Improving and Evaluating Results with Social Network Analysis
National Public Health Leadership Network
The National Public Health Leadership Network brings together 7,000 graduates from statewide, regional, national, and international leadership programs that seek to strengthen the public health infrastructure around the world. These efforts were initially begun by the Center for Disease Control in 1990 in response to an Institute of Medicine report calling for increased leadership skills among the nation's (and the world's) public health leaders. The initial leadership program enrolled senior leaders from local, state and federal levels of public health, as well as public health academia, health care organizations and national health organizations. A number of the 800 leaders that this program trained went on to establish state and regional leadership programs around the country, providing a multiplier effect (Umble et al., 2007)Last Updated on Sunday, 18 April 2010 10:06
The network first formed among state and regional leaders who had graduated from the program and who offered each other "plenty of advice and support from those who had already walked the road." These are the characteristic qualities of peer leadership networks. Over time, however, the network began to weave independent state and regional strands together "into a common 'rope' or movement that could pull the field forward together." (Umble et al., 2007). Network members took up a number of initiatives to benefit the field of public health. They developed documents on ethical practice in public health that were officially adopted by the American Public Health Association in 2002. They issued white papers on the public health workforce, workforce development, and leadership. The Network has also developed a "Public Health Leadership Competency Framework" that includes transformational, political, trans-organizational, and team building competencies. All of these activities have strengthened the leadership of the public health field.
Social Capital of Twenty-First Century Leaders
[The following was originally posted to Connectedness February 1, 2006. Reprinted with permission of the author.]
How do effective leaders manage relationships with their team members? Dan Brass and David Krackhardt put this question in context:
"Despite the acknowledgment that leaders spend most of their time interacting with others, the social capital of leaders is perhaps the most ignored, under-researched aspect of leadership."They go on to explain the critical role of social capital in their article "The Social Capital of Twenty-First-Century Leaders" (which is the ninth chapter of Elsevier's Out of the Box Leadership):
"In contrast to human capital (traits and behaviors), social capital refers to relationships with other actors, and the accompanying access to information, resources, opportunities, and control. Because organizational leadership involves accomplishing work through others, it is critical that we assess the social capital of leaders."What does the social capital of leaders tell us? Brass and Krackhardt summarize the state of the art with two broad strategies: "strong tie" and "weak tie."
Connecting to Central Others—The Strong Tie StrategyStrong ties refer to close, trusting relationships. To motivate the strong tie strategy, consider which connections will most help Greg to lead ABC Associates, pictured below. Each link represents the regular sharing of work-related information:
With the additional insight that Stephanie, Stewart, and Kathy each manage separate sub-teams within the above organization, it becomes clear that Greg should establish strong ties to these three central figures.
The strong tie strategy helps Greg support a cohesive team with just a few direct relationships. Notice that it also makes Greg completely dependent on Stephanie, Stewart, and Kathy. This may not be a problem, since Greg is able to build loyal and trusting relationships with each of these highly connected others. However, this loyalty comes with a price: Greg's tight-knit team of managers will be relatively impervious to outside ideas and highly susceptible to group-think.
Connecting Others Who Are Not Connected—The Weak Tie StrategySuppose that ABC Associates flattens its organizational hierarchy by eliminating middle managers and increasing collaboration among their former direct reports. At such a re-engineered ABC Associates, Greg faces a more complex terrain of relationships where the strong tie strategy breaks down:
With responsibility for such a dispersed team, Greg can no longer establish his leadership with just a few loyal lieutenants. Instead, Greg must rely also on weak ties, which refer to acquaintances and other distant colleagues.
In the example above, Greg connects to each of his seven work groups with a strategic combination of strong ties (dark edges) and weak ties (light edges). With his weak tie strategy, Greg wins influence over the organization through his unique access to a wide variety of novel resources and information. As long as he can continue to maintain relationships with such a diverse group of contacts, Greg is in an ideal position to promote innovation within his organization.
Choosing Between the Strong Tie and Weak Tie StrategiesBrass and Krackhardt devote the second half of their paper to contingencies that influence the relative effectiveness of the strong tie and weak tie strategies. Briefly, these include
Commonwealth Software (a pseudonym) is an example of an organizational leadership network. It is a young company with 50 employees that is planning for rapid growth. In order to grow effectively, the CEO of Commonwealth started an "Emerging Leaders" program. About a year later, Commonwealth used social network analysis to evaluate all its leadership development efforts, including the Emerging Leaders program.
The figures below display some of the basic data that informed the evaluation: the weekly advice network of Commonwealth Software. Node shapes correspond to formal departments: circles are members of administration, sales, and marketing departments; triangles are members of the software engineering department; and squares are members of the product development department. Two distinct teams within the product development department are indicated as two different shades of squares. Node sizes indicate betweenness centrality (i.e., bridging).
The first goal of the Commonwealth Software evaluation — and the goal most directly related to its Emerging Leaders program — was to check the CEO's assumptions about who were the emerging leaders. The overall view confirmed the emerging leadership perceived by the CEO, especially the leadership of one new employee, node 36 in the map.
Overall View: Weekly advice network of Commonwealth Software
Node size corresponds to betweenness (bridging).
Further analysis of the advice network and related data revealed an overtaxed senior leader in the organization (node 24 in the overall view, above). Members of this leader's own department reported that he was frequently busy and inaccessible. Members of other departments reported relatively little interaction with him. These data highlighted a specific area for improvement in the existing leadership network and underscored the need for the Emerging Leaders program.
Another goal of the social network analysis was evaluating the collaboration between the two product development teams of Commonwealth, one of which occupied a different office than the rest of the company. The departmental view below shows the advice network of Commonwealth, with specific focus on formal organizational boundaries. The network map is drawn so that employees in the same department are close to each other; thick links show intra-departmental advice and thin links show inter-departmental advice.
The network map of the departmental view revealed a surprising gap in the company's existing leadership development efforts. Because Product Development Team A worked in the same office as the rest of the company, and Product Development Team B worked in an office many miles away, the CEO was concerned that Team B might be isolated. In fact, the network evaluation revealed that Team B was not isolated from the rest of the company; instead, members of Team B were relatively isolated from each other.
The CEO used the results of the network evaluation to reorganize the formal leadership structure of Commonwealth Software. The overtaxed senior leader was repositioned and additional leadership resources were added in order to help wean employees from their dependence on him. Realigning Team B — a team with strong historical ties to this same senior leader — was an integral part of this effort.
When interviewed after the evaluation and restructuring, the CEO of Commonwealth reported that, even though the evaluation proceeded as he had hoped, its most important benefits to him were entirely unanticipated. Maps such as those above gave him a new perspective on how the pieces of Commonwealth fit together. He reoriented his leadership development focus from individuals to the collective leadership system. We will speak more about the CEO's appreciation of the collective leadership system in our next installment of this case study, when we see how SNA revealed to the CEO that a key leader of his company was someone outside the formal organization altogether.
Departmental View: Weekly advice network of Commonwealth Software
Node size corresponds to betweenness (bridging).
Organizational Network Analysis Utility
[The following was originally posted to Connectedness November 30, 2005. Reprinted with permission of the author.]
Organizational network analysis provides intuitively compelling pictures of how work really happens, giving us a handle on slippery intangibles that drive the future success of an enterprise.
Any questions? Please post them on this thread in our discussion forum.
[The following was originally posted to Connectedness September 8, 2005. Reprinted with permission of the author.]
How can we contain the AIDS epidemic in Africa? How do new ideas take hold in an organization? Social network analysts frame both of these questions in terms of diffusion. And as experts in the field, social network analysts might be expected to practice optimal strategies in combatting STDs and spreading good ideas fast. But we are human too (thank god). So perhaps it is not surprising that I can still impress my colleagues as an innovator by sharing an article published twelve years ago in the Harvard Business Review.
When rivalries flare among his four divisions, Leers decides to bring key leaders from each division together into a strategic task force. His initial choice of leaders turns out to be off the mark, but a careful look at information, advice, and trust networks in the company helps him see how to resolve the situation.
As I mentioned, several colleagues of mine have been positively blown away by this article. So the big question is, how can an HBR article from 1993 still strike people as "the next big thing"? I went straight to the source and asked David Krackhardt that very question. Here's what he had to say:
"I think that one of the reasons is that there are a lot more software packages that support analyzing these networks in ways that were not possible before. Up until fairly recently, all of us in the field were also programmers, having to write our own code to do the analysis (and and visual renderings). Coupled with the high profile interest that physicists have taken in the area, and you have lots more exposure. And some of this work is also tapping into large scale networks, like the WWW or email connections, which in some ways is much easier data to collect and write about."
Speaking of software packages for SNA, I put David's HBR case study into Steve Borgatti's NetDraw and came up with this image of the advice network:
Each division is a different color. The size of the node corresponds to the influence that individual has over the advice network. Then, since the trust network also plays a key role in this case study, I represented influence over the trust network by scaling the size of the node labels (i.e., trusted individuals literally have big names). Near the middle of the network, you can see a big yellow node illegibly labelled "Calder" because Calder is so professionally competent and at the same time personally abrasive. Not necessarily good leadership material.
Unlike 1993, the problem with SNA today is preventing information overload. There are so many networks we can collect data on, and so many gee-whiz tools to display them, that it's important to keep focused on the big picture.
MomsRising.orgMomrising.org is a national on-line network of over 150,000 citizen members who are concerned about building a more family-friendly America. MomsRising.org uses the power of on-line organizing to mobilize citizen advocates around motherhood and family issues. Two recent successes included the passage of a paid family leave bill in Washington in 2007 and the prevention of the Consumer Product Safety Commission from requiring toxic chemicals to be sprayed on furniture to make it flame-retardant. MomsRising.org wants to change the culture that tolerates workplaces, policies, and social priorities that do not support families.
One of MomsRising.org's core strategies is partnering with aligned organizations on joint campaigns and getting the word out on breaking issues to members who can take rapid action. They currently have 85 partner organizations who work together to enact family-friendly policies by mobilizing their vast grassroots networks. The partnerships enable each organization to accomplish much more with the resources they have and to more effectively promote a full-range of motherhood and family policy initiatives.
In a recent Stanford Social Innovation Review article about MomsRising.org (Gehl, 2008), featured in the magazine's "What Works" section, the author summarizes what works to mobilize moms.
The Sierra Health Leadership Network
The Sierra Health Leadership NetworkThe Sierra Health Leadership Network includes 130+ nonprofit executive leaders from 21 northern and central California counties who work on health-related issues. All of these leaders have participated in the Sierra Health Leadership Program, a nine-month program of retreats, leadership training sessions, team action learning projects, and an alumni network. Leaders bond with each other during their leadership program experience which includes many opportunities for self-reflection, clarification of core values, and finding one's "noble cause" in conversation with others. The foundation hosts retreats three times a year to reinvigorate relationships; reinforce core learning from the program; explore new topics and ideas; and expand connections to other cohorts. In an evaluation of the alumni network, Reinelt, Kubo and Hoppe (2005) found the most important outcomes to be: