An Eight-Step Action Research Network Analysis Approach to Complex Global Systems:

The Case of Global Finance

By:

Steve Waddell

and

Verna Allee

July 28, 2009

Presented to:

Applications of Social Network Analysis Annual Meeting

August 27-28, 2009

University of Zurich

This paper draws from two projects with the authors and following researchers:

Buck Child,

Randy Kemp,

Sanjeev Khagram,

Catrina Lucero,

Judith McCrory,

Subarna Mathes,

Oliver Schwabe

The authors wish to thank iScale/GAN-Net ( for their generous sponsorship of this project.

For more information contact:

Steve Waddell – PhD, MBA

Boston, MA, USA

+1 (617) 388-7658

Abstract

Introduction

The Three Network Analysis Methodologies

Web Crawl Mapping

Social/Organizational/Inter-Organizational Network Analyses (SNA/ONA/IONA)

Value Network Analysis (VNA)

Table 1: Comparison of Methodologies

A Comparison of the Mapping Approaches

Two Action Research Cases

The Overall Research Process

The Steps

Step 1: Initial Analysis

Step 2: Web Crawls

Step 3: Data Review

Step 4: Website Mining

Step 5: Interviews

Step 6: Participant Validation & Stewardship Team Formation

Step 7: Surveys, Interviews, and Maps

Step 8: Participant Validation and Community Initiation

Conclusion

Endnotes

Eight Step SNA Methodology

An Eight-Step Action Research Network Analysis Approach to Complex Global Systems:

The Case of Global Finance

July 28, 2009

Abstract

SNA methodologies continue to evolve both in number and sophistication. But how might they be used together to deepen understanding of a field? This paper discusses the application of three network analysis methods and proposes an integrated model for their use in eight sequential steps. The methodologies applied are (1) web crawls analyzing linkages between web sites, (2) organizational network analysis looking at the relationships between legally independent entities, and (3) value network analysis that describes relationships in terms of what is being exchanged between nodes. The final model integrates these SNA methodologies with interviews, surveys, and focus groups.

This model was developed in an investigation of a global issue system: the global public policy finance system. The goal of the investigation was to develop a description of the system that could be used by people who are aiming to reform it, but who have very modest understanding of global finance. The project was designed as a community-organizing intervention that would help people identify key strategic points of focus.

The model described here is not what was done in the project – although the project made use of all the methodologies mentioned. Rather, the model proposed arises from lessons that emerged from the project. The paper, therefore, explains the model’s rationale in reference to both the process and product of the global public policy finance system investigation.

Introduction

Global networks of business, government and civil society organizations are the organizing frontier of the 21st century.[i] In complexity and potential they far surpass the 20th century organizational highlights: welfare state governments, multi-national corporations and local-global NGOs. The 21st century needs networks in order to address complex questions such as climate change and poverty, because such challenges are beyond the resources, skills and capacity of any one organization or type of organization.

Describing such networks is a huge challenge, but critically important if we are to realize their power. Network analysis methodologies offer promise, but how to use them effectively in issue arenas of such scale and complexity requires further development. This paper draws from experimental application of three network analysis methodologies to describe the global financial system, to propose an eight-step methodology for issue domains of such complexity and scale. The eightsteps and their sequence do not describe the way the investigation of the global finance system proceeded; rather, they arise from reflection upon the experience of the investigation.

The Three Network Analysis Methodologies

There are several types of network analysis. This global finance investigation required ones that could help gather large amounts of information and identify actors in the issue arena, and that could handle large amounts of information to describe their relationships from various perspectives. These requirements suggest that relatively sophisticated software would be needed. Three mapping methodologies came to the fore and are first discussed generically.

Web Crawl Mapping

The internet has an increasing role in communications and daily life. People have personal as well as organizational web-sites. Sites bringing people together around shared interests and concerns are numerous. The importance of the internet in political and other campaigns is unquestioned. Although web presence is not uniform around the world, certainly for global issues and increasingly for local ones the internet presents an incredible information resource. Web crawls produce maps using this information.

The internet is structured around sites that have unique URL addresses. Most sites also have (hyper)links that you click on to take you to other sites or pages. These are inserted because they have more detailed information with regards to a topic (including, of course, ads), because the host wants to connect people to allies or colleagues, or because they may be foes on an issue.

These connections between unique URLs provide the basis for mapping relationships by doing a web crawl. A software program can draw the relationships between organizations’ web links, to give a description of the virtual network of the organization. The global finance project generated web crawler maps using the Issue Crawler (issuecrawler.net) from the Govcom.org Foundation, Amsterdam, directed by Prof. Richard Rogers, Chair in New Media & Digital Culture, University of Amsterdam.

Of course the weakness of this mapping approach is that the sophistication of organization’s web presence varies enormously.

Social/Organizational/Inter-Organizational Network Analyses (SNA/ONA/IONA)

These are essentially the same type of classic social network analysis, but applied at different levels: individuals, parts of organizations, and an issue system or subset of it. All produce maps that describe the relationships with a single lines or arrow between nodes, where nodes are individuals, parts of organizations, organizations, or groups of organizations. The nature of all links is the same, showing that people are linked through the same kind of activity or affiliation.

Given the scale ofissue systems like global finance, the need is for analysis at the inter-organizational level, and for groups of organizations. Traditionally, data is gathered by interviewing individuals about their organization’s relationships with other organizations. The critical weakness of ONA is that it gathers data from individuals, and extrapolates as though the individuals represent the organization. Naturally, individuals’ knowledge of their organization’s relationships with other organizations is limited and draws largely from their own work experience and personal interactions. Therefore, of particular importance is that interviews are with senior individuals with responsibility as directly related to the issue domain being investigated.

Value Network Analysis (VNA)

VNA is related to the idea of “value chain”, but the latter is a linear whereas VNA is a system-level analysis. Although VNA is usually applied to organizations, it can also be applied to open systems and networks. The approach is associated with visual representations of systems in terms of roles and exchanges between roles (“deliverables”). This is very useful when you want to understand what roles and exchanges are necessary for a healthy issue system, which roles need more attention and which might be so well resourced that it is creating problems such as needless redundancy or bottlenecks.

Organizations inter-act with other organizations through certain exchanges or deliverables. These include formal, explicit and contracted tangibles such as reports, workshops, and money. But the exchanges are often intangible, such as influence, strategic information, planning knowledge, and referrals to other people.

This data results in maps about how an organization creates value in a system. A “participant-based” VNA map portrays organizations and their exchanges; in effect, these are much like SNA/ONA/IONA, but with more information. It allows for multiple links or ties between organizations or individuals and each tie is unique in terms of directionality and the specific nature of deliverable. The exchanges and the activities suggest roles or functions that an organization plays within a system. For example, an organization may act as a researcher, policy developer, or investor. Most organizations play more than one role.

But perhaps the greatest value is with “role-based” VNA maps. In these, the nodes are roles rather than organizations that might play those roles. For example, there may be a role of “policy advocate” or “evaluator”; although the role may be played by more than one entity, it will only appear once, with linkages to it raising important insights about what is needed for that role to be played well. This moves fully into the 21st century network-centric rather than organization-based world. It also allows moving beyond the confusion and emotional loyalties that organization-based analysis in a network can produce.

ValueNet Works™ is a mapping methodology developed by Verna Allee, and the global finance project used it both for the VNA and ONA. It was supported by the ValueNetworks.com software application (

Comparison of Methodologies

Table 1 compares these methodologies, in order to help explain why using all three is much more powerful that using just one. Web crawls are the most easily produced maps, since they do not require any one-on-one data gathering. But they are also the most crude and are not appropriate for a more surgical analysis. Nevertheless, they provide an excellent way to identify organizations that are active in a particular issue domain and a rough description of their relationships. The fact that the crawls only describe web-site relationshipswas not a particular concern for the investigation into global finance, since any organization that is a player of import could be assumed to have a relatively well-developed web-site.

The IONA methodology is good for describing links between organizations from the viewpoint of particular people within them. It illustrates explicit and implicit alliances and “neighborhoods” of commonality, where there may be over-crowding of attention or gaps where relationships should be built. It is good for helping to identify key players who can be brought together to represent a “system”.

Table 1:

A Comparison of the Mapping Approaches

Output / Goal / Use when you want to: / Limitations
Web Crawls /
  • Map of URL connections
/
  • Identify orgs. in an issue arena
  • Identify sub-groups, central orgs, bridging orgs in virtual space
/
  • Identify key stakeholder and influencers in an issue area.
  • Get a general sense of relationships within the system
/
  • Only good if issue arena orgs have good web-sites
  • Usually good for global arenas
  • Remember: virtual reality does not equal reality

SNA/
ONA /IONA /
  • Map of links between individuals, parts of an org or orgs
/
  • Identify work flows, alliances, who is central
  • Understand the network structure of the system
/
  • Initiate a network
  • Strengthen a network by better understanding its social “connective tissue”
  • Strengthen specific orgs and their relationships
  • Reduce duplication/redundancy
/
  • Must have good starting list of people/orgs
  • In ONA/IONA data is gathered from individuals, but applied to orgs

VNA /
  • Maps of value creation process
  • List of roles in the network
  • Definition of value outputs
/
  • Understand how issues can be addressed in terms of:
  • Formal and informal exchanges between orgs
  • Roles and interactions between them
/
  • Initiate a network
  • Strengthen a network by supporting specific roles, relationships, or interactions
  • Reduce duplication/redundancy
  • Shift from org-based to role-based network development
/
  • Challenging to identify “boundaries” and “roles” in a network for manageable analysis
  • Institutional models and mindsets get in the way.
  • Maps can initially appear complicated so output needs to be managed carefully for communication.

VNA is useful in understanding how a system works in terms of exchanges and roles. Perhaps most importantly in terms of network development, it helps shift people from thinking of networks in terms of organizations (the 20th century construct) to thinking of them in terms of roles.

Two Action Research Cases

This paper draws from the experiences of two action research projects undertaken by iScale and Global Action Networks Net (GAN-Net). These two organizations focus upon large system change challenges, and the projects were undertaken specifically to advance the development of the systems. One goal was to develop a way to analyze an issue domain or system, with data gathered at the organizational level.

Analyzing a system is much more difficult than analyzing an organization or even a network. For iScale projects, “system” means all the stakeholders who have an interest in, or are affected by, a particular issue; it is almost always much bigger than any particular network. In contrast to “organizations”, systems are more open, there are not necessarily exchanges between the various parts (stakeholders, roles) and defining who is in and who is out is often more difficult.

One project focused upon the global financial policy domain (the Global Finance Initiative or GFI). Given almost everyone in the modern world can be considered part of the finance system, the problems of boundary-setting are particularly obvious.

The second project focused upon the mass atrocities (MA) domain (think Darfur, the Balkans with systemic rape, civilian murder, human mutilation, and recruitment of child soldiers). In this case, simply labeling the system demanded significant attention – defining it as “peace”, “mass conflict”, “genocide” and “human security” would produce different arrays of stakeholders.

In both projects the goals were to explore the value of, and the way to create, a multi-stakeholder global network to address the issue collectively. This goal was driven by a theory of change about the need to build “connective tissue” between stakeholders, networks and organizations to address these types of challenges globally. The two systems are seen as relatively immature, as defined by the awareness of the stakeholders of their participation in the system and their interdependence, and of the actual value of creating more robust interactions.

For brevity and clarity, the rest of this paper will reference the GFI only. The global finance system is characterized by two distinct groups of stakeholders. There are traditional actors – governments and commercial financial institutions – with quite a strong sense of being “the system” and who are quite well connected. A second group of stakeholders are pushing for innovation, but are poorly connected between themselves and with the traditional actors. Moreover, the two groups have competing perspectives of the system goals. The traditionalists’ goals concern economic growth and stability narrowly defined in economic/financial terms; the second group is focusing upon issues of development and environmental-social impact.

The projects are framed as “action research” for three reasons. One was that the goal of the project was not the data and its quality in a scientific sense, but rather development of powerful perspectives to help participants in the issue arenas have new types of conversations and define a set of actions to address the challenges of the field. Data was simply the means to the end, and its social construction was well-recognized.

The second reason for the action research frame is that the projects aimed to engage people in the field as co-participants, rather than subjects. The data is socially constructed, rather than driven by some external “objective” tests. Eight months along in GFI (August 2008), a stewardship group was formed in order to formalize input, share control, build legitimacy and enhance connections to others in the issue arenas. A second stewardship meeting was held in January 2009 at HarvardBusinessSchool.

The third and most important reason that the project was termed “action research” is that a goal was to investigate how the methodologies could be applied to large systems change challenges and help advance addressing the issue. These projects were undertaken with great enthusiasm for mapping as a methodology that could help the stakeholders “see” their system and identify how to further develop its efficiency and effectiveness. The project team included Verna Allee and Oliver Schwabe, who were also interested in further developing the capacity of VNA to support the project goals. However, the scope and complexity of the issue domain was beyond anything they had done before.