娀 Academy of Management Journal

2005, Vol. 48, No. 5, 741–763.

UNBUNDLING THE STRUCTURE OF INERTIA:

RESOURCE VERSUS ROUTINE RIGIDITY
CLARK G. GILBERT
Harvard University

I work to unbundle the structure of inertia into two distinct categories: resource
rigidity (failure to change resource investment patterns) and routine rigidity (failure to
change organizational processes that use those resources). Given discontinuous
change, a researcher’s failure to recognize these distinctions can generate conflicting
findings regarding effects of threat perception on inertia. Using field data on the
response of newspaper organizations to the rise of digital media, I show that a strong
perception of threat helps overcome resource rigidity but simultaneously amplifies
routine rigidity. I develop an interpretive model exploring mechanisms for overcoming
these divergent behaviors.

The inability of incumbent firms to overcome distinct categories (1) failure to change resource
organizational inertia when threatened with dis- investment patterns (Christensen & Bower, 1996;
continuous technological change has been a topic Henderson, 1993) and (2) failure to change the or-
of repeated scholarly inquiry (Hannan & Freeman, ganizational processes that use those resource in-
1977; Henderson & Clark, 1990; Levinthal, 1992; vestments (Leonard-Barton, 1992; Nelson & Winter,
Tushman & O’Reilly, 1996). One of the reasons this 1982). The first category is referred to here as re-
topic receives so much attention is that incumbent source rigidity. The second is referred to here as
failure is extremely prevalent (Christensen & routine rigidity.
Rosenbloom, 1995; Tushman & Anderson, 1986). This article will show that the failure to differen-
Perhaps even more intriguing is that failure often tiate the aspects of inertia is an oversight in the
occurs even when managers are aware of the need literature, partly because it leaves the underlying
to change (Johnson, 1988). I propose that some of phenomenon inadequately described. But the lack
the difficulty observed (both in management prac- of specificity can also lead to conflicting, even con-
tice and scholarly research) occurs because of a tradictory, findings. I will show how differentiating
failure to differentiate between two very distinct between resource and routine rigidity can be par-
forms of inertia. Previous definitions of incumbent ticularly important when exploring the role of
inertia describe it as the inability to enact internal threat perception under conditions of discontinu-
change in the face of significant external change ous change. Whereas some scholars suggest that
(Miller & Friesen, 1980; Tushman & Romanelli, threat perception enables response, others argue
1985). I divide this general phenomenon into two that it constrains response. For example, studies
show that threat can unlock inertia by motivating
change (Huff, Huff, & Thomas, 1992; Lant, Mil-
liken, & Batra, 1992). And yet, threat perception has
I would like to thank my dissertation committee, Jo-
also been shown to increase inertia by narrowing
seph Bower, Clayton Christensen, and Tom Eisenmann,
as well as acknowledge the generous financial support of
alternatives and focusing response on previously
the Doctoral Programs of the Harvard Business School. learned routines (Dutton & Jackson, 1987; Staw,
Mike Tushman also provided tremendous support in de- Sandelands, & Dutton, 1981). I show here how re-
veloping these ideas and preparing them for publication. conceptualizing inertia as having two distinct
Other useful comments have come from Jane Dutton, forms can help explain this seeming contradiction.
Kathy Eisenhardt, Ranjay Gulati, Nicolaj Siggelkow, Using longitudinal field data, I compare percep-
Mary Tripsas, and three anonymous AMJ reviewers. I am tions in eight newspaper organizations of the emer-
likewise grateful to the participants of the Wharton Tech-
gence of digital publishing. Findings show that the
nology Mini-Conference and the BYU Winter Strategy
way managers perceive the threat of discontinuous
Conference for their suggestions. An earlier version of
this paper was the winner of the Robert J. Litschert Best change creates paradoxical links between resource
Doctoral Student Paper in Business Policy at the Acad- and routine rigidity. This article also examines
emy of Management annual meeting in 2001 and a dis- mechanisms for overcoming both types of rigidity
sertation finalist in 2002. simultaneously.
741
742 Academy of Management Journal October

THEORETICAL BACKGROUND forces blocking incumbent investment in discon-

tinuous change.
The Problem of Incumbent Inertia
Routine rigidity. Even when incumbent firms
In this section, I first specify how resource and invest, there is a second inertial problem. The lit-
routine rigidity represent different types of inertia. erature has repeatedly revealed the persistence and
I then explore how interpretation might have a inflexibility of firm routines. Routines are defined
different effect on each type, looking specifically at here as repeated patterns of response involving in-
conditions of discontinuous change. For this arti- terdependent activities that become reinforced
cle, discontinuous change refers to external through structural embeddedness and repeated use
changes that require internal adaptation along a (Feldman & Pentland, 2003; Nelson & Winter,
path that is nonlinear relative to a firm’s traditional 1982). Part of the explanation for routine rigidity is
innovation trajectory (Christensen & Bower, 1996; that organizational processes that are tightly
Tushman, Smith, Wood, Westerman, & O’Reilly, aligned with one environment can be difficult to
2003). change, because they are self-reinforcing and are
Resource rigidity. Of the many reasons that not built to adapt to discontinuities (Miller &
firms might underinvest in discontinuous change, Friesen, 1980; Siggelkow, 2001; Teece, Pisano, &
resource dependency and incumbent reinvestment Shuen, 1997; Tushman & Anderson, 1986). Further,
incentives are two important ones. The theory of exploitation processes can drive out exploration
resource dependency originated with the work of processes, making it difficult to develop new capa-
Pfeffer and Salancik (1978), which showed that a bilities (Benner & Tushman, 2001; Burgelman,
firm’s external resource providers shape and con- 1994; March, 1991). Another part of the explana-
strain its internal strategic choices. Resource pro- tion is that the original motivation for designing an
viders can include both capital markets and cus- organizational routine can be separated from the
tomer markets. For example, Noda and Bower people executing the routine (Nelson & Winter,
(1996) demonstrated how public equity markets 1982). The underlying logic pervades the thinking
place performance requirements that constrain of the organization, often manifesting as deeply
changes in business models and product architec- ingrained cognition (Prahalad & Bettis, 1986; Trip-
tures to conform to positions that originated at the sas & Gavetti, 2000). The challenge is increased
time of their initial funding. Christensen and because this cognition also becomes tacit (Schein,
Bower (1996) demonstrated how customers can 1985), making it even more difficult to recognize
penetrate a firm’s internal resource allocation sys- the sources within the routines that are creating the
tems (Bower, 1970). “When the initial price/perfor- difficulties. Thus, managers often rely on a learned
mance characteristics of emerging technologies pattern of response that is structurally and cogni-
render them competitive only in emerging market tively reinforced, instead of employing new search
segments, and not with current customers, resource efforts (March & Simon, 1958).
allocation mechanisms typically deny resources to
such technologies” (Christensen & Bower, 1996:
The Effect of Threat Perception on Inertia
198). In these studies, resources were both financial
and attention-based (Ocasio, 1997). Although the literature yields considerable in-
The second general argument concerning re- sight into the sources of both resource and routine
source rigidity relates to market power. A debate rigidity, less attention has been paid to their differ-
persists in the economics literature as to whether ences and possible interactions (Henderson, 1993).
incumbent firms, strongly positioned with a given Awareness of these differences is particularly im-
technology, will invest sufficient resources in dis- portant when exploring potential contradictions re-
continuous change (Arrow, 1962). Gilbert and garding the role of threat perception in discontin-
Newberry (1982) found that if entry to a new tech- uous change—specifically, whether the perception
nology is blocked, incumbent firms have strong of a threat increases or decreases inertia. Threat
incentives to reinvest in their current market posi- perception is defined as a deep sense of vulnerabil-
tions and not in the new technology. Reinganum ity that is assumed to be negative, likely to result in
(1983) argued that if incumbent investment in- loss, and largely out of one’s control (Dutton &
creases the probability of market adoption in a way Jackson, 1987; Jackson & Dutton, 1988). A well-
that alters a firm’s otherwise dominant position, developed literature on strategic change suggests
the firm has strategic incentives not to invest. Thus, threat perception is a response catalyst. Huff and
whether constraints stem from a desire to preserve his colleagues (1992) showed how threat-related
market power or from blinders created by resource stress could overcome inertia. Lant and her col-
dependence, they represent powerful inertial leagues (1992) found that companies threatened
2005 Gilbert 743

with performance decline were more willing to pact of threat perception on resource and routine
commit to strategic change. Barr and Huff (1997) rigidity.
showed that managers must view external changes
as having a negative impact on performance before
internal changes are possible. METHODS
But although the strategic management literature The present research might best be described as
suggests that threat is a catalyst in reducing inertia, theory elaboration (Lee, 1999; Lee, Mitchell, & Sa-
other research shows that threat can actually in- blynski, 1999) in that it elaborates theoretical links
crease inertia. Management scholars who apply so- not previously addressed in the literature. For ex-
cial-psychology ideas to threat-motivated response ample, previous studies on the role of interpreta-
discuss three intermediate behaviors that can in- tion have not distinguished between resource and
crease inertia. These are contraction of authority, routine rigidity, resulting in the apparent contra-
reduced experimentation, and focus on existing re- dictions described in the previous section. Thus, I
sources. For example, Staw, Sandelands, and Dut- attempt here to “simplify, reconnect, and redirect
ton found threat led to “increased centralization of theory” (Lee et al., 1999: 166) on the role of inter-
authority, more extensive formalization, and stan- pretation in response to discontinuous change in a
dardization of procedures” (1981: 513). Hermann way that differentiates between these two sources
(1963) noted a contraction of authority in crisis of inertia.
situations. Other research shows that threat nar-
rows the number of alternatives firm leaders are
willing to consider and reduces the level of exper- Research Setting
imentation in firm response (Ross & Staw, 1993; The research employed a multicase design that
Staw et al., 1981). Finally, because threat-induced supports a “replication logic,” whereby a set of
behavior is concerned with averting loss, it is di- cases is treated as a series of experiments, each
rected toward preserving current resources rather serving to confirm or disconfirm a set of observa-
than toward creating new opportunities (Dutton, tions (Yin, 1994). The primary unit of analysis was
1992; Hartman & Nelson, 1996; Mittal & Ross, an online venture of a newspaper organization re-
1998). sponding to digital media; the embedded units
In an effort to understand the conflicting conclu- were the sponsoring newspapers and their corpo-
sions regarding the impact of threat on inertia, I rate managements. Table 1 describes the four com-
propose two related research questions: (1) How panies and eight newspapers studied. Each com-
does threat perception affect incumbent inertia in pany owned two of the sampled newspapers,
the face of discontinuous change? and (2) Is the though each newspaper operated with significant
effect of threat perception different for resource autonomy. Print circulation across the sample
rigidity and routine rigidity? After describing this ranged from 200,000 to more than 500,000 average
study’s research methods, I articulate a set of for- daily readers. All of the newspapers but one were
mal observations (propositions) based on compar- locally distributed only. The competitive dynamics
ative data from eight field sites. These observations across markets were similar; each paper was the
yield an interpretive model of response to dis- largest competitor in a low-rivalry market. Internet
continuous change that addresses the varying im- penetration across each market varied, but not by

TABLE 1
Description of the Eight Newspapers Studied

Number of Online
Daily Print Circulation Online Launch Employees in
Newspaper Parent Organization Circulation Range Date 2000

Beacon A The Beacon Company 250,000 Local 1994 45

Beacon B The Beacon Company 200,000 Local 1995 20
Press A The Press Company ⬎500,000 Local/regional 1994 ⬃100
Press B The Press Company 400,000 Local/regional 1995 60
Expositor A The Expositor Company ⬎500,000 Local/regional 1995 ⬃100
Expositor B The Expositor Company 200,000 Local 1996 32
Morning News A The Morning News Company ⬎500,000 National 1994 ⬃100
Morning News B The Morning News Company 300,000 Local 1996 41
744 Academy of Management Journal October

more than 10 percentage points from highest to & Strauss, 1967) along a series of polar types that were
lowest. Each of the newspapers launched a Web likely to extend the emerging theory (Eisenhardt,
site between 1994 and 1996. Because of the sensi- 1989a). Thus, I targeted four innovative ventures and
tivity of the data, the names of the organizations four rigid ventures for examination.
and the newspapers are disguised.
I chose to examine the response of newspaper
Data Sources
firms to digital publishing for two reasons. First,
the effect of electronic publishing on the newspa- Data were collected from three main sources:
per industry generally matched the research ques- open-ended interviews, archival documents, and
tions. Online publishing was a discontinuous direct observations. Table 2 summarizes these
change in that it presented external changes that sources. I also collected over 150 public docu-
required nonlinear internal adaptations. For exam- ments, including press releases, annual reports, an-
ple, the features of the Internet that early online alyst reports, and industry articles.
users valued—access to breaking news, searchable Interviews. Of a total 62 interviews, 51 were
databases, live weather and traffic— differed con- in-depth, one- to two-hour in-person interviews
siderably from the features that were available in with the senior executives at the corporate, news-
print. Similarly, the online business model was paper, and online venture levels of the sampled
also considerably different, because it was driven businesses. I used semistructured interview tem-
by an altered cost structure, new categories of rev- plates concerning what motivated a manager to
enue, and different customer requirements. And commit to digital media, how that commitment
yet, this discontinuity threatened to supplant both evolved over time, the relation between print and
print readership and advertising over time, despite online efforts, and so forth. Additionally, 11 30-
the lack of initial customer overlap. The second minute follow-on telephone interviews were used
reason for selecting the newspaper industry is that to expand on the specific question of why a corpo-
singling out one industry helps control for extrane- rate management chose to separate its digital ven-
ous variation (Eisenhardt, 1989a). The selection of ture from its print business, or chose to keep the
case sites was based on theoretical sampling (Glaser venture integrated with the print newspaper. I at-

TABLE 2
Sources of Data

Interviewsa Archival Documents Direct Observations

Newspaper Corporate Newspapers Ventures Total Number Examples Number Examples

Beacon A 4 (2) 2 3 9 7 Business plan, customer list, 5 Planning meeting, sales

internal memo, strategic calls
plan, sales collateral
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, 5 Content development
internal memo, sales meetings, budgeting
collateral meetings
Press B 4 (2) 2 3 9 6 Business plan, customer list, 4 Sales calls, newsroom
strategic plan, sales planning meetings
collateral
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, 5 n.a.
internal memo
Morning News A 3 (1) 2 3 8 2 Business plan, customer lists, 5 Budgeting meetings,
internal memo 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 and
11 telephone interviews that were 30-minute follow-up conversations with the corporate executives on their decisions to separate their
online ventures from the parent newspapers or keep them integrated. The total figure includes both the extended in-person interviews and
the shorter telephone interviews. The telephone interviews are also shown in parentheses.
2005 Gilbert 745

tempted to triangulate by using multiple infor- was blind to the original coding and purpose of the
mants and cross-checking information against ar- study, cross-checked my coding efforts for each site
chival and public documents to avoid retrospective over all periods (Miles & Huberman, 1984). In no
bias in the interviews. Strict case study protocol instance was there a conflict between the second
was followed. More than 90 percent of the inter- reader’s coding and the original coding.
views were transcribed and entered into a case I then developed a set of formally stated obser-
study database. vations that were based on early case analysis of a
Archival documents. The 33 internal archival set of matched-pair, polar cases— one innovative
documents that I collected at seven of the eight and the other rigid. Below, these formal observa-
sites included online business plans, strategy pro- tions are presented as research propositions. I ana-
posals, internal memos, annual strategic planning lyzed the case data and “enfolded” a set of relevant
documents, customer lists, and historical sales col- literatures, following methods for inductive theory
lateral material. These documents constituted a development (Eisenhardt, 1989a; Glaser & Strauss,
valuable primary source of data and offered a way 1967). I then used analytical replication to deter-
to cross-check the interviews and to control for mine whether the emerging relationships were con-
retrospective bias. I focused on documents that dis- firmed or disconfirmed in the rest of the sample
cussed whether or not to fund an online business. (Eisenhardt, 1989; Yin, 1994). To confirm or dis-
These included business plans starting as early as confirm the relationships identified in each obser-
1990. vation, I arrayed the data following techniques for
Direct observations. Over a year and a half, from cross-case pattern sequencing (Eisenhardt, 1989b)
2000 to 2001, I recorded meetings among online and tabular displays (Miles & Huberman, 1984). As
venture, newspaper, and corporate executives, at- with deductive hypothesis testing, the formal ob-
tended planning meetings for the online ventures, servations fit a consistent pattern, though they did
observed story creation for the newspapers and the not always conform perfectly (Eisenhardt, 1989a;
Web sites, monitored sales calls for both the news- Sutton & Callahan, 1987). I used the individual
papers and the online products, and visited plan- observations to construct an interpretive model of
ning meetings. In all, I observed 24 field events response to discontinuous change that differenti-
from five of the research sites and recorded them ated inertia into resource and routine rigidity.
into the case study database.

ANALYSIS OF DATA
Research Process
Response in the Absence of Threat: Resource
My investigation was informed by three charac-
Rigidity
teristics that have been observed as being associ-
ated with threat perception: negative focus, empha- The field analysis confirmed that without a per-
sis on loss, and sense of a lack of control (Dutton & ception of threat, there was considerable resource
Jackson, 1987; Jackson & Dutton, 1988). I grouped rigidity around discontinuous change. This was
interview and archival data into time-sequenced manifested in the data by the failure of established
arrays to see how threat perception evolved over newspaper firms to invest financial and attention-
time. I also tried to identify instances in which field based resources in digital publishing. I explored
observations did or did not fit these categoriza- this occurrence of rigidity by examining the forces
tions. For example, I coded the motivation to com- of resource dependence and position reinvestment
mit to the Internet as either a threat or an opportu- incentives.
nity. When a business plan stated that “online Resource dependency. Much of the initial re-
revenue would cannibalize print revenue” (nega- source rigidity stemmed from resource dependen-
tive, loss) or “we can slow it down, but we can’t cies related to the demands of the established print
stop it” (lack of control), the motivation of the newspaper customers—both advertisers and read-
manager who had presented the plan was coded as ers. These demands were difficult to reconcile with
a threat perception. In contrast, if a manager de- the requirements of an emerging set of online cus-
scribed online revenue as largely “additive to print tomers. The data in Table 3 show that business
revenue” (positive, gain) or as susceptible to “in- proposals for online often stalled for more than two
fluence through involvement” (in control), I coded years in review in these newspaper organizations.
the manager’s motivation as opportunity percep- Even when money was provided, operating atten-
tion. Although sample size and the varying charac- tion could be equally difficult to secure. An online
teristics of data sources at the different sites did not sales representative at the Beacon A recalled this:
allow statistical comparison, a second reader, who “Print reps could sell the online product, but with
746 Academy of Management Journal October

TABLE 3
Resource Rigidity

Years in Initial Date of Examples: Financial and Operational

Newspaper Evidence Proposal Commitment Launch Constraints

Beacon A Interviews, archival documents, 2 Publisher 1994 Sales: “Print reps could sell the online product,
public documents sponsorship but with varying degrees of success. Their
margins were higher on other products that
were easier for them to sell. Online was
really just a novelty to them.” (sales
representative)
Beacon B Interviews, archival documents 2 Forecast 1995 Sales: Staff provided training, but then dropped
profitability the program multiple times owing to lack of
client interest.
Press A Interviews, archival documents, 2 Forecast 1994 Sales: “I occasionally sell a bundled print and
public documents profitability online package . . . it is hard to really know
what print advertisers would want.” (sales
representative)
Press B Interviews, archival documents, 2 CEO 1995 Newsroom: Editors call online staff “low-brow
public documents sponsorship content” for using radio feeds for breaking
news, personals, and unedited user-posted
content and refuse to work together.
(newspaper editor)
Expositor A Interviews, archival documents, 3 CEO 1995 Budgeting: “And in the end, the only real value
public documents sponsorship is cash and cash creation. You can’t build a
business just on potential or hope.” (CEO)
Expositor B Interviews, archival documents, 2 Forecast 1996 Budgeting: “Look, when we roll these up into
public documents profitability our budgets we miss our targets.” (vice
president, product development)
Morning News A Interviews, archival documents, 1 Publisher 1994 Newsroom: “I will be goddamned if some
public documents sponsorship online reporter is going to call my sources
and say they are from our paper.” (publisher)
Morning News B Interviews, public documents 2 Forecast 1996 Sales: “We bundled print and online, but there
profitability were clients who wanted online only. These
were less interesting to sales reps and the or-
ganization wasn’t ready to deal with that
reality.” (vice president, marketing)

varying degrees of success. Their margins were these newspaper sites fell almost entirely into the
higher on other products that were easier for them category of resource dependence. But even as these
to sell. Online was really just a novelty to them.” “I incumbent organizations came to realize the risks
occasionally sell a bundled print and online pack- incurred by depending so heavily on their tradi-
age,” explained a print rep at the Press A. “There is tional customers, they did not necessarily over-
no standard package, and it is hard to really know come their resource rigidity. Market position incen-
what the print advertisers would want.” I gathered tives to continue reinvesting in the core newspaper
data on the overlap between print and online ad- businesses remained strong. Nearly every research
vertisers at five of the research sites, asking manag- site conducted huge studies to estimate the canni-
ers to estimate how many of their top 25 online balization effect of providing their news informa-
advertisers were represented among the top 25 tion online. This concern was particularly acute at
newspaper advertisers. Customer overlap was only the Expositor A and the Press A. The director of the
7 of a possible 125. The travel category at the Press Press Company’s Internet group reflected, “I think
A was a good example. Of the top ten booking the notion that people would start reading their
agents online, only four advertised in print. In sum- newspapers on the screen was quite prevalent.”
mary, the lack of customer overlap, a different sell- Managers at both of these sites wanted to stall or
ing process, and the lower relative gross margins delay any investment in a technology that might
combined to limit the amount of time and attention cannibalize the core newspaper business. One se-
print sales reps were willing to invest in online. nior manager described his strategy as “wait and
Position reinvestment incentives. The resource see,” preferring not to take his paper online before
rigidity that I observed in the initial response of adoption became inevitable.
2005 Gilbert 747

Threat Perception and Resource Rigidity publisher of the Beacon A. “We can slow it down,
but we can’t stop it.”
I next examine how a perception of threat af-
In seven of the cases, this mounting concern
fected the resource rigidity at each field site, first
about the Internet threat led to expanded financial
presenting the field data and then summarizing the
observed behavior more formally. The data indicate and organizational commitment. The archival data
that a strong perception of threat was associated also showed evidence that threat perception gener-
with increased financial and organizational com- ated resource commitment. For example, a 1997
mitment to an online venture. At the various sites, business plan for the Beacon B estimated that 15 to
the perception of threat built gradually and at 20 percent of its print classified revenues might be
slightly different paces. But by 1997–98, most lost to online. One business plan explained, “If we
newspapers’ managements had become genuinely don’t cannibalize ourselves, someone else will.”
concerned about the imminence of online newspa- While earlier proposals had emphasized financial
per publishing, even though the advertisers in their returns and new market opportunities, arguments
mainstream markets still did not view digital media for increased funding now stressed the growing
as a valuable advertising outlet. The director of threat to print revenue from online competition.
marketing at the Morning News B described the Despite mounting losses in the online business,
evolution of their thinking: frequently exceeding 100 percent of online reve-
nues, expenditures grew by as much as 400 percent
You felt like Chicken Little screaming “The sky is during the years that threat perceptions were build-
falling,” but after a while people started listening ing (see Table 4). The number of employees allo-
when they saw what the other competitors were cated to the online ventures also increased. For
doing. We made watch lists for TV, radio, vertical
example, dedicated online staff at one site ex-
start-ups, telephone companies, and Citysearch. City-
search was poaching people. . . . The publisher was
panded from 5 to 40 people during an eight-month
unlike some in that he saw the threat. period in 1998. Staffing levels reached or exceeded
100 individuals at many sites (see Table 4).
Whereas in earlier funding discussions, propo- The expanding resource commitment was not
nents tended to describe online publishing as a limited to financial expenditures and headcount; it
growth opportunity meriting investment (often extended to operating commitments such as man-
with little success), in 1997–98 these discussions agement time and effort. The threat posed by digital
turned to the perceived threat posed by the Inter- media overcame customer dependencies that might
net. Data in Table 4 show that threat perception otherwise have pulled investment away from the
developed across the sample. Threat perception new technology. “Look, it didn’t make any sense
was evidenced as negative statements about the for us to try to sell this stuff, but we began to feel
lack of control managers had in preventing losses to that if we didn’t work on it, it might come back to
the Internet. For example, the CEO of the Press haunt us,” said a sales manager at the Beacon A.
Company’s Internet group explained, “There were Reporters were asked to summarize articles and
people who thought we would lose half of our stories before they were published in print, and
circulation.” Others worried about losing classified many were encouraged to write follow up on sto-
advertising products, including employment, real ries just for the Web.
estate, and auto listings. These products (represent- Indeed, threat perception was observed to have a
ing as much as 60 percent of the profit of most powerful catalytic effect on both types of resource
metropolitan newspapers) appeared to be very por- rigidity, a situation that can be summarized in my
table to the searchable database format available first two propositions:
online. One executive recalled his organization’s
perception as follows: “McKinsey had come in and Proposition 1a. The perception of an imminent
had done a rather startling analysis of the classified threat in the face of discontinuous change en-
business. They predicted that 20 –30 percent of our ables managers to overcome sources of re-
classified revenue would disappear by 1998. That source rigidity that stem from resource
raised enormous alarm bells in some people.” Fi- dependence.
nally, managers were concerned that the impact of
digital media was largely out of the newspaper Proposition 1b. The perception of an imminent
companies’ control and that it would run its course threat in the face of discontinuous change en-
with or without their participation. “What if we do ables managers to overcome sources of re-
every damn thing we can think of and execute source rigidity that stem from incumbent posi-
flawlessly and we still don’t make it?” lamented the tion reinvestment incentives.
 TABLE 4
Perception of Threat Helps Overcome Resource Rigidity

Stated Source of Year Threat Pattern of Expenditure Employee

Newspaper Threat Emerged Evidence Examples of Threat Behaviora Expansionb Expansion

Beacon A Readership, display 1996 Interviews, archival documents, Lack of control: “What if we do every Literal 1996–98: 1996–98:
advertising, public documents damn thing we can think of and replication 400% 15340
classified execute flawlessly and we still don’t
advertising make it? We can slow it down, but
we can’t stop it.” (publisher)
Beacon B Classified 1997 Interviews, archival documents Focus on loss: 1997 internal estimates Literal 1997–99: 1997–99:
advertising of 15–20% loss of classified share by replication 300% 5318
2001 (1997 strategic plan)
Press A Readership, display 1997 Interviews, archival documents, Focus on loss: “There were people who Literal 1997–1999: 1997–99:
and classified public documents thought we would lose half of our replication 300% 503100⫹
advertising circulation.” (CEO, Internet group)
Press B Readership, display 1997 Interviews, archival documents, Negative effect: “People were telling us Literal 1997–99: 1997–99:
and classified public documents that newspapers were heading to the replication 250% 30350
advertising graveyard and we were beginning to
believe them.” (publisher)
Expositor A Stock price, 1997 Interviews, archival documents, Focus on loss: “We were worried about Literal 1997–99: 1997–99:
classified public documents classifieds primarily. These new replication 300% 503100⫹
advertising firms were set to come in and take
our most profitable piece of
business.” (early online president)
Expositor B Readership, display 1998 Interviews, archival documents, Negative effect: “We [middle managers] Literal 1998–2000: 1998–2000:
and classified public documents had been concerned for a while, replication 250% 15331
advertising but . . . the reason we finally got into
the market was that our CEO was
taking heat from Wall Street.” (vice
president, product development)
Morning News A Primary motivation n.a. Interviews, archival documents, Not threat: “Eighty percent of my costs Theoretical 1997–99: 1997–99:
distribution/production public documents are production and distribution. Now replication 200% 603100⫹
savings, not threat all of a sudden I have a solution. It is
not a content play, but a major cost
reducer and product expander.”
(publisher)
Morning News B Classified 1998 Interviews, public documents Negative effect: “You felt like Chicken Literal 1998–2000: 1998–2000:
advertising Little screaming “the sky is falling,’ replication 400% 5340
but after a while people started
listening.” (vice president, marketing)

a
These categories are based on Yin’s (1994) analytical technique of replication. In theoretical replication, the behavior was not observed, but its absence was consistent with the theory
argued above. The cost structure at the Morning News A revealed an opportunity motivation existed, but for reasons that were theoretically consistent with the general pattern observed.
b
Period based on relevant expansion of threat perception at a site. Numbers represent company estimates from the start of year 1 of the study period to the end of year 3.
2005 Gilbert 749

Observations at every primary research site con- Contraction of authority. Data confirmed that
firmed the negative relationship between threat corporate management asserted its control over de-
perception and resource rigidity (see Table 4). cision making, withdrawing considerable authority
In Table 4, which summarizes these observa- from operating divisions, in each newspaper adop-
tions, I have noted whether the data observed fol- tion of online publication studied here. This con-
lowed Yin’s pattern of replication (1994). When an traction of authority involved transferring more
observed site replicated the pattern, I called the control to corporate officers, such as the head of
case a literal replication. When a site did not follow business development, the CEO, or a newly ap-
the pattern, but for reasons that were not inconsis- pointed online director. At nearly every site I stud-
tent with the proposition, I called the case a theo- ied, these individuals or groups assumed control of
retical replication. If sites had not followed the strategy. Table 5 gives examples of the observed
pattern for other reasons, the exceptions would also contraction of authority. Sometimes this shift was
have been noted. Notice in Table 4 that seven of the accomplished by imposing business plan templates
eight sites showed behavior that was a literal rep- on local site managers. According to the vice pres-
lication of the pattern: threat helped overcome re- ident of technology and operations at the Beacon A,
source rigidity. The exception to this pattern was
It was very centralized in the beginning, which was
the Morning News A. At this national newspaper,
very uncharacteristic, because the culture is very
the Internet actually helped to solve an existing much to let these guys run their own businesses. We
organizational problem. Thus, it was not discontin- had a basic business model for every site. We gave
uous to its operating requirements in the same way them money. We told them they could hire people,
that it was for other locally distributed newspaper but we told them exactly how to do it.
organizations. Whereas the other papers had huge
economies of local production and distribution, the The contraction of authority also took the form of
Morning News A did not. According to the presi- centralized decision approval. The Morning News
dent and publisher: Company’s CEO personally directed online strategy
at the flagship paper, and the head of business
This was a wonderful opportunity from the start. If development made strategic, financial, and hiring
you are a national newspaper with a 3 percent pen- decisions for all the other newspapers. Similarly,
etration, all of a sudden you have an opportunity for the Press A required the newspaper CEO to approve
virtually no cost to distribute the product. . . . The all senior hires. Sites had to follow detailed budgets
Internet creates huge opportunities to deliver prod- and adhere to strict marketing plans. The ventures
uct in areas that were uneconomical before. . . . were allowed little autonomy in local planning.
Eighty percent of my costs are production and dis-
This restriction increased reliance on existing rou-
tribution. Now all of a sudden I have a solution.
tines because it limited the alternatives considered.
Further, operating managers were less likely to
Also, because it had a limited classified product,
change corporate-imposed routines. As one operat-
this newspaper did not share the fear of cannibal-
ing manager explained, “We felt like even though
ization with many of the other sites in the sample.
we were all very focused on the online business,
The Internet matched a set of sustaining needs for
the corporate folks kind of had a plan and stuck
the print newspaper business and was accepted
with it, even though we could see the failures oc-
accordingly. The key distinction was that the op-
curring in our local market.” The effect of this
portunity was not discontinuous for the Morning
behavior can be formally stated:
News A, as it was for the other research sites. Thus,
the newspaper does not follow the pattern of rep- Proposition 2a. Perception of an imminent
lication observed in the other sites, but for theoret- threat leads to a contraction of authority that
ically consistent reasons (Yin, 1994). amplifies routine rigidity.
Reduced experimentation. The contraction of
authority had a feedback effect on the level of ex-
Threat Perception and Routine Rigidity
perimentation in online strategy. For example, one
Although threat perception reduced resource ri- of the papers at the Beacon Company wanted to
gidity, it increased routine rigidity. In keeping with experiment with other forms of revenue generation,
previous research, my findings show that threat but it felt constrained by the corporate strategy that
perception produces three intermediate behaviors imposed a site template for sales strategy, business
that amplify routine rigidity: contraction of author- models, and product plans. This reduced experi-
ity, reduced experimentation, and focus on existing mentation was not solely a function of corporate
resources (Staw et al., 1981). control; the data show that the aggressive pace of
 TABLE 5
Evidence of Threat-Motivated Behaviors Linked to Routine Rigidity, 1996 –98a

Contraction of Authority Reduced Experimentation Focus on Existing Resources

Newspaper Behavior Examples Behavior Examples Behavior Examples

Beacon A Literal replication Breaking from an early Literal Local market experiments and Literal replication In 1996, corporate beings tracking
policy of local replication variation discontinued print readership and online
development, corporate- because each site was forced audience overlap tracked monthly
developed business plan to follow corporate plan
templates for all sites,
including budgets, and
organization charts
Beacon B Literal replication Product template, including Literal Nonstaged deployment of Literal replication Product development based on value
content categories, replication capital created lock-in on to the print product
determined by corporate initial corporate plan
Press A Literal replication Whereas earlier Literal Despite increasing losses, Literal replication Tremendous concerns about
involvement was replication resource commitment cannibalization, readership studies
minimal, the CEO expands without pushing on done asking whether people would
becomes involved in all the drivers of missed targets stop reading the paper
online budget approval
decisions
Press B Literal replication Corporate staff select initial Theoretical Some local experimentation Theoretical Focus moves away from the
online director and early replication occurs after outside online replication newspaper to the business model
strategy director is hired; different and customer requirements that are
content strategies, revenue unique to online media
stream tested and changed
based on market tests
Expositor A Literal replication CEO drives entry into Literal Online given directive not to Literal replication Advertising products built around
online, taking over replication lose money, and funding for the needs of print advertisers
previous role of local experiments very difficult
publishers
Expositor B Literal replication All decisions regarding Literal Corporate driven classified Literal replication Cannibalization surveys conducted
classified strategy replication plan enforced over previous
governed by corporate local experiments
Morning News A Theoretical CEO runs the online Theoretical Experimentation occurred Literal replication All product development
replication business at flagship paper Replication within the current business expenditures built around impact
for extended period model and print advertiser studies on print advertisers
customer requirements
Morning News B Literal replication Corporate business Literal Local market experiments and Literal replication Selling strategy built around “up-
development director replication variation discontinued, selling” to existing advertisers
creates site strategy and forced to follow corporate
personally approves all
funding

a
These categories are based on Yin’s (1994) technique of replication. In theoretical replication, the behavior was not observed, but its absence was consistent with the theory argued
above. The Morning News A was opportunity, not threat, motivated. However, the opportunity the newspaper saw was related to cost savings in the core business, not to creating a new
business (see Table 4). Therefore, the experimentation that did occur was focused on the core business. Thus, the case varies from the pattern, but for theoretically consistent reasons. The
Press B did not follow the observed pattern for reasons discussed subsequently in the text.
2005 Gilbert 751

resource commitment also made it more difficult to Proposition 2c. Perception of an imminent
step back and change behavior. The sample firms threat leads to a focus on existing resources
expanded resource commitments by more than 100 that amplifies routine rigidity.
percent per year during the period of accelerated
Table 6 summarizes evidence on the effects of
threat perception. Recall that employment at the
each of these threat-motivated behaviors on routine
Morning News A went from 5 to 40 people in less
rigidity. In almost every instance, threat-motivated
than eight months. Expenses at the Beacon B dou-
response led to aggressive replication of the news-
bled, then nearly doubled again in 1997 and 1998,
paper product and business model. Seven of the
contradicting original forecasts of profitability by
eight research sites turned out a product that was
1996 (see Table 5). Whereas initial disappoint-
merely an extension of the newspaper. Indeed,
ments might have prompted managers to regroup,
many of the sites republished more than 85 percent
the perception of threat caused them to press ahead
of their Web site content directly from the local
on the same misdirected paths. To paraphrase San-
newspaper product. The publisher of the Beacon A
tayana’s (1905) definition of a fanatic, it was as if
observed this: “We learned [from early involve-
the organizations were doubling their speed upon
ment with new media] that there wasn’t very much
losing their direction. If an initial response is
appetite for an ‘electronic’ newspaper. . . . But that
wrong, then expanding resources may only solidify
is exactly what we did with the Internet.” Common
those initial tendencies. The Beacon A, for exam-
features such as discussion boards, site-searching
ple, hired more than 40 people to implement an
tools, breaking news from third-party sources, com-
expansion strategy that closely resembled that of
munity forums, and other content features com-
the newspaper business model. Because the expan-
monly provided by the many new entrants were
sion occurred so quickly, the resources invested
largely absent from newspaper sites. Ironically, the
reinforced rather than reshaped established rou-
technologies to develop these products were
tines of the parent. The effect of this behavior can
largely available and relatively easy to deploy. The
be formally stated:
chairman and CEO of the Beacon Company
reflected:
Proposition 2b. Perception of an imminent
threat leads to a reduced level of experimenta- Where I think we missed the boat is that we saw it as
tion that amplifies routine rigidity. an extension of the newspaper. In other words,
something richer and deeper than the newspa-
Focus on existing resources. The fear of canni- per. . . . Our Internet operations were really run by
balizing the core newspaper business prompted people who came out of the newsroom, so they were
editors who tended to look at this more as a
managers to focus on their existing resources,
newspaper.
rather than consider new options presented by the
new technology (see Table 5). One manager at the Most sites simply reproduced the newspaper.
Press A expressed fear that the online publication The online director at one site remarked, “Remem-
would drain revenues from the print publication: ber that I had said to the CEO at the time that it
“Cannibalization was a huge concern for everyone made absolutely no sense to replicate the newspa-
initially. . . . We asked questions about readership per on the Internet. Then I saw the product and it
overlap and whether they would stop reading the was just that.”
paper.” Functional managers often retained over- The rigidity was expressed not only in the prod-
sight for the online products. For example, the uct but also in the business model. I compared the
president of the Expositor A Internet site explained income statements of the eight research sites
that “because the classified organization was so against a panel set of five competing online entrant
worried about defending the print classifieds busi- firms. This analysis identified six categories of rev-
ness, that group held onto the online business.” enue associated with digital media that differed
Because they were focused on the existing busi- from those built around a print newspaper (e.g.,
ness, they responded with routines that worked e-mail and interactive advertising). Whereas a com-
well in that business. Development of the new tech- parative set of entrants averaged more than five
nology was often driven by the effect it would have new revenue categories per site, most of the news-
on the newspaper, despite the parallel growth of a paper sites had only one (see Table 6). These new
separate online ad market. The head of sales for the categories accounted for more than 40 percent of
Press Company’s Internet group stated, “On the one many of the entrant firm revenue streams. As the
hand I should go do whatever we need to do, but on CEO of the Expositor Company explained, “We
the other hand there is concern about the paper.” couldn’t see any models that we were familiar with,
The effect of this behavior can be formally stated: nor any we knew how to make money with.”
752 Academy of Management Journal October

TABLE 6
Routine Rigidity Associated with Threat-Motivated Behavior

Early Percent Print New Revenue

Newspaper Product Contenta Categoriesb Examples

Beacon A Extension of ⬎75% 1 “We learned . . . that there wasn’t very much of an appetite for
newspaper an ‘electronic’ newspaper . . . But that’s exactly what we did
with the Internet.” (publisher)
Beacon B Extension of ⬎85% 1 “Where I think we missed the boat is that we saw it as an
newspaper extension of the newspaper, in other words, something richer
and deeper than the newspaper.” (CEO)
Press A Extension of ⬎85% 3 “Remember that I had said to the CEO at the time that it made
newspaper absolutely no sense to replicate the newspaper on the
Internet. Then I saw the product and it was just that.” (CEO,
Internet group)
Press B Multisourced ⬍50% 3 “We are really becoming a separate company from the
interactive newspaper. I came from there. I love the paper, but we are
media now a different group with a very different way of working.”
(online editor)
Expositor A Extension of ⬎85% 1 “We couldn’t see any models that we were familiar with, nor
newspaper any we knew how to make money with.” (CEO)
Expositor B Extension of ⬎85% 1 “We failed to recognize the importance of tools such as search,
newspaper but rather presented this in the layout of a printed
newspaper.” (vice president, product development)
Morning News A Extension of ⬎90% 1 “Where we made our mistake was we missed the next wave of
newspaper opportunity. We could have said we want to be a national
classified source. We could have become different content
verticals. But we have done very little on content verticals.”
(publisher)
Morning News B Extension of ⬎80% 1 “I don’t see this as that different than what the newspapers
newspaper currently 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, and
CNET. Six categories of revenue were identified as being new for print newspapers: fee-based archival access, e-mail marketing, e-mail list
rental, fee-based data analysis, behavioral targeting, and demographic targeting.

The Notable Exception—Outside Influence and initial impulses to replicate the newspaper. This
Separate Structure at the Press B shift originated with the suggestion by the CEO that
someone from outside of the newspaper industry
The most notable exception to the pattern of rou-
look at the online business. The CEO’s view had
tine rigidity was the Press B. And although this site
was similar to the others in most contextual factors, been shaped from the recommendation of a long-
it was the only site that launched a structurally time friend and personal advisor who was based in
differentiated venture from the outset—a decision Silicon Valley and had been observing some of the
that was largely influenced by the CEO’s external changes created by the Internet. The Press B senior
personal network. Like the rest of the sample, the management then launched a search for outside
Press B struggled with resource rigidity early on, advice on strategy. A Silicon Valley business exec-
and threat perception ultimately became the means utive with a background in new media was hired to
to overcome that constraint. Recall that in initial write the original business plan. That original plan
proposals managers had argued, “If we don’t can- called for an organizational design with significant
nibalize ourselves, someone else will.” Similarly, autonomy from the newspaper. The new venture
the Press B management’s early thinking about the was subsequently set up as a wholly owned sub-
form of its Internet product did not significantly sidiary and hired outside managers with new me-
differ from the assumptions of the managers at the dia experience. The management team then re-
other newspapers in the sample. The Press B news- cruited a separate sales force to price and sell
paper managers assumed that the product would be online ads. They also developed a separate brand to
a “newspaper in electronic form.” But unlike the signal that the product, although owned by the
other incumbents, the Press B did not follow these parent, would be distinct from the newspaper. Fi-
2005 Gilbert 753

nally, the Internet group was located in offices ment at the Morning News A stated: “Our basic
more than a mile away from the print offices. Some strategy is an integrated strategy. . . . In the local
newspaper staff were allocated to the venture, and information market, the newspaper has an advan-
important links to the newsroom were preserved. tage. To separate the online unit from the newspa-
But the locus of the new venture, both physically per is to give away a lot of that advantage.”
and operationally, was distinct. Nevertheless, four of the research sites in the
Early on, the site evolved into a regional source sample did eventually differentiate the organiza-
of news and information that was distinctly differ- tional structure of their online ventures. The un-
ent from the newspaper. More than 50 percent of derlying motivation in each case was related to
site content originated from sources other than the outside influence on decision making. Two of the
newspaper (see Table 6). Lead stories were differ- sites separated in response to the adamant de-
ent, and they rotated throughout the day; sections
mands of online employees hired from outside the
that did not exist in print were added; and users
newspaper. In the case of the Press A, an influential
were provided with a host of tools that enabled
board member from the technology industry helped
them to take advantage of digital media, including
introduce the company to a high-profile new media
traffic Web cameras, searchable event databases,
expert who was eventually offered a job as the
interactive discussion forums, and new forms of
content. Just how different the new product became online media director. His condition for joining the
was described by one online editor: venture was that it be separated from the parent
organization. Thus, though the venture was
Page views from the newspaper are now barely more launched from within the newspaper, it was sepa-
than one-third of the available pages on our site. We rated from it in the summer of 1997. At the Beacon
are really becoming a separate company from the
Company the process was more evolutionary and
newspaper. I came from there. I love the paper, but
we are now a different group with a very different
driven by interactions with outside partners. For
way of working. They are one source of informa- example, an outside partner that marketed online
tion—an important source. But we buy our content directory services through the newspaper’s Web
from them like we buy it from anywhere else. site had asked to be allowed to build its own agency
to sell the online ads. Within a year, the sales
The site captured new and different categories of agency had tripled the performance of the print
revenue, and it developed its own business model. sales reps, despite an identical headcount. The
Though its site was not as innovative as those of
newspaper eventually purchased the agency, but
competing nonnewspaper entrants, the Press B ev-
kept it separate at the demand of the sales director.
idenced considerably less routine rigidity than the
The agency kept adding online advertising catego-
sites of other online newspapers in the sample (see
ries beyond the directory ads through 2000, by
Table 6). Whereas, for example, most of the others
captured only one new category of revenue, the which time it was selling close to two-thirds of the
Press B captured three: fee-based archival access, online ads for the Beacon A site.
e-mail marketing, and fee-based usage data analy- Each site that separated its online venture from
sis. Similarly, less than 50 percent of the Press B’s the parent organization cited the influence of out-
site content was reused newspaper content. side parties in the decision process. The process of
differentiation often occurred over several years,
owing to the reticence of functional units in the
Outsiders, Structural Differentiation, and Threat newspaper organization. As mentioned earlier, the
Perception Press A separated from the newspaper organization
Though all the research sites debated whether to in the summer of 1997. The Beacon A separated in
structurally differentiate their online ventures, a the spring of 1999, and the Beacon B followed in
desire to leverage the assets of the print business the summer of 1999. To determine whether a ven-
motivated many to stay integrated with their parent ture was structurally differentiated from its parent
newspaper organizations. In the Beacon A’s origi- organization, I performed analyses that included
nal 1990 online business plan, the publisher wrote: units’ own reported structural classifications as
“The power of the newspaper to provide thrust for well as other metrics, such as reporting lines and
the new services can be harnessed only if it physical locations (see Table 7). Note that in the
achieves deep levels of integration with the news- four sites that separated, corporate managers men-
paper. Structuring the experiment as an enterprise tioned outside influence, while managers who
separate from the newspaper would be crippling if chose to remain integrated did not. This pattern of
not fatal.” The vice president of business develop- observation leads to a formal proposition:
754 Academy of Management Journal October

TABLE 7
Outside Influence and Structural Variation, 2000

Outside Influence on Timing of Divisional Reporting Physical

Newspaper Structural Choice Changes Unita, b Linesc Newsd Salese Locationf Description

Beacon A Outside partner Spring 1999 D D H H D Set up as a differentiated unit

1997; outside in a different building from
online CEO, 1999 the parent with primary
reporting through the online
general manager.
Beacon B Outside advisor, Summer 1999 D D H H H Set up as a differentiated unit,
1998 but shared responsibilities
with some parent functional
staff. Primary reporting
responsibility still through
online general manager.
Press A Outside board Summer 1997 D D H D D Set up as a differentiated unit
member, 1997; in a different building from
outside online the parent with primary
CEO, 1997 reporting through the online
general manager. All
functional staff hired
separately for online.
Press B CEO’s personal Spring 1995 D D H D D Set up as a differentiated unit
advisor, 1994; in a different building from
outside business the parent with primary
plan consultant, reporting through the online
1995; online CEO, general manager. All
1995 functional staff hired
separately for online.
Expositor A No key outside Remained I I I I H Kept integrated with the parent.
influence integrated Online general manager and
mentioned small support staff hired and
located in a separate
building, but all functional
staff kept integrated with
parent.
Expositor B No key outside Remained I I I I I Kept integrated with the parent
influence integrated organization. Online general
mentioned manager hired, but all
functional responsibilities
coordinated through parent.
Morning News No key outside Remained H I H I I Structure was hybrid.
A influence integrated Newsroom was clearly
mentioned hybrid, with some online
content going back to the
newspaper, but most
functional reporting went
through the parent.
Morning News No key outside Remained I I H H I Kept integrated with the parent
B influence integrated organization. Online general
mentioned manager hired, but all
functional responsibilities
coordinated 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).
2005 Gilbert 755

Proposition 3. Involving outside influence opportunity perception was given a chance to de-
when deciding how to respond to discontinu- velop simultaneously in the ventures. Stated
ous change will increase the likelihood that formally:
managers will structurally differentiate a new
Proposition 4. Structural differentiation can
venture from its parent organization.
help decouple threat perception in a parent
Even before sites differentiated structurally, from an opportunity perception in a new
some managers had begun to perceive that the In- venture.
ternet, while perhaps still representing a threat,
In the four sites that did not differentiate their
could become a source of new opportunity for their
ventures from the parent organizations, managers
newspapers. The president of the Press Company
continued to be preoccupied with the threat to their
explained:
organizations. The vice president of the Expositor B
We were worried about the Web in that it would insisted, “We continue to see this as a way to pro-
alter the way in which people would get informa- tect classifieds, and that if we don’t do it someone
tion, but it was not purely defensive. We had else will.” Comparing the sites that separated with
launched into entertainment years ago as a defen- those that remained integrated reveals that oppor-
sive move. It eventually became a new source of tunity perception emerged only where there was
growth for us. Many of the threats eventually be-
structural differentiation. Table 8 summarizes com-
come opportunities. The Internet may be the same
parative data for Proposition 4.
way.
Not only did outside influence and structural
This dualistic view was, however, hard to main- differentiation help to decouple the cognitive per-
tain in an environment in which operating respon- ceptions in the newspaper organizations from those
sibilities for the newspaper predominated. Thus, in the online venture, but also all three of these
the data reveal that one reason for creating a differ- variables were correlated with relaxed routine ri-
entiated structure is to decouple the motivation at gidities. This relaxation was driven largely by the
the parent from the motivation at the venture. As effect of each of these variables on the three inter-
the former head of new media at the Beacon Com- mediate behaviors that increase routine rigidity.
pany reflected, “I didn’t focus people on the threat, For example, structural autonomy lowers the ten-
especially those managing the new business. dency for a parent organization to assert authority
Where I did emphasize the threat was in working over a new venture. Structural autonomy expanded
with the print folks to get them off their butts and in the ability of venture management to run local ex-
arguing for resources.” Data from the other sites periments that would not have been possible in a
confirm that structure affects how managers per- world of business model and product templates.
ceive their motivation. Outside influence also helped expand the alterna-
The data show that the differentiated structures tives considered in the new ventures. One manifes-
helped to create environments where motivation tation of this was the previously noted impact out-
could be built entirely around the separate oppor- siders had on the choices of structure. But outside
tunity that existed for the online model. “When we influence also expanded the product and business
simply changed our name from the newspaper model ideas in these ventures. The external partner
name to ‘the city.com’ . . . it changed people’s ex- at the Beacon A proposed pricing and product cat-
pectations of what would be on the site. This, in egories that the newspaper had not thought of. As a
turn, changed how people in our online organiza- Web site editor who came from the newspaper
tion viewed who they were and what they were commented about the online director who was
producing.” The new president of the Beacon Com- brought in from outside: “He is constantly seeing
pany’s new media group observed, “Now that we digital media in different ways than I am used to or
are separate, we own the opportunity in a way we appreciate. At first, it bothered me, but now that I
never did when we were still with the newspaper.” see it working, I increasingly endorse the input.” In
Even as the concept of the opportunity was chang- addition, structural autonomy and a renewed op-
ing in the differentiated units, the parent organiza- portunity mind-set freed venture managers from a
tions remained focused on the threat to the core newspaper focus, because their responsibilities in
business; discussions there centered on cannibal- the parent newspapers were no longer immediate.
ization and the inevitable path of digital media. The four sites that differentiated structurally also
“This,” cautioned one CEO, “could be the death of substantially increased their innovation. Whereas
our entire franchise.” That threat perception re- the integrated sites continued to derive as much as
mained high in the newspaper organizations was a 90 percent of their Web site content from their
critical factor in overcoming resource rigidity, but sponsoring newspapers, all of the differentiated
 TABLE 8
Decreasing Routine Rigidity, 2000
Local
Content Market New
Managerial from Penetration Revenue
Newspaper Outside Influence Structure Framing Examples Printa Scoreb Categoriesc

Beacon A Outside partner, 1997; Differentiated Opportunity “Now that we are separate, we own the 45% 1.9 4
outside online CEO, emerges opportunity in a way we never did when we
1999 were still with the newspaper.” (new head of
new media)
Beacon B Outside advisor, 1998 Differentiated Opportunity “We were all set to let people buy ads 50% 1.4 4
emerges online. . . . The papers didn’t want to buy into
it. This is one area where we will do better as
a separate company.” (vice president,
technology and operations)
Press A Outside board member, Differentiated Opportunity “They just sit there and make us figure it out. 50%c 2.1 6
1997; outside online emerges They don’t make the decisions for us.” (vice
CEO, 1997 president, sales)
Press B CEO’s personal advisor, Differentiated Opportunity “Page views from the newspaper are now barely 35% 1.7 5
1994; outside business emerges more than one-third of the available pages on
plan consultant, 1995; our site. . . . The newspaper is one source of
online CEO, 1995 information, an important source. But we buy
our content from them like we buy it from
anywhere else.” (online editor)
Expositor A No key outside Integrated Threat “Functional reporting relationships are 75% 1.1 2
influence mentioned persists extremely time-consuming. It’s not just that
the groups think like the newspaper. It takes a
lot longer to make collective decisions.” (early
online president)
Expositor B No key outside Integrated Threat “We continue to see this as a way to protect 70% 0.9 3
influence mentioned persists classifieds, and that if we don’t do it someone
else will.” (vice president, product
development)
Morning News A No key outside Integrated Threat “Our basic strategy is an integrated strategy . . . 90% 1.6 2
influence mentioned persists In the local information market the newspaper
has an advantage. To separate the online unit
from the newspaper is to give away a lot of
that advantage.” (vice president, business
development)
Morning News B No key outside Integrated Threat Discussions with management remain centered 75% 0.5 2
influence mentioned persists on defending classified products.

a
Based on internal estimates and income statement analysis at each site.
b
Ratio of monthly Web site users to daily newspaper readers. Data collected from newspaper sites and checked against Media Metrix and Nielson NetRatings as well as circulation data
from the Audited Bureau of Circulation.
c
Based on 1998 comparisons to entrant income statements. Entrants were selected by new media experts and included Citysearch, Monster.com, Yahoo, i-village, and CNET. Six
categories of revenue were identified as new to the newspapers: (1) fee-based archival access, (2) e-mail marketing, (3) e-mail list rental, (4) fee-based data analysis, (5) behavioral targeting,
and (6) demographic targeting.
2005 Gilbert 757

sites borrowed only 50 percent or less of their con- example, Johnson (1988) showed that inertia is
tent from their sponsoring newspapers (see Table comprised of both a motivational determinant and
8). This evolution had a positive effect on market a procedural determinant. Using data from the re-
adoption of the new products; sites that had sepa- sponse of a clothing retailer who was threatened
rated and shifted their emphasis to the emerging with industry change, Johnson showed investment
opportunity enjoyed much higher local market can be highly motivated, but also deployed rigidly
penetration than the other sites in the sample. I through traditional business routines. Faced with
created a local market penetration score on the the performance decline of its core business in
basis of public data, calculating the ratio of “down-market niche” men’s wear, the retailer ini-
monthly Web site users to daily newspaper readers; tially sought to diversify into new markets. But it
A score of 1.0 implied that a Web site’s monthly then reapplied its existing business model, rather
users were equal to the newspaper’s daily readers; than adapting or repositioning its main business.
anything above that amount implied more users per Threat motivated resource commitment, but rou-
month relative to daily readership. Whereas the tines remained locked on the traditional business
differentiated sites averaged a score of nearly 1.8, model.
the integrated sites averaged barely 1.0 (see Table The current study identifies these unique deter-
8). The separated sites also introduced more inno- minants of inertia as resource and routine rigidity.
vation to their underlying business models, averag- Both constrain adaptation, but they have very dif-
ing close to five new categories of revenue, com- ferent underlying causal mechanisms. Data show
pared with just more than two new categories in the that resource rigidity stems from an unwillingness
integrated sites (see Table 8). Both the autonomy to to invest, while routine rigidity stems from an in-
focus on the separate business and the perception ability to change the patterns and logic that under-
of an independent opportunity seemed to facilitate lie those investments. The first relates to the moti-
greater divergence from the traditional routines of vation to respond, the second to the structure of
the core business. Thus, a final proposition can be that response. Recall that the issues regarding re-
stated: source rigidity in the data dealt with allocating
financial resources or management time to projects
Proposition 5. Outside influence, structural that supported the venture. These inertial forces
differentiation, and opportunity framing com- were very different from those related to the rou-
bine to relax routine rigidity in a new venture. tines and logics traditionally used to develop news
Data confirming this observation were also con- content and run the newspaper business. Both the
sistent across research sites (see Table 8). current study and the previous research by Johnson
(1988) reveal that not only are these determinants
of inertia different, but also that they can move
DISCUSSION independently. By analogy, resource rigidity is
concerned with movement along a line, while rou-
Differentiating the Structure of Inertia
tine rigidity deals with the trajectory of the line. A
In the introduction to this article, I noted the manager could invest aggressively and still fail to
apparent contradiction in the literature regarding adapt underlying routines. Sull (1999) called such
the impact of threat perception on organizational a pattern of behavior “active inertia.”
inertia. Whereas a number of scholars have ob- Ironically, a closer examination of the previous
served threat to be a catalyst that enables organiza- literature reveals that earlier measures of inertia
tions to overcome inertia (Barr & Huff, 1997; Cyert actually align with the observed categorizations of
& March, 1991; Lant et al., 1992), others have found resource and routine rigidity, even if the definitions
evidence that threat actually increases inertia (Dut- lacked the underlying specificity. Specifically, re-
ton & Jackson, 1987; Staw, Sandelands, & Dutton, examination of previous research shows that threat
1981). Acknowledging conflicting observations in decreases resource rigidity but increases routine
related research on risk, Sitkin and Pablo (1992) rigidity in a predictable, repeated pattern across the
hinted that the phenomenon itself might be under- previous studies. For example, the literature that
specified. They suggested that part of the confusion views threat as a catalyst to response typically mea-
stems from viewing outcomes along a single deter- sures behavior as a willingness to commit re-
minant of behavior and argued for a model based sources. Kahneman and Tversky’s (1984) study of
on a more complex set of determinants. Scholars risk showed that individuals are more often willing
have also observed the need for a more theoreti- to commit financial resources when they perceive
cally complex view of inertia in research on re- the issues of concern as being in the domain of loss
source commitment and organizational change. For rather than in the domain of gain. Mittal and Ross
758 Academy of Management Journal October

(1998) observed that individuals who had been gest that when discontinuities are led by noncore
given threat scenarios exhibited a significantly customers, established firms do not allocate re-
greater willingness to spend than participants who sources to a new business or technology. And yet
had been given opportunity scenarios. Similarly, the current study shows that threat perception can
the literature on strategic change suggests that lead to intense resource commitment, even in the
threat-driven response unlocks resources for in- absence of core customer demand—recall the rapid
vesting in new strategic initiatives (Lant et al., growth in expenditures observed in the present
1992). sample, which occurred despite the virtual lack of
This is very different from the view that threat is overlap between print and online customers. As
a constraint, as is seen in research that measures one sales representative noted at the Beacon A,
changes in organizational process and operational “Look, it didn’t make any sense for us to try to sell
logic. For example, Herman (1963) measured com- this stuff, but we began to feel that if we didn’t
munication and information patterns within an or- work on it, it might come back to haunt us.” Threat
ganization. Staw, Sandelands, and Dutton (1981) was the catalyst to overcoming resource rigidity.
and Nutt (1984) examined changes in the underly- The data also reveal that the ability to overcome
ing business logic of organizations. Thus, variables one type of inertia appears to increase problems
that measure a willingness to invest financial or with the other. The observed increase in routine
attention-based resources (Lant et al., 1992) have rigidity stems from three intermediate behaviors
been confused with variables that measure change that arise from threat-induced response— contrac-
in “dominant logic” (Prahalad & Bettis, 1986) or tion of authority, reduced experimentation, and fo-
change in operating routines (Staw et al., 1981). cus on existing resources (Propositions 2a, 2b, 2c).
The reason these subtypes of inertia move in dif- These behaviors were shown to be self-reinforcing.
ferent directions is that their underlying causal For example, threat perception led to a rigid focus
mechanisms differ, a factor that ironically shows on the existing business. This focus on existing
up in these previous measurement efforts in the resources was hardened by the aggressive pace of
relevant research. commitment, which created lock-in effects and re-
duced the ability to experiment. The aggressive
deployment of resources required increasing corpo-
Toward an Interpretive Model of Response to
rate oversight. And corporate leaders’ contraction
Discontinuous Change
of authority further reinforced the focus on the
Failure to recognize the difference between re- established business at the expense of the new op-
source rigidity and routine rigidity risks more than portunity (see Figure 1). These self-reinforcing be-
just contradictory results. When researchers do not haviors led to intense routine rigidity, causing
recognize this distinction, they may also fail to managers to adhere more closely to familiar rou-
observe important interactions between the two tines and behavioral patterns.
subtypes of inertia. For example, research on over-
coming inertia should focus on response motiva-
Observations on Structural Differentiation:
tion and resource commitment. And yet, while
Sources and Implications
variance perceptions and performance declines of-
ten decrease motivational constraints (Barr & Huff, Previous research has repeatedly demonstrated a
1997; Cyert & March, 1963), they can simulta- link between structural autonomy and innovation
neously increase constraints on the underlying (Christensen, 1997; Tushman & Anderson, 1986;
logic of an organization’s operating routines (John- Tushman & O’Reilly, 1996). The current study
son, 1988; Sull, 1999). By differentiating types of helps expand understanding of the forces that lead
inertia, one can focus on their unique implications to the decision to structurally separate a new ven-
for organizational response. I used my observations ture. The data show that outside influence shapes
from the study, formalized above as Propositions the choice to structurally differentiate (Proposition
1a– 4, to develop a longitudinal model that maps 3) and that structural differentiation cultivates an
how the perception of discontinuous change im- environment in which managers are more likely to
pacts overall inertia. Figure 1 contains this model. turn their attention to the independent opportunity
Propositions 1a and 1b support the notion that associated with a discontinuity (Proposition 4).
the perception of an imminent threat can unlock And while outside influence does appear to be
resource rigidity (Cyert & March, 1963; Lant et al., linked to the decision to separate a venture from its
1992). Note that threat perception leads to behavior parent, there are still questions as to why some of
that is different from the behavior observed by the newspaper companies studied here incorpo-
Christensen and Bower (1996). Their findings sug- rated such external influence, while others did not.
 FIGURE 1
An Interpretive Model of Inertia in Response to
Discontinuous Changea, b

a
Primary question variables are shaded gray.
760 Academy of Management Journal October

Note that the data do not seem to indicate that the ferent parts of an organization simultaneously—
sites that failed to harness outside influence had threat framing overcomes resource rigidity in the
access to outsiders and then ignored them. In the parent, while opportunity framing eases routine ri-
interviews with senior managers regarding the gidity in the autonomous venture. Structure’s decou-
choice of organizational structure, only those who pling role further reinforces the key contribution of
structurally separated their ventures mentioned the study: the recategorization of inertia into resource
outside influence. I asked the managers at the four and routine rigidity.
sites that remained integrated, “What outsiders did Alternative explanations. Taken together, these
you involve in your decision?” None of them men- observations suggest that the conflicting findings in
tioned outside individuals or organizations in their the literature regarding threat’s effect on inertia can
responses. Thus, the first relevant question appears be explained by the failure to differentiate between
to be how outsiders entered firm and individual resource and routine rigidity. An alternative expla-
networks. On this point the data are not entirely nation might simply be that the competing streams
conclusive, but all of the companies that eventually of findings are referring to different levels of threat
detached their online ventures seemed to draw perception. For example, the research describing
their outside influence from external networks: an threat perception as a catalyst might have exam-
outside friend of the CEO (the Press B); a board ined only moderate levels of threat, while the stud-
member and new hire (the Press A); an apparently ies that describe threat as a source of inertia might
serendipitous set of comparative successes with have been looking only at crisis situations. If so,
outside partners (the Beacon A); and an outside previous findings of a U-shaped relationship be-
advisor (the Beacon B). That the reach of formal tween stress and inertia might explain the observed
and informal networks would influence the firms’ differences—moderate levels of stress decrease in-
internal decision making is consistent with net- ertia, but extreme levels increase inertia (Yerkes &
work theories and team composition studies (Gelet- Dodson, 1908). However, one can derive reasons to
kanycz & Hambrick, 1997; Podolny, 2001; Stuart & rule out this alternative explanation both from the
Podolny, 1996;). This observation leads to ques- literature and from the data in the current study.
tions regarding how such nodes in a firm’s external First, the literature on threat as a catalyst contains
network develop and how they influence the firm. examples of extreme crisis (Mintzberg, Raising-
Influence may be a function of status (as indicated hani, & Théorêt, 1976; Papadakis, Kaloghirou, &
by the role of one CEO’s personal network) or of the Iatrelli, 1999). Moreover, the empirical literature on
intensity of an interaction (as in the case of the threat rigidity does not delineate levels of threat
external partner of the Beacon A). Given the origi- (Dutton & Jackson, 1987; Jackson & Dutton, 1988).
nal research questions and the design of this study, More importantly, a U-shaped relationship be-
these questions cannot be immediately answered. tween threat and inertia cannot explain the behav-
Nonetheless, the observed link between outside in- ior of different types of inertia. The data in this
fluence and structural choice should help sharpen study show settings of intense resource commit-
subsequent research on the sources of and mecha- ment that accompany low levels of routine adapta-
nisms leading to structural autonomy. tion. Further, if a U-shaped relationship were in
The study also provides a more refined view into play, one wouldn’t expect variance in behavior
the mechanisms by which structural autonomy when threat levels remain constant. The data in the
helps relax routine rigidity. The data confirm that study show that threat perceptions emerged be-
outside influence, structural independence, and tween 1996 and 1998 and remained high in all the
opportunity orientation combine to relax routine print organizations in my sample throughout the
rigidity and encourage innovation (Proposition 5). period of the study, both the innovative and non-
Conversely, when the companies studied here did innovative sites. Recall that even in the late periods
not access outside influence and remained inte- at the Beacon A, for example, the newspaper CEO
grated and focused on the threat to the parent organ- continued to say, “This could be the death of our
ization, the rigidity was perpetuated. Again, note that entire franchise.” Thus, it appears more likely the
the role of structural autonomy is consistent with variance in routine rigidity can be accurately
existing structural arguments regarding innovation linked to the changes in structure that enable firms
(Christensen, 1997; Tushman & O’Reilly, 1996). But to decouple the effects of cognitive framing be-
what makes the observations in this study unique is tween resource and routine rigidity.
how structure was seen to be the mechanism that Limitations and future research. Like any
decouples resource and routine rigidity. The data model intended to capture the complexity of an
show that structural autonomy allows threat and op- organizational response process, the model devel-
portunity cognition to have different impacts on dif- oped here has elements that need further explora-
2005 Gilbert 761

tion. For example, the study showed that the ob- influence and external networks. Moreover, the
served threat perception was largely a response to benefit of structural autonomy is more than simply
external analysts’ forecasts of the demise of the providing a separate setting for innovation. Struc-
newspaper industry. The question remains tural autonomy helps decouple the effects of cog-
whether threat-based motivation could have been nition on different types of inertia—separate struc-
triggered without this external pressure. A second ture allows threat perception to overcome resource
limitation in the study is that threat’s link to rigid rigidity in a parent company, while opportunity
behavior emerges in part because changes are dis- perception relaxes routine rigidity in a new ven-
continuous with a firm’s traditional routines. ture. Distinguishing between resource and routine
While these discontinuous settings helped specify rigidity not only helps explain response to discon-
the differences between resource and routine rigid- tinuous change, but also opens up future research
ity, they do impose boundary conditions on some exploration of the differences and interactions be-
of the findings regarding threat’s link to routine tween these categories of inertia.
rigidity. For example, the threat-induced behaviors These findings should also have relevance to man-
of contracted authority, reduced experimentation, agement practice. Managers can draw on the power
and focus on existing resources are problematic in threat has as a catalyst for commitment. And yet, this
discontinuous settings because they reinforce a study shows that response to discontinuous change
firm’s existing routines. When routines do not re- requires more than just the commitment of resources.
quire this nonlinear adaptation, threat-induced re- The underlying organizational routines that use those
sponse is not as likely to be as maladaptive. resources must also adapt when change is discontin-
Another limitation, which was noted earlier, is uous. These findings might encourage managers to
that the source of outside influence appears to be draw more heavily on their external networks to in-
linked to firm and CEO networks. But that obser- volve managers with experience outside of an exist-
vation is only indicative, not conclusive, and needs ing business. Managers might also structurally sepa-
to be tested more formally. Furthermore, questions rate their new ventures to restore opportunity frames
remain as to how a node in a network might be among venture managers while maintaining threat
more or less influential on a firm’s internal deci- framing in the parent organization.
sion process. I have presented potential mecha- In conclusion, the data show that threat percep-
nisms, such as status and intensity, that might be tion is a powerful interpretive force that affects firm
explored. Research should also look at whether the response to discontinuous change. And yet this
design of these networks is explicit or serendipi- influence is very different when one considers two
tous. Again, the evidence from the current study is distinct types of inertia: resource and routine rigid-
not conclusive. Finally, the current study reveals a ity. Recognizing these distinctions has been shown
tight correlation between outside influence, struc- to be significant for two fundamental reasons. First,
tural differentiation, and opportunity framing. the literature rarely recognizes these categories,
Each of these variables is also correlated with in- and by simply specifying their characteristics one
novation. The question remains as to whether these can better describe the underlying phenomenon of
variables can vary independently. inertia. Second, under conditions of discontinuous
change, not only are these types of inertia different,
but also, the mechanisms for overcoming one type
CONCLUSIONS
can amplify problems with the other. Further re-
This study began with an effort to unwind the search needs to be done on how and why some
structure of inertia. I was able to show that the firms are more likely than others to structurally
subcategories of resource and routine rigidity are decouple resource and routine rigidity, but I hope
discrete and have different causal mechanisms. Us- that these initial findings will open new paths of
ing this recategorization, researchers can resolve an inquiry and inform future research on inertia and
inconsistency in the literature concerning whether organizational change.
threat perception is a catalyst or a constraint to
discontinuous change. In building an interpretive
model of response, I saw how threat perception REFERENCES
releases constraints on resource rigidity while am- Arrow, K. J. 1962. Economic welfare and the allocation of
plifying problems with routine rigidity. Further, resources for inventions. In R. Nelson (Ed.), The rate
although previous studies have shown the benefits and direction of inventive activity: Economic and
of structural autonomy for innovation, the current social factors. 609 – 625. Princeton, NJ: Princeton
research broadens knowledge of what leads firms to University Press.
structurally differentiate by making links to outside Barr, P. S., & Huff, A. S. 1997. Seeing isn’t believing:
762 Academy of Management Journal October

Understanding diversity in the timing of strategic the negative domain. Group and Organization
response. Journal of Management Studies, 34: 337– Management, 21: 146 –162.
370. Henderson, R. 1993. Under-investment and incompe-
Benner, M. J., & Tushman, M. L. 2001. Exploitation, tence as responses to radical innovation: Evidence
exploration, and process management: The produc- from the photolithographic alignment equipment in-
tivity dilemma revisited. Academy of Management dustry. RAND Journal of Economics, 24: 248 –270.
Review, 28: 238 –256. Henderson, R., & Clark, K. 1990. Architectural innova-
Bower, J. L. 1970. Managing the resource allocation tion: The reconfiguration of existing product tech-
process. Boston: Harvard Business School Press. nologies and the failure of existing firms. Adminis-
trative Science Quarterly, 35: 9 –30.
Burgelman, R. 1994. Fading memories: A process theory
of strategic business exit in dynamic environments: Hermann, C. F. 1963. Some consequences of crisis which
Administrative Science Quarterly, 39: 24 –56. limit the viability of organization. Administrative
Science Quarterly, 8: 61– 82.
Christensen, C. M. 1997. The innovator’s dilemma. Bos-
ton: Harvard Business School Press. Huff, J. S., Huff, A. S., & Thomas, H. 1992. Strategic
renewal and the interaction of cumulative stress and
Christensen, C. M., & Bower, J. L. 1996. Customer power,
inertia. Strategic Management Journal, 13: 55–75.
strategic investment, and the failure of leading firms.
Strategic Management Journal, 17: 197–218. Jackson, S. E., & Dutton, J. E. 1988. Discerning threats and
opportunities. Administrative Science Quarterly,
Christensen, C. M., & Rosenbloom, R. S. 1995. Explaining
33: 370 –387.
the attacker’s advantage: Technological paradigms,
organizational dynamics and the value network. Re- Johnson, G. 1988. Rethinking incrementalism. Strategic
search Policy, 24: 233–257. Management Journal, 9: 75–91.
Cyert, R. M., & March, J. G. 1963. A behavioral theory of Kahneman, D., & Tversky, A. 1984. Choice, values, and
the firm. New York: Prentice-Hall. frames. American Psychologist, 39: 341–350.
Dutton, J. 1992. The making of organizational opportuni- Lant, T. K, Milliken, F. J., & Batra, B. 1992. The role of
ties—An interpretive pathway to organizational managerial learning and interpretation in strategic
change. In B. M. Staw & L. L. Cummings (Eds.), reorientation. Strategic Management Journal, 13:
Research in organizational behavior, vol. 14: 195– 585– 608.
226. Greenwich, CT: JAI Press. Lee, T. W. 1999. Using qualitative methods to organize
Dutton, J., & Jackson, S. 1987. Categorizing strategic is- research. Newbury Park, CA: Sage.
sues: Links to organizational action. Academy of Lee, T. W., Mitchell, T. R., & Sablynski, C. J. 1999. Qual-
Management Review, 12: 76 –90. itative research in organizational and vocational psy-
Eisenhardt, K. M. 1989a. Building theories from case chology. Journal of Vocational Behavior, 55: 161–
study research. Academy of Management Review, 187.
14: 532–550. Leonard-Barton, D. 1992. Core capabilities and core ri-
Eisenhardt, K. M. 1989b. Making fast strategic decisions gidities: A paradox in managing new product devel-
in high-velocity environments. Academy of Man- opment. Strategic Management Journal, 13: 111–
agement Journal, 32: 543–576. 125.

Feldman, M. S., & Pentland, B. T. 2003. Reconceptualiz- Levinthal, D. 1992. Surviving Schumpeterian environ-
ing organizational routines as a source of flexibility ments: An evolutionary perspective. Industrial and
and change. Administrative Science Quarterly, 48: Corporate Change, 1: 427– 443.
94 –118. March, J. G. 1991. Exploration and exploitation in organ-
Geletkanycz, M. A., & Hambrick, D. C. 1997. The external izational learning. Organization Science, 2(1): 71–
ties of top executives: Implications for strategic 87.
choice and performance. Administrative Science March, J. G., & Simon, H. A. 1958. Organizations. New
Quarterly, 42: 654 – 681. York: Wiley.
Gilbert, R. J., & Newberry, D. M. 1982. Preemptive pat- Miles, M., & Huberman, A. M. 1984. Qualitative data
enting and the persistence of monopoly. American analysis. Newbury Park, CA: Sage.
Economic Review, 72: 514 –526. Miller, D., & Friesen, P. H. 1980. Momentum and revo-
Glaser, B., & Strauss, A. L. 1967. The discovery of lution in organizational adaptation. Academy of
grounded theory. Chicago: Aldine. Management Journal, 23: 591– 614.
Hannan, M., & Freeman, J. 1977. The population ecology Mintzberg, H., Raisinghani, D., & Théorêt, A. 1976. The
of organizations. American Journal of Sociology, structure of “unstructured” decision processes. Ad-
82: 929 –964. ministrative Science Quarterly, 21: 246 –275.
Hartman, S. J., & Nelson, B. H. 1996. Group decision in Mittal, V., & Ross, W. T. 1998. The impact of positive and
2005 Gilbert 763

negative affect and issue framing on issue interpre- Sull, D. 1999. The dynamics of standing still: Firestone
tation and risk taking. Organizational Behavior and Tire and Rubber and the radial revolution. Business
Human Development, 76: 298 –324. History Review, 73: 430 – 464.
Nelson, R., & Winter, S. 1982. An evolutionary theory of Sutton, R. I., & Callahan, A. 1987. The stigma of bank-
economic change. Cambridge, MA: Harvard Univer- ruptcy: Spoiled organizational image and its man-
sity Press. agement. Academy of Management Journal, 30:
Noda, T., & Bower, J. L. 1996. Strategy making as iterated 405– 436.
processes of resource allocation. Strategic Manage- Teece, D. J., Pisano, G., & Shuen, A. 1997. Dynamic
ment Journal, 17: 169 –192. capabilities and strategy management, Strategic
Nutt, P. C. 1984. Types of organization decision pro- Management Journal, 18: 509 –553.
cesses. Administrative Science Quarterly, 29: 414 – Tripsas, M., & Gavetti, G. 2000. Capabilities, cognition,
451. and inertia: Evidence from digital imaging. Strategic
Ocasio, W. 1999. Towards an attention based view of the Management Journal, 21: 1147–1161.
firm. Strategic Management Journal, 18: 187–206. Tushman, M. L., & Anderson, P. 1986. Technological
Papadakis, V. M., Kaloghirou, Y., & Iatrelli, M. 1999. discontinuities and organizational environments.
Decision making: From crisis to opportunity: Busi- Administrative Science Quarterly, 31: 439 – 465.
ness Strategy Review, Spring: 29 –57. Tushman, M. L., & O’Reilly, C. 1996. Ambidextrous or-
Pfeffer, J., & Salancik, G. R. 1978. The external control of ganizations: Managing evolutionary and revolution-
organizations. New York: Harper & Row. ary change. California Management Review, 38(4):
8 –30.
Podolyny, J. M. 2001. Networks as the pipes and prisms
of the market. American Journal of Sociology, 107: Tushman, M. L., & Romanelli, E. 1985. Organizational
33– 60. evolution: A metamorphosis model of convergence
Prahalad, C. K., & Bettis, R. A. 1986. The dominant logic: and reorientation. In L. L. Cummings & B. M. Staw
A new linkage between diversity and performance. (Eds.), Research in organizational behavior, vol. 7:
Strategic Management Journal, 7: 485–501. 171–122. Greenwich, CT: JAI Press.

Reinganum, J. F. 1983. Uncertain innovation and the Tushman, M. L., Smith, W., Wood, R., Westerman, G., &
persistence of monopoly. American Economic Re- O’Reilly, C. 2003. Innovation streams and ambi-
view, 73: 741–748. dextrous organizational designs: On building dy-
namic capabilities. Working paper no. 03-106, Har-
Ross, J., & Staw, B. M. 1993. Organizational escalation vard Business School, Boston.
and exit: Lessons from the Shoreham nuclear power
plant. Academy of Management Journal, 36: 701– Yerkes, R. M., & Dodson, J. D. 1908. The relation of
732. strength of stimulus to rapidity of habit formation.
Journal of Comparative and Neurological Psychol-
Santayana, G. 1905. The life of reason. New York:
ogy, 18: 459 – 482.
Charles Scribner & Sons.
Yin, R. K. 1994. Case study research: Design and meth-
Schein, E. H. 1985. Organizational culture and leader-
ods (2nd ed.). Newbury Park, CA: Sage.
ship. San Francisco: Jossey-Bass.
Sigglekow, N. 2001. Change in the presence of fit: The
rise, the fall, and the renaissance of Liz Claiborne.
Academy of Management Journal, 44: 838 – 867.
Sitkin, S. B., & Pablo, A. L. Reconceptualizing the deter- Clark Gilbert (cgilbert@hbs.edu) is an assistant profes-
minants of risk behavior. Academy of Management sor of entrepreneurial management at the Graduate
Review, 17: 9 –39. School of Business at Harvard University. He earned
Staw, B. M., Sandelands, L., & Dutton, J. 1981. Threat his DBA in the area of business policy at Harvard in
rigidity effects in organizational behavior. Adminis- 2001. His current research interests include corporate
trative Science Quarterly, 26: 501–524. innovation, discontinuous change, and strategy adap-
tation in new ventures.
Stuart, T. E., & Podolny, J. M. 1996. Local search and the
evolution of technological capabilities. Strategic
Management Journal, 17: 21–38.