Wednesday, November 30, 2005

Organizational network analysis utility

[Ed note: Originally written in 2005, this post was updated June 2009 so that the featured spreadsheet utility is freely downloadable here. Previously the spreadsheet was distributed via email.]

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.

Although this kind of intuitive analytical power has very wide appeal, its usefulness is limited right now by the unwieldy software tools currently available.

Deep down, making good simple network pictures is inherently complicated, but using network visualization software doesn't have to be. Progress is being made every day. See the newly updated list of SNA software in the right sidebar for some great examples. (And please let me know if I'm missing something.)

Even with the simplest of these tools, my non-technical clients often get hung up right away with the basic task of getting the data in. We power-users can easily forget how hard it was to build our first network, until we see someone else learning for the first time.

Here's an Excel spreadsheet utility my clients and I find helpful. I now make it freely available, in the hopes that more people will enjoy the benefits of seeing the big picture of the network perspective.

The spreadsheet includes three worksheets. One worksheet is the actual survey, which can be modified to suit the specific project. It automatically incorporates the names of the survey population into a drop-down list.
After distributing the survey via email, collected responses can be pasted in any order into a "compiled survey" worksheet:

Then an "automatrix" worksheet converts the compiled results into square matrices that can easily be pasted into available network analysis tools. The matrix calculator makes it easy to manage who opts in or out of the survey, and it provides access to multiple relationships.

If you'd like a copy of the spreadsheet, which includes a copy of a great California Computer case study (permission granted by David Krackhardt), you can download it here.

Copyright (c) 2005-2009 Connective Associates, except where otherwise noted.

Monday, November 28, 2005

The "Steinway" of networking strategies

I consolidated my rolodex last weekend in anticipation of sending holiday cards. As I was double-checking contact information, I reflected on how I originally met each person in my life. Did I meet them randomly, as part of a group, or perhaps through a specific introduction?

Looking at the big picture of these relationships can be enlightening. Here you see part of my business introduction network, mapped with NetDraw. My favorite part of the network is the long chain that begins with a random encounter with Kathy Kram (a professor of organizational behavior, whose son competed in a piano competition I organized) and leads ultimately to Steve Borgatti, guru of social network algorithms.

The origin of this notable chain is actually not all that random. In my own networking practice, I draw continually on lessons learned from my piano-selling former colleagues. Steinway reps, who think nothing of massaging several hundred separate leads through a two-year sales cycle, are some of the smoothest and most persistent networkers around. How do they do it? It takes a special breed. But it also helps to be anal about tracking contact information, including especially the date and context of each communication, and a specific reminder of when next to follow up. When each introduction takes two months (about average in my example chain above), tracking these data religiously is essential to good networking.

For a helpful complement to my Steinway ethnography, visit the website of networking guru Diane Darling, including "How to work a room."

Except where otherwise noted, Copyright (c) 2005 Connective Associates.

Sunday, November 27, 2005

More social network pictures of government scandal in the NY Times

As much as I appreciate when the NY Times publishes a network map, I wish they weren't going downhill in quality. Today's paper features this fairly uninteresting example:
See also last month's view of government scandal and this view of emails at Enron.

Except where otherwise noted, Copyright (c) 2005 Connective Associates.

Friday, November 25, 2005

Measuring all-time greatest hits with Google Analytics

Google released its web analytics as a free service about a week ago. I registered as soon as I heard about it, and already I have learned a lot of useful information that I never found through my previous web monitor, SiteMeter. Thanks to Google Analytics I have a clearer sense than ever of which pages of Connectedness draw the most hits.

Most hit-on pages of Connectedness:
Compare the above list with these most highly ranked pages of Connectedness and you will see a few worthy additions:

Wednesday, November 23, 2005

Making blogs more usable with page ranking and clustering

Happy Thanksgiving everybody!

Dennis Smith (notable team/project guru) sent me this handy list: "Weblog Usability: The Top Ten Design Mistakes." Two of them really got me--"the calendar is the only navigation" and "classic hits are buried."

The answer to the navigation question would seem to be categories, a feature that users like myself have long coveted. But after giving it some more thought this week, I decided to let go of my category-craving as just another vestigial attachment to the old-school org-chart mentality.

Having let go of categories, I now embrace the Google-ness of and simultaneously tackle the "classic hits" question. Those of you who subscribe to my blog may want to stop by the actual website and see the navigational features I have added to the right sidebar. Foremost is an "All-time greatest hits" link, which turns Google onto Connectedness and shows the highest-ranked pages on the site. In addition, I have selected a few notable keywords and provided links to scan my most popular postings on human capital, innovation, and visualization, among other topics.

It was only by necessity that I provided specific keywords to seed this new and dynamically ordered table of contents. I would much prefer to turn the entire task over to a clustering engine like Clusty or Grokker. Unfortunately, those tools cluster over multiple URLs and so don't help anyone navigate the interconnected web of postings filed away here. Maybe someday soon.

Monday, November 21, 2005

Our networks are privileged information

I recently did a social network analysis for a membership association as part of a presentation on "professional vs. personal networking." (E-mail me for a copy of the slides.) Afterwards, the coordinator of the group commented how quite of few of the most talented networkers in the room were precisely those who did not return their surveys. Whether personal or professional, relationships are privileged information, and responsible social network analysts must respect those who choose not to divulge. For an in-depth discussion of SNA ethics, see Borgatti and Molina's "Toward ethical guidelines for network research in organizations."

In stark contrast to those secretive networkers, many other seemingly judicious people become blatant exhibitionists in their online personas. Wendy McClure's essay "Mysteries of the Amazon," which appeared a week ago in the NY Times, is a highly entertaining confessional about the awkwardness that results when secrets meant for e-strangers fall into the wrong hands (in her case, because she is a digital snoop on her friends).

In regular life, we have developed sophisticated strategies for when to share which parts of our life stories. The online version of this is far from being worked out, but impressive progress is being made. For a look at the leading edge, see the Eclipse Higgins Project, led by Paul Trevithick of Parity Communications.

Sunday, November 20, 2005

New Yorker cartoon on bridging structural holes

I couldn't resist scanning this from The New Yorker of November 21, 2005, page 75.

For a more serious look at tapping diversity for the benefit of your business, see this overview I wrote a while back about Ron Burt's seminal work: "Structural Holes and Collaborative Innovation."

Wednesday, November 16, 2005

Expertise from Google rankings and CiteSeer

Not long ago I wrote about how browser add-ons like Outfoxed combine Google rankings with personal factors. Let me clarify my own thoughts and point out that Google rankings are really not so distinct from personal factors. That's because each time someone adds a link from one website to another, that is effectively a personal recommendation, which gets automatically incorporated into Google's rankings.

The implications of this are huge. See "Just Googling it is striking fear into the hearts of companies," an article from the NY Times of Nov 6, 2005, which aptly describes how Google and its ilk are "delivering results that are more and more like the advice of a trusted expert." Or, as Professor David B. Yoffie at Harvard Business School says in the same article, "Google is the realization of everything we thought the Internet was going to be but wasn't until Google."

For an example of how Google-rankings (or something similar) are providing trusted expert advice directly relevant to my readership, check out CiteSeer. This website of computer science publications tracks citations from one paper to another, giving it remarkable insight into the degree of relatedness and relative significance of the nearly one million papers in its archives.

That means that average people are much less dependant than they used to be on the real experts of the world to find just the information they're looking for. For example, anyone with an Internet connection can go to CiteSeer, look up what's been published about "social networks," and then automatically zoom in on the most influential "hub" publications that introduce major sub-topics--whatever those sub-topics may be. See my new "Technically Speaking" link at right to do just that. You can also ask CiteSeer to list publications according to how "hot" they are--specifically how many times are they expected to be cited by other papers this year, based on past trends. To try that yourself, just click on the link, "Tech II: What's Hot," which is also included in the right sidebar.

Monday, November 14, 2005

"Human capital" in NY Times Op-Ed by David Brooks

Notable more for its prominent placement than the actual content, a column by David Brooks proclaims the importance of human capital to America's international competitiveness.

Brooks breaks human capital into five underlying components, which he defines as
  • Cultural capital: the habits, assumptions, emotional dispositions and linguistic capacities we uncon­sciously pick up from families, neigh­bors and ethnic groups - usually by age 3.
  • Social capital: the knowl­edge of how to behave in groups and within institutions.
  • Moral capital: the ability to be trustworthy.
  • Cognitive capital: This can mean pure, inherited brainpower. But important cognitive skills are not measured by IQ tests and are not fixed.
  • Aspirational capital: the fire-in-the-belly ambition to achieve.
Brooks' first three components and their definitions seem a bit arbitrary to me, but I do appreciate his sense of urgency. Brooks particularly laments the growing gap between our understanding and use of human capital:
"Over the past quarter-century, researchers have done a lot of work trying to understand the different parts of human capital. Their work has been almost completely ignored by policy makers, who continue to treat human capital as just skills and knowledge. The result? A series expensive policy failures."
For more on human capital on these pages, see my recent post on maximizing the return on your human capital investment, as researched by Watson Wyatt.

Wednesday, November 09, 2005

IEEE Internet Computing: special Networked issue

The last time someone called me "Hopper" was third grade, until now. See last month's IEEE Internet Computing special Networked issue, which features the article "Social Networks and Social Networking" by Elizabeth F. Churchill and Christine A. Halverson. It looks like a good article, modulo a typo or two.

Thanks to Rajmohan Rajaraman for the tip!

Sunday, November 06, 2005

CAPTCHAs and googling (the pros and cons of smart machines)

Many thanks to Dave Koelle of Charles River Analytics for alerting me to the new "CAPTCHA" feature on For those of you unfamiliar with the term (as I was a couple days ago), CAPTCHA stands for "Completely Automated Public Turing test to tell Computers and Humans Apart." And let me add that a "Turing test" is simply a way to tell computers and humans apart solely based on their behavior, not their body chemistry.

If you surf the net at all, you have probably already passed many CAPTCHAs by reading distorted text (as on the right) and re-typing it. Humans can read "gpidhr" easily but computers, including spam-bots, cannot. Hence I can re-open comments on Connectedness to everybody, provided you can make it past the CAPTCHA.

The really interesting thing about CAPTCHAs is that the boundary between human behavior and computer algorithm is in fact less of a boundary than it is an arms race. See the CAPTCHA project at Carnegie Mellon for one side of this arms race, and read this post by a certified spam-botter for a glimpse at the other side.

Which side of the arms race are you on? I guess most of us want to feel better than computers and protect ourselves from spam and so root for the home team. But hold that thought while you read this article, "Just Googling It Is Striking Fear Into Companies," from today's NY Times. The Internet now makes it so easy to find the best price, even Wal-Mart is nervous. Makes you want to cheer for the forces of the digital age, doesn't it?

As I reported here earlier, researchers are augmenting Google with social network methods, resulting in systems like "Outfoxed" that combine Google-rankings with personal factors-- like what do my trusted friends and colleagues have to say about my web query? That combination will be more than enough to keep the CAPTCHA crew working overtime for as long as they want to defend the ever-blurring boundary between human expertise and machine intelligence.

Weekend edition: silkscreen prints by C. Barbour

The first piece of art I bought after I moved to Boston was a poster-sized silkscreen print, buried under a pile of seconds at Carol Barbour's booth at a Lexington arts and crafts fair. Ten dollars later I was the proud owner of the print, and ten years later it still hangs in a place of honor in my home. Occasionally I still see Carol at art fairs around the area. Last Christmas I bought a few cards from her, including the one below, which evokes in me everything I love about what I do:
I bought five of these and just mailed my last copy to a dear friend a couple weeks ago. I scanned it before putting it in the mail.

Wednesday, November 02, 2005

VisualComplexity by Manuel Lima

I spent some time today catching up on my favorite blogs (such as elearnspace, MeshForum, and Centrality). Almost all of them had recent links to visualcomplexity, a wonderful site by Manuel Lima, noted by Valdis Krebs on SOCNET. For those of us interested in visualizing networks, Manuel's compilation (a true labor of love) is a must-see. I have added it to my right hand column as well.