Christakis, Gladwell, and Catalyzing Movements

Last month, Malcolm Gladwell wrote a provocative essay on online networks and social movements. I personally thought it was a brilliant articulation, but he made one problematic claim: “The platforms of social media are built around weak ties.”

This was a gross and unnecessary simplification. Folks on the Internet — including much of my close community — predictably threw a hissy fit, which was justified, but unfortunate. We missed the opportunity to use Gladwell’s piece as a launching pad for thoughtful deliberation on how to make social media platforms even better at facilitating meaningful connections.

I skirted the issue then, but I’m compelled to enter the fray now, thanks to Nicholas Christakis and James Fowler, the authors of Connected and two of the leading researchers in the field of social networks.

I am a card-carrying member of their fan club. And so it pains me to say that their essay, published today, on Tweeting and behavior change completely mystified me.

They start with an interesting anecdote. The actress, Alyssa Milano, who has over a million followers on Twitter, tweeted about their book and posted a link to purchase it. The net effect? Zero extra books sold.

Alyssa Milano MLB 2008

As an experiment, they decided to get someone “more influential” — in this case, Tim O’Reilly — to tweet about their book to his 1.5 million followers. Net result? A few extra books sold.

Finally, they had Susannah Fox of the Pew Research Center’s Internet and American Life Project tweet about their book to her 4,000 followers. This time, they sold three extra books.

What were their takeaways from this sordid experiment?

  1. The nature of your ties is what matters, not the number. Furthermore — and they directly call out Gladwell here — online connections can be either strong or weak.
  2. “Second, it is not just “influentials” who matter, but also “influenceables.” O’Reilly and Milano can both be persuasive, and they are connected to millions of people. But it was the thousands of followers of Fox who ultimately won –granted, not by much — our little contest. To make change happen, we need sheep as well as shepherds.”
  3. Online interactions must feel real in order to cultivate change.

I’m completely on board points 1 and 3. I have no idea of what 2 means, which is why I just quoted them.

My problem with their essay is that none of these points follow from their anecdote.

Their experiment centered around a simple set of desired actions: Click on the link, buy the book. In all three cases, basically no one completed the second step. If anything, this result supports Gladwell’s argument: online connections are mostly weak ties.

But that doesn’t follow either. Case in point: In 1999, the day after Barbara Walter’s interview with Monica Lewinsky aired, the lipstick Lewinsky was wearing (Club Monaco in Glaze) completely sold out. I can’t give you a definitive explanation for why this happened. What’s certain is that everyone who was watching the interview that night was a weak tie, and yet somehow, Lewinsky was able to get thousands of her “followers” to go out and make a purchase.

Here’s my takeaway: Yes, the nature of your ties matter. Yes, authenticity matters. But space matters also. We react to space in visceral, often irrational ways. Rational or not, if we understand how space affects our behavior, we can leverage that space accordingly.

This is why architects installed long stairwells leading up to the entrances of the great cathedrals. (People look up when they walk up stairs, and looking up elicits a feeling of reverence.) This is why you won’t find clocks or windows in Las Vegas casinos. This is why I chose to use a picture of Alyssa Milano in this blog post, even though she is only peripherally relevant here.

And this brings me back to my original point. What we — people interested in leveraging these new tools for social change — really need to start talking about is how the tools themselves can do better at eliciting desirable behaviors, how to transform transactions into meaningful action.

Twitter, Facebook, and Social Boundaries

Speaking of tweets and Twitter, I finally succumbed and activated my Twitter account a few weeks ago. Come follow me! I had many reasons for doing so, but the kicker was probably learning that Twitter is in fact for old people. Seriously, my main reason was that I’ve been blog-free for many months, and I wanted to maintain a lighter-weight Visible Pulse for sharing ideas and letting people know that I was still alive.    (MTL)

Unlike my experience with this blog, I initially found it hard to start tweeting. I love to share ideas, but I don’t like talking about myself. My blog has been great for that, and I figured I’d use Twitter the same way. But it’s hard to do in 140 characters. It’s much easier to mention what I had for breakfast than it is to share some brilliant new insight, although simply tweeting “Eureka!” probably fulfills the latter need quite nicely.    (MTM)

So to get going on Twitter, I used a trick that I never had to use with my blog. I built an audience. I started following people, and many of them reciprocated. Now tweeting was about having a conversation with the people in my network. I didn’t have to do this with my blog, because I was motivated enough to start sharing my ideas without it. Getting that audience simply furthered that motivation (the past few months not withstanding).    (MTN)

This is obvious stuff to people who already blog or tweet regularly, but it’s not obvious to those who don’t, and when it’s explicitly understood, it can be used to your advantage. This is why Pattern Languages are so useful.    (MTO)

It would be interesting to do some analytics on tweeting based on size of social networks. For example, do people with larger social networks tweet more?    (MTP)

Another nice Twitter pattern is the character limit: Constrained Space. Many people have told me that they find blogging intimidating, because they feel a lot of pressure to actually write something. The character constraint relieves that pressure. It’s easier to come up with a 140 character message than it is to fill a blank slate. Size of space matters.    (MTQ)

The issue I’m now having with Twitter is with social boundaries, and not surprisingly, the cause of this isn’t Twitter at all. It’s Facebook. My close colleagues know that I am obsessed with Facebook, not because of some deep seeded need to feel like I have a lot of friends, but because I think it is one of the most well-designed online tool I have ever seen. There are all of these well thought out social patterns deeply embedded in the tool.    (MTR)

Because of this, it’s no surprise that so many people across so many different networks use it. My pattern for studying Social Network sites has always been to only connect to people who connect to me first, then to watch what happens. With the vast majority of sites, it’s the same group of people end up pinging me. In other cases, certain niches become apparent. Facebook is the first Social Network tool I’ve used where people across almost all of the different communities of which I’m a part have found and reached out to me.    (MTS)

The problem is that Social Networks are not frictionless. You can’t just mix them all up and expect everything to be wonderful. A while back, Pamela Dingle told a great anecdote that wonderfully portrayed some of the unsavory consequences of these boundary issues.    (MTT)

One of the things that exacerbates Facebook’s challenges with Social Network friction is its open API. For a lot of reasons, Facebook has encouraged people across many different networks to intentionally come together in a shared space. However, its API allows people to bring new networks to the same shared space unintentionally.    (MTU)

I enjoy the status updates on Facebook, and so when I started tweeting, I decided to sync Twitter with my Facebook status. By doing so, my audience went from a somewhat closed community of folks who speak the same Shared Language to a much larger community of folks, many of whom think I’ve gone nuts. These include people like high school friends, most of whom find the idea of posting a picture that’s not hidden behind a password absolutely ludicrous.    (MTV)

Those of us who have been part of Identity Commons for a long time have been talking about these issues for ages, yet it’s fascinating to experience them firsthand. I don’t find them that big a deal, because I have well-defined boundaries that I think work with my different networks. I don’t mention people by name unless I know they’re comfortable with it or I’m attributing an idea to someone. I’ll happily write about what I had for breakfast (Tartine bread and gruyere this morning), but I won’t write about my dating life.    (MTW)

I don’t know how long the Twitter experiment will last. It still feels a bit uncomfortable, but it’s been fascinating, and I probably won’t stop anytime soon.    (MTX)

Social Implications of Data Sharing

A few weeks ago, Evan Henshaw-Plath was explaining to me his epiphany about 43people, where an aggregation of different services tied together by Social Networks was starting to look very compelling. He then said that the next natural step for the folks at The Robot Co-op was calendaring. If anyone knows about calendaring, it’s Evan, who started a calendaring company with Kellan Elliott-McCrea in a past life. (Evan, you need Purple Numbers in your blog. Everyone needs them, dammit!) Trying to find relevant event information works much better when tied to Social Networks. Folks are starting to recognize this en masse, and the industry is reacting accordingly.    (JV3)

In an ideal world, Social Networks wouldn’t be tied to a particular site. Instead, that information would be distributed, and users would control the distribution. I’d love to tie my Google searches with my social network profile at LinkedIn, for example. That’s not going to happen unless Google acquires LinkedIn, or unless there are specs and a culture for distributed data sharing.    (JV4)

The culture is the tricky part. For as long as I’ve known him, Ross Mayfield has had an email signature that says what recipients are allowed to do with that email. (The choices are “bloggable,” “ask first,” and “private.” I think “ask first” is his default.) That, my friends, is a link contract.    (JV5)

Andy Dale has been working on the technical details of distributed data sharing with his XDI work. XDI is hairy stuff, but it’s graspable. Recently, Andy blogged about the form that link contracts and data sharing agreements might take. Victor Grey has said many times that we need a Creative Commons for data sharing agreements — simple, understandable, reusable, and legally enforceable contracts for our data.    (JV6)

That’s just a start. What will the user interface for specifying these agreements look like? Will users pay any more attention to these then they already do to Terms of Agreement on web sites?    (JV7)

Emergent Learning Forum Gathering on Social Networks

I’ve been following Alex Gault‘s blog for several months now. It is an outstanding source of articles and information on collaboration and Knowledge Management. Earlier this past week, I learned that Alex had organized the program for last Tuesday’s Emergent Learning Forum meeting, “Social Networking, Relationship Capital and Expertise Management.” Both the speaker list (Spoke Software‘s Andy Halliday, Intel’s Anita Lo, and Tacit Knowledge SystemsDavid Gilmour) and the opportunity to meet yet another blogger in person were too much for me to resist.    (140)

Alex kicked off the meeting with an excellent introduction to Social Networks, providing some background material (see Social Networks) and a concise overview of the current marketplace.    (141)

Andy Halliday followed by talking a bit about Spoke Software, although the scope of his talk was much more general. Spoke’s tool identifies social networks within the company (including people’s contacts outside of the company) by analyzing outbound email, and then acts as a referral broker. Andy emphasized individuals’ abilities to control their profile and protect their data. Afterwards, I asked Andy whether they had identified a threshold for how large an organization usually is before it can benefit from such a tool. He answered 1,000 people, but added that Spoke’s extended search capabilities (for contacts outside of the company) increased the tool’s utility for smaller companies.    (142)

Spoke is marketing the tool to salespeople, but it is clearly cognizant of the wider opportunities. Andy cited a few, including search results based on your social network (essentially Brian Lincoln‘s Collab:GrassRootsPeerReview idea) and a tool for sorting your inbox (including spam filtering). Spoke hosts a free online version of its tool called the Spoke Network.    (143)

Anita Lo, Intel’s Productivity Program Manager, gave a remote presentation on Intel’s recently deployed expert locator service. There were some technical difficulties and the talk was cut short before Anita could talk in-depth about the system, but a few points caught my ear. Intel conducted an internal survey to identify its most salient knowledge management need, and expert location was the top priority. The result was a system called People Yellow Pages, based on a tool that they purchased but that Anita did not identify. The system seemed to depend on people keeping their profiles updated as well as an overall taxonomy for categorization, which was managed by a librarian and validated periodically via surveys. This approach is in stark contrast with Spoke Software‘s and Tacit Knowledge Systems‘s, so it would have been nice to discuss how well it worked and whether Intel had evaluated any tools that automatically built and maintained people’s profiles.    (144)

(Seb Paquet posted some comments on Anita’s talk as well. Interestingly, Seb watched the talk remotely from his perch in Canada, and he may have been able to follow the talk more clearly than those in attendance!)    (145)

David Gilmour closed out the morning’s talks. I wrote about David’s excellent Harvard Business Review article before. His talk mirrored that article in some ways — an impassioned belief in the knowledge brokering model, coordinating collaboration rather than trying to force people to work together — but he provided much more detail in several areas. In particular, he spent much time discussing his company’s emphasis on individual privacy (as had Andy earlier) and noted that Tacit Knowledge Systems had several patents on ways to protect privacy, one indication of the value the company places on it.    (146)

David observed that individuals liked to use the tool to see their own profiles, which are automatically constructed based on their behavior (email, documents, Web, etc.). On the business end, he noted that pharmaceutical companies (which make up many of his clients) needed no persuasion regarding the ROI of his approach; the only question was whether or not the tool worked as advertised.    (147)

Jay Cross, the CEO of Emergent Learning Forum, posted some notes on the day’s talks as well. He also mentioned Alex’s other blog, the Collaboration Cafe, which I have added to my aggregator as well.    (148)

Huberman on Communities of Practices and Forecasting

Bernardo Huberman, the director of the Information Dynamics Lab at Hewlett-Packard, gave a talk at Stanford on January 8 entitled, “Information Dynamics in the Networked World.” Huberman covered two somewhat disparate topics: Automatically discovering Communities Of Practice by analyzing email Social Networks, and a general forecasting technique based on markets.    (TX)

The first part of his talk was a summary of his recent papers. I’ve alluded to this work here many times, mostly in reference to Josh Tyler and SHOCK. The premise of the work is similar to that posed by David Gilmour in his recent Harvard Business Review article.    (TY)

Huberman described a technique for discovering clusters of social networks within an organization by analyzing email. The key innovation is an algorithm for discovering these clusters in linear time. The algorithm, inspired by circuit analysis, treats edges between nodes as resistors. By solving Kirchoff’s equations (which gives “voltage” values for each node), the algorithm determines to which cluster a node belongs.    (TZ)

The second part of Huberman’s talk was enthralling: Predicting the near-term future using markets. This is not a new idea. A very topical example of a similar effort is the Iowa Electronic Markets for predicting the outcome of presidential elections.    (U0)

The methodology Huberman described (developed by Kay-Yut Chen and, Leslie Fine) aggregates the predictions of a small group of individuals. It works in two stages. The first provides behavioral information about the participants, specifically their risk aversion. Huberman remarked that risk aversion is like a fingerprint; an individual’s level of risk aversion is generally constant. In the second stage, participants make predictions using small amounts of real money. The bets are anonymous. Those predictions are adjusted based on risk aversion levels, then aggregated.    (U1)

Huberman’s group set up a toy experiment to test the methodology. There were several marbles in an urn, and each marble was one of ten different colors. Participants were allowed to observe random draws from the urn, and then were asked to bet on the likelihood that a certain color would be drawn. In other words, they were guessing the breakdown of colors in the urn.    (U2)

Although some individuals did a good job of figuring out the general shape of the distribution curve, none came close to predicting the actual distribution. However, the aggregated prediction was almost perfect.    (U3)

The group then tried the methodology on a real-life scenario: predicting HP’s quarterly revenues. They identified a set of managers who were involved in the official forecast, and asked them to make bets. The official forecast was significantly higher than the actual numbers for that quarter. The aggregated prediction, however, was right on target. Huberman noted that the anonymity of the bets was probably the reason for the discrepancy.    (U4)

The discussion afterwards was lively. Not surprisingly, someone asked about the Policy Analysis Market, the much maligned (and subsequently axed) brainchild of John Poindexter’s Information Awareness Office. Huberman responded that the proposal was flawed, not surprisingly suggesting that this technique would have been the right way to implement the market. It was also poorly framed and, because of the vagaries of politics, will most likely never be considered in any form for the foreseeable future.    (U5)

I see many applications for this technique, and I wonder whether the Institute for the Future and similar organizations have explored it.    (U6)

As an aside, I love how multidisciplinary the work at HP’s Information Discovery Lab is. I’ve been following the lab for about a year now, and am consistently impressed by the quality and scope of the research there.    (U7)