SNA of a “collective impact” initiative: Shape Up Somerville

Overview of collective impact

A few weeks I posted an article from the NYTimes about emerging community-based efforts to address social problems, sometimes referred to as “collective impact”. The idea is that by bringing a variety of public, nonprofit and private sector partners to the table, agreeing to a common vision and measurable goals, and pursuing evidence-based actions toward those goals, communities are capable making a difference in addressing problems that no single institution can accomplish alone. The idea of multiple community actors cooperating toward a common goal is nothing new. What is new is the depth of collaboration and sharing of resources across different actors, facilitated by several factors: technology and communications technology have greatly increased the ease of information and data sharing; trends in the nonprofit and public sectors toward quantitative methods of evaluation and “best practices” programming; the emergence of the public health field; the growth and mainstreaming of CSR.

Intro to Shape Up Somerville

One of the most lauded recent collective impact efforts is underway in Somerville through an initiative called Shape Up Somerville (SUS), which is coordinated by the SUS Task Force in the City’s Health Department. Recognizing the need to address childhood obesity, SUS was started in 2002 and designed largely by Friedman School professor Christina Economos. The central SUS goal was to decrease the average student intake by 125 calories a day by increasing their physical activity in and outside the classroom, increasing their intake of healthy foods, decreasing consumption of less healthy foods, and enhancing students’ knowledge of the link between diet, exercise, and health. (http://bit.ly/naJGbM)

Some of the activities coordinated under the SUS banner include improving the city’s walkability to encourage students to walk to school, increasing the number of bike paths, adding healthier options to the school lunch menu, incorporating physical activity in the classroom, increasing access to healthy foods outside of the schools through farmers markets, and promoting healthy meal options at local restaurants. (http://bit.ly/rithQM)

It takes a village: the network of shareholders

To execute all of these activities, the project involved over 25 stakeholders in the project, including teachers, parents, school nurses, school food staff, after school and nonprofit program managers, local restaurants, farmers markets, and city policymakers.

Evaluating SUS: the role of SNA

The program is recognized as a success, with published studies showing that Somerville students in grades 1-3 gained significantly less – about 1 lb. per student – than students in comparable schools without this type of intervention. (http://bit.ly/pOcLLb)

This success is chalked up to the fact that so many different players are committed to achieving the same goal: it’s a networked effort. But what is this network? The SUS Task Force might produce an organizational chart with them as the hub of a spoke-like network, connecting them to all the different players, and maybe some of them to each other. But is this a realistic depiction of how the network really functions? And what of the dynamics within the network that help make it so successful? No one has yet examined these questions, and this is why I would like to conduct an SNA of the SUS initiative, which I believe can make a meaningful contribution to the understanding of what’s working here, and what might be done to make it even better.

SNA study design

I propose to design and distribute a survey. Ideally, I would distribute the survey to all members of the SUS network, though there would be far too many respondents for that to be practical, so instead I will select representative organizations of each stakeholder (school, nonprofit, restaurant, etc.), and distribute the survey to each employee within that organization whose role is relevant to the SUS initiative (ex. School custodial staff would not be surveyed).

The network question:

“Please note how often you have discussed Shape Up Somerville with this person in the last three months”.

Underneath the header for each representative shareholder organization, I will list every relevant employee along with their position, and include the following options:

1. Never
2. Occasionally
3. Frequently

Note on a limitation of the network question: it does not distinguish between favorable and critical discussions about SUS. This distinction is important, and follow-up would be required to determine the type of discussion about SUS one has had (a clue about the likely nature of one’s conversations re: SUS could be found in a respondent’s answer to the attribute question about their satisfaction with SUS described below). However, I am making the assumption that the measure of one’s conversations about SUS is an indicator of one’s interest and stake in the outcome of SUS.

The survey would include a number of attribute-based questions that could be helpful in the analysis of the network data, including:

• Place of work and job position(s)
• Knowledge of SUS – how familiar are you with the SUS initiative (scale of 1 to 3)
• Do you read the publications distributed by the SUS Task Force?
• Please rate your satisfaction with SUS (scale of 1 to 3).
• Personal exercise routine/diet habits

Analyzing the network data

There are several questions I would like to explore through this network survey:

1. What does this network look like? How are people connected to one another?
I would like to explore what the pattern of communication ties reveals about the configuration of the SUS initiative on a whole-network level. Does it appear to have a more centralized, decentralized, or distributed form? I would expect there to be either components or communities (in the clustered Girvan Newman sense) formed around certain attributes, particularly place of work and job position. What would be very interesting is to see what connections might occur between communities.

I can also analyze the data to test certain assumptions about what distinguishes successful collective impact efforts. FSG is a nonprofit consulting firm that assists communities design collective impact initiatives, and they argue that successful collective impact efforts share some basic elements, including: regular communication within and amongst partners, particularly among players at the chief executive level; an organization serving as the “backbone” dedicated full-time to managing the partnership. These elements assume certain patterns of communication between partner organizations and people within those organizations, which I can investigate. For instance, is the SUS Task Force really serving as a backbone, or is the spoke-like organizational chart a myth?

2. How much collective interest is there about the initiative? Along with respondents’ answers to some of the attribute questions, looking at density of ties within, and perhaps across, different communities may provide an indirect indicator of how prevalent the issue is on the minds of all the partners.

3. Who are the go-to people in the network? In identifying node-central people (particularly high in-bound, out-bound centrality, and high Eigenvector), we’d be able to pinpoint influential members who are also likely among those with the greatest interest in SUS. It would be interesting to look at any commonalities among these people. Do they occupy positions of leadership within the SUS structure or their respective organizations? Are the people who occupy formal positions of leadership actually the thought leaders?

4. How might the SUS leadership facilitate connections within and between different communities? / How might information about SUS might be more efficiently disseminated across the network? Identify and strategize how to better incorporate the influential people within a component/community, as well as strategizing how to utilize the in-betweeners to connect different communities and spread information across the network.

SNA in a Governmental Organization

The Turkish Ministry of Economy consists of 8 independent units. The Directorate General of Exports is the one that I have been working for almost 6 years.

I prepared a survey (which is very similar to our Consulting Firm case) so as to initiate a SNA on the institutional structure of the Directorate General of Exports (the total sample size is 217).

In order to reach maximum participation level and avoid having possible statistical problems, I have tried to form a sample that averagely represents the structure of the unit. I have chosen 7 Departments (out of 11) under 3 different Deputy Director Generals (out of 4). We have 110 participants all told.

Survey Questions:

1) Gender
2) Date of birth
3) Place of birth
4) Education
5) Title
6) Department
7) Room number
8) Date of entry
9) Main Question (how often you have turned to this person for formal paperwork, information or advice on work-related topics in the past one year? Please enter an answer for everyone, and leave yourself blank)

Never - Seldom - Sometimes - Frequently

Names
.
.
.
.

In light of the possible findings of this SNA, we will not only able to increase efficiency in our institution but also find opportunity to apply it to all other governmental organizations.

"The Capitalist Network that Runs the World"

It's not exactly a "social" network, but theorists at the Swiss Federal Institute of Technology have mapped the connections between 43,000 transnational corporations to identify the small number of companies that have a disproportionate amount of influence over the global economy.

Their process and findings:

From Orbis 2007, a database listing 37 million companies and investors worldwide, they pulled out all 43,060 TNCs and the share ownerships linking them. Then they constructed a model of which companies controlled others through shareholding networks, coupled with each company's operating revenues, to map the structure of economic power.

The work, to be published in PloS One, revealed a core of 1318 companies with interlocking ownerships (see image). Each of the 1318 had ties to two or more other companies, and on average they were connected to 20. What's more, although they represented 20 per cent of global operating revenues, the 1318 appeared to collectively own through their shares the majority of the world's large blue chip and manufacturing firms - the "real" economy - representing a further 60 per cent of global revenues.

And the potential implications:

Crucially, by identifying the architecture of global economic power, the analysis could help make it more stable. By finding the vulnerable aspects of the system, economists can suggest measures to prevent future collapses spreading through the entire economy. Glattfelder says we may need global anti-trust rules, which now exist only at national level, to limit over-connection among TNCs. Bar-Yam says the analysis suggests one possible solution: firms should be taxed for excess interconnectivity to discourage this risk.

SNAs of terrorist networks

Pages 178 to 186 of this report (CLICK HERE) has an excellent paper on the use of social network analysis to identify nodes in a terrorist network, either to disrupt or prevent an attack from being carried out successfully.

Amongst other examples, the report cites the use of SNA -- or in this case, link analysis -- to track down Saddam Hussein and break up the "Virginia Jihad Network." SNA can also be applied going forward to track the development of the use of improvised explosive devices (IEDs) in Iraq.

Finally, the report places deserved importance on the use of post-hoc SNA, citing the famous Krebs analysis of the 9/11 hijackers and drawing on additional studies to suggest the 1995 sarin gas attack on the Tokyo subway by the terrorist group Aum Shinrikyo could have been prevented with the use of SNA.

Overall a fascinating report, and certainly worth a skim -- if you don't have the time or heart to read the entire 400-page paper, that is.

Using SNA to map critical skills, expertise, and communication in a corporate security department

The security directors of multinational companies must manage risks related to physical security, business continuity, intellectual property protection, local corruption and crisis response, among others. Given that many MNC's are now larger economic entities than some countries, managing these types of threats at the corporate level is becoming comparable to running the security of a small country.

 

How can SNA help a corporate security department whose success relies on cross-department and international collaboration?

Hypothesis: SNA will help the department (1) know what its resources are in the first place; (2) Facilitate collaboration by bringing together resources that should be connected sooner.

The ultimate goal is to make the company more prepared for and resilient against crisis situations.

Data required: 

The data-collection phase should begin with an assessment of the skills that are critical to the corporate security department achieving its goals. What kind of people should the department have inside it? Speak to frequently? Have access to?

I anticipate surveying along the following attributes and connection topics:

1.     Identifying and defining your resources:

a.     Identifying attributes:

                                               i.     Department

                                              ii.     Location

                                            iii.     Gender

                                            iv.     Title

                                              v.     Age

b.     Experience attributes:

                                               i.     Years with company

                                              ii.     Undergraduate degree

                                            iii.     Graduate degree

                                            iv.     Journalism

                                              v.     Law enforcement

                                            vi.     Military

                                           vii.     Government

                                         viii.     Law

                                            ix.     Academia

                                              x.     Corporate security

                                            xi.     Residence in countries where the business has operations

c.      Skill attributes:

                                               i.     Languages

                                              ii.     Certifications (CFE, CAMS, PPI, etc.)

                                            iii.     Investigations

                                            iv.     Security consulting

                                              v.     Political risk analysis

d.     Expertise attributes:

                                               i.     Industries (oil & gas, pharmaceuticals, mining, legal, etc.)

                                              ii.     Regions (Middle East, South Asia, etc.)

                                            iii.     Issues that matter to critical business functions or corporate strategy (business continuity, travel security, corruption, problem solving etc.)

2.     Describing the existing network:

a.     Patterns of communication:

                                               i.     Have never communicated

                                              ii.     Communicate with occasionally (once per week or less, group meetings)

                                            iii.     Communicate with frequently (more than once per week)

b.     Patterns of collaboration:

                                               i.     Have worked with on a client engagement in the past year

                                              ii.     Have not worked with on a client engagement in the past year

Important network measures:

• Network density: Cross-departmental collaboration on project work (density of networks on issues/regions/collaborations that make sense, such as Middle East political analysis, oil & gas security)
• Ego networks: which employees do the best job of connecting departments on projects.
• Organizational holes: lack of personnel with attributes necessary to satisfy specific corporate requirements; missing connections between employees who should be speaking more frequently (is there a security analyst with years of experience in China but now based in Wisconsin who has never spoken with the China country manager?)

What will this help the company do?

Ideally, the SNA will result in intelligence that allows the department to:

• Create more effective project teams.
• Identify people who aren’t talking who should be.
• Identify areas where the company is over or under-resourced on critical skills and personnel.
• Identify organizational vulnerabilities (bottlenecks; central people who if removed would have a significant negative effect on the company’s ability to get things done).

Mitt Romney vs. Rick Perry

Another interesting possible use for social network analysis, albeit outside the scope of international relations, is confirming whether Governor Rick Perry's charge (namely, that Romney's healthcare advisors advised Obama's health care "czar") is accurate. Using the White House visitors log, we could identify those who were in Romney's inner circle and then see whether they visited Nancy Deparle and other senior White House Officials. We could find out the frequency and whether Romney's advisors met with multiple Obama officials.

Who knows, a social network graph could be the basis for the next attack ad!

Incredible...now that we can identify the links between social networks, we can actually identify the link between our facebook friends and our brain functions!



http://mashable.com/2011/10/19/facebook-friend-count-brain/

Women on Corporate Boards

Several social network analysis studies have been done on the interlocking interlocking nature of corporate boards. Theyrule.net is a particularly interesting site, where visitors can search for individuals, companies, institutions, and instantly get a sense of the interlocked nature of the world’s most powerful corporate leaders. The Corporate Library, a corporate governance research firm in Portland, ME, now part of GMI (GovernanceMetrics International), has also done a lot of work to show the “interlocks” between boards of the world’s most powerful corporations.

This is the so-called “old boys” network. It would be interesting to get a better sense, through social network analysis, of some of the reasons why women are so excluded from this network. Even more interesting would be to better understand how the inclusion of more women in board memberships and executive positions could change the nature of a company’s practices. A 2007 report by Mckinsey & Co. found that companies with a higher proportion of females represented in top management also show better economic performance.

These findings have not gone unnoticed, and the push to incorporate more women on corporate boards has gained traction in the past couple of years. GMI recently launched a Diverse Director Datasource (3D) in order to provide companies with a database of qualified, diverse candidates for upcoming board appointments. A Boston-based non-profit, 2020 Women on Boards has formed around a campaign to increase the percentage of women on corporate boards to 20% by the year 2020. Many European countries have gone so far as to pass legislation requiring that a certain percentage of corporate board members be female by a certain year, 10, 15, or 20 years into the future.

It would be interesting, using social network analysis, to take a look at the interlocking nature of boards as it pertains to women. Are women by nature less “connected” in these circles than men? If so, is connectivity a main driver behind women’s lack of inclusion in corporate boards? What is the effect of the “connectivity” and “interlocks” of boards on economic performance?

Finally, it would be interesting to analyze the women who have made it onto the boards of major corporations and to see what is different about them. Are they more or less connected than their male counterparts? How are they connected to each other? There would be a lot to consider and analyze if given the chance. It would also be worthwhile to analyze GMI’s 3D database and compare the connectivity of the more diverse candidates who perhaps have yet to get a foot in the boardroom door with those already established in the board world.

- HP

Traditional research vs Social Media research

In a world that nothing remains intact over time, business and its related parties perform drastic changes in their operations in order to adjust to the constantly evolving status quo.

Based on that, would it be safe to assume that traditional research would be replaced by social media research? Social media are the future of market research since they allow the extraction of customer insights from information that individuals have voluntarily shared on the World Wide Web. Even though it sounds very convenient, accurate and cost effective, social media research cannot replace the necessity of traditional research - at least not at the moment.

Despite their efficiency, social media lack the capability to transmit in depth qualitative data that enhance the researcher's comprehensive understanding of customer needs. Apart from inability to generate accurate demographic data, social media do not offer information about non-internet users. In addition, the use of customer data available on the web are not necessarily suitable for research purposes since the individuals have not consented in such use of their personal information.

Will surveys ever be replaced? There is a lot to be debated, but it is definite that the combination of the two research approaches would provide invaluable information to businesses that will assist in the making of products - even more - tailored to the increasing needs of the modern consumer.  



 http://news.yahoo.com/comparing-social-media-traditional-research-maritz-research-randy-071227383.html

Who Has Had Tea with POTUS: A Network Analysis of the White House Visitors Log

Blog Post (note rough draft): One of the most fascinating applications for social network analysis is in the political realm. Network analysis is uniquely well suited to determine who has the ear of public policymakers and whether that access has a clear correlation with preferred outcomes for that well-connected individual(s). Earlier we discussed the “structural holes” organizational model that plagued President Carter’s ability to effectively lead. We propose to use social network analysis to gain insight into the Chief of Staff and key foreign policy professionals in President Barack Obama’s Administration. Since mapping over one million unique visitors to the White House would span a lifetime, we’ll isolate the visitors of a few key “nodes.” In fulfilling a campaign pledge, the Administration has took the unprecedented step of publicizing the White House Visitors log which documents the 1.7 million guests of the White House, Eisenhower Executive Office Building and the New Executive Office Building (among others). While there is a wealth of information, we will explore the obstacles that face a detailed analysis of the network as a whole.

Whitehouse.gov allows you to download the visitors log into an Excel Document that simplifies the digestion of the data. Before continuing, we will jump into terms of reference. First, a macro level social network analysis of the Visitors Log has already been conducted. However, it is by no means comprehensive and focuses solely upon the President’s “visitors.” Speaking of which, there is ample reason to suspect that this visitors record is not complete. For an instance, the “visitors” of the President of the United States (POTUS) are usually not one-on-one visitors. Typically these type of “visitors” have been invited to receptions, the Roosevelt Garden or other type of ceremonial events. Therefore, while certain individuals may have strong ties with “POTUS” based on the frequency of visits, they aren’t necessary intimately connected—there are definitely alternative explanations for their access. Likewise, while the natural inclination is to zero in on visitors to the “Oval Office” or “West Wing,” further analysis shows that those visitors were simply on West Wing tours led (in more cases than not) by junior level staffers.

For an instance, while I worked in the Executive Office of the President and visited “POTUS” four times, I only met him fleetingly on one of those occasions. Much to my disappointment our ties are not strong, but a network analysis based on frequency would overstate our relationship.

NAMELAST	NAMEFIRST	visitee_namelast	Description
Narendra	Blake	Cizner
NARENDRA	BLAKE	CUBA	BASKETBALL COURT
NARENDRA	BLAKE	ECKHOUSE
NARENDRA	BLAKE	MACK	BOWLING ALLEY
NARENDRA	BLAKE	MACK	BOWLING
NARENDRA	BLAKE	NORTHERN
NARENDRA	BLAKE	OFFICE	OPEN HOUSE TOUR/
NARENDRA	BLAKE	OFFICE	GROUP TOUR
NARENDRA	BLAKE	POTUS	STATE ARRIVAL - MEXICO**
NARENDRA	BLAKE	POTUS	ASIAN AMERICAN HERITAGE MONTH RECEPTION./
NARENDRA	BLAKE	POTUS	STATE ARRIVAL - MEXICO**
NARENDRA	BLAKE	POTUS	GREEK INDEPENDENCE DAY RECEPTION
NARENDRA	BLAKE	POTUS	ALL APPOINTEE EVENT
Narendra	Blake	POTUS	GERMANY STATE ARRIVAL
NARENDRA	BLAKE	WACKLER
NARENDRA	BLAKE	WINTERS

How to shape the survey: To get around the difficulty of ascertaining whether individuals are visiting the President in an official or social capacity, we propose to create an Ego Net of former Chief of Staff Rahm Emanuel’s visitors. As the President’s “gatekeeper,” the Visitors Log would provide unique insight into who had access to the position Presidential scholars have called “the 2^nd most powerful person in the world.” Our network may, or may not, help us to graphically show that the issue affiliations of Emanuel’s visitors match the President’s issue interests (eg healthcare reform). In the period of 2009-2010, Emanuel had over 500 recorded visitors to his West Wing Office. We first task would be to identify those who are unique visitors and then sort them by frequency of visit. For an instance, 1.) “two visits or less”; 2.) “between 3-5 visits”; 3.) “more than 5 visits.”

Then we would attempt to create a survey based on their unique attributes of the visitors. For an instance, we’d explore whether they are from the public or private sector. We also want to know if they are White House Staff or Government Agency staff. One illuminating possibility is whether Emanuel met frequently with a particular White House staffer. Obviously this could be indicative of that individual’s important role or it could suggest they were repeated victims of Emanuel’s well-documented wrath. Assuming most are public sector officials, we’d create an attribute value for hierarchy (note: position title for all visitors not readily known). We especially would like to identify what Cabinet level officials visited Emanuel the most often. Then, to the extent possible, we’d identify the issue specialty of the visitor. Much of this analysis is predicated on whether the surveyor can identify the affiliations of Emanuel’s visitors; fortunately, many are prominent figures. While this small number of attributes above just scratches the surface, this is what our sample survey would look like working backwards from the Visitors Log data:

How often have you visited Rahm Emanuel in the survey period?

1=1-2 times

2=Between 3-5 times

3=More than 5 times

Do you Represent the Public or Private Sector?

1=Public

2=Private

(If applicable) are you White House Staff or an employee of a Government agency?

1=White House Staff

2=Government Agency Employee

From the below options, what is the closest approximation your leadership level?

1=Junior Level, no management of paid employees

2=Mid Level, degree of management but accountable to one or more superior(s)

3=Senior Level or Principal, Cabinet or Subcabinet Level Official (President or head of company/NGO)

Please Identify the two issues you most identify with or are an expert on?

1=Environment/Energy Policy

2=National Security (ie military/diplomacy)

3=Healthcare Policy

4=Trade Policy

5=Electoral Politics

6=Financial Sector Policy

7=Country Specific Expertise

8=Other

As I’m apparently not a domestic politics student, we could internationalize the White House Visitors log by analyzing the visitors of John Brennan, the President’s Deputy National Security Advisor. Similar to the Chief of Staff, Brennan had 454 visitors in the span of 2009-2011. We could similarly create a survey to identify what individuals or nodes and interests are represented vis a vis Brennan’s West Wing visitors (many of whom are Deputy or Principal Level officials). Our most laborious task is identifying key attributes of Brennan’s visitors with nothing but their names. To illustrate this network, we could create a survey based on the existing data.

What Agency do you Represent?

1=OVP/National Security Staff

2=State

3=DOD Civilian

4=DOD Uniform

5=Energy

6=USAID

7=DHS

8=DNI

9=DOJ

10=Other

What is your level of leadership?

1=Junior Official, no supervisory functions

2=Mid Level Official, some supervisory functions

3=Subcabinet Level Official

4=Cabinet Level Official

Identify one or two greatest areas of expertise?

1=Nonproliferation

2=Nuclear (Civilian)

3=Nuclear, Biological, Chemical (Weapons)

4=Iraq

5=Afghanistan

6=Pakistan

7=China

8=Syria/Lebanon

9=Israel/Palestine

10=Border Security

11=Counterterrorism

12=Cyber Security

13=Disaster Recovery (Domestic or International)

14=Other

International Relations Curriculum is Not Chalk and Blackboard Anymore

Most of you have probably seen this, but just in case - Fletcher is featured in the CS Monitor in an article about incorporating social media in the classroom! The article weighs the merits of using wikipedia, blogs and youtube in academia.

SNA for Marketing

In this article the author reflects on the potential use of SNA analysis for marketing.

He draws parallel between the reasearch done for epidemics and how that can be used for discovering trends in marketing and averting PR crisis.

He also brings up the quality over quantity argument to imply that a smaller number of like-minded participants in a network can be better information disseminators than a large number of diverse audience. This idea sits well with the notion of using targeted marketing campaigns for relevant communities that can then help the message go viral.

The blog in general offers some interesting ideas for those interested in using Social <insert appropriate word here> for marketing.

Shashank

Gen. McChrystal: It Takes a Network

This article, written by Gen. Stanley McChrystal, follows his realization and understanding of the terrorists in Afghanistan as a network, what type of network, and also how his team needed to work as a network. He even draws a picture! The main thought driver behind this was that the US Army needed to be a network in order to defeat a network.

Inside Darpa’s Secret Afghan Spy Machine

An interesting (though somewhat long ) story about how DARPA, with the help of a few M.I.T. grad students, is transforming the way intelligence and networks are visualized in Afghanistan.

Using Social Network Analysis to Drive Sales

From a brief article in the Harvard Business Review last month:

Salespeople have always prized the individual at the center of a web of prospects, because once that person becomes a customer, peer influence may cause the rest to follow.

The trick, of course, is finding that well-connected prospect. Traditionally the search has required all the salesperson’s black arts: sniffing, calling, schmoozing, and guessing. But network theory is beginning to turn it into a science. 

It's an interesting discussion because it touches on combining data from different kinds of sources (insurance claims, drug sales, locations of doctors, who talks to whom) to help pharmaceutical companies identify their most influential clients and concentrate their sales efforts accordingly.

Dubious Value of Social Networks

Published in the New Yorker, this article casts doubt on the indispensable value of social networks in affecting change by using recent (attempted) political revolutions as cast studies. Malcom Gladwell interestingly begins by explaining that social movements such as the sit ins of the 1960's occurred with great success before the advent of Twitter, Facebook, and other internet mediums.

However, Gladwell concedes that use of Facebook does offer the United States a unique advantage over terrorists; supporting this idea, he quotes a former State Department advisor in saying, "the Internet is now about interactivity and connectivity" (and terrorists networks are behind the curve in this respect).

Gladwell goes on to cite the Iranian protests in 2009 as evidence of Twitter's overstated impact. The vast majority of tweets were in English not Farsi, implying that those tweeters were Western sympathizers not Iranians.

Blake Narendra

Occupy Wall Street: Social Media's Role In Social Change

I heard a piece on the radio about how well-organized the Occupy Wall Street protesters are - neatly lined-up shoes, shower rotations, medical and security teams, free food distribution. That and the facts that there apparently aren't any leaders, and that the folks at AdBusters - who put out the original call for this protest- said they were inspired by the Arab Spring, made me wonder about the role that social media is playing in this (maybe) burgeoning movement.

This piece doesn't get to the bottom of it, though it does acknowledge that social media is very much present and has certainly played a role in recruitment: "The best way to get people away from their computer is through the computer".

Social media has also be instrumental in catching the mainstream media's attention, which in turn has helped feed the protestor's ranks.

Still, I'm curious about if and how social media has played a role in the organizing logistics.

How many members does it take to make a functioning network?

The New York Times reviewed a smartphone app called Sportaneous this weekend. Sportaneous is designed to connect people who want to play pick-up sports games in New York City, but it ran into some challenges. The reviewer wrote,

"The only problem is that the app is not always useful. I tried it repeatedly this summer in hopes of finding a basketball game in Greenwich Village or Astoria, Queens, but never set foot on a court. Internet services like this rely on something called the network effect, meaning that they work well only after they have become popular. With only about 100 people in its system, Sportaneous was stuck at zero."

 In order to deal with the "network effect," the app's creators started promoting the app in specific neighborhoods, hoping to get enough members to support pick-up games. Still, the reviewer said, "People are much more likely to use Sportaneous as a way to supplement and expand the social circles they already belong to than they are to compete against people they have never met."

I thought this was an interesting example of location-based social networking and the challenges involved in trying to get people to move beyond their own strong-tie networks.

SNA on Savings Groups in Sub-Saharan Africa

The following link describes dependence in various African countries on savings groups as a means to finance important needs. Some microfinance institutions are also leveraging this concept by making loans to groups of people rather than to individuals. These small microloan and savings groups are essentially networks, usually of women, in developing countries. It occurred to me that a social network analysis of these kinds of groups might be interesting to see if there were any structural commonalities between those groups that had better repayment rates or used their funds to generate higher returns. I'd hypothesize that groups with stronger ties may be more successful at working together to use funds more efficiently. It would be interesting to ask the women in the group to rank how well they know the others in the group or to ask them to assign a confidential ranking based on how close their friendship is with other women to get a sense of how strong the bonds are between members of the group.

Furthermore, if we were able to do a social network of several small villages where, for example, local women had a savings group, we could also see if the returns on their use of funds correlates with the average Eigenvector or closeness scores for the group. This would be logical because the better connected women are to others in their community, perhaps the better able they are to sell goods or provide services to their community with the funds that their group gives them. I'd imagine that this could also be used to analyze individual sales people in a village or in a souq to see if having a robust social network in the community translates to financial gains for the business. It would then be interesting to weigh the strength of the social network versus other buyer care-abouts like price, quality, etc. in determining how strongly consumer behavior is influenced by social networks. While I initially thought of this idea based on an interest in developing markets, I'd assume that the takeaways could well be applicable and relevant to businesses in the developed world also, and that could be a really meaningful insight for marketing professionals everywhere.

Ravi