I've been in Washington, DC for the Network Effectiveness and Social Media Strategy Map working session for Packard Foundation Grantees convened by Monitor Institute. This is a post to help me identify what I don't know about social network analysis and mapping tools with the hope that you'll fill in the gaps in the comments.
I flew back from DC last night on Virgin America. Unfortunately, the weather in DC kept us on the ground and in the airplane on the runway for a couple of hours. They permitted the use of cell phones, so I tweeted that even if you are delayed, it's a pleasant experience on Virgin America. A minute later I noticed Pistachio's tweet about flying Virgin America from Boston and Glenn Stratchan direct messaged me asking what seat I was in because he thought we might be on the same Virgin America flight.
Since I couldn't use my computer to get on the wifi, I took out my session workbook to review my notes on the module about "Understanding Your Network" which included the basics of social network analysis and mapping.
Source: Monitor Institute
The visual above is a social network. Each dot represents a person or in network jargon, a node. A connection between people is called link. The definition of links or how you're connected is defined however you want. Some examples:
- A follower on Twitter or someone you're following
- Someone on your mailing list or rolodex
- Someone you know well enough to call
- Family members
- Organizational reporting
- Communications flows
- Information flows
Basically, connections can describe anything you want to gather data about to make decisions to improve your network or reach your goals. I've summarized some suggested processes below which simply good research/decision-making frameworks.
Connections can have directions flows, one-way or two-way. For example, a one-way connection on Twitter is someone who follows you, but you don't follow back. A two-way connection is someone you follow on Twitter and follows you back. (BTW, there's an interesting discussion about this on David Armano's blog). Connections also be strong or loose. For example, on Twitter, a strong tie could be two nodes or people that engage consistently in two-way conversation. (See this map from Malina). David Armano likes to talk about how Twitter can strengthen loose ties (two people who are connected on social networks, but don't know each other in real life or very well)
A Hub is a person who has a lot of connections to other people who may not be connected or know one another. Hubs look like starbursts on the visualization. A cluster is a bounded group of people who are connected, but have few connections to other nodes. (Think of a Tower-like organization where staff aren't encouraged to network or within siloed departments.) The Core are people who do most of the work (think wikipedia editors.) The periphery is edge, people who may more connected in other networks and are lurking in your network or dip in. The nodes on the edge can infuse the network with new ideas or energy if they brought into the fold. (Read Power to the Edges)
How would you describe Virgin America in social networking analysis and mapping terminology?
That's not a real question, just the idle thoughts of a slightly exhausted airline passenger stuck on a runaway on a Virgin America flight who can't use the wifi to get back to the clouds to summarize the rest of the of learnings/advice about social network analysis that came from Twitter while I was live tweeting.
(Interesting, one of the back channel comments was the definition of an expert. My definition is that your knowledge lives in the clouds. Dave Witzel tweeted it. At that moment, I realized the limitation of knowledge in the clouds and the definition.)
Draw your network
After the presentation on Network Basics (materials here), there were two exercises. The first was to look at the different social network analysis maps posted on the walls and figure out what was going on. The next exercise participants worked in small groups using a low-tech method to "doodle out" their networks (magic markers and flip chart paper.) This exercise reminded me of an exercise I used to do in workshops 12 years ago with nonprofits to introduce the Internet, "Draw A Picture of the Internet"
Some of the maps in the workshop were created with social analysis mapping software. This is an area of tools and techniques that I have wanted to explore in further depth, but haven't.
One of the questions that also came up in my mind was: What's the difference between a community of practice and a network? I tweeted this and Valdis Krebs had a great answer:
The network is the "structure" upon which the community dances and self-organizes... builds network as it needs it.
I used the tag #packnet and asked for the "Poor Man Dummies Guide to Social Network Analysis and Mapping Tools." Va;dis Krebs and other Twitter users pointed to some great resources which are listed below along with others that I discovered.
One question:
The best resource I discovered that answers this questions was "Building Smart Communities Through Network Weaving" by Valdis Krebs and June Holley. The maps are used to improve the results of your network. The network map is a snapshot of where your network is "as is." Network maps are "talking documents" or prompts for reflection and strategy brainstorms. Network maps support "what if conversations."
The paper goes on to describe five general patterns:
- Birds of a feather flock together: nodes link together because of common attributes, goals or governance
- Diversity is important. Though clusters form around common attributes and goals, vibrant networks maintain connections to diverse nodes and clusters.
- Robust networks have several paths between any two nodes.
- Some nodes are more prominent than others. These are critical to network health
- Most nodes are connected by an indirect link in the network.
The paper also describes how networks evolve over time - and you can observe this visual patterns. The paper discusses network weaving techniques to help evolve a more effective network - one that is more woven.
Some questions related to the network analysis tools that I don't have the answers to, but have started scanning and asking for answers.
- What is a good framework for designing a social network analysis mapping project so you get the most out of it?
- What is a good step-by-step approach for analyzing your map and making decisions?
- Why map your network with a marker and paper? Why map your network with a software tool?
- What information can a social network analysis software tool provide that simple low tech method cannot?
- What are the different tools available? What are the features? How technical? How expensive? Are there low cost techology tools for less technical people that provide a resulting map that is useful?
Process Steps
One important lesson is that you should really spend the time framing the problem or your goal. Also, social network analysis mapping may not be your only source of research - so if you are looking at a complex network and multiple research sources think about staging and phasing.
Angus Parker over at WiserEarth suggests the following process for using simple network mapping tools - these steps could be used with low tech, low cost or more complex technology tools. This is adapted from a framework developed by Roberto Cremonini from the Barr Foundation illustrated in the diagram below:
I love these social network maps. My concern is that they don't necessarily help you highlight who's missing -- who could/should be a part of your network (because their interests are at stake, because they're buy-in is required for your effectiveness, etc.) but isn't. Do you know anyone who's working specifically to draw attention to the "holes" in our networks? June Holley comes to mind -- need to explore her work more closely -- but any other resources would be appreciated, too.
When selling to executives, many times I start by showing a picture of a produce section of a grocery store and ask: What is missing here? (and then move to the analogous question that is perhaps close to their heart) What are the gaps that if found could signal opportunities? How to you reallocate resources to the most important gaps? This set of questions is very difficult to answer, of course -- ultimately, I am asking them "what is that you do not know that you do not know?" (and this is not meant as a play with words).
Within a bounded problem space we have figured out a way to consistently answer that question -- and have been using the result for acquisition integration, sales, knowledge transfer, and much more.
Some Off the Wall Ideas To Synthesize into the Process
I've been intrigued by social data exploration and wonder what offline processes might be adapted to working this with a software tool?
Also, I wonder how you can integrate some of the thinking for external communication that uses social networks/social media?
The tools
I was not able to find a taxonomy for social network analysis tools. So, I'm roughly dividing these into two categories: Low-Tech and High-Tech
Low-Tech Approaches
These use tools like crayons and paper and process exercises.
- Marty Kearns has a diagnostic tool over in his Advocacy 2.0 wiki that gives you a good set of questions to ask after you've created a descriptive drawing.
- Net-Map Tool Kit is an interview-based low-tech mapping tool that helps people understand, visualize, discuss, and improve situations in which many different actors influence outcomes in a community or network. It includes a step-by-step guided approach.
High Tech Approaches
-
KeyHub: Angus Parker from Wiser Earth has written a review for KeyHubs. It looks like a social network mapping software that can analyze informal networks within an organization. There is a tour on the site, free trial, and some case studies.
-
TouchGraph: Marty Kearns recommended TouchGraph because it is an easy tool if you have a relational dataset.
-
Issue Crawler: Rob Stuart recommended Issue Crawler. You need a good set of URLs to feed it. More information can be found here.
- Valdis Krebs - Orgnet Software
- June Holley's Smart Network Analyzer: Sarah Carr from EBM a demo from a webinar (here is the demo). Software designed by Valdis Krebs and colleagues.
Additional Resources
Visual Mapping and Diagnostics for Scaling Change: From Steve Waddell, Iscale - this paper has done the impossible! It has explained social network analysis and techniques in a way that is easy to understand for those who are not experts. With the onslaught of information and greater complexity in our work, visual diagnostics are more and more important. There's also a paper that Steve co-authored about a methodology for social network analysis.
Stephen Baker, BusinessWeek - How Much Is A Friend Worth?
Gaurav Mishra, Using Network and Influence Analysis to Map Social Media Consumer Behavior
Valdis Krebs, Managing Core Competencies of the Corporation: Organizational Network Mapping
Valdis Krebs, Social Network Analysis: A Brief Introduction
Peter Plastrik and Madeleine Taylor, Network Power for Nonprofits and Philanthropy
Bruce Hoppe and Claire Reinelt, Social Network Analysis and Leadership Development
Bruce Hoppe Blog - Visualization Category
Vladis Krebs, Building Sustainable Communities Through Network Weaving
Mark Surman, Hybrid Organizations
Peter Morville's Social Network Analysis Synthesis
John Kleinberg's Course on Networks
How Kevin Bacon Cured Cancer
CD3: Community Impact
Visual Complexity
Great write-up, Beth. I've used some of the mapping and analysis tools on social networks. I tend to like the software options because it makes it easy to visually show patterns by tweaking colors, shapes, line weights, etc.
Let me know (IM?) if you want to talk about it further. I love the way you ask questions and make me think of things in new ways.
Posted by: Ruby Sinreich | May 29, 2009 at 06:52 PM
Sometimes I think that people promoting and considering using SNA research methods for organizational development get overly fascinated with or hung up on the "mapping" and visualization components of SNA. One can only learn so much looking at a visualization, but it seems their aesthetic appeal is the major draw to the practice.
Actual social network analysts derive far more value from analyzing network metrics, such as centrality, cohesion, and degree. Some of these metrics provide information about the entire system being studied, some of them can show how the network attributes of individuals are distributed through the system. Many of these metrics represent different aspects of influence and power. One gets a sense of the structure from the visualization, but comes close to actually learning about that structure from the numbers, the metrics.
As for applications in management or organizational development, these metrics can provide important information about problems such as individuals who have unintentionally become bottlenecks or people whose departure would have serious consequences for the structure of flows in the network.
This is to answer one of your questions re: low-tech, paper SNA vs data-driven tools. Paper is great if you want to lead a workshop to "get people thinking" about networks. If you want to really get close to knowing something about any particular network, data and computation are essential.
Some resources on network metrics:
http://en.wikipedia.org/wiki/Social_networks#Metrics_.28Measures.29_in_social_network_analysis
http://www.orgnet.com/sna.html
By the way, your post is a really great article that I am going to file away for purposes of introducing SNA.
If you're interested, I have been stashing SNA resources on delicious for about 3 years now: http://delicious.com/james.nathaniel/sna
Posted by: Nathaniel James | May 29, 2009 at 08:54 PM
Interesting! Tweets of where we are geographically, work in a similar way to network maps that show where we are relationally. Both provide opportunities for meeting/connecting that were not obvious before.
- "Oh, I didn't know you were on this flight, let's meet/sit together"
- "Oh, I didn't know you were connected to X, can you introduce me?"
Posted by: Valdis Krebs | May 30, 2009 at 11:22 AM
Valdis, thanks for your comment and pointing me to some excellent resources.
On Twitter, a colleague Marty Kearns said:
http://twitter.com/MartyKearns/status/1974469300
"I am not sure if you answered all those open questions on network analysis if it would get you anything in an advocacy context."
Makes me wonder about the best application and context for a SNA.
What are your thoughts?
Posted by: Beth Kanter | May 30, 2009 at 02:55 PM
Hi Beth, I have the same thing as you, have been relucted to dive into SNA.. I will follow you as you seem to speed up :)
Did you know this page on the km4dev wiki about SNA?
http://www.km4dev.org/wiki/index.php/Social_Network_Analysis
Posted by: Joitske Hulsebosch | May 31, 2009 at 03:52 AM
The network is the "structure" upon which the community dances and self-organizes... builds network as it needs it.
Our agreements are our structures ?
Posted by: Jon Husband | May 31, 2009 at 06:27 PM
Hey, Beth ... per your email request.
I've written about the theme "our agreements are our structures" here and there for about the last 5 years. Remember TA .. (generality alert) well, old-school hierarchical org'ns are pretty much parent-child structures and dynamics, no ? Dave Weinberger once said "if conversations are between nominal equals, well, you never really have a conversation with your boss" (yes, well he IS a philosopher, after all ;-)
Moving on to networks, esp. networks of purpose .. seems reasonably clear to me that much of the P2p-ish dynamics we are experiencing are pretty much adult-adult, a la TA.
Extrapolating from that, my friend Chris Corrigan caught a piece I wrote last August titled "In Networks, Our Agreements With Each Other Will Be Our Structures", and added some thoughts, here:
http://chriscorrigan.com/parkinglot/?p=1531
Posted by: Jon Husband | May 31, 2009 at 08:24 PM
Here's the original blog post:
In Networks, Our Agreements With Each Other Will Be Our Structures
Posted by: Jon Husband | May 31, 2009 at 08:26 PM
Beth, I like the way that you explore all different avenues of SNA, high and low tech etc. With Net-Map I am on the low tech, pen-and-paper side of things and one thing that I like about drawing networks of actors with them, is that you can immediately ask the "why" and "how" questions as you see something interesting emerging on the paper. For example, we drew a network map about reporting and interventions concerning avian influenza in an African country and asked right there: So where are the corruption hot-spots and why?
Also, in our governance work, we often deal with networks where some of the actors are individuals, some organisations or parts of organisations. Strictly speaking it doesn't make sense to calculate quantitative network indicators for these situations, but still, mapping them is a great tool for getting a better understanding of how things work and where the problems are.
In our experience it is often very instructive to map both formal and informal links and to assess the influence (as perceived by interviewees) and goals of actors. This gives the network maps further depth and helps everyone to think strategically about their networks.
Posted by: Eva Schiffer | June 01, 2009 at 06:14 AM
I'd love to have the capability to colour-code the connections I have with people in my networks with respect to (say) four levels of trust (eg. implicitly, pretty much, beginning to, and not sure / new) within some element / component of the technical protocols that carry the information (you know, the FriendConnect kind of technical stuff whereby apps and services "know" your address book or connections).
That way, over time maps and visual representations of networks might well be more useful in terms of being able to identify trusted territory, trusted space, lands wherein one might enter and not need to be quite so wary, or in the flip sense, know that they will be held to account for what they say and do.
I know the carefulness about trusted connections and space already exists through peoples' care with introductions and via the various ways of checking other people out, but in terms of the technical / explicitness aspect of SNA, my guess is that some kind of capability like that would be quite useful ?
Posted by: Jon Husband | June 01, 2009 at 07:52 AM
Interesting discussion of social networks! I would also like to point you to an excel template that allows you to do sna in excel. I teach social networking and I find that it is much easier for people learning about the method.
http://www.codeplex.com/NodeXL - its free and can import your twitter network through excel ribbon.
Posted by: Brandy Aven | June 03, 2009 at 09:47 AM
@ BRANDY AVEN - Excellent template.Thanks for posting such a useful information to us…
Posted by: SEO | June 17, 2009 at 11:50 PM