO’s personal network) or of theintensity of an interaction (as in the case of theexternal partner of the Beacon A). Given the origi-nal research questions and the design of this study,these questions cannot be immediately answered.Nonetheless, the observed link between outside in-fluence and structural choice should help sharpensubsequent research on the sources of and mecha-nisms leading to structural autonomy.
The study also provides a more refined view intothe mechanisms by which structural autonomyhelps relax routine rigidity. The data confirm thatoutside influence, structural independence, andopportunity orientation combine to relax routinerigidity and encourage innovation (Proposition 5).Conversely, when the companies studied here didnot access outside influence and remained inte-grated and focused on the threat to the parent organ-ization, the rigidity was perpetuated. Again, note thatthe role of structural autonomy is consistent withexisting structural arguments regarding innovation(Christensen, 1997; Tushman & O’Reilly, 1996). Butwhat makes the observations in this study unique ishow structure was seen to be the mechanism thatdecouples resource and routine rigidity. The datashow that structural autonomy allows threat and op-portunity cognition to have different impacts on dif-
ferent parts of an organization simultaneously—threat framing overcomes resource rigidity in theparent, while opportunity framing eases routine ri-gidity in the autonomous venture. Structure’s decou-pling role further reinforces the key contribution ofthe study: the recategorization of inertia into resourceand routine rigidity.
Alternative explanations. Taken together, theseobservations suggest that the conflicting findings inthe literature regarding threat’s effect on inertia canbe explained by the failure to differentiate betweenresource and routine rigidity. An alternative expla-nation might simply be that the competing streamsof findings are referring to different levels of threatperception. For example, the research describingthreat perception as a catalyst might have exam-ined only moderate levels of threat, while the stud-ies that describe threat as a source of inertia mighthave been looking only at crisis situations. If so,previous findings of a U-shaped relationship be-tween stress and inertia might explain the observeddifferences—moderate levels of stress decrease in-ertia, but extreme levels increase inertia (Yerkes &Dodson, 1908). However, one can derive reasons torule out this alternative explanation both from theliterature and from the data in the current study.First, the literature on threat as a catalyst containsexamples of extreme crisis (Mintzberg, Raising-hani, & Theoret, 1976; Papadakis, Kaloghirou, &Iatrelli, 1999). Moreover, the empirical literature onthreat rigidity does not delineate levels of threat(Dutton & Jackson, 1987; Jackson & Dutton, 1988).More importantly, a U-shaped relationship be-tween threat and inertia cannot explain the behav-ior of different types of inertia. The data in thisstudy show settings of intense resource commit-ment that accompany low levels of routine adapta-tion. Further, if a U-shaped relationship were inplay, one wouldn’t expect variance in behaviorwhen threat levels remain constant. The data in thestudy show that threat perceptions emerged be-tween 1996 and 1998 and remained high in all theprint organizations in my sample throughout theperiod of the study, both the innovative and non-innovative sites. Recall that even in the late periodsat the Beacon A, for example, the newspaper CEOcontinued to say, “This could be the death of ourentire franchise.” Thus, it appears more likely thevariance in routine rigidity can be accuratelylinked to the changes in structure that enable firmsto decouple the effects of cognitive framing be-tween resource and routine rigidity.
Limitations and future research. Like anymodel intended to capture the complexity of anorganizational response process, the model devel-oped here has elements that need further explora-
760 OctoberAcademy of Management Journal
tion. For example, the study showed that the ob-served threat perception was largely a response toexternal analysts’ forecasts of the demise of thenewspaper industry. The question remainswhether threat-based motivation could have beentriggered without this external pressure. A secondlimitation in the study is that threat’s link to rigidbehavior emerges in part because changes are dis-continuous with a firm’s traditional routines.While these discontinuous settings helped specifythe differences between resource and routine rigid-ity, they do impose boundary conditions on someof the findings regarding threat’s link to routinerigidity. For example, the threat-induced behaviorsof contracted authority, reduced experimentation,and focus on existing resources are problematic indiscontinuous settings because they reinforce afirm’s existing routines. When routines do not re-quire this nonlinear adaptation, threat-induced re-sponse is not as likely to be as maladaptive.
Another limitation, which was noted earlier, isthat the source of outside influence appears to belinked to firm and CEO networks. But that obser-vation is only indicative, not conclusive, and needsto be tested more formally. Furthermore, questionsremain as to how a node in a network might bemore or less influential on a firm’s internal deci-sion process. I have presented potential mecha-nisms, such as status and intensity, that might beexplored. Research should also look at whether thedesign of these networks is explicit or serendipi-tous. Again, the evidence from the current study isnot conclusive. Finally, the current study reveals atight correlation between outside influence, struc-tural differentiation, and opportunity framing.Each of these variables is also correlated with in-novation. The question remains as to whether thesevariables can vary independently.
CONCLUSIONS
This study began with an effort to unwind thestructure of inertia. I was able to show that thesubcategories of resource and routine rigidity arediscrete and have different causal mechanisms. Us-ing this recategorization, researchers can resolve aninconsistency in the literature concerning whetherthreat perception is a catalyst or a constraint todiscontinuous change. In building an interpretivemodel of response, I saw how threat perceptionreleases constraints on resource rigidity while am-plifying problems with routine rigidity. Further,although previous studies have shown the benefitsof structural autonomy for innovation, the currentresearch broadens knowledge of what leads firms tostructurally differentiate by making links to outside
influence and external networks. Moreover, thebenefit of structural autonomy is more than simplyproviding a separate setting for innovation. Struc-tural autonomy helps decouple the effects of cog-nition on different types of inertia—separate struc-ture allows threat perception to overcome resourcerigidity in a parent company, while opportunityperception relaxes routine rigidity in a new ven-ture. Distinguishing between resource and routinerigidity not only helps explain response to discon-tinuous change, but also opens up future researchexploration of the differences and interactions be-tween these categories of inertia.
These findings should also have relevance to man-agement practice. Managers can draw on the powerthreat has as a catalyst for commitment. And yet, thisstudy shows that response to discontinuous changerequires more than just the commitment of resources.The underlying organizational routines that use thoseresources must also adapt when change is discontin-uous. These findings might encourage managers todraw more heavily on their external networks to in-volve managers with experience outside of an exist-ing business. Managers might also structurally sepa-rate their new ventures to restore opportunity framesamong venture managers while maintaining threatframing in the parent organization.
In conclusion, the data show that threat percep-tion is a powerful interpretive force that affects firmresponse to discontinuous change. And yet thisinfluence is very different when one considers twodistinct types of inertia: resource and routine rigid-ity. Recognizing these distinctions has been shownto be significant for two fundamental reasons. First,the literature rarely recognizes these categories,and by simply specifying their characteristics onecan better describe the underlying phenomenon ofinertia. Second, under conditions of discontinuouschange, not only are these types of inertia different,but also, the mechanisms for overcoming one typecan amplify problems with the other. Further re-search needs to be done on how and why somefirms are more likely than others to structurallydecouple resource and routine rigidity, but I hopethat these initial findings will open new paths ofinquiry and inform future research on inertia andorganizational change.
REFERENCES
Arrow, K. J. 1962. Economic welfare and the allocation ofresources for inventions. In R. Nelson (Ed.), The rateand direction of inventive activity: Economic andsocial factors. 609–625. Princeton, NJ: PrincetonUniversity Press.
Barr, P. S., & Huff, A. S. 1997. Seeing isn’t believing:
2005 761Gilbert
Understanding diversity in the timing of strategicresponse. Journal of Management Studies, 34: 337–370.
Benner, M. J., & Tushman, M. L. 2001. Exploitation,exploration, and process management: The produc-tivity dilemma revisited. Academy of ManagementReview, 28: 238–256.
Bower, J. L. 1970. Managing the resource allocationprocess. Boston: Harvard Business School Press.
Burgelman, R. 1994. Fading memories: A process theoryof strategic business exit in dynamic environments:Administrative Science Quarterly, 39: 24–56.
Christensen, C. M. 1997. The innovator’s dilemma. Bos-ton: Harvard Business School Press.
Christensen, C. M., & Bower, J. L. 1996. Customer power,strategic investment, and the failure of leading firms.Strategic Management Journal, 17: 197–218.
Christensen, C. M., & Rosenbloom, R. S. 1995. Explainingthe attacker’s advantage: Technological paradigms,organizational dynamics and the value network. Re-search Policy, 24: 233–257.
Cyert, R. M., & March, J. G. 1963. A behavioral theory ofthe firm. New York: Prentice-Hall.
Dutton, J. 1992. The making of organizational opportuni-ties—An interpretive pathway to organizationalchange. In B. M. Staw & L. L. Cummings (Eds.),Research in organizational behavior, vol. 14: 195–226. Greenwich, CT: JAI Press.
Dutton, J., & Jackson, S. 1987. Categorizing strategic is-sues: Links to organizational action. Academy ofManagement Review, 12: 76–90.
Eisenhardt, K. M. 1989a. Building theories from casestudy research. Academy of Management Review,14: 532–550.
Eisenhardt, K. M. 1989b. Making fast strategic decisionsin high-velocity environments. Academy of Man-agement Journal, 32: 543–576.
Feldman, M. S., & Pentland, B. T. 2003. Reconceptualiz-ing organizational routines as a source of flexibilityand change. Administrative Science Quarterly, 48:94–118.
Geletkanycz, M. A., & Hambrick, D. C. 1997. The externalties of top executives: Implications for strategicchoice and performance. Administrative ScienceQuarterly, 42: 654–681.
Gilbert, R. J., & Newberry, D. M. 1982. Preemptive pat-enting and the persistence of monopoly. AmericanEconomic Review, 72: 514–526.
Glaser, B., & Strauss, A. L. 1967. The discovery ofgrounded theory. Chicago: Aldine.
Hannan, M., & Freeman, J. 1977. The population ecologyof organizations. American Journal of Sociology,82: 929–964.
Hartman, S. J., & Nelson, B. H. 1996. Group decision in
the negative domain. Group and OrganizationManagement, 21: 146–162.
Henderson, R. 1993. Under-investment and incompe-tence as responses to radical innovation: Evidencefrom the photolithographic alignment equipment in-dustry. RAND Journal of Economics, 24: 248–270.
Henderson, R., & Clark, K. 1990. Architectural innova-tion: The reconfiguration of existing product tech-nologies and the failure of existing firms. Adminis-trative Science Quarterly, 35: 9–30.
Hermann, C. F. 1963. Some consequences of crisis whichlimit the viability of organization. AdministrativeScience Quarterly, 8: 61–82.
Huff, J. S., Huff, A. S., & Thomas, H. 1992. Strategicrenewal and the interaction of cumulative stress andinertia. Strategic Management Journal, 13: 55–75.
Jackson, S. E., & Dutton, J. E. 1988. Discerning threats andopportunities. Administrative Science Quarterly,33: 370–387.
Johnson, G. 1988. Rethinking incrementalism. StrategicManagement Journal, 9: 75–91.
Kahneman, D., & Tversky, A. 1984. Choice, values, andframes. American Psychologist, 39: 341–350.
Lant, T. K, Milliken, F. J., & Batra, B. 1992. The role ofmanagerial learning and interpretation in strategicreorientation. Strategic Management Journal, 13:585–608.
Lee, T. W. 1999. Using qualitative methods to organizeresearch. Newbury Park, CA: Sage.
Lee, T. W., Mitchell, T. R., & Sablynski, C. J. 1999. Qual-itative research in organizational and vocational psy-chology. Journal of Vocational Behavior, 55: 161–187.
Leonard-Barton, D. 1992. Core capabilities and core ri-gidities: A paradox in managing new product devel-opment. Strategic Management Journal, 13: 111–125.
Levinthal, D. 1992. Surviving Schumpeterian environ-ments: An evolutionary perspective. Industrial andCorporate Change, 1: 427–443.
March, J. G. 1991. Exploration and exploitation in organ-izational learning. Organization Science, 2(1): 71–87.
March, J. G., & Simon, H. A. 1958. Organizations. NewYork: Wiley.
Miles, M., & Huberman, A. M. 1984. Qualitative dataanalysis. Newbury Park, CA: Sage.
Miller, D., & Friesen, P. H. 1980. Momentum and revo-lution in organizational adaptation. Academy ofManagement Journal, 23: 591–614.
Mintzberg, H., Raisinghani, D., & Theoret, A. 1976. Thestructure of “unstructured” decision processes. Ad-ministrative Science Quarterly, 21: 246–275.
Mittal, V., & Ross, W. T. 1998. The impact of positive and
762 OctoberAcademy of Management Journal
negative affect and issue framing on issue interpre-tation and risk taking. Organizational Behavior andHuman Development, 76: 298–324.
Nelson, R., & Winter, S. 1982. An evolutionary theory ofeconomic change. Cambridge, MA: Harvard Univer-sity Press.
Noda, T., & Bower, J. L. 1996. Strategy making as iteratedprocesses of resource allocation. Strategic Manage-ment Journal, 17: 169–192.
Nutt, P. C. 1984. Types of organization decision pro-cesses. Administrative Science Quarterly, 29: 414–451.
Ocasio, W. 1999. Towards an attention based view of thefirm. Strategic Management Journal, 18: 187–206.
Papadakis, V. M., Kaloghirou, Y., & Iatrelli, M. 1999.Decision making: From crisis to opportunity: Busi-ness Strategy Review, Spring: 29–57.
Pfeffer, J., & Salancik, G. R. 1978. The external control oforganizations. New York: Harper & Row.
Podolyny, J. M. 2001. Networks as the pipes and prismsof the market. American Journal of Sociology, 107:33–60.
Prahalad, C. K., & Bettis, R. A. 1986. The dominant logic:A new linkage between diversity and performance.Strategic Management Journal, 7: 485–501.
Reinganum, J. F. 1983. Uncertain innovation and thepersistence of monopoly. American Economic Re-view, 73: 741–748.
Ross, J., & Staw, B. M. 1993. Organizational escalationand exit: Lessons from the Shoreham nuclear powerplant. Academy of Management Journal, 36: 701–732.
Santayana, G. 1905. The life of reason. New York:Charles Scribner & Sons.
Schein, E. H. 1985. Organizational culture and leader-ship. San Francisco: Jossey-Bass.
Sigglekow, N. 2001. Change in the presence of fit: Therise, the fall, and the renaissance of Liz Claiborne.Academy of Management Journal, 44: 838–867.
Sitkin, S. B., & Pablo, A. L. Reconceptualizing the deter-minants of risk behavior. Academy of ManagementReview, 17: 9–39.
Staw, B. M., Sandelands, L., & Dutton, J. 1981. Threatrigidity effects in organizational behavior. Adminis-trative Science Quarterly, 26: 501–524.
Stuart, T. E., & Podolny, J. M. 1996. Local search and theevolution of technological capabilities. StrategicManagement Journal, 17: 21–38.
Sull, D. 1999. The dynamics of standing still: FirestoneTire and Rubber and the radial revolution. BusinessHistory Review, 73: 430–464.
Sutton, R. I., & Callahan, A. 1987. The stigma of bank-ruptcy: Spoiled organizational image and its man-agement. Academy of Management Journal, 30:405–436.
Teece, D. J., Pisano, G., & Shuen, A. 1997. Dynamiccapabilities and strategy management, StrategicManagement Journal, 18: 509–553.
Tripsas, M., & Gavetti, G. 2000. Capabilities, cognition,and inertia: Evidence from digital imaging. StrategicManagement Journal, 21: 1147–1161.
Tushman, M. L., & Anderson, P. 1986. Technologicaldiscontinuities and organizational environments.Administrative Science Quarterly, 31: 439–465.
Tushman, M. L., & O’Reilly, C. 1996. Ambidextrous or-ganizations: Managing evolutionary and revolution-ary change. California Management Review, 38(4):8–30.
Tushman, M. L., & Romanelli, E. 1985. Organizationalevolution: A metamorphosis model of convergenceand reorientation. In L. L. Cummings & B. M. Staw(Eds.), Research in organizational behavior, vol. 7:171–122. Greenwich, CT: JAI Press.
Tushman, M. L., Smith, W., Wood, R., Westerman, G., &O’Reilly, C. 2003. Innovation streams and ambi-dextrous organizational designs: On building dy-namic capabilities. Working paper no. 03-106, Har-vard Business School, Boston.
Yerkes, R. M., & Dodson, J. D. 1908. The relation ofstrength of stimulus to rapidity of habit formation.Journal of Comparative and Neurological Psychol-ogy, 18: 459–482.
Yin, R. K. 1994. Case study research: Design and meth-ods (2nd ed.). Newbury Park, CA: Sage.
Clark Gilbert ([email protected]) is an assistant profes-sor of entrepreneurial management at the GraduateSchool of Business at Harvard University. He earnedhis DBA in the area of business policy at Harvard in2001. His current research interests include corporateinnovation, discontinuous change, and strategy adap-tation in new ventures.
2005 763Gilbert
