Title: The Hidden Power of Social Networks: Developing Knowledge Maps Using Social Network Analysis
1The Hidden Power of Social Networks Developing
Knowledge Maps Using Social Network Analysis
- Dr. Jay Liebowitz
- Professor
- Graduate Division of Business and Management
- Johns Hopkins University
- jliebow1_at_jhu.edu
2Sharing Knowledge is Power?!(December 10, 2004)
3Where is the Business Intelligence?
4How to Battle the Coming Brain Drain(Anne
Fisher, Fortune, March 21, 2005)
- By 2010, more than half of all workers in the
U.S. will be over 40. Tens of millions of
baby-boomers turn 60 this year, and the decade
ahead will see vast numbers of people retiring,
or at least leaving their current full-time
careers. - General Electric, Dow Chemical, Northrop Grumman
capture their oldsters knowledge disseminate
it to younger workers before its too late
5Lost Knowledge
- In sectors such as government, manufacturing,
energy, health care, education, and aerospace,
knowledge retention will become an overriding
concern in the years ahead. David DeLong, Lost
Knowledge Confronting the Threat of an Aging
Workforce, Oxford University Press, 2004
6Keynote Address Medical Informatics and
Emergency Medicine (Dr. Feied, 2004 AEM Consensus
Conference on Emergency Medicine Informatics)
- Work to reduce institutional dependence on
specialized personnel with secret knowledge
that allows them to complete tasks nobody else
can perform. If the unit secretary is the only
one who knows how to place or cancel an order,
every coffee break can put a congested department
further behind.
7The Hidden Power of Social Networks (Cross, 2004)
- Networks of informal relationships have a
critical influence on work and innovation - Research shows that appropriate connectivity in
well-managed networks within organizations can
have a substantial impact on performance,
learning, and innovation - Actors/nodes (individuals/units) and links/arcs
(relationships/ties)
8Conducting a Social Network Analysis
- Step 1 Identify a Strategically Important Group
- Integrating networks that cross core processes
- Promoting innovation through connectivity in new
product development or process improvement
initiatives - Facilitating post-merger integration and
large-scale organizational change - Supporting CoPs
- Forming strategic partnerships and alliances
- Improving learning and decision making in top
leadership networks
9Size of the Network
- Typically limit to 250 people for visualization
purposes (but could have more for analysis
purposes) - Ask each person to rate his/her interactions with
other members of the group
10Step 2 Assess Meaningful and Actionable
Relationships
- Relationships that reveal collaboration in a
network (e.g., Communication, Information,
Problem solving, Innovation) - Relationships that reveal the information sharing
potential of a network (e.g., access) - Relationships that reveal rigidity in a network
(e.g., decision making, influence) - Relationships that reveal well-being and
supportiveness in a network (e.g., liking,
friendship, trust)
11Survey
- Confidentiality
- Full disclosure of the results via an
all-employee debriefing or other communication - Allow only one person or a select group to see
the full results so that helpful action can be
taken - Disclose no names whatsoever
12Step 3 Visually Analyze the Results
- UCINet-Netdraw
- Pajek
- Netminer
- Inflow
- International Network of Social Network Analysts
(INSNA)
13Staff for General Advice Questions
14Junior-Senior Staff Relationships
15Network Diagram General Advice (Individual)
16Network Diagram Subject Matter Expertise
(Department)
17Social Network Analysis (www.orgnet.com)--Org.
Mapping
18Step 4 Quantitatively Analyze the Results
(Indiv. Network Measures)
- In-degree centrality (the number of incoming ties
a person has for a given relationship (such as
communication or trust) - Out-degree centrality (the of outgoing ties a
person has for a given relationship) - Betweenness centrality (the extent to which a
particular person lies between the various
other people in the networknetworks that contain
individuals with high betweenness are vulnerable
to having info flows disrupted by power plays or
key individuals leaving)
19Indiv. Network Measures (cont.)
- Closeness centrality (the extent to which a
person lies at short distances to many other
people in the network. Persons highly central
with respect to closeness tend to hear info
sooner than others) - Brokerage measures
- Coordinators (people who broker connections
within the same group) - Gatekeepers (people who broker connections
between their own group and another) - Liaisons (those who broker connections between 2
different groups)
20Group Measures
- Density (the of individuals who have a given
type of tie with each other, expressed as a of
the maximum possibleif each person were
connected to every other person in the network,
the density would be 100) - Cohesion (the average of the shortest paths
between every pair of people in the
networkaverage cohesion score should be around 2
in groups where managers are interested in
employees leveraging each others expertise)
21Step 5 Create Meaningful Feedback Sessions
- Conduct interviews with a select number of people
(8-10 network members) to better understand the
dynamics behind the network - Look at the network analysis to determine who to
interview (central and peripheral roles) - Documentation or workshop with all network
participants
22Step 6 Assess Progress and Effectiveness
- Follow-up (post-audit) 6-9 months after network
analysis is first conducted - SNA research areas
- Improving the calculation of interval/ratio
scales for measuring the strength of ties between
actors (AHP) - Effective representation of complex network
information (IBMs Raison, data mining tool)
23Identifying Types of Individuals in the Network
- Central Connectors most arrows pointing to them
- Unsung Hero Invisible work but critical to
the network and can consume many hours each day
(may be completely unrecognized by senior
management) - The Bottleneck some people become so central to
the organization that they end up holding the
group back
24Boundary Spanners
- Provide critical links between 2 groups of people
that are defined by functional affiliation,
physical location, or hierarchical level - Can play important roles in large-scale change
efforts
25Information Brokers
- Indirect connections
- For example, Ian and Frank have no direct
connection, but each has a relationship with
Gayle (so they are connected through her) people
such as Gayle play a brokering role that can hold
together entire groups
26Peripheral Specialists
- Have one connection each and are not linked to
each other - Some people are stuck (e.g., newcomers) on the
periphery and others may choose (e.g., experts)
to be on the periphery
27Case Study Examination of Raw Data
- All data downloaded to Excel spreadsheet
- 225 respondents answered questions 1, 2, and some
portion of 13, identifying other employees from
whom they sought a form of knowledge advice - All others were removed, including duplicates
- All cells containing names were examined
identifying abbreviated names, the use of
initials, variations in spelling, etc. - Numbering of all names yielded 698 employees
28Examination of Raw Data
- The data identified 1,621 knowledge advice
connections or communications between the 698
employees - Each connection between employees was assigned a
weight based on the frequency of the knowledge
being sought and the importance of the knowledge - Complete employee data for respondents,
non-respondent employee numbers, employee
connections and connection weights were loaded
into NetMiner 2.5 for social network analysis. - Analysis was conducted across 6 defined advice
communication areas
29Social Network Analysis Schematic
Employee
Adjacency Advice Communication Matrix
Advice Communications
Employee
Employee Attributes
Communication Measures
30Social Network Analysis Schematic
Employee
Adjacency Advice Communication Matrix
- Advice Communications
- Context K Advice
- Expert Process K Advice
- General K Advice
- Process K Advice
- Strategic K Advice
- Relationship K - Advice
Employee
31Knowledge Types
Context K of what applications
Expert Process K of how networks and systems work
General K of non-work related questions
Process K of how the business works
Relationship K of who has information
Strategic K of why business opportunities reduce cost and other strategic decisions
32Social Network Analysis Schematic
Employee
Adjacency Advice Communication Matrix
- Employee Attributes
- Department
- Level (Position Hierarchy)
- Tenure
Employee
33Social Network Analysis Schematic
Employee
Adjacency Advice Communication Matrix
- Communication Measures
- Centrality (in and out degree)
- Closeness
- Betweenness
- Cliques
- Density
- Brokerage
Employee
34Is there much intra-departmental communication?
- This can be measured by examining the amount of
communication within each department in two
areas density and cohesiveness - Density is the proportion of possible lines that
are actually present in the network. It is the
ratio of the number of lines present to the
maximum possible. - Cohesion Index is the extent to which ties are
concentrated within a subgroup, rather than
between subgroups.
35Is there much inter-departmental communication?
- Departments blocked and then examined for
degree and density measures - As with individuals, departments yielded low
density of communication with other departments - Some departments have no communication with
others within certain knowledge areas
(structural holes) - XYZ department unconnected from all others in 4
of the 6 knowledge types, including the knowledge
of who has the organizational information - JKL department not connected within 3 of the
knowledge types including the what of
applications and the how of network and systems
work
36(No Transcript)
37(No Transcript)
38Is the organization well connected among
employees?
- Most communication is occurring between
non-managerial employees - Of the 1,621 knowledge connections, 1,199 were
connections to non-management employees - 74 of reported advice communications were
employee to employee
39Is the organization well connected among the
employees and the managers, directors and
executives?
- 336 of the 1,621 knowledge connections (21) were
to persons in management positions - 36 of 62 Managers, Directors, and Executives were
named as sources of knowledge advice 5 or more
times - Communication rate from employee to employee is
the highest
40Are the junior employees interacting with the
senior employees?
- Junior employees have limited contact with senior
employees (executives) - Employee contact is greater with other employees
- Greatest direct connection between executives and
non-managerial employees was in the Process K
area - Lowest level Context K
41(No Transcript)
42(No Transcript)
43(No Transcript)
44(No Transcript)
45(No Transcript)
46(No Transcript)
47Are the directors and executives in the Power
positions centrally located?
- Executives and directors are not present or are
relatively weak in most power or central
positions in knowledge advice communications - A single executive was prominent in Strategic,
General, and Relationship knowledge areas - Three others in managerial positions appeared in
central positions, with a single one of these
appearing in 4 of the 6 K types - Most power or central positions are held by
non-managerial employees or a single expert
48(No Transcript)
49(No Transcript)
50(No Transcript)
51(No Transcript)
52(No Transcript)
53(No Transcript)
54Are there correlations as to those employees
sought based on the different types of knowledge?
- No evident correlation between knowledge type and
employee attributes in central or power positions
within each knowledge area - While individual actors hold multiple positions
across K area, no single department appears
strongly central with a given K type - Common factor for all employees sought for advice
(with complete attribute data) was tenure greater
than 1 year (with the exception of employee 188,
significantly in the power position for Strategic
Knowledge)
55Are there more isolates, transmitters, receivers
or carriers in the organization?
- In all knowledge areas, the greatest number of
employees are isolates - Receivers are generally greater in number than
transmitters in all knowledge areas - Carriers are the fewest in number in all
knowledge areas
56Node Types by Area
K Area Isolate Transmitter Receiver Carrier
Context 329 133 190 46
Exp. Pr 361 126 171 40
General 378 118 162 42
Process 442 110 109 37
Relation 445 94 121 38
Strategic 396 113 151 38
57How can communication and collaboration be
improved?
- Further departmental examination of results
- Follow-up interviews with persons in central or
power positions - Task analysis of these persons performance in
the positions - Positive reinforcement program should be
developed to create behavioral change in other
positions, based upon benchmarks set by central
or power positions - Planned insertion of employees or departments to
fill all structural holes should be considered
for succession planning and workforce development
58Organizational Individual Constraints
(Q.25,26,27)
- 37 Lack of Time/Work Overloads
- 14 Too Many Silos/Consolidate into Central
Repository - 10 Not knowing who to ask/who has the
information - 6 Job Insecurity/Downsizing
- 6 Getting People to Share Their
Knowledge/Culture Doesnt Encourage it/Tie to
IPEP - 6 Search Engine Needs to be Improved/No Search
Engine for Internal Info - Other (21)(in order)
- Business processes needed/Standardization/People
to follow processes - Politics
- Rigid hierarchical structures
- Supervisor is seated away from Dept/Geographic
Dispersion - Limited access, due to security, to access
intranet from home - More communication needed between people
- Need knowledge capture tool
- Constantly changing organization with unclear
roles and responsibilities - Need to create a continuous learning culture
- Need to capture rationale why things dont work
59KM-Related Recommendations Based on SNA,
Q25,26,and 27
- Portal
- Central repository
- Google-type search engine
- Expertise locator
- Recognize knowledge sharing in performance
reviews (extrinsic and intrinsic motivators) - Lessons learned/best practices
- Online communities of practice
- More cross-training/cross-functional teams
- Mentoring program
- Knowledge sharing forums (storytelling)
- Knowledge fairs around certain key topics
60Summary
- Low volume of communications due to an unbounded
network and low levels of advice seeking - Departments have relative cohesions of
communication but often fail to communicate with
other divisions across all knowledge areas,
creating structural holes - Within the Process K area, executive level actors
are most strongly connected to non-managerial
employees - Experts, while present in the analysis, do not
appear in powerful or central positions (with the
exception of employee 158)
61Summary (cont.)
- Other individuals appear in powerful or central
positions and may be considered experts relative
to the knowledge area (the most significant of
these is employee 157) - All knowledge areas demonstrate high numbers of
isolated actors and fewer numbers of transmitters
and carriers (employees not communicated with
others within certain knowledge types) - Apply knowledge management/knowledge sharing
approaches
62Final Comments
- Social network analysis is a wonderful technique
to identify knowledge flows and knowledge gaps in
organizations to help in knowledge
mapping/knowledge audits - The grapevine effect is stronger than the formal
org-chart effect (i.e., informal networks are
stronger than formal ones) - SNA helps build a basis for developing a
knowledge management and human capital strategy
63Questions and Answers