Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
1)
Foundations:Growth versus Governance Brokerage versus Closure
Appendices:I. Example Network Questionnaire for a Web Survey (pages 42-45, from 2010 Neighbor Networks, 2017 Management and Organization Review)II. Measuring Network Closure/Embedding (page 46, from 2007 "Closure & Stability)III. Closure/Embedding and the Theory of the Firm (page 47, from 1992 Structural Holes, 1924 Legal Foundations of Capitalism)IV. Measuring Access to Structural Holes (pages 48-54, from 1992, Structural Holes, 2010 Neighbor Networks)V. Closure and Learning Curves (pages 55-59, from 1919 Psychological Monographs, 1965 Review of Economics and Statistics, 1992, Upside, 1998
Perspectives on Strategy, 1999 Organizational Learning)VI. Snipits on Business Culture (pages 60-62, 1998 Financial Times, other)VII. Network Endogeneity -- Bavelas-Smith-Leavitt experiments (pages 63-64, 1949 Leavitt dissertation, 1951 Leavitt, “Some effects of certain com-
munication patterns upon group performance”)
This handout was prepared as a basis for discussion in executive education (Copyright © 2018 Ronald S. Burt, all rights reserved). To download work referenced here, or research/teaching materials on related topics, go to http://faculty.chicagobooth.edu/ronald.burt.
For text on this session, see Chapters 1, 2, and 3 in Brokerage and Closure (including adjunct bits from Neighbor Networks).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
2)
from Burt (2018, Structural Holes in Virtual Worlds).
CEO C-Suite Heir Apparent
Other, RespondentOther, NonRespondent
Bill
Bob
Sociogram of Formal Network in a Large EU Healthcare Company (Org Chart)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
3)Social Networkat the Top of theCompanyLines indicate frequent and substantive work discussion; heavy lines especiallyclose relationships.
CEO
C-Suite
Heir Apparent
Other Senior Person
Bill
B
B
B
B
B
BB
Bob
Asia US
EU andEmerging Markets
R&D
FrontOffice
Back Office
Figure 1 in Burt and Soda, "The social origins of great strategies" (Strategy Science, 2017)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
4)Coordination across groups is the source of competitive advantage from social networks — an advantage often difficult to realize, as illustrated by the below opinions from a senior executive education program. The difficulty most often cited as needing to be improved: Coordination across groups.
SURVEY QUESTION: Every person in the [company] program has a perspective on [company] operations and clients. Drawing on your experience, what would you like to see changed within [the company] to increase the company's current value as a provider of high-quality, effective, attractive service to clients? RESPONSES: At least two-thirds of responses are variations on "improve coordination across groups." For example (idea ratings are based on independent pile-sort evaluations by the head of HR and a regional P&L leader):
HIGH RATED: We need a "DNA" change - we are so deeply ingrained in the P&L, near-term financial results mindset that we limit our potential for changing the world. It goes beyond P&L reporting or annual goals - we need more people talking and thinking and acting and making decisions that are focused on one voice, delivering for clients and colleagues. We are not a "clients first" firm - we say we are, but we are not. We are a Wall Street/investor first firm. Need to change that.
HIGH RATED: (1) Improve team work in global network. Global clients want to see us acting more as one team (e.g. global P&L for TOP Global clients). (2) Better understanding of capabilities and solutions of other business units, and improving colleague network cross business units (one voice). (3) Best practice sharing cross countries, esp. for same roles/job titles to copy/paste success stories and solutions for our clients. (4) Improve incentive/bonus structure to drive x-Selling (in country between BUs, also globally within BUs). (5) Better understanding of [company]'s global capabilities and whom to talk to for a special solution.
HIGH RATED: (1) “One Client View." Ability to easily understand [company]-wide relationships with any client or prospect in the world. Comprehensive, single instance client account data system. Ability to pull up information on a client (any business unit) from one system. Data should include: revenue, oppty amount, current solutions provided, account owner(s), history of account, etc. (2) Connect the dots between business initiatives and commitments, and finance/budget. When the business makes commitments to do something, sometimes we need to do a better job at ensuring those commitments are resourced appropriately (otherwise we end up with frustrated and over-worked colleagues with compensation not commensurate with their scope and/or efforts).
ILLUSTRATIVE RELATED IDEAS: - More coordination among different parts of the company. A better understanding of each other's business and the value it brings to clients. A clearer internal structure that allows us to deliver to clients the best of [company] with internal P&L driven issues.
- Continue to endeavor to remove our P&Ls as barriers to cross-practice collaboration. - A far more joined-up [company]. We don't speak to client yet routinely about our true capability across [company]. We are
moving in that direction, but this would be a genuine game changer and add material value. Not a new idea I know!- Better integrated client management activities across the operating groups.
- Find an easier way to work across businesses and geographies to really bring the best of [company] to all clients.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
5)
Rule 1 of Social Capital
CEO
C-Suite
Heir Apparent
Other Senior Person
Bill
B
B
B
B
B
BB
Bob
Now the Social Network
Lines indicate frequent and substantive work discussion; heavy lines especially close relationships.
Asia US
Front Office
R & D
Back Office
EU and Emerging Markets
RULE 1: For bottom-line growth, closed networks facilitate and maintain trust and reputation within the network, promoting reliable, efficient operations within the network (Sherif, 1935; Festinger et al., 1950; Asch, 1951; Granovetter, 1985, 1992; Burt, 1987; Coleman, 1988; Ellickson, 1991; Bernstein, 1992, 2001; Krackhardt, 1992; Barker, 1993; Burt & Knez, 1993; Burt, 2005; Putnam, 1993; Uzzi, 1997).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
6)
Trust Builds withinRelationships Slowly
TRUST — committing to an exchange before you know how the other will behave.
REPUTATION — extent to which you are known as trustworthy.
I. Good Behavior as theSource of Trust
third parties irrelevant to trust & distrusttoo slow (graph to right), too dangerous
(Burt, 1999, "Private games are too dangerous")
II. Network Closure and Structural Embedding as the Source,
Bandwidth Storythird parties enhance information and
enforcement, and so facilitate trust (next page)
from Figure 3.2 in Brokerage and Closure
JJ
J
J J
J
J
JJ J
J J
J JJ
J
J J
J
J
J J J
J
JJ
JJ
JJJ J
J
JJ
J JJ
J
J
JJ J
J
JJ J
J
J
JJ
J
J J
J
J J
JJ
J
(710)10 ormore
Percent citedfor trust
Percent cited for distrust
(378)1 orless Years Known
(number of relationships in parentheses)
(370)2
(395)3
(121)9
(183)8
(145)7
(195)6
(275)5
(243)4
(170)(33) (76) (48) (85)(37)(48)(46)(82)(31)
(509)(782) (576) (311) (43)(89)(85)(180)(198)(190)
40%
30%
20%
10%
30%
20%
10%
30%
20%
10%
217 Staff Officersin Two Financial Services Firms
Cite 3,324 Colleagues
60 Senior Managers in a Chemicals FirmCite 656 Colleagues
284 Senior Managers in anElectronics & Computer Firm
Cite 3,015 Colleagues
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
7)
Closure creates "bandwidth:" more channels of communication allow more accurate and rapid communication, so poor behavior is
more readily detected and managed.
1985: Granovetter (1985 AJS) on the risk of trust reduced by third-party enforcement (discussed as structural embeddedness, 1992:44): "My mortification at cheating a friend of long standing may be substantial even when undiscovered. It may increase when the friend becomes aware of it. But it may become even more unbearable when our mutual friends uncover the decit and tell one another." (also Tullock, 1985 QJE, pp. 1076, 1080-1081)
1988: Coleman (1988:S107-108 AJS, 1990 book) on the risk of trust reduced by third-party enforcement (discussed as network closure) with respect to rotating-credit associations: "The consequence of closure is, as in the case of the wholesale diamond market or in other similar communities, a set of effective sanctions that can monitor and guide behavior. Reputation cannot arise in an open structure, and collective sanctions that would ensure trustworthiness cannot be applied." E.g., Putnam's (1993 book) explanation of higher institutional performance in regional Italy attributed to the trust, norms, and dense networks that facilitate coordinated action.
1989: Maghribi traders in North Africa during the 1000s, respond to strong incentives for opportunism in their trade between cities by maintaining a dense network of communication links which encouraged them to protect their positive reputations and facilitated their coordination in ostracizing merchants with negative reputations (Greif, 1989 JEH; and for other applications, such as guilds, see Greif, 2006, Institutions and the Path to the Modern Economy).
CLOSURE — the lackof structural holeswithin a network
Third Parties Are an Early-Warning System that Protects Nice from
Nasty in the Initial Games of a Relationship.
Third parties enhance communication and enforcement, and so
create reputation costs which facilitate trust.
For discussion, see pages 127-130
in Brokerage and Closure, and for detailed discussion
with respect to specific markets, see Lisa Bernstein
on diamonds, cotton, and supplier relations
(respectively 1992 Journal of Legal Studies, 2001 Michigan
Law Review, and 2016 Journal of Legal Analysis).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
8)
Figure 3.1 in Brokerage and Closure (for discussion, see pages 105-111). See Appendix III on network embedding in the theory of the firm.
Robert Jessica Robert Jessica Robert Jessica
Situation ARobert New Acquaintance
(no embedding)
Situation BRobert Long-Time Colleague
("relational" embedding)
Situation CRobert Co-Member Group
("structural" embedding)
More connections allow more rapid communication, so poor behavior can be more readily detected and punished. Bureaucratic authority was the traditional engine for coordination in organizations (budget, head count). The new engine is reputation (e.g., eBay). In flattened-down organizations, leader roles are often ambiguous, so people need help knowing who to trust, and the boss needs help supervising her direct reports. Multi-point evaluation systems, often discussed as 360° evaluation systems, gather evaluative data from the people who work with an employee. These are "reputational" systems in that evaluations are the same data that define an employee's reputation in the company. In essence, reputation is the governance mechanism in social networks.
Closed Networks within the ClustersFacilitate Trust and Shared Beliefs
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
9)
from Burt (2018, Structural Holes in Virtual Worlds). See Appendix II on measuring network closure/embedding.
Prob
abili
ty th
at R
elat
ions
hip
is C
ited
Nex
t Yea
r as
Goo
d or
Out
stan
ding
All Colleagues(z = 14.88)
Number of Third PartiesLinking Employee
with Colleague this Year
10+
A. Analysts and Investment Bankers
Mean Number of Third PartiesConnecting Banker with
Colleagues This Year
Mea
n C
orre
latio
n fo
rB
anke
r’s R
eput
atio
nfr
om th
is Y
ear t
o N
ext
(13-
pers
on s
ubsa
mpl
e)
Bold line through black dots describes aboveaverage reputations (8.1 routine t-test). Dashed
line through hollow dots describes reputationsaverage and below (6.1 routine t-test).
banker banker
1
2
3
4
1 2 3 4
1
2
3
4
1 2 3 4
10+
B. Investment Bankers
“Reputation cannot arise in an open structure.”(AJS, Coleman, 1988:S107)
Closure-Trust Associations, ManagementDots are average Y scores within intervals of X. Graph A describes 46,231 observed colleague relations with analysts and
bankers over a four-year period (adapted from Burt, 2010: 174-175). Vertical axis is the proportion of relations cited next year as good or outstanding. Horizontal axis is number of mutual contacts this year. Logit z-score test statistics are estimated with controls for differences in network size and adjusted for autocorrelation between relationships (Stata "cluster" option). Graph B describes for the bankers subsample correlations between positive (above average) and negative (below average) reputa-tions this year and next year (adapted from Burt, 2010:166; routine t-tests reported across 1,179 banker-year observations).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
10)
More
EverQuest II prediction from closure in social network (n = 216,677) Y = 5.23 X - .287 X2 + other (30) (-19)
EverQuest II prediction from closure in economic network (n = 199,118) Y = 1.31 X - .058 X2 + other (20) (-11)
Second Life prediction from closure in social network (n = 2,218,770) Y = 5.17 X - .276 X2 + other (104) (-56)
And the Same Holds for Online Social RelationshipsDots are average Y scores within intervals of X. Second Life trust is friendship rights granted to contact as predicted in Table 3.1 by Model 2. EverQuest II trust is housing rights granted to contact as predicted in Table 3.2 by Model 4 for social relations and Model 5 for economic relations. Standard errors in parentheses are adjusted for autocorrelation between relations from same character using STATA “cluster” option.
from Burt (2018, Structural Holes in Virtual Worlds).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
11)
The Learning Curve: Build for Network Closure to Cut Costs, Delivering on a Known Value StreamLEARNING CURVE (also known as experience curve) — increased efficiency associated with cumulative volume produced by group (e.g., timing & locating supplies, scheduling, tacit knowledge between colleagues, etc.). THE MECHANISM — With its dense social ties providing wide bandwidth for information flow, closure enhances communication and enforcement within a group, (1) which creates reputation, facilitating trust within a group division-of-labor, (2) which enhances performance as people become self-aligning between tasks, pushing one another to extraordinary efforts down the learning curve. The result is lower costs, and so higher productivity. Reputation is the engine. Closure delivers value through peer pressure on reputation within a group (else exogenous shocks disrupt the alignment of even personally dedicated individuals).
“Costs characteristically decline 20 to 30 percent in real terms each time accumu-lated experience doubles. This means that when inflation is factored out, costs should always decline.”
Associated with BCG and Bruce Henderson (1974, “The experience curve reviewed: why does it work?” reprinted in Stern and Stalk, 1998, Perspectives on Strategy), but more with Liberty Ships, e.g., Rapping, "Learning and World War II production functions"(1965, Review of Economics and Statistics) and Argote et al., "The persistence and transfer of learning in industrial settings" (1990, Management Science). Also see Thurstone "The learning curve equation" (1919, Psychological Monographs). For review of industrial research largely preceding Henderson, see Yelle "The learning curve" (1979, Decision Sciences). For discussion, see Appendix V on closure and example learning curves.
$0.40
$0.60
$0.80
$1.00
1 2 4 8
Ave
rag
e C
ost
per
Un
it
Cumulative Unit Volume
80¢
64¢
51¢
20% cost reductionwith each
doubling of volume
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
12)
Bottom-Line Performance Advantage of Closed Networks: Reputation Mechanism Generates Trust and Efficiency
By creating a wide bandwidth for information flow, closure enhances communication and personal visibility within a group,
(1) which creates reputation costs for individuals who express opinion or behavior inconsistent with group standards,
(2) which makes it less risky to trust within the group,
(3) which enhances productivity as people become self-aligning in extraordinary efforts (lowering costs for labor, monitoring, quality, and speed).
Reputation is the mechanism by which closure has its effect. Closure delivers value by creating a reputation cost for deviation from cooperative, extraordinary effort. In other words, closure grows the bottom line. As illustrated by the examples just discussed, you often see closure in the teamwork associated with successful efficiency programs such as TQM, SixSigma, and Lean Manufacturing.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
13)
Stratified Random Sample of 700 Chinese Entrepreneurs from
Seven Cities in Three Provinces of China’s Yangtze River Delta
Region. The three provinces account in 2013 for 20.2% of China’s GDP and 31.9%
of China’s foreign trade.
Sample Characteristics N % Small (10 - 100) 468 67% Medium (101 - 300) 169 24% Large (> 300) 63 9% Textile 170 24% Transportation Equipment 171 24% Machinery 180 26% Pharmaceutical 77 11% Electronics 102 15% Respondent is Founder 559 80% Year Born 1967 median, 8.4 sd, 1938-1988 Yr Founded 2001 median, 4.6 sd, 1982-2011
The map is taken from the Wikipedia entry for “Yangtze River Delta” with the delta proper indicated in green. Bold lines separate provinces. Bars indicate small, medium, and large firms in the sample 100 entrepreneurs from each city (respectively, light, dark grey, and black areas of city bar).
Shanghai
(municipality)
Nanjing (capital)
Changzhou
Hangzhou
(capital)
Wenzhou
Ningbo
Nantong
Jiangsu Province
Zhejiang Province
ShanghaiProvince
Figure A1 in Burt and Burzynska, "Chinese entrepreneurs, social networks, and guanxi," (2017 Management and Organization Review)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
14)
Figure 4 in Burt and Burzynska, "Chinese entrepreneurs, social networks, and guanxi," (2017 Management and Organization Review)
Network Closure and Trust in China
NOTE — Dots are average Y scores at each level of X. Graph A describes 46,231 observed colleague relations with analysts and bankers over a four-year period (adapted from Burt, 2010:174-175). Vertical axis is the proportion of relations cited next year as good or outstanding. Horizontal axis is number of mutual contacts this year. Graph B describes 4,463 relationships cited by the 700 Chinese entrepreneurs. Vertical axis is mean respondent trust in the contact, measured on a five-point scale. Horizontal axis is the number of other people in a respondent’s network connected with the contact being evaluated for trust. Test statistics are estimated in both graphs with controls for differences in network size and adjusted for autocorrelation between relationships (Stata "cluster" option, see Table 4 for estimates with further controls).
Pro
bab
ilit
y t
hat
Rela
tio
nsh
ip is C
ited
Next
Year a
s G
oo
d o
r O
uts
tan
din
g
All Colleagues
(z = 14.88)
Continuing Colleague
(first cited two years
ago, z = 0.81)
10+
A. Western Analysts
and Bankers
6+
Resp
on
den
t E
valu
ati
on
of
Tru
st
in C
on
tact
(1 f
or
su
sp
ect,
5 f
or
co
mp
lete
tru
st)
B. Chinese
Entrepreneurs
Network Closure
Number of Third Parties
Linking this Year
Evaluator with Evaluated
Network Closure
Number of Third Parties
Linking Respondent
with Contact
All Contacts
(t = 22.56)
Founding Contact (t = 3.21)
Other Event Contact (t = 10.09)
NonEvent Contact (t = 22.73)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
15)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 1 2 3 4 5 6 7 8 9 10
Bridge relationships(employee and colleaguehave no mutual contacts)
Relationships embedded in closed network of one or more mutual colleagues
One of the 25%most embeddedrelationships
Relationship Duration(in years, up through this year)
Prob
abili
ty th
at R
elat
ions
hip
Dec
ays
befo
re N
ext Y
ear
(rel
atio
n ci
ted
this
yea
r is
not c
ited
next
yea
r)
Over the Longer Run, Closure Slows Decay,
Especially in New Relations.
Here, closure has its effect through the
first two years of a relationship.
These are decay functions for colleague relations
with investment bankers and analysts during the
1990s. Logit z-scores in parentheses below (based
on 46,231 relations).
from Figure 4.8 in Brokerage and Closure. For general discussion of structural embedding primarily facilitating the formation of relations rather than their long-term survival, see Dahlander & McFarland (2013 ASQ), "Ties that last: tie formation and persistence in research collaborations over time."
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
16)
Same Network Mechanisms in Different Composition, Make for Different Business Environments
Among the Western Analysts and Bankers,
1,233 Ties Are Guanxi of 13,780 at Risk of Being Guanxi
Among the Chinese Entrepreneurs, 2,905 Ties Are Guanxi
of 4,464 at Risk of Being Guanxi
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
17)
Summary,Tie Strength
In and Beyondthe Network
Around a T-Shaped
Manager("strong relation" is an
aggregate of trust, responsibility, cooperation, coordination)
Inside the network (dark dots — ego network, personal network, first-order zone): Relational Embedding: Relations stronger with frequent contact over many years, some of which develop into guanxi. Structural Embedding: Relations stronger with many mutual contacts, i.e., within closed network. RULE 1: Close the network to promote trust, responsibility, cooperation, coordination (i.e., efficiency on known task).
Beyond the network (hollow dots — friends of friends, neighbor networks, second-order zone): Friends of Friends: Relations stronger with friends of closest contacts. Homophily: Relations stronger with people of distinguishing attributes (attributes you share that few others share).
Further Out (strangers undistinguished by attributes [no homophily], with no known connections into your network): The more connected the inside, the more suspicious the outside.
you
Central tenet in network analysis: How to do a thing depends on the social context in which you do it.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
18)
Prob
abili
ty o
f Clo
se F
eelin
gsbe
twee
n th
e Tw
o M
anag
ers
Close Feelings withina Person’s Network(N = 1,390; t = 8.60)
Mutual Colleagues (Third Parties)Linking Two Managers
Close Feelings withinand Across Networks
(N = 78,210; z = 237.33)
Third Parties Freq. Percent --------+-----------------------
1 | 74,856 95.71 2 | 2,155 2.76 3 | 721 0.92 4 | 285 0.36 5 | 133 0.17 6 | 35 0.04 7 | 17 0.02 8 | 5 0.01 9 | 3 0.00
-------+----------------------- Total | 78,210 100.00
Number of Relations
Averaged in Hollow-Dot Line
Positive Sentiment Associated with ClosureThese data come from a network survey of senior management in a large financial organization. The horizontal axis is a count of the number of people found to connect respondent and colleague. The line through the solid dots describes the probability that a survey respondent says he or she feels emotionally close to a colleague cited as a "frequent and substantive work contact." This is a probability within the network around a respondent. The line through the hollow dots describes the probability two people in the 396-person population feel emotionally close, given the strength of their connection through mutual colleagues (connection between i and j through mutual colleagues k is ∑k zikzkj, where zkj is the strength of connection strength between employees k and j, 0 ≤ zkj ≤ 1, which is rounded to the nearest integer here).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
19)
Homophily: Relations develop more easilyas a by-product of felt similarity. When two people feel they are socially similar on certain issues, they can more easily interact on other issues potentially difficult.
From Figure 3 in Wojcieszak & Mutz (2009, J of Communication), “Online Groups and Political Discourse: Do Online Discussion Spaces Facilitate Exposure to Political Disagreement?” For a review of relations associated
with people sharing similar backgrounds or interests, see McPherson, Smith-Lovin & Cook (2001, Ann. Rev. Sociology), “Homophily in social networks” (usual dimensions are people in the same place at the same time, same
age, gender, religion, occupation, income, social class). For some tactical guidance on your network, see Uzzi & Dunlap (2006, HBR), “How to build your network,” and Cassario et al. (2016, HBR), “Learn to love networking.”
Graph shows that the potential for discussion across political differences occurs primarily in online groups where politics is not the purpose of the discussion space. Horizontal is purpose of discussion space. Vertical axis is an index of extent to which space draws many users, often discussing politics, and encountering high levels of political disagreement. Leisure includes groups based on shared hobbies/activities, social support, socializing, romance, fan groups for a TV show, actor, musical group, or sports team, and general trivia groups. Responses are from a national probability sample of 1028 people who report participating in one or more chat rooms or message boards.
Active (vs Passive)StructuralEquivalence:highway experimentsingles bar rhetoric
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
20)
From Figure 4 and Table 1 in Opper, Burt, and Holm (2018), "Social network and cooperation with strangers."
Network Closureand
Cooperationwith Strangers
The more closed the inside, the more suspicious the outside,
Especially for people who have been successful with a closed network.
A Behavioral Measure of Cooperation “Like you, the other player is CEO of a Chinese firm, and a citizen of China.”
Move by Other Player
Your Move: Cooperate Defect
Cooperate 250, 250 50, 400
Defect 400, 50 100, 100
Network Closure(measured by network constraint)
Observations are averages for 5-point intervals on X, with tails of X truncated for infrequency. Correlation is computed from
data in the graph. Solid/hollow dots are averages for more/less successful entrepreneurs (respectively distinguished by
above/below median profit last year).
r = -.87
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
21)
Rule 2 of Social Capital
CEO
C-Suite
Heir Apparent
Other Senior Person
Bill
B
B
B
B
B
BB
Bob
Now the Social Network
Lines indicate frequent and substantive work discussion; heavy lines especially close relationships.
Asia US
Front Office
R & D
Back Office
EU and Emerging Markets
RULE 2: For top-line growth, large open networks facilitate innovation and achievement via information breadth, timing, and arbitrage advantages from bridging structural holes between closed networks (Granovetter, 1973; Freeman, 1977; Burt, 1980, 1992, 2005; Lin et al., 1981; Lin, 2002; Cook et al., 1983; Gould & Fernandez, 1989, 1994; Aral & Van Alstyne, 2011).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
22)
To begin, the "network" around a person is a pattern of relationships with and between colleagues.This worksheet is completed in four steps:
(1) In the oval, write the name of a significant colleague. The colleague could be your most valuable subordinate, your most difficult, your boss, an important source of support, or a key contact in another organization. Who doesn't matter. It just has to be someone you know well enough to know their key contacts.
(2) In the squares, write the name of the five contacts with whom the person in the oval has the most frequent and substantial business contact.
(3) Draw a line between any pair of contacts that are connected in the sense that the two people speak often enough that they have some familiarity with current issues in one another's work.
(4) Compute network density. Count the number of lines between contacts (TIES). Divide by the number possible (n[n-1]/2, where n is the number of contacts, which is 5 if you entered five contacts). Multiply by 100 and round to nearest percent.
DENSITY = _____________
SOCIOGRAMgraphic image of
a network in which dots represent
nodes (a person, group, etc.) and lines represent
connections
Appendix I contains an illustrative survey webpage used to gather network data.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
23)
Robert JamesFrom Figure 1.8 in Brokerage and Closure. Data pooled across eight
study-population graphs in Appendix IV on measuring network constraint.
Circles are average z-score performance (Z) for a five-point interval of network constraint (C) within each of eight study populations. Dashed line goes through mean values of Z for
intervals of C. Bold line is performance predicted by the natural log of C.
Social Capital of Brokerage Manifest as better ideas, more-positive evaluations, higher compensation,
earlier promotion, and faster teams.
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
5 15 25 35 45 55 65 75 85 95
Network Constraint (C)many ——— Structural Holes ——— few
Z-S
core
Rel
ativ
e P
erfo
rman
ce(c
ompe
nsat
ion,
eva
luat
ion,
pro
mot
ion)
First, establish the empirical fact that the people we will discuss as "network brokers" enjoy achievement and rewards higher than their peers.
Brokers are to the left on the horizontal axis contrasting open with closed networks.
small, closedlarge, open
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
24)
Social Capital of Brokerage Manifest as better ideas, more-positive evaluations, higher compensation,
earlier promotion, and faster teams.
E
E
E
E
E
E
E
E
E
E
E
E
E
EEE
E
EE
E
EE
E
E
E
E
EE
EE
E
E
EE
E
E
EEE
E
EE
EE
E
E
E
E
EE
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
EE
E
EEE
E
E
E
EE
E
E
E
E
E
E
E
E
E
E
EE
E
E
E
EE
E
E
E
E
E
E
E
E
J
JJ
J
J
J J JJJJ
JJ
J
J J
J
J
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
5 15 25 35 45 55 65 75 85 95
Z = 2.78 - .82 ln(C)r = -.53
Network Constraint (C)many ——— Structural Holes ——— few
Z-S
co
re R
ela
tive P
erf
orm
an
ce
(com
pensation, evalu
ation, pro
motion)
median network constraint (35 points)
From Figure 1.8 in Brokerage and Closure. Data pooled across eight study-population graphs in Appendix IV on measuring network constraint.
Circles are average z-score performance (Z) for a five-point interval of network constraint (C) within each of eight study populations. Dashed line goes through mean values of Z for
intervals of C. Bold line is performance predicted by the natural log of C.Robert James
Per
form
ance
Ind
icat
or
(com
pens
atio
n, e
valu
atio
n, p
rom
otio
n ra
te)
Human Capital et al.(e.g., job rank, age, geography, kind of work, division, education, etc.)
performancelower thanexpected
performancehigher thanexpected
Achievement and rewards are distinguished on the vertical axis,
measuring the extent to which a person is doing better than his or her peers.
small, closedlarge, open
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
25)
Graph is from Figure 1.8 in Brokerage and Closure. Data are pooled across eight management populations. Pie charts are from Figure
2.4 in Neighbor Networks. On causal order, see Appendix VII.
Robert James
Social Capital of Brokerage Manifest as better ideas, more-positive evaluations, higher compensation,
earlier promotion, and faster teams.Brokerage is a large percentage of explained performance differences.
E
E
E
E
E
E
E
E
E
E
E
E
E
EEE
E
EE
E
EE
E
E
E
E
EE
EE
E
E
EE
E
E
EEE
E
EE
EE
E
E
E
E
EE
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
EE
E
EEE
E
E
E
EE
E
E
E
E
E
E
E
E
E
E
EE
E
E
E
EE
E
E
E
E
E
E
E
E
J
JJ
J
J
J J JJJJ
JJ
J
J J
J
J
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
5 15 25 35 45 55 65 75 85 95
Z = 2.78 - .82 ln(C)r = -.53
Network Constraint (C)many ——— Structural Holes ——— few
Z-S
co
re R
ela
tiv
e P
erf
orm
an
ce
(co
mp
en
satio
n,
eva
lua
tion
, p
rom
otio
n)
median network constraint (35 points)
55%
17%
28%
10%
81%
9%
64%2%
33%
Brokerage Contributes "Slightly More than Half" of Predicted Variance in Performance Differences between Managers:
Network constraint (white), job rank (red), and other factors (striped). First pie is investment banker compensation and analyst election to the All-America Research Team. Second pie is supply-chain and HR manager compensation in corporate bureaucracies. Third pie is early promotion to senior job rank in a large electronics firm.Circles are average z-score performance (Z) for a five-point interval of network constraint
(C) within each of eight study populations. Dashed line goes through mean values of Z for intervals of C. Bold line is performance predicted by the natural log of C.
small, closedlarge, open
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
26)
Returns to Brokerage Aggregate to Companies, Industries, and Communities
People with phone networks that span structural holes live in communities higherin socio-economic rank Networks are defined by land-line & mobile phone calls (map to left). Socio-economic rank is UK government index of multiple deprivation (IMD) based on local income, employment, education, health, crime, housing, and environmental quality (graph below). Units are phone area codes.
figures from Eagle, Macy, and Claxton (2010, Science), “Network diversity and economic development”
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
27)
Returns to Brokerage Are Evident Online in theNetwork-Achievement Connection within Virtual Worlds
Dots are average Y scores within integer (left) or five-point (right) intervals on horizontal axis. EverQuest II achievement variable is the predicted character level in Model 8, Tables 3.4 and 3.5. Second Life achievement is the canonical correlation dependent variable in Model 15, Tables 3.5 and 3.6.
Effective Size(Number of NonRedundant Contacts)
Pred
icte
d A
vata
r Z-s
core
Ach
ieve
men
t
Network Constraint (x 100)
Pred
icte
d A
vata
r Z-s
core
Ach
ieve
men
t
Second Life prediction fromconstrained social network
Second Life prediction fromnonredundant social contacts
EverQuest II predictionfrom constrained social (upper)versus economic (lower) network
EverQuest II prediction fromnonredundant social (upper)
versus economic (lower) contacts
25+
from Burt (2018, Structural Holes in Virtual Worlds).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
28)
Returns to Brokerage Are Evident in Low Returns to Student Specialization
Figures and text are from Merluzzi and Phillips (2016, Administrative Science Quarterly),“The Specialist Discount." For more applied discussion, see Merluzzi, (June 2016, HBR),
"Generalists get better job offers than specialists." Looking later in the career, Keinbaum (2012, ASQ) "Organizational misfits," shows with email data that managers with unusual patterns of communication are most likely to emerge the valued network brokers.
Recent scholarship on the returns to labor market specialization often claims that being specialized is advantageous for job candidates. We argue, in contrast, that a specialist discount may occur in contexts that share three features: strong institutionalized mechanisms, candidate profiles with direct investments that signal their value, and a high supply of focused candidates relative to demand. We then test whether there is a specialist discount for graduating elite MBAs, as it is a labor market that exemplifies these conditions under which we expect specialists to be penalized. Using rich data on two graduating cohorts from a top-tier U.S. business school (full-time students, 2008-2009), we show that elite MBA graduates who established a focused (specialized) market profile of experiences relating to investment banking before and during the program were less likely to receive multiple job offers and were offered less in starting-bonus compensation than similar MBA candidates with no exposure or less-focused exposure to investment banking. Our theory and findings suggest that the oft-documented specialist advantage may be overstated.
Figure 1 displays predicted (marginal) probabilities of receiving multiple offers for candidates who have mean values for each of the control variables but different profiles.
Figure 2 compares the starting bonuses of hypothetical job candidates with different profiles. Each hypothetical candidate is a single white male who graduated from a top-20 undergraduate institution, has above a 3.8 GPA, received more than one job offer, has the mean age and work experience characteristics (months, number of firms), accepts a job in I-banking, and earns the mean base salary for I-banking jobs in his 2008 cohort year. The only difference is the candidate’s profile in terms of exposure to I-banking.
FOCUSED (career history in finance before mba, concentration in finance, joined an i-banking club during mba, and i-banking internship; 61% of students who graduate to a job in i-banking were focused on i-banking)NON-SEQUENTIAL exposure (neither of the above categories, but some mba program contact with i-banking)PARTIAL sequential exposure (prior experience in finance + concentration in finance or participation in i-banking club)PRE-MBA exposure (only exposure before mba program)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
29)
Figure 3 in Burt and Burzynska, "Chinese entrepreneurs, social networks, and guanxi," (2017 Management and Organization Review)
Network Brokerage and Business Success in China
NOTE — Dots are average scores for a five-point interval of network constraint in a study population. Lines are vertical axis predicted by the natural logarithm of network constraint. Statistics in the graphs are computed from the displayed data. Graph A shows success (measured by evaluation, compensation, or promotion) increasing with more structural holes in the networks around 1,989 analysts, bankers, and managers in American and European companies, with controls for differences between the individuals (from Burt, Kilduff, and Tasselli, 2013:535; Burt, 2010:26; cf. Burt 2005:56). Graph B shows business success increasing with more structural holes in networks around the 700 Chinese entrepreneurs running each business. Business success is measured by the first principal component of patents, employees, and sales adjusted for having a research and development department (see Table 1).
Network Constraint (x 100)many ——— Structural Holes ——— few
Z-Sc
ore
Bus
ines
s Su
cess
(em
plo
ye
es,
sa
les,
pa
ten
ts,
ad
juste
d f
or
R&
D d
ep
art
me
nt)
Z-Sc
ore
Bus
ines
s Su
cess
(po
sitiv
e e
va
lua
tio
n,
hig
h c
om
pe
nsa
tio
n,
fast
pro
mo
tio
n)
Network Constraint (x 100)many ——— Structural Holes ——— few
A. Executives in Americanand European Companies
B. ChineseEntrepreneurs
r = -.58 r = -.82
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
30)
Cumulative Advantagefor the Chinese Entrepreneurs
Network Constraint in 2012 (x 100)many ——— Structural Holes ——— few
More successful firms in 2012(r = -.59)
Less successful firms in 2012(r = -.02)
The displayed association shows the 2012 network around an
entrepreneur predicting survival five years later. All 700 Chinese
entrepreneurs are included.
No controls are applied here. The association is modeled with
controls in first three rows of Table 2. Survival probabilities are computed within five-point intervals of network constraint.
Displayed correlations displayed are computed from the plotted
data.
Figure 1 in Zhao and Burt, "A Note on Business Survival and Social Network," (2017 working paper)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
31)
from Figures 1.1 and 1.3 in Burt (1992, Structural Holes) and Figure 1.2 in Brokerage and Closure
HOW IT WORKS: Recombinant Sticky InformationContacts as Source vs. Portal
Network A Network B Network C
YOU YOU YOU
Redundancyby Cohesion YOU
Redundancyby StructuralEquivalence
(cf. felt similarityon page 18)
YOU
ContactRedundancy
James
Robert
1
23
54
6
7
A
B
C & D 25
1
0
100
29
Group A
Group B
Group C
Group D
Density Table
0
85
5
0
0
person 3: .402 = [.25+0]2 + [.25+.084]2 + [.25+.091]2 + [.25+.084]2
Robert: .148 = [.077+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2
Network Constraint(C = Σj cij = Σj [pij + Σq piqpqj]2, i,j ≠ q)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
32)
James
Robert
1
23
54
6
7
A
B
C & D 25
1
0
100
29
Group A
Group B
Group C
Group D
Density Table
0
85
5
0
0
person 3: .402 = [.25+0]2 + [.25+.084]2 + [.25+.091]2 + [.25+.084]2
Robert: .148 = [.077+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2
Network Constraint(C = Σj cij = Σj [pij + Σq piqpqj]2, i,j ≠ q)
Networkindicates
distributionof sticky
information, which defines
advantage.
From Figure 1.1 in Brokerage and Closure. For an HBR treatment of the network distinction between Robert and James, see in the course packet Kotter's classic distinction between "leaders" versus "managers." Robert ideally corresponds to the image of a "T-shaped manager," nicely articulated in Hansen's HBR paper in the course packet.
Bridge & Cluster: Small World of Organizations & Markets
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
33)
BEFORE
1
2
3
4
5
2
1
3
4
5
The employee AFTER is more positioned at the crossroads of communication between social clusters within the firm and its market, and so is better positioned to craft projects and policy that add value across clusters.
Here is the core network for a job BEFORE and AFTER the employee expanded the social capital of the job by reallocating network time and energy to more diverse contacts.
Research shows that employees in networkslike the AFTER network,spanning structural holes,are the key to integratingoperations across functional and business boundaries. In research comparing senior peoplewith networks like these BEFORE and AFTER networks, it is the AFTER networks that are associated with more creativity, fasterlearning, more positive individual and teamevaluations, faster promotions,and higher earnings.
*Network scores refer to direct contacts.
It is the weak contact connections (structural holes) in the AFTER network that provides the expanded social capital.
AFTER
53.6 constraint
20.0 constraint*
From Figure 1.4 in Burt (1992, Structural Holes), and Figure 1.2 in Brokerage and Closure. See Appendix I on survey network data, Appendix IV on measuring network constraint, and Pfeffer's note in packet for a readable overview.
Create Valueby BridgingStructural
Holes
STICKY INFORMATIONInformation expensive to move because: (a) tacit, (b) complex, (c) requires other knowledge to absorb, or (d) interaction with sender, recipient, or channel.
STRUCTURAL HOLEdisconnection between two groups or clusters of people
BRIDGErelation across structural hole
NETWORK ENTREPRENEURor "broker," or "connector:" a person who coordinates
across a structural hole
BROKERAGEact of coordinating across
a structural hole
information breadth, timing, and arbitrage
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
34)
Competitive Advantage in Social Networks Begins withStigler’s “Economics of Information,” JPE 1961
"The expected saving from given search will be greater, the greater the dispersion of prices.” When price varies greatly between sellers, it is worth a buyer’s time to search for the lowest price. It makes no sense to search for the lowest price of a commodity good; all prices are similar.
The potential value of search is an incentive for entrepreneurs to aggregate price information by enforcing localized transactions, as in medieval markets, or by becoming "specialized traders whose chief service, indeed, is implicitly to provide a meeting place for potential buyers and sellers.”
In short, the value of search is proportional to information variation, and search is more productive for people more exposed to the variation.
As referenced in Stigler’s Nobel acceptance speech: "The proposal to study the economics of information was promptly and widely accepted, and without even a respectable minimum of controversy." "All I had done was to open a door to a room that contained many fascinating and important problems."
Discussed in Burt and Soda, "The social origins of great strategies" (Strategy Science, 2017)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
35)
HOW IT WORKS: Creativity and InnovationAre at the Heart of It
from Burt, "The social capital of structural holes" (2002, The New Economic Sociology). The consequences of the information diversity associated with network brokerage is productively elaborated at length in economist Scott
Page's 2007 book, The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools and Societies.
Brokerageacross
Structural Holes
Creativity & Innovation(What should be done?)
Achievement & Rewards(What benefits?)
Adaptive Implementation(How to frame it & who should be involved?)
Alternative Perspective (how would this problem look from the perspective of a different group, or groups — thinking “out of the box” is often less valuable than seeing the problem as it would look if you were inside a specific “other box”)
Best Practice (something they think or do could be valuable in my operations)
Analogy (something about the way they think or behave has implications for how I can enhance the value of my operations; i.e., look for the value of juxtapositioning two clusters, not reasons why the two are different so as to be irrelevant to one another — you often find what you look for)
Synergy (resources in our separate operations can be combined to create a valuable new idea/practice/product)
What in your workimproves the odds
that you will discover the value of something
you don't know you don't know?
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
36)
Illustration: Where did the M-16 come from?
*Photos are from the video shown during the session. For discussion and references, see page 73 in Brokerage and Closure. For sampling on the dependent variable, see Rosenzweig, “Misunderstanding the nature of
company performance: the halo effect and other business delusions,” 2007 California Management Review.
Discussion Question*
Consequential ideas are typically attributed to special people, geniuses, in part to make us feel less uncomfortable about our own ideas. True to form, an American armament expert describes Eugene Stoner, the engineer who developed the M-16 assault rifle, as "an engineering genius of the first order." Another describes him as "the most gifted small-arm designer since Browning." (Browning patented the widely-adopted BAR and 45 automatic.) 1. Based on the brief history video, how would you describe Stoner's genius? 2. What circumstances might allow you or your colleagues to be as creative?
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
37)
from Figure 2.1 in Brokerage and Closure (or Figure 5 in Burt, "Structural holes and good ideas," 2004 American Journal of Sociology, point is elaborated in Burt and Soda, "The social origins of great strategies," 2017 Strategy Science).
Brokerage, Good Ideas, and Innovation,Digging a Little Deeper
^
Network Constraint (C) on Manager Offering Idea
Man
agem
ent
Eva
luat
ion
of
Idea
's V
alu
e
Y = a + b ln(C)
Pro
bab
ility
". . . for those ideas that wereeither too local in nature,incomprehensible, vague,
or too whiny, I didn't rate them"
P(dismiss)
E
E E
E
E
E
E
E
E
E
E
E
E
E
E
E
G
G
G
G
G
G
G
GG
G
G
G G
G
G
G
1
1.5
2
2.5
3
3.5
10 20 30 40 50 60 70 80 90 100
C
C
C
C C
C
C
C
C
C
C
C C C
C
C
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
10 20 30 40 50 60 70 80 90 100
a
6.42
4.08
5.51
b
-1.04
-.63
-.91
t
-5.8
-3.9
-7.4
Judge 1
Judge 2
Combined
^ P(no idea)11.2 logit test statistic
^
^
5.5 logittest statistic
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
38)
Episode Creativity
Role Creativity
Constraint -0.018 (-6.73)
-0.018 (-6.44)
# Episodes 0.003 (0.51)
-0.005 (0.36)
# in Fallow Period -0.013 (-0.99)
-0.014 (-0.73)
Intercept 5.32 5.30
R2 .31 .29
N observations 200 200
Max
imum
Epi
sode
Cre
ativ
ity
Maximum episode creativity
Mean of the other two measures
Maximum role-creativity in episode
Network Constraint(lack of structural holes within and between
producer-director-writer teams in which person worked)
Maximum Career Creativity by Career Access to Structural HolesGraph is from Mannucci, Soda, and Burt (in production). The observations are all 200 people who worked as producers, directors, or writers in
any of the 273 episodes of the BBC series, Dr. Who. The horizonal axis is a person’s network constraint score for the network of people with whom the person worked. High scores indicate the person worked with people who primarily worked with one another. Low scores indicate the
person worked with many different people, who themselves came together from working with many different people. Constraint and creativity are averaged within 5-point intervals on the horizontal axis (two intervals containing a single person are combined with the closest adjacent
interval). Creativity is measured on the vertical axis in two ways: (1) maximum creativity score a person ever received for an episode on which s/he worked (mean 1-5 creativity score from two expert critics, hollow circles), and (2) maximum creativity score a person ever received
for his or her role as producer, director, or writer (mean 1-5 creativity score from two expert critics, hollow squares). The table to the right contains OLS regression models showing the strong creativity-network association after holding constant the number of episodes on which a person worked, and the person’s number of episodes during a fallow period in the Dr. Who series (coefficients presented with test statistics in
parentheses). Picture is an evil alien in the series.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
39)
Network Constraint(lack of structural holes within and between
producer-director-writer teams in which person worked)
Number highly creative episodes
Mean of the two measures
Number of high role-creative episodes
Num
ber o
f Hig
h-C
reat
ivity
Epi
sode
s
AggregateCareer Creativity
by Career Accessto Structural Holes
Graph is from Mannucci, Soda, and Burt (in production). The observations are all 200 people who worked as producers, directors, or writers in any of all 273 episodes of the BBC series, Dr. Who. The horizonal axis is a person’s network constraint score for the network of all people with
whom the person worked. High scores indicate the person worked with people who primarily worked with one another. Low scores indicate the person worked with many different people, who themselves came together from working with many different people. Constraint and creativity are averaged within 5-point intervals on the horizontal axis (two intervals containing a single person are combined with the closest adjacent interval). Creativity is measured on the vertical axis in two ways: (1) number of a person’s episodes that were judged highly creative by one or both of two expert critics (hollow circles), and (2) number of episodes in which a person was judged by either or both of the two expert critics to have played
their role as producer/director/writer in a highly creative way (hollow squares). To be highly creative in multiple episodes, one has to work on multiple episodes, so the table to the right contains Poisson regression models showing the strong creativity-network association after holding constant the number of episodes on which a person worked, and the person’s number of episodes during a fallow period in the Dr. Who series
(coefficients presented with test statistics in parentheses).
Episode Creativity
Role Creativity
Log Constraint -1.11 (-8.11)
-0.92 (-6.32)
# Episodes 0.04 (5.81)
0.05 (6.89)
# in Fallow Period -0.04 (-5.19)
-0.05 (-5.19)
Intercept 4.18 3.27
Pseudo R2 .46 .42
N observations 200 200
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
40)
Four Summary Points on FoundationsNetwork Structure Is a Proxy for the Distribution of Information
For reasons of opportunity, shared interests, experience — simple inertia — organizations and markets drift toward the bridge-and-cluster structure known as a “small world.”
RULE 1: Closure-Trust AssociationNetwork closure enhances communication and individual visibility within a group, (a) which creates reputation costs for individuals who express opinion or behavior inconsistent with group standards, (b) which makes it less risky to trust within the group, (c) which enhances productivity as people become self-aligning in extraordinary efforts. Value comes from lower costs for labor, management, and time. Closure delivers that value by creating a reputation cost for deviation from colleague opinion and practice. Over time, information becomes "sticky" within clusters, different between clusters.
RULE 2: Brokerage-Achievement Association Bridge relations across the structural holes between clusters provide information breadth, timing, and arbitrage advantages, such that network brokers managing the bridges are at higher risk of “productive accident” in detecting and developing good ideas. By clearing the sticky-information market across organizations, brokers tend to be innovation leaders, better compensated than peers, more widely celebrated than peers, and promoted more quickly to senior rank.
Three Points Follow from the Link between Network Brokerage and Good Ideas- Closed networks do not identify unintelligent managers so much as expert specialists.- Innovation is an import/export process. Value is not created at the innovation source. It is created
each time productive knowledge produces innovation in a target audience.- Innovation depends on the network as well as the person. Innovation does not depend on individual
genius so much as it depends on employees finding opportunities to broker knowledge from where it is routine to where it would create value.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
41)
AppendixMaterials
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
42)
Appendix I:
Example Network
Questionnaire for a
Web Survey
for discussion of these slides
and how to collect network data,
see Appendix A, "Measuring the
Network," in
Neighbor Networks.
Figure A1 in Neighbor Networks
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
43)
Appendix I,continued
Figure A2 in Neighbor Networks
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
44)
from Burt and Burzynska, "Chinese entrepreneurs, social networks, and guanxi," (2018 Management and Organization Review)
Figure A2. Business Event Name Generator
The next five questions generate a summary picture of the business network. To draw the picture, you will be asked about people, but we do not want to know any one's name. I will go through this network worksheet with you, asking about people who were useful to your business in one way or another. Without mentioning anyone's name to me, please write on your worksheet the names of people who come to mind in response to the questions. We will create a list of names then refer to people by their order on the list. No names. You will keep the worksheet to yourself.
Q1. Let me begin with an example so you can see how the interview protects your confidentiality at the same time that a picture of the business network emerges. Your business time line shows that your firm was founded in _(say founding year)_. Please think back to your activities in founding the firm. Who was the one person who was most valuable to you in founding the firm?
Q2. Now please do the same thing for each of the significant events you listed on your business time line. The first significant event you listed was __(say first event)__ in _(say year)_. Who was the person most valuable to you during that event? Please write on the first line below the person's name. The person most valuable in this event could be the same person who was most valuable to you in founding the firm. You would just enter the name again.
Confidential
Time Line for an Example Firm
today2012
|_____
|_____
businessfounded_____
|_____
today2012
|_____
|_____
businessfounded_____
|_____
1992
1993, secured technology partner
1999, first bank loan
2008, secured currentprimary export customer
2004, first export contract
1997 2002 2007
2000, critical supplierno longer available
Time Line for Your Firm
Business Time Line Worksheet
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
45)
from Burt and Burzynska, "Chinese entrepreneurs, social networks, and guanxi," (2017 Management and Organization Review)
Figure A3. Name Interpreters Flesh Out Relationships and Define Connections among Cited Contacts
§ Contact Gender (male, female)
§ Emotional Closeness to Contact (especially close, close, less close, distant)
§ Duration of Connection with Contact (years known)
§ Frequency of Contact (daily, weekly, monthly, less often)
§ Trust (1 to 5, low to high trust) “Consider the extent to which you trust each of the listed people. For example, suppose one of the people asked for your help. The help is not extreme, but it is substantial. It is a level of help you cannot offer to many people. To what extent would you trust each person to give you all the information you need to decide on the help? For example, if the person was asking for a loan, would they fully inform you about the risks of them being able to repay the loan? If the person was asking you give a job to one of their relatives, would they fully inform you about their relative's poor work attitude or weak abilities, or other qualities that would make you prefer not to hire the relative?”
§ Role (all that apply: family, extended family, neighbor, party, childhood, classmate, military, colleague, business association)
§ Matrix of Connections between Contacts (especially close, distant, or something in between)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
46)
Appendix II. Measuring Network Closure/EmbeddingLet a 2-step connection refer to a connection between two people through a mutual contact. For example, the “1” under “D” for Jim in the first row of the table refers to person 4 in the sociogram. Person 4 is the only contact linked directly to Jim and person 1. The “3” underneath the “1” in the table refers to three mutual contacts between Jim and person 2. The mutual contacts are persons 4, 6, and 7. Two-step connections are this chapter’s measure of direct structural embedding. Indirect structural embedding is measured in this chapter with 3-step connections. For example, the “1” under “I” for Jim in the second row of the table refers to persons 5 and 3 in the sociogram. Jim’s connections to 2 through persons 4, 6, and 7 are 2-step connections. Jim’s fourth contact, person 5, is not connected to person 2, but is connected to 3 who is connected to 2, so Jim has a 3-step connection to person 2 via person 5. In graph theoretic terms, I am looking for geodesics linking two people through one intermediary (direct structural embedding) or two intermediaries (indirect structural embedding). Since I want to know how indirect embedding adds to direct embedding, I only count distant connections in the absence of closer connections. For example, Jim is connected to person 6 who is connected to 3 who is connected to 2, which is an 3-step connection between Jim and person 2. However, Jim reaches 2 through 6 directly, so the table reports one 3-step connection (the 5-3-2 connection).
This is Figure 2 in Burt, "Closure and stability" in The Missing Links: Formation and Decay of Economic Networks, edited by J. Rauch (2007 Russell Sage Foundation). For elaboration and illustration of indirect connections, see Chapter 7 in the on-line network textbook, Introduction to
Social Network Methods, by Robert A. Hanneman and Mark Riddle (http://faculty.ucr.edu/~hanneman/nettext).
1
2
3
4
5
6
7
8
9
Jim
James
Mean per
Contact
(in box)
D
1
3
3
0
0
0
0
3
1
—
1
I
3
1
1
3
2
3
2
0
2
—
3
D
3
3
3
3
1
2
1
2
1
1
—
I
0
0
0
0
3
2
3
2
3
3
—
Jim James
Figure 2.
Network Closure from Direct and Indirect Embedding
Number of 2-Step (Direct) and
3-Step (Indirect) Connections
0.0 2.5 3.0 0.0
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
47)
Appendix III: Closure/Embedding and the Theory of the Firm
The Source is John Commons’ Five-Player Unit
for Transactional Analysis
(1) MAY — range of behaviors allowed in relationship
(2) MUST — minimum obligations of relationship
(3) CAN — minimum rights in relationship
(4) CANNOT — behaviors prohibited in relationship
fifth player
Graphic is from Figure 7.1 in Structural Holes (Burt, 1992), see John R. Commons (1924), Legal Foundations of Capitalism, chapter on transactions, which set a stage for Coase's (1937) nobel-winning "The Nature of the Firm" in Economica, and subsequent work on "competitive strategy."
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
48)
Appendix IV: Measuring Access to Structural Holes*from Burt, "Formalizing the argument," (1992, Structural Holes); "Gender of social capital" (1998, Rationality and Society); Appendix B "Measuring Access to Structural
Holes," (2010, Neighbor Networks). See the Pfeffer note in the packet for a readable overview ("A note on networks and network structure").
Network brokerage is typically measured in terms of opportunities to connect people. When everyone you know is connected with one another, you have no opportunities to connect people. When you know a lot of people disconnected from one another, then you have a lot of opportunities to connect people. “Opportunities” should be emphasized in these sentences. None of the usual brokerage measures actually measures brokerage behavior. They index opportunities for brokerage. Reliability and cost underlie the practice of measuring brokerage in terms of opportunities. It is difficult to know whether or not you acted on a brokerage opportunity. One can know with more reliability whether or not you had an opportunity for brokerage. Acts of brokerage could be studied with ethnographic data, but the needed depth of data would be expensive, if not impossible, to obtain by the practical survey methods used to measure networks. Good reasons notwithstanding, the practice of measuring brokerage by its opportunities rather than its occurrence means that performance has uneven variance across levels of brokerage opportunities. Performance is typically low in the absence of opportunities. Performance varies widely where there are many opportunities: (1) because some people with opportunities do not act upon them and so show no performance benefit, (2) because it is not always valuable to move information between disconnected people (e.g., explain to your grandmother the latest technology in your line of work), or (3) because the performance benefit of brokerage can occur with just one key bridge relationship. A sociologist might do more creative work because of working through an idea with a colleague from economics, but that does not mean that she would be three times more creative if she also worked through the idea with a colleague from psychology, another from anthropology, and another from history. The above three points can be true of brokerage measured in terms of action, but under the assumption that people invest less in brokerage that adds no value, the three points are more obviously true of brokerage measured in terms of opportunities. It could be argued that people more often involved in bridge relations are more likely to have one bridge that is valuable for brokerage, and to understand how to use bridges to add value, but the point remains that the network measures discussed below index opportunities for brokerage, not acts of brokerage.
Bridge CountsBridge counts are an intuitively appealing measure. The relation between two people is a bridge if there are no indirect connections between the two people through mutual contacts. Associations with performance have been reported measuring brokerage with a count of bridges (e.g., Burt, Hogarth, and Michaud, 2000:Appendix; Burt, 2002).
ConstraintI measure brokerage opportunities with a summary index, network constraint. As illustrated on the next page, network constraint begins with the extent to which manager i’s network is directly or indirectly invested in the manager’s relationship with contact j (Burt 1992: Chap. 2): cij = (pij + Σqpiqpqj)2, for q ≠ i,j, where pij is the proportion of i’s network time and energy invested in contact
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
49)
Illustrative Network and Computation
Constraint measures the
extent to which a network doesn't span structural
holes
A
B
C
DE
F
contact-specificconstraint (x100):
= aggregate constraint (C = Σj cij)
network data
A . 1 0 0 1 1 1B 1 . 0 1 0 0 1C 0 0 . 0 0 0 1D 0 1 0 . 0 0 1E 1 0 0 0 . 0 1F 1 0 0 0 0 . 1 1 1 1 1 1 1 .gray dot
A 15.1 B 8.5 C 2.8 D 4.9 E 4.3 F 4.3
total 39.9
cij = (pij + Σq piqpqj)2 q ≠ i,j
100(1/36)
Network constraint measures the extent to which your network time and energyis concentrated in a single group. There are two components: (direct) a contactconsumes a large proportion of your network time and energy, and (indirect) acontact controls other people who consume a large proportion of your networktime and energy. The proportion of i’s network time and energy allocated to j, pij, is the ratio of zij to the sum of i’s relations, where zij is the strength of connectionbetween i and j, here simplified to zero versus one.
Figure 2.2 in Structural Holes.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
50)
j, pij = zij / Σqziq, and variable zij measures the strength of connection between contacts i and j. Connection zij measures the lack of a structural hole so it is made symmetric before computing pij in that a hole between i and j is unlikely to the extent that either i or j feels that they spend a lot of time in the relationship (strength of connection “between” i and j versus strength of connection “from” i to j; see Burt, 1992:51). The total in parentheses is the proportion of i’s relations that are directly or indirectly invested in connection with contact j. The sum of squared proportions, Σjcij, is the network constraint index C. I multiply scores by 100 to discuss integer levels of constraint. The network constraint index varies with three network dimensions: size, density, and hierarchy. Constraint on a person is high if the person has few contacts (small network) and those contacts are strongly connected to one another, either directly (as in a dense network), or through a central, mutual contact (as in a hierarchical network). The index, C, can be written as the sum of three variables: Σj(pij)2 +2Σjpij(Σqpiqpqj) + Σj(Σqpiqpqj)2. The first term in the expression, C-size in Burt (1998), is a Herfindahl index measuring the extent to which manager i’s relations are concentrated in a single contact. The second term, C-density in Burt (1998), is an interaction between strong ties and density in the sense that it increases with the extent to which manager i’s strongest relations are with contacts strongly tied to the other contacts. The third term, C-hierarchy in Burt (1998), measures the extent to which manager i’s contacts concentrate their relations in one central contact. See Burt (1992:50ff.; 1998:Appendix) and Borgatti, Jones, and Everett (1998) for discussion of components in network constraint.
SizeNetwork size, N, is the number of contacts in a person's network. In graph-theory discussions, the size of the network around a person is discussed as “degree.” For non-zero network size, other things equal, more contacts mean that a manager is more likely to receive diverse bits of information from contacts and is more able to play their individual demands against one another. Network constraint is lower in larger networks because the proportion of a manager’s network time and energy allocated to any one contact (pij in the constraint equation) decreases on average as the number of contacts increases.
DensityDensity is the average strength of connection between contacts: Σ zij / N*(N-1), where summation is across all contacts i and j. Dense networks are more constraining since contacts are more connected (Σqpiqpqj in the constraint equation). Contact connections increase the probability that the contacts know the same information and eliminate opportunities to broker information between contacts. Thus, dense networks offer less of the information and control advantage associated with spanning structural holes. Density is only one form of network closure, but it is a form often discussed as closure. Hypothetical networks in the figure on page 52 illustrate how constraint varies with size, density, and hierarchy. Relations are simplified to binary and symmetric in the networks. The graphs display relations between contacts. Relations with the person at the center of the network are not presented (that person at the center is referenced by various labels such as "you," "ego," or "respondent"). The first column in the figure contains examples of sparse networks (zero density). No contact is connected with other contacts. The third column of the figure contains maximum-density networks (density = 100). Every contact has a strong connection with each other contact. At each network size, constraint is lower in the sparse-network column.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
51)
HierarchyDensity is a form of closure in which contacts are equally connected. Hierarchy is another form of closure in which a minority of contacts, typically one or two, stand above the others for being more the source of closure. The extreme is to have a network organized around one contact. For people in job transition, such as M.B.A. students, that one contact is often the spouse. In organizations, hierarchical networks are sometimes built around the boss. Hierarchy and density both increase constraint, but in different ways. They enlarge the indirect connection component in network constraint (Σqpiqpqj). Where network constraint measures the extent to which contacts are redundant, network hierarchy measures the extent to which the redundancy can be traced to a single contact in the network. The central contact in a hierarchical network gets the same information available to the manager and cannot be avoided in manager negotiations with each other contact. More, the central contact can be played against the manager by third parties because information available from the manager is equally available from the central contact since manager and central contact reach the same people. Network constraint increases with both density and hierarchy, but density and hierarchy are empirically distinct measures and fundamentally distinct with respect to social capital because it is hierarchy that measures social capital borrowed from a sponsor. To measure the extent to which the constraint on a person is concentrated in certain contacts, I use the Coleman-Theil inequality index for its attractive qualities as a robust measure of hierarchy (Burt, 1992:70ff.). Applied to contact-specific constraint scores, the index is the ratio of Σj
rj ln(rj) divided by N ln(N), where N is number of contacts, rj is the ratio of contact-j constraint
over average constraint, cij/(C/N). The ratio equals zero if all contact-specific constraints equal the average, and approaches 1.0 to the extent that all constraint is from one contact. Again, I multiply scores by 100 and report integer values. In the first and third columns on the next page, no one contact is more connected than others, so all of the hierarchy scores are zero. Non-zero hierarchy scores occur in the middle column, where one central contact is connected to all others who are otherwise disconnected from one another. Contact A poses more severe constraint than the others because network ties are concentrated in A. The Coleman-Theil index increases with the number of people connected to the central contact. Hierarchy is 7 for the three-contact hierarchical network, 25 for the five-contact network, and 50 for the ten-contact network. This feature of hierarchy increasing with the number of people in the hierarchy turns out to be important for measuring the social capital of outsiders because it measures the volume of social capital borrowed from a sponsor, which strengthens the association with performance (this point is the focus of the later session on outsiders having to borrow network access from a strategic partner). Note that constraint increases with hierarchy and density such that evidence of density correlated with performance can be evidence of a hierarchy effect. Constraint is high in the dense and hierarchical three-contact networks (93 and 84 points respectively). Constraint is 65 in the dense five-contact network, and 59 in the hierarchical network; even though density is only 40 in the hierarchical network. In the ten-contact networks, constraint is lower in the dense network than the hierarchical network (36 versus 41), and density is only 20 in the hierarchical network. Density and hierarchy are correlated, but distinct, components in network constraint.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
52)
Network Constraintdecreases with number of contacts
(size), increases with strength of connections between contacts
(density), and increases with sharing the network (hierarchy).
This is Figure 1 in Burt, "Reinforced Structural Holes," (2015, Social Networks, an elaboration of Figure B.2 in Neighbor Networks). Graph above plots density and hierarchy for 1,989 networks observed in six management populations (aggregated in Figure 2.4 in Neighbor Networks to illustrate returns to
brokerage). Dot-circles are executives (MD or more in finance, VP or more otherwise). Hollow circles are
lower ranks. Executives have significantly larger, less dense, and less hierarchical networks.
To keep the diagrams simple, relations with ego are not presented.
E B
D C
A
CliqueNetworks
3100
093
3131311.00.0
5100
065
13131313131.00.0
10100
0361.00.0
PartnerNetworks
3677
84
4420201.70.5
5402559
366666
3.43.0
102050418.2
18.0
BrokerNetworks
30033
1111113.03.0
50020
44444
5.010.0
100010
10.045.0
SmallNetworks
contactsdensity x 100
hierarchy x 100constraint x 100
from:ABC
nonredundant contactsbetweenness (holes)
LargerNetworks
contactsdensity x 100
hierarchy x 100constraint x 100
from:ABCDE
nonredundant contactsbetweenness (holes)
Still LargerNetworks
contactsdensity x 100
hierarchy x 100constraint x 100
nonredundant contactsbetweenness (holes)
E B
D C
A
A
C B
A
C D
A
C B
A
C B
E B
D C
A
Network Density
Net
wor
k H
iera
rchy
Partners
CliquesBrokers
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
53)
The Network Measures of Access to Structural HolesAre Strongly Correlated
These are network metrics for 801 senior people in two organizations analyzed in Burt, "Reinforced structural holes" (2015, Social Networks). One organization is a center-periphery network of investment bankers (circles). The other is a balkanized network of supply-chain managers in a large electronics company (squares). The point is that networks rich in structural holes by one measure tend to be rich in the other measures.
-.90 correlation with log constraintN
onR
edun
dant
Con
tact
s
Network Constraint (x 100)many ——— Structural Holes ——— few
-.71 correlation with log constraint
Ego-
Net
wor
k B
etw
eenn
ess
(Num
ber M
onop
oly-
Acc
ess
Hol
es)
Network Constraint (x 100)many ——— Structural Holes ——— few
NonRedundant Contactsfew ——— Structural Holes ——— many
Ego-
Net
wor
k B
etw
eenn
ess
(Num
ber M
onop
oly-
Acc
ess
Hol
es) R2 = .99
R2 = .92
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
54)
This is Figure 3 in Burt, "Reinforced structural holes" (2015, Social Networks), based on the above networks in Figure 1 of Vedres and Stark, "Structural folds: generative disruption in overlapping groups" (2010, American Journal of Sociology). Correlations to the right are across the 801 bankers and managers analyzed in the 2015 article.
Structural Folds Indicate Access to Structural Holes
Kind of Network Network
Size (Contacts)
Effective Size (NonRedundant
Contacts)
Network Constraint
Ego-Network Betweenness
(Structural Holes)
Reinforced Holes (RSH)
Ego-Network Modularity
(Newman Q) Raw Normalized
Closed (3, 4, 5, 7, 8, 9, 11, 12, 13,
14, 15, 16) 3 1.0 92.6 .00 .00 0% .00
Broker (1) 2 2.0 50.0 1.00 .75 75% .00
Broker (2, 6) 4 2.5 58.3 3.00 1.75 29% .00
Fold Broker (10) 6 4.0 46.3 9.00 6.00 40% .50
3
4
5
2
9
6
1
7
8 12
11
13
10
14
15
16
Log Constraint 1.00
Effective Size -.90 1.00
EN Betweenness -.71 .88 1.00
RSH -.71 .93 .91
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
55)
Appendix V: Closure and Learning Curvesby Michael Rothschild
Bruce Henderson certainly didn’t look like a revolutionary. No tattered army fatigues. No fiery rhetoric. He favored starched white shirts and pinstripe suits. He spoke softly, in the measured, almost halting, manner of a southern gentleman. But Bruce Henderson had the “right stuff” of a revolutionary — profoundly new ideas that change the way society works. The originator of modern corporate strategy and founder of The Boston Consulting Group (BCG), Bruce Henderson died this summer in his hometown of Nashville, Tennessee. He was 77.
Trained as an engineer, Bruce Henderson became fascinated with economic ideas for terribly practical business reasons. Back in the days before he established the discipline of corporate strategy, making the big decisions about a company’s long-term future was pretty much a “seat of the pants” affair. The CEO, with perhaps a few senior executives and board members, would sit around and talk until they came up with a plan that seemed sensible. “Bet-your-company” decisions like launching a new product line, acquiring a subsidiary, or shutting down a factory, were made on little more than intuition.
A rigorous analytical approach to making key decisions was impossible, because there were no guiding strategic prin-ciples, no theories that could be turned into quantifiable models. Standard economic models existed, of course, but every sophisticated businessman knew that the economists’ mythical kingdom of “perfect competition” bore no relationship to reality. To turn corporate strategy into a credible discipline — and consulting assignments that major clients would pay major money for — Henderson had to find a hard link between business and underlying economic forces.
Henderson’s search began with highly detailed analyses of production costs. Early in his career, while a purchasing manager for a Westinghouse division, he wondered why suppliers who produced their goods in virtually identical factories often put in bids at dramatically different prices. Economic theory said it wouldn’t happen. Producers using similar capital equipment were supposed to have similar unit costs and offer roughly the same prices. But economic theory was wrong. In case after case, actual unit costs varied dramatically among suppliers. Henderson didn’t know why, but he had zeroed in on the crucial question.
Then, in 1966, shortly after he founded BCG, a study for Texas Instruments’ semiconductor division revealed the answer. When TI’s unit cost data for a particular part was plotted against the company’s accumulated production experience, the cost of the part declined quite predictably. For example, if the 1000th unit off the line had cost $100 to make, the 2000th unit would cost 80% as much, or $80. By the time the 4000th unit was produced, it would cost just $64 ($80 x 80%). Every time cumula-tive experience doubled, unit costs dropped about 20%. Though it’s “old hat” among today’s high-tech managers, the notion of predictably declining costs was a radical concept when Bruce Henderson began teaching companies about the “experience curve” a quarter century ago.
(over)
This article appeared in a longer form in Upside (December 1992, Copyright 1992 The Bionomics Institute).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
56)
During the 1970s, Henderson’s concept became the foundation of modern corporate strategy. For the first time, it was possible to explain why building a factory just like your competitor’s didn’t mean you could match his costs. If he had a head start in experience, you could wind up chasing him down the experience curve. If you both sold at the market price, he’d make money on every unit, while you’d be lucky to break-even.
Once the experience curve was understood, the importance of being the first one to enter a new market became clear. Properly executed, the preemptive strike could mean long-term market leadership and long-term profits. Similarly, the experience curve explained why defending market share mattered. Raising prices to boost short-term profits sold off market share, slowed experience growth, and often handed over low cost leadership to an aggressive competitor. It’s a scenario that’s been played out hundreds of times as “experience conscious” Japanese competitors overtook their “profit conscious” American rivals.
Simply put, Bruce Henderson’s experience curve explained how an industry’s past shapes its future. Where conventional economics banished history by blithely assuming that “technology holds constant,” Henderson used the experience curve to show how the new insights generated by practical experience translated into higher productivity and lower costs. Where conventional economics taught the “law of diminishing returns,” Bruce Henderson taught the “law of increasing returns.” Where mainstream economics taught that marginal unit costs must rise at some point, Henderson showed that marginal unit costs can continually fall.
When the cost/performance potential of a particular technology is nearly exhausted, an industry will shift to a substitute technology and begin a new “experience curve.” For example, even as the airlines have shifted from one aircraft technology to the next, their cost/seat-mile keeps falling, opening up air travel to the entire population. By substituting new knowledge for labor and materials, experience-driven innovation keeps pushing costs down. As Henderson put it, when a firm is properly managed, its “product costs will go down forever.”
Though he concentrated on the practical problems of clients, Henderson knew full well that the experience curve had undermined the intellectual foundation of mainstream economics. In 1973, he wrote: The experience curve is a contradiction of some of the most basic assumptions of classical economic theory. All economics assumes that there is a finite minimum cost which is a function of scale. This is usually stated in terms of all cost/volume curves being either L shaped or U shaped. It is not true except for a moment in time. . . Our entire concept of competition, anti-trust, and non-monopolisitc free enterprise is based on a fallacy.
I’m often asked whether the work of the great Austrian economist F.A. Hayek inspired me to write Bionomics. Despite my unending admiration for Hayek, the short answer is no, I’d never read him. Bruce Henderson inspired me to rethink the received economic wisdom. Without his “experience curve,” there is no final and fully satisfying explanation for falling costs, rising incomes, and the phenomenon of economic growth. More than anyone else, he made it both possible and necessary for economic thinkers to break free of the static, zero-sum mentality that has gripped the “dismal science” for 200 years. Bruce Henderson gave us the key to “positive-sum” economics. Thanks for the revolution, Bruce.
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
57)
Detail on Learning Curves,Productivity on the WW II Liberty Ships
from Figure 3.7 in Brokerage and Closure (see pages 50-61 below for Financial Times pieceon kinds of industries in which a strong culture is a competitive advantage.
Ñ
[[[
[[[[[
[
[
[[
[[[[[[[[[[[[[[
[[
[[
[[[[[[[[[[[
[[[[[[[[[[[[[
[
[
[
[[[[[[[[[[[[[[[
[[[[[[[[[[[[[[[[[[[
[
E
E
EEE
EE
EEEEEEEEEEEEEEEEEEEEEEEEE
Ñ
Ñ
Ñ
Ñ
Ñ
ÑÑ
ÑÑÑÑÑ
ÑÑÑÑÑÑÑÑÑÑÑÑÑÑÑÑÑÑ
BB
B
B
BBB
BBBBB
BBBBBBBBBBB
BBBBBBBBBBBB
BBBBB
BBB
JJ
JJ
JJ J
J
J
JJ
J
JJJJJJJ
JJJJJJJJJJJJ
JJJJJJJJJJ
Apr
Jul
Oct
Jan
Apr
Jul
Oct
Jan
Apr
Jul
Oct
Jan
Apr
Jul
Oct
Jan
Apr
Jul
0
50
100
150
200
250
300
Da
ys
fro
m S
tart
to
De
liv
ery
1941 1942 1943 1944 1945
Month that Work on Ship Began(Keel Laid)
E
E
E
EEE
EEEEEEEEEEEEEEEEEE
EEE
EEEE
E
EEEEEEEEEEEEEEEEEEEEEEEEEE
E
EEEEEEEEEEEEEEEEEEEEEEE
EE
EEEE
E
E
E
EEEEEEEEEEEEEEEEEEEEEEEEEEEE
E
EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE
EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE
EEEEEEEEEEE
[[[[[[
[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[
[[[[[
[[[
[[[
[[[[[[[[[[[[[[[[[[[[[[[[
[[[[[[
[[[[[[[[[
[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[
B
BB
B
BBB
B
BBBBB
BBB
B
B
BBBB
BB
BBBBBBB
BBBB
BBBBBBBBBB
BBBBBBBBBBBBBBBBBBBBB
BBBBBBBBBBBBBBBBBBBB
BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB
BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB
BBBBBBBBBBBBBBBBBBBBBB
BBB
B
B
B
B
JJ
JJJJ
J
J
JJ
J
J
J
J
J
J
JJ
JJJJJJJJJJJJJ
JJ
JJJ
JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
JJJ
JJJJJJJJJJJJJJJJ
JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
JJ
JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
T
T
TT
TTTT
T
TT
TTTTT
TTTTTTTTTTTTTTT
TTT
TTT
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
TTTTTTTTTTTTTT
T
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
[[[[
[
[
[[[[[[[[[[[[[[[[[[[
[
[[
[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[
[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[
0
50
100
150
200
250
300
0 50 100 150 200 250 300 350 400
E California (Permanente Metals #2, 4-day Robert E. Peary)
B Maine (New England)
J Maryland (Bethlehem Fairfield)
T Oregon (10-day Joseph N. Teal)
Sequence of Shipin Shipyard Production
Da
ys
fro
m S
tart
to
De
liv
ery
Texas (Todd Houston)
Florida(St. Johns River)
[ Other (Florida, Georgia, Texas)
Georgia(Southeastern)
Robert E. PearyJoseph N. Teal
At the start of the programme some of Liberty
Ships suffered catastrophic fracture, though not necessarily so dramatically as the Schenectady. The stern of SS
John P Gaines is pictured here after the vessel split in two off the Aleutians in 1943. As noted in
website, later design changes reduced the fracture rate to 5%.
30%
Peter
Thompson's
The general technological fraternity was unaware of Fracture Mechanics principles when these ships were designed, and the reason for the disastrous fractures was a mystery since conventional safety assessments were unremarkable and the extremely short lives ruled out conventional fatigue as the culprit. It later became clear that the failures could be attributed to:
Recent recoveries from the Titanic suggest that poor steels in association with low temperatures might have contributed to that disaster too, although this vessel was
riveted throughout.
- the all- welded construction which eliminated crack- arresting plate boundaries which are present in riveted joints - the presence of crack- like flaws in welded joints performed by inexperienced operators pressed into service by the
exigencies of the programme - the use of materials whose low resistance to crack advance ( toughness ) was further reduced by low temperatures.
Closer to home, these photographs of the SS Bridgewater a 29,000 ton Liberian tanker, were shot by the author in Fremantle harbour early in 1962. The vessel though not a Liberty Ship broke in two after a cyclone in the Indian Ocean 400km west of Fremantle. All members of the crew were rescued, and the stern towed back to port.
(click to enlarge),
The first shot shows the what's left of the vessel, down at the stern and tied up at the wharf with a
crowd of curious onlookers examining the fracture just for'ard
of the ventilator.
A close-up of portion of the transverse fracture is shown on the
right. The author also was ignorant of Fracture Mechanics at
the time, so the shots are perhaps not so illuminating as they
might have been. However the photograph suggests strongly that the deck plating fractured instantaneously in
a brittle fashion with none of the ductile tearing which is evident elsewhere.
(arrowed)
10/31/05 10:02 AMDANotes: Fracture mechanics: Maritime examples
Page 2 of 3http://www.mech.uwa.edu.au/DANotes/fracture/maritime/maritime.html
Stern of the SS John P. Gainesin the Aleutian Islands
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
58)
SOURCE: Graphs to the left are from Stern and Stalk (1998: pp. 14, 19), Perspectives on Strategy. The one below is from Thurstone (1919, p. 45) "The learning curve equation," Psychological Monographs. The association below can be described as Y = aXb, where Y is words typed in four minutes, X is cumulative words typed (at 250/page), and the estimated slope b is .42 (cf. slope estimates of .11 to .29 for ship production in Rapping, 1965, p. 65, "Learning and World War II production functions" Review of Economics and Statistics).
Research on semiconductor learning curves shows 20% decrease in cost with cumulative volume doubling, learning three times faster from one's own experience than from experience in another organization, and spillovers between organizations as likely within as between
countries (Irwin and Klenow, "Learning-by-doing spillovers in the semiconductor industry," 1994, Journal of Political Economy).
Some Example Learning Curves
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
59)
Learning curves fromArgote’s 1999 book,
Organizational Learning.Units of cumulative output are omitted to protect confidentiality. Pizzadelivery is from page 9 in book. Squawks per aircraft is from page 8(originally in 1993 British Journal of Social Psychology, “Group andorganizational learning curves”). Hours per vehicle is from page 21
(originally in 1990 Science, “Learning curves in manufacturing.”).
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
60)
Ronald S
. Burt is the
Hobart W
. William
sP
rofessor of Sociology
and Strategy at the
University of C
hicagoG
raduate School of
Business, and the S
hellP
rofessor of Hum
anR
esources at INS
EA
D.
His w
ork describesthe social structure ofcom
petition: network
mechanism
s that ordercareers, organizations,and m
arkets.
When is C
orporate Culture
a Com
petitive Asset?
Sum
mary
-----------------------------------------------------------------------------------A
dvocates speak of corporate culture affecting the bottom line, but
the cited evidence is rarely more than anecdotes, and then
inconclusive. Som
e companies doing w
ell have strong cultures, butother com
panies do well w
ith nothing in the way of shared beliefs
that could be termed a corporate culture. S
o why w
orry about it? Itis to be w
orried about because in certain industries, a strong culturecan be a pow
erful advantage over competitors. T
he complication is
that in other industries, culture is irrelevant to performance. T
hetrick is to know
when culture is a com
petitive asset and when it is
not. Ron B
urt explains with em
pirical evidence how and w
here astrong corporate culture can be a com
petitive asset. Know
ing thecontingent value of culture can be a guide to deciding w
hen to investin the culture of your ow
n organization, when to protect the culture
of an organization merged into your ow
n, and when not to w
orryabout culture.
Culture is to a corporation w
hat it is to any othersocial system
, a selection of beliefs, myths, and
practices shared by people such that they feelinvested in, and part of, one another. P
uttingaside the specific beliefs that em
ployees share,the culture of an organization is strong to theextent that em
ployees are strongly held togetherby their shared belief in the culture. C
ulture isw
eak to the extent that employees hold w
idelydifferent, even contradictory, beliefs so as to feeldistinct from
one another.
Culture effect in theory
In theory, a strong corporate culture can enhancecorporate econom
ic performance by reducing
costs.T
here are lo
wer m
on
itorin
g co
sts. Th
eshared beliefs, m
yths, and practices that definea co
rpo
rate cultu
re are an in
form
al con
trol
mech
anism
that co
ord
inates em
plo
yee effo
rt.E
mployees deviating from
accepted practice canbe detected and adm
onished faster and less vis-ibly by friends than by the boss. T
he firm’s goals
and practices are more clear, w
hich lessens em-
ployee uncertainty about the risk of taking in-ap
pro
priate actio
n so
they
can resp
on
d m
ore
qu
ickly
to ev
ents. N
ew em
plo
yees are m
ore
effectively brought into coordination with es-
tablished employees because they are less likely
to hear conflicting accounts of the firm’s goals
and practices. Moreover, the control of corpo-
rate culture is less imposed on em
ployees thanit is socially constructed by them
, so employee
motivation and m
orale should be higher thanw
hen control is exercised by a superior throughbureaucratic lines of authority.
There are low
er labor costs. For reasons of
social p
ressure fro
m p
eers, the attractio
n o
fpursuing a transcendental goal larger than theday-to-day dem
ands of a job, or the exclusionof em
ployees who do not feel com
fortable with
the corporate culture, employees w
ork harderand for longer hours in an organization w
ith astrong corporate culture. In other w
ords, a strongco
rpo
rate cultu
re extracts u
np
aid lab
or fro
mem
ployees.T
hese savings mean that com
panies with a
stronger corporate culture can expect to enjoyhigher econom
ic performance. W
hatever them
agnitude of the economic enhancem
ent, it isthe "culture effect."
Evidence is m
ixedT
he most authoritative evidence of the culture
effect comes from
a study by Harvard B
usinessS
cho
ol p
rofesso
rs Joh
n K
otter an
d Jam
esH
eskett, based on data published in the appendixo
f their 1
99
2 b
oo
k, C
orp
orate C
ultu
re and
Perfo
rman
ce. Measu
res of p
erform
ance an
dstrong culture are listed for a large sam
ple offirm
s in a variety of broad industries analogousto the industry categories in F
ortune magazine.
To
measu
re relative stren
gth
of cu
lture,
Kotter and H
eskett mailed questionnaires in the
early 1980s to the top six officers in each sample
company, asking them
to rate (on a scale of 1 to5) the strength of culture in other firm
s selectedfor study in their industry. T
hree indicators ofstrong culture w
ere listed: (1) managers in the
firm com
monly speak of their com
pany’s styleor w
ay of doing things, (2) the firm has m
adeits values know
n through a creed or credo andh
as mad
e a seriou
s attemp
t to en
cou
rage
managers to follow
them, and (3) the firm
hasb
een m
anag
ed acco
rdin
g to
lon
g-stan
din
gpolicies and practices other than those of justthe incum
bent CE
O. R
atings were averaged to
define the strength of a firm's corporate culture,
which can be adjusted for the industry average
to make com
parisons across industries.
For exam
ple, Johnson & Johnson is cited as
benefiting from its strong culture in the rapid
recall of Tylenol w
hen poisoned capsules were
discovered on shelves. In the Kotter and H
eskettstudy, Johnson &
Johnson received an averagera
ting
o
f 4
.61
, th
e
hig
he
st g
ive
n
to
ap
harm
aceutical firm
in th
e stud
y, 1.0
7 p
oin
tsabove the 3.51 average for pharm
aceutical firms,
so you see the company to the far right of the
graph below (G
raph 1).R
elative economic perform
ance is plottedon the vertical axis of the graph. K
otter andH
eskett list th
ree measu
res repo
rted to
yield
similar conclusions about the culture effect: net
inco
me g
row
th fro
m 1
97
7 to
19
88
, averag
ereturn on invested capital from
1977 to 1988,and average yearly increases in stock prices from1977 to 1988. F
or illustration here, I use averagereturn on invested capital.
For exam
ple, Johnson & Johnson enjoyed a
17
.89
% rate o
f return
ov
er the d
ecade, b
ut
ph
armaceu
ticals is a hig
h-retu
rn in
du
stry in
which 17.89%
was slightly below
average, soyou see Johnson &
Johnson below zero on the
vertical axis of the graph (17.89 minus 20.21
equals the Johnson & Johnson score of -2.32).
The point is the lack of association betw
eeneconom
ic performance and corporate culture.
Graph 1 contains pharm
aceutical firms, along
with
samp
le firms fro
m b
everag
es, perso
nal
care, and comm
unications — a total of 30 firm
s.N
o extreme cases obscure an association. T
hereis sim
ply no association. The correlation of .06
is almost the .00 you w
ould get if performance
were perfectly independent of culture. K
ottera
nd
He
ske
tt rep
ort a
sligh
tly h
igh
er .3
1co
rrelation
across all o
f their firm
s, bu
t the
correlation was still sufficiently w
eak for themto conclude in their book that: "the statem
ent'stro
ng
cultu
res create excellen
t perfo
rman
ce'appears to be just plain w
rong."
Contingent value of culture
There is a pow
erful culture effect in fact, but itoccurs elsew
here in the economy. G
raph 2, atthe top of the next page, has the sam
e axes asG
raph
1 b
ut p
lots d
ata on
samp
le com
pan
iesfrom
other industries — airlines, apparel, m
otorvehicles, and textiles. T
he 36 sample firm
s fromth
ese ind
ustries sh
ow
a close asso
ciation
between perform
ance and culture; the strongerthe corporate culture, the higher the return oninvested capital.
The key point is illustrated in G
raph 3, which
sho
ws a p
redictab
le shift fro
m cu
lture b
eing
economically irrelevant (G
raph 1) to it being acom
petitive asset (Graph 2). N
ineteen industriesfrom
the Kotter and H
eskett study are orderedon the vertical axis of G
raph 3 by the correlationbetw
een performance and culture. A
pparel isat the top of the graph w
ith its .76 correlationb
etw
ee
n
cu
lture
a
nd
p
erfo
rma
nc
e.
Co
mm
un
ication
s is at the b
otto
m w
ith its
negligible -.15 correlation.T
he horizontal axis of Graph 3 is a m
easureof m
arket competition in each industry. U
singd
ata in th
e pu
blic d
om
ain (p
rimarily
the
benchmark input-output tables published by the
U.S
. Departm
ent of Com
merce; sim
ilar data areav
ailable fo
r agg
regate in
du
stries in m
ost
adv
anced
econ
om
ies), mark
et com
petitio
n is
deriv
ed fro
m th
e netw
ork
effect on
ind
ustry
profit margins of industry buying and selling
with
su
pp
liers
an
d
cu
stom
ers
(thu
s th
e"effective" level of com
petition). The effective
level of market com
petition is high in an industryto the extent that producers show
lower profit
margins than expected from
the network of their
transactions with suppliers and custom
ers (form
easurement details see, under F
urther Reading,
my 1999 paper on com
petition and contingencyw
ith Miguel G
uilarte at the Fielding Institute,
Holly R
aider at INS
EA
D, and Y
uki Yasuda at
Rikkyo U
niversity). G
raph 3 shows that m
arket and culture arecom
plements. T
o the left, where producers face
an effectively low level of m
arket competition,
culture is not a competitive asset. T
hese are the30 sam
ple firms in G
raph 1 taken from the four
Relative C
ulture Strength
(firm score - industry average)
0.0-1.0
-2.01.0
2.0
Y =
.00 + .34 X
r = .06
t = 0.3
Beverages
Pharm
aceuticalsP
ersonal Care
Com
munications
Relative Return on Invested Capital(firm score - industry average)
15%
10%5%0%
-5%
-10%
-15%
Johnson &
Johnson
Graph 1
Scheduled to
appear in anA
utumn, 1999
series in theF
inancial Times
on Mastering
Strategy
from the
Autum
n, 1999Financial Tim
esseries on“M
asteringStrategy”
Appendix VI: Snipits on B
usiness Culture
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
61)
Fu
rther read
ing
:
J. P. Kotter and J. L
Heskett (1992)
Corporate C
ulture andP
erformance.
R. S
. Burt, S
. M.
Gabbay, G
. Holt, and
P. Moran (1994)
"Contingent
organization as anetw
ork theory: theculture-perform
ancecontingency function,"A
cta Sociologica
37:345-370.
J. B. S
ørensen (1998)
"The strength of
corporate culture andthe reliability of firmperform
ance," (http://gsbw
ww
.uchicago.edu/fac/jesper.sorensen/research).
R. S
. Burt, M
. Guilarte,
H. J. R
aider and Y.Y
asuda (1999)"C
ompetition,
contingency, and theexternal structure ofm
arkets," (http://gsbw
ww
.uchicago.edu/fac/ronald.burt/research; also here isthe industry appendixfrom
which the results
in the box are taken).
industries en
closed
by a d
otted
line in
the lo
wer-
left of G
raph 3
. These are co
mplex
, dynam
icm
ark
ets su
ch
as th
e c
om
mu
nic
atio
ns a
nd
ph
arm
aceu
tical in
du
stries, in
wh
ich
pro
fitm
argin
s are good, b
ut co
mpan
ies hav
e to stay
nim
ble to
take ad
van
tage o
f the n
ext sh
ift in th
em
arket. T
here is co
mpetitio
n to
be su
re (see the
1999 p
aper), b
ut th
e poin
t here is th
at a strong
co
rpo
rate
cu
lture
is n
ot a
sso
cia
ted
with
econ
om
ic perfo
rman
ce. (My
colleag
ue at th
eU
niv
ersity
of C
hic
ag
o, Je
sper S
øren
sen
, has
studied
these firm
s over tim
e, and d
escribes in
his 1
998 p
aper o
n reliab
le perfo
rman
ce how
the
cultu
re effect is weak
er for firm
s more su
bject
to m
arket ch
ange.)
At th
e oth
er extrem
e, to th
e right in
Grap
h3, w
here p
roducers face an
effectively
hig
h lev
el
of m
ark
et c
om
petitio
n, c
ultu
re is c
lose
lyasso
ciated w
ith eco
nom
ic perfo
rman
ce. These
are the 3
6 sam
ple firm
s in G
raph 2
taken
from
the fo
ur in
dustries en
closed
by a d
otted
line in
the u
pper-rig
ht o
f Grap
h 3
. In th
ese industries
of
effe
ctiv
ely
h
igh
m
ark
et
co
mp
etitio
n,
pro
ducers are easily
substitu
ted fo
r one an
oth
er,su
ppliers, cu
stom
ers or fo
reign p
roducers are
strong, an
d m
argin
s are low
.
Co
ntin
gen
cy fun
ction
Be
twe
en
th
e
two
m
ark
et
ex
trem
es,
the
perfo
rman
ce effect of a stro
ng co
rporate cu
lture
incre
ase
s with
mark
et c
om
petitio
n. T
he
nonlin
ear regressio
n lin
e in G
raph 3
(the so
lidbold
line), can
be u
sed as a co
ntin
gen
cy fu
nctio
nd
esc
ribin
g h
ow
cu
lture
's effe
ct v
arie
s with
mark
et com
petitio
n. F
or an
y sp
ecific level o
fm
arket co
mpetitio
n o
n th
e horizo
ntal ax
is, the
co
ntin
gen
cy
fun
ctio
n d
efin
es a
n e
xp
ecte
dco
rrelation o
n th
e vertical ax
is betw
een cu
lture
strength
and eco
nom
ic perfo
rman
ce.S
ince in
dustry
scores o
n th
e horizo
ntal ax
isare co
mputed
from
data p
ublicly
availab
le on
all in
du
stries, th
e e
xp
ecte
d v
alu
e o
f a stro
ng
co
rpo
rate
cu
lture
in a
ny
ind
ustry
can
be
ex
trap
ola
ted
from
the c
on
ting
en
cy
fun
ctio
n.
Resu
lts for a selectio
n o
f industries are g
iven
inth
e box to
the rig
ht.
Th
e h
igh
co
rrela
tion
for th
e c
on
ting
en
cy
functio
n sh
ow
s that th
e functio
n is an
accurate
desc
riptio
n o
f cu
lture
's effe
ct in
the d
iverse
mark
ets (r = .8
5, fo
r details o
n d
erivin
g, an
dex
trapolatin
g fro
m, th
e contin
gen
cy fu
nctio
n see
my 1
994 article o
n co
ntin
gen
t org
anizatio
n w
ithS
hau
l Gab
bay
at T
ech
nio
n, G
erh
ard
Ho
lt at
INS
EA
D, a
nd
Pete
r Mo
ran
at th
e L
on
do
nB
usin
ess Sch
ool).
At th
e level o
f indiv
idual firm
s, 44%
of th
evarian
ce in co
mpan
y retu
rns to
invested
capital
can b
e pred
icted b
y th
e industry
in w
hich
they
prim
arily o
perate, an
d th
eir relative stren
gth
of
corp
orate cu
lture acco
unts fo
r anoth
er 23%
of
the v
ariance. C
ultu
re accounts fo
r half ag
ain
the p
erform
ance v
ariance d
escribed
by in
dustry
differen
ces!
Th
inkin
g strateg
ically abo
ut cu
lture
Co
ntin
gen
t valu
e is th
e m
ain
po
int h
ere
. Astro
ng
co
rpo
rate
cu
lture
is neith
er a
lway
sv
alu
ab
le, n
or a
lway
s irrele
van
t. Valu
e is
contin
gen
t on m
arket. A
strong co
rporate cu
lture
can
be a
po
werfu
l co
mp
etitiv
e a
sset in
aco
mm
od
ity m
ark
et. In
a c
om
ple
x, d
yn
am
icm
arket, o
n th
e oth
er han
d, cu
lture is irrelev
ant
to eco
nom
ic perfo
rman
ce.T
he c
on
ting
en
t valu
e o
f cu
lture
can
be a
guid
e to th
inkin
g strateg
ically ab
out cu
lture. T
he
more y
our co
mpan
y’s in
dustry
resembles a co
m-
modity
mark
et, the m
ore eco
nom
ic return
you
can ex
pect fro
m in
vestin
g in
a strong co
rporate
cultu
re. Furth
er, when
you m
erge w
ith a n
ewco
mpan
y, ask ab
out its in
dustry. If th
e industry
resembles a co
mm
odity
mark
et and th
e com
-p
an
y h
as n
o c
orp
ora
te c
ultu
re, th
en
the
com
pan
y's p
erform
ance w
ou
ld b
e hig
her if a
strong cu
lture w
ere instilled
. But if th
e indus-
try resem
bles a co
mm
odity
mark
et and th
e com
-pan
y alread
y h
as a strong co
rporate cu
lture, p
ayatten
tion to
the cu
lture b
ecause th
e com
pan
y's
perfo
rman
ce is som
e part d
ue to
its cultu
re. On
the o
ther h
and, if th
e com
pan
y o
perates in
a com
-plex
, dynam
ic mark
et, you are free to
integ
rateth
e com
pan
y in
to y
our o
wn w
ithout co
ncern
for
whatev
er cultu
re existed
befo
re becau
se cultu
reis irrelev
ant to
perfo
rman
ce in su
ch m
arkets.
Fin
al illustratio
n: co
nsid
er two co
nsu
ltants
assemblin
g resu
lts on th
e perfo
rman
ce effectsof a stro
ng co
rporate cu
lture. O
ne selects 1
0teleco
mm
unicatio
n firm
s for case an
alysis b
e-cau
se he w
ork
ed in
the in
dustry, an
d so
has g
ood
perso
nal co
ntacts th
ere. The o
ther co
nsu
ltant
selects 10 tex
tile firms.
Th
ese
are
two
reaso
nab
le a
nd
inte
restin
gpro
jects, with
a relatively
large n
um
ber o
f firms
for case an
alysis.
There is n
o n
eed to
read th
eir reports. T
he
first consu
ltant selected
an in
dustry
with
a low
effe
ctiv
e le
vel o
f mark
et c
om
petitio
n (th
eco
mm
un
icatio
ns in
du
stry is to
the fa
r left in
Grap
h 3
). A stro
ng co
rporate cu
lture is n
ot a
co
mp
etitiv
e a
sset in
such
co
mp
lex
, dy
nam
icin
dustries. T
his co
nsu
ltant w
ill find n
o ev
iden
ceof h
igher p
erform
ance in
strong-cu
lture firm
s,w
ill gen
eralize his resu
lts to co
nclu
de th
at the
cultu
re effect does n
ot ex
ist, then
earnestly
(since
he h
as research to
support h
is conclu
sion) ad
vise
clie
nt firm
s ag
ain
st wastin
g re
sou
rces o
nin
stitutio
nalizin
g a stro
ng co
rporate cu
lture.
The seco
nd co
nsu
ltant selected
an in
dustry
at the o
ther ex
treme o
f the co
ntin
gen
cy fu
nc-
tion. T
extile p
roducers face an
effectively
hig
hlev
el of m
arket co
mpetitio
n (th
ey ap
pear at th
efar rig
ht o
f Grap
h 3
). A stro
ng co
rporate cu
l-tu
re is a com
petitiv
e asset in su
ch in
dustries.
Th
is seco
nd
co
nsu
ltan
t will fin
d e
vid
en
ce o
fhig
her p
erform
ance in
strong-cu
lture firm
s, will
gen
eralize her resu
lts to co
nclu
de th
at perfo
r-m
ance d
epen
ds o
n d
evelo
pin
g a stro
ng co
rpo-
rate cultu
re, then
earnestly
(since sh
e too h
asresearch
to su
pport h
er conclu
sion) ad
vise cli-
ent firm
s to co
ncen
trate on in
stitutio
nalizin
g a
strong co
rporate cu
lture.
When
these co
nsu
ltants ap
pro
ach th
e same
clients, clien
ts will h
ear earnest, co
ntrad
ictory
results, an
d co
nclu
de th
at the ju
ry is still o
ut o
nco
rporate cu
lture. A
ll of th
ese peo
ple are d
raw-
ing reaso
nab
le conclu
sions w
ithin
the lim
its of
their ex
perien
ce. Nev
ertheless, all are w
rong;
simplistic in
their ig
noran
ce of th
e contin
gen
tvalu
e of a stro
ng co
rporate cu
lture.
-.47
Real estate &
rental
-.08
*C
om
mu
nicatio
ns (n
ot rad
io o
r TV
)-.0
8T
ob
acco0
.06
Bu
siness serv
ices0
.13
Op
tical, op
hth
almic &
ph
oto
grap
hic eq
uip
.0
.15
Ord
nan
ce & accesso
ries0
.16
*F
oo
d (b
everag
es)0
.19
Rad
io &
TV
bro
adcastin
g0
.20
Electric, g
as, water &
sanitary
services
0.2
2H
otels, p
erson
al & rep
air services
0.2
2*
Dru
gs, clean
ing
& to
ilet prep
aration
s0
.26
Sto
ne &
clay p
rod
ucts
0.2
7*
Aircraft &
parts
0.2
7A
mu
semen
ts0
.33
Co
nstru
ction
& m
inin
g eq
uip
men
t0
.38
*P
etroleu
m refin
ing
0.3
9*
Prin
ting
& p
ub
lishin
g0
.43
*P
aper &
allied p
rod
ucts (n
ot co
ntain
ers)0
.44
Wh
olesale trad
e0
.44
Rad
io, T
V &
com
mu
nicatio
n eq
uip
.0
.45
Electric lig
htin
g &
wirin
g eq
uip
.0
.47
Tran
spo
rtation
& w
areho
usin
g (n
ot airlin
es)0
.47
Eatin
g &
drin
kin
g p
laces0
.47
Mach
ines, m
aterials han
dlin
g0
.48
*C
hem
icals0
.48
Fu
rnitu
re (no
t ho
useh
old
)0
.48
Heatin
g, p
lum
bin
g &
struc. m
etals pro
du
cts0
.49
Farm
& g
arden
mach
inery
This is a selectio
n o
f industries fro
m th
e 1982 b
ench
mark
input-o
utp
ut tab
le publish
ed b
y th
e U.S
.D
epartm
ent o
f Com
merce. In
dustries are listed
in o
rder o
f the ex
tent to
which
a strong co
rporate
cultu
re is a com
petitiv
e asset. The fractio
n n
ext to
each in
dustry
is the co
rrelation (p
redicted
by th
eco
ntin
gen
cy fu
nctio
n) in
the in
dustry
betw
een cu
lture stren
gth
and eco
nom
ic perfo
rman
ce. Kotter
and H
eskett in
dustries are m
arked
with
an asterisk
(note h
ow
similar th
e pred
icted co
rrelations
belo
w are to
the co
rrelations in
Grap
h 3
that w
ere observ
ed in
the in
dustries).
0.4
9S
cientific &
con
trollin
g in
strum
ents
0.4
9*
Lum
ber &
wo
od p
rod
ucts (n
ot co
ntain
ers)0.4
9P
aints &
allied p
rod
ucts
0.5
3*
Fin
ance (b
ankin
g)
0.5
3*
Rubb
er & m
iscellaneo
us p
lastic pro
ducts
0.5
4*
Office, co
mp
utin
g &
accoun
ting m
achin
es0.5
7*
Plastics &
syn
thetic m
aterials0.5
8*
Foo
d (n
ot b
everag
es)0.5
8Jew
elry, spo
rts, toys &
oth
er misc. m
anu
.0.6
0M
edical/ed
ucat. serv
ices & n
onp
rofit o
rgs.
0.6
2*
Retail trad
e (no
t eating
& d
rinkin
g p
laces)0.6
3F
inan
ce (bro
kers an
d in
suran
ce)0.6
5M
achin
es, metalw
ork
ing
0.6
6E
ngin
es & tu
rbin
es0.6
7H
ou
sehold
app
liances
0.6
9F
ootw
are & o
ther leath
er pro
ducts
0.7
0M
achin
es, gen
eral industry
0.7
0*
Mo
tor v
ehicles &
equ
ipm
ent
0.7
2E
lectrical ind
ustrial eq
uip
men
t0.7
2F
urn
iture (h
ou
seho
ld)
0.7
3*
Airlin
es0.7
4*
Ap
parel
0.7
4G
lass & g
lass pro
du
cts0.7
5E
lectronic co
mpon
ents &
accessories
0.7
9*
Fab
rics, yarn
& th
read m
ills0.7
9*
Tex
tile goo
ds &
floo
r coverin
gs
0.8
0S
crew m
achin
e pro
du
cts & stam
pin
gs
0.8
7M
achin
es, special in
dustry
Relative Return on Invested Capital(firm score - industry average)
Relative C
ulture Strength
(firm sco
re - industry
averag
e)
0.0
-1.0
-2.0
1.0
2.0
Y =
-.60 +
4.9
0 X
r = .7
2t =
5.8
Apparel
Tex
tilesM
oto
r Veh
icles
Airlin
e
15%
10%
5%
0%
-5%
-10%
-15%
Graph 2
Effective M
arket Co
mp
etition
with
in In
du
stry
Correlation within Industrybetween Performance and Strong Culture
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
textiles
airlines
apparel
ban
kin
g
food
chem
icals
perso
nal
care
publish
ing
com
municatio
ns
bev
erages
aerosp
ace
retail(o
ther)
lum
ber &
pap
er
retail(fo
od-d
rug)
com
puters
petro
leum
moto
r veh
iclesru
bber
pharm
aceuticals
-0.2 0
0.2
0.4
0.6
0.8
These four industries
contain the 30 sample firm
sdisplayed in G
raph 1.G
raph 3
Y =
.941 +
.312 ln
(1-X
)r =
.85
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
62)
from Burt, Hogarth, and Michaud "The social capital of French and American managers" (2000, Organization Science)
X
X
X
XX
XX
XXX
X
XX
X
X
X
X
XXX
X
XXXX
X
XXX
XX
XX
XXX
XX
X
X
XX
XX
XX
X
X
XXX
X
X
X
X
X
X
X
X
X
XXX
X
XX
X
X
XX
X
X
X
X
X
X
XXXXX
X
XXX
X
X
XX
X
X
X
X
X
X
X
X
X
XX
X
X
XX
X
X
X
X
X
XXX
XXX
X XX
X
X
X
X
X
X
XX
X
X
X
X
X
XXX
XX
X
X
X
XX
X
X
X
XXX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XXX
X
X
XX
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XXX
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
XX
X
XX
X
X
X
X
XX
X
XX
X
X
X
XX
X
X
XXX
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
XX
X
XX
X
XX
XX
X
XXX
X
X
X
X
XXX
X
X
XX
XXXXX
X
X
XX
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
XX
X
XX
X
XX
X
X
XX
X
XX
X
X
X
X
X
X
X
X
X
X
XX
X
XX
XX
X
XXX
X
X
X
X
XX
X
X
X
X
X
X
XX
X
X
XXX
X
XXXX
X
X
X
X
XX
X
X
X
X
X
XXX
X
X
X
X
X
X
X
X
XX
X
XX
X
X
X
XX
XX
XX
X
X
XXX
X
XXXXX
X
XXX
XX
XXX
X
X
X
X
X
X
X
X
X
XXXXX
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
XX
X
XXXXXX
XXX
XXXX
X
X
XXX
XX
X
XX
XXX
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
XXXX
X
XX
X
X
X
X
XXXX
X
XX
X
XX
X
XX
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XXX
X
X
X
X
XXX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
XX
XXX
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
XXXXX
X
X
XX
XX
X
X
X
X
X
X
XX
X
X
X
X
XX
XX
X
X
X
X
X
X
X
X
X
X
X
XX
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XXXXXXXXXXX
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XXXX
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
XX
X
XX
X
X
X
XX
XXX
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
X
X
XX
X
X
X
X
X
X
X
X
XXX
XX
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
XXXXX
XX
XX
XX
X
X
X
XX
X
X
X
X
X X
XX
XX
X
XX
X
X
XX
X
X
X
XX
XX
XX
X
X
X
XX
X
X
X
X
X
XX
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
XX XX
X
X
X
X
X
X
X
X
X
X
X
X
XXX
XX
X
X
X
XX
X
X
X
XX
X
X
XX
X
X
X
X
X
XX
X
X
XXXXXXX
X
X
X
X
XXX
X
XX
X
X
X
X
X
XX
X
X
X
XXX
X
X
X
X
X
X
XX
X
X
X
XXX
XX
XX
XX
X
X
XX
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
XXX
X
XXX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
XX
X
XX
X
X
X
X
XX
X
X
X
X
X
XXXX
X
X
X
X
X
X
X
X
X
X
XX
X
X
XX
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
XX
X X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
X
X
X
X
XX
X X
XX
X
X
X
XX
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
XXX
X
X
XX
X
X
XX
X
X
X
X
XXX
X
X
X
XX
X
X
XXX
X
X
X
X
X
X
X
X
X
XX
X
X
X
XX
X
X
X
X
XX
XX
X
X
X
XX
XX
X
X
X
X
X
X
X
XX
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
XXXX
XX
X
X
X
X
X
XX
X
X
X
X
X
X
XX
XXXX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
XX
X
X
X
X
X
XXX
X
X
X
XXXXX
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
XX
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
XX
XX
X
XX
X
X
XX
X
X
X
X
X
XXX
X
X
X
X
X
X
X
XX
X
XXX
X
X
X
X
X
XXX
XX
X
X
X
XX
X
X
X
X
X
XX
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XXX
X
X
X
X
X
X
X
X
X
X
X
XXX
X
X
X
X
XX
XX
X
X
X
X
X
X
XX
X
XX
X
X
XX
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
X
XX
X
X
X
X
X
X
X
X
X
XX
XX
XX
X
X
X
X
X
X
X
X
X XXX
X
X
X
X
X
XX
X
X
X
X
XX
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
XX
XX
X
X
X
X
X
X
X
X
XXX
X
X
X
X
X
XX
X
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
X
X
X
X
XXX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
XX
XX
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
XX
XX
X
X
X
X
X
XX X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
XX
X
X
X
XXX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
XX
X
X
XXX
X
XXXX
XX
XX
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
XX
X
X
X
X
X
X
X
X
X
X
XX
XXXXX
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
XX
X
XX
X
XXX
XX
X
XX
X
X
X
XX
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
XX
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
XXX
X
X
X
XX
XX
X
XX
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
XXX
X
X
X
X
X
X
XXX
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
XXX
XX
XX
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
XX
X
X
X
XX
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
XX
X
X
X
X
X
X
X
X
XX
X
X
XX
XX
X
X
X
XX
X
XXX
X
XX
X
XX
X
X
X
X
X
X
X
XX
X
XX
X
X
X
X
XXX
X
X
X
X
X
X
X
X
X
XX
X
XXX
XX
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XXX
XX
X
X
X
X
X
XX
X
X
X
X
X
X
XX
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
XX
XX
X
XX
X
X
X
X
X
X
XX
X
X
X
XX
XXX
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
3010 25201550
French Manager Years in the Firm3010 25201550
American Manager Years in the Firm
Year
s A
cqua
inte
d w
ith C
onta
ct
30
10
25
20
15
5
0
30
10
25
20
15
5
0
Years inthe Firm
0 to 10
11 to 20
Over 20
Total
NumberColleagues
105
160
391
656
French Managers
% KnownBefore Firm
26%
15%
5%
11%
Mean YearsKnown
5.2
8.2
10.3
9.0
NumberColleagues
691
875
129
1695
American Managers
% KnownBefore Firm
81%
42%
6%
55%
Mean YearsKnown
12.6
13.5
14.9
13.0
Distinctions Between
Inside and Outside the
Firm(colleague relations
pre-dating entry into the firm)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
63)
A A
A
A
B B
B B C C
C C
D D
D
D E E E
E
WHEEL (32.0 sec)
N NC Happy
A 1 100 37.5
B 1 100 20.0
C 4 25 97.0
D 1 100 25.0
E 1 100 42.5
Avg 1.6 85.0 44.4
Most Distributed Leadership
(slow, happy)
Most Centralized Leadership
(fast, unhappy)
Y-NETWORK (35.0 sec)
N NC Happy
A 1 100 46.0
B 1 100 49.0
C 3 33 95.0
D 2 50 71.0
E 1 100 31.0
Avg 1.6 76.7 58.4
CHAIN (53.2 sec)
N NC Happy
A 1 100 45.0
B 2 50 82.5
C 2 50 78.0
D 2 50 70.0
E 1 100 24.0
Avg 1.6 70.0 59.9
CIRCLE (50.4 sec)
N NC Happy
A 2 50 58.0
B 2 50 64.0
C 2 50 70.0
D 2 50 65.0
E 2 50 71.0
Avg 2.0 50.0 65.6
Appendix VII: Network Endogeneity
The four networks are from the Bavelas-Leavitt experiments on leadership in task groups. The WHEEL is a traditional bureaucracy in which C is in charge. The other three networks involve distributed leadership (all five people in the CIRCLE; B, C, and D in the CHAIN; C and D in the Y-NETWORK). More distributed leadership is associated with more messages, slower task completion, and greater enjoyment. Speed, messages, and enjoyment scores are from Leavitt (1951). Number of contacts (N) and network constraint (NC) are computed from binary ties in the sociograms (number of contacts equals number of non-redundant contacts in these structures).
Figure 2.4 in Burt (2018, Structural Holes in Virtual Worlds)
Stra
tegi
c Le
ader
ship
Foun
datio
ns (p
age
64)
Behavioral and Opinion Correlates of Network Brokers
Figure 2.5 in Burt (2018, Structural Holes in Virtual Worlds)
Network Constraint
Mea
n En
joym
ent S
core
Network Constraint ( )
Tim
es C
ited
as G
roup
Lea
der
Network Constraint
Mea
n M
essa
ges
Sent
Answer messagesInformation messages
Enjoyment after first trialEnjoyment after last trial
A. Network brokers tend to distribute answers, people in moderately constrained positions tend to be conduits for informational messages.
Data are from Leavitt (1949: Table 30, following page 62).
B. Network brokers are least happy initially, but eventually become the most pleased with the experience.
Data are from Leavitt (1949:Table 29, pages 60-61; "How did you like your job in the group?).
C. The final outcome, by the end of the experiment, is that network brokers are most likely to be recognized as the unofficial group leader.
Data are from Leavitt (1949: Table 8, page 38; “Did your group have a leader? If so, who?”).