Throughout the 2004 campaign, I participated in the emerging conversation of civic activists engaged in using Internet technology to drive the campaigns. Most of the software components– mailing lists, websites, forums, blogs– were already designed and ready to go, so what was left was to debate how to properly use it. Part of that debate was driven by what values were inherent in the network models, a discussion that I felt was high on rhetoric and short on substance. So I thought I’d take time to sort out what network models mean and how they ought to be considered.

Network Models

A network defines the connections between things: computers, concepts, cities, citizens. There are some basic shapes, or topologies, of networks; common ones are the mesh and star networks. A mesh is where each node has multiple connections to different nodes; in a star network, each node connects through a single hub.

Most large networks are a hybrid. The most familiar electronic network is, of course, the Internet, which at the core is a mesh, and resilient to attack. But at a local level, it’s a hub, with a single connection drawn out to each residential subscriber. Adding a redundant line would be an unnecessary expense.

The Internet itself hosts many “virtual” networks which don’t need physical connections to be tied together. As the links are cheap to maintain, the mesh structure proliferates– in examples such as the world wide web, peer-to-peer networks for file sharing, and social networking sites. The egalitarian nature of meshes has been embraced by progresive activists as a way to connect to others independent of the party structure. In addition, it’s seen as a driving anti-authoritarianism.

Jock Gill, a contributor for GreaterDemocracy, and developer of the first White House email system, wrote an essay Pyramid Marketing vs Peer-to-Peer back in April. (I first met Jock in February at a Boston meetup) He was very turned off by what he saw in the Kerry campaign:

The problem I see is that the old political pros in the Kerry campaign only know the Roman Army model, adopted by the Catholic Church: A standard plan enforced from the top down. The old story imbedded in the technical and architectural limitations of the broadcast model. To be fair, how could they “know” any other model? It is what they grew up with. Kerry is the product in part of the Catholic church’s approach to organization and military world view.

If you are a big iron, proprietary software, client server engineer, all you see is the superior role of the center and the dependence of the thin (weak) client. If you are an Open Source, P-2-P [peer-to-peer] person, you see a world of vastly different. solutions and possibilities. Power at the user end point, not in the middle.

At the time I pointed out to Jock that Kerry’s time in the Church and in the military was rather insignificant relative to his work in government at the county, state, and federal levels. Then again, looking back over the campaign imagery reknown of “reporting for duty” and “I was an altar boy”, it’s possible these associations may stick. In the inside account of what really happened on the campaign, however, the mistake was was not too much top-down discipline but not enough between the warring factions (see Newsweek’s post-mortem).

Would the “peer-to-peer” model of Howard Dean’s campaign have fared any better? Chris Suellentrop anticipated in July 2003 the problems that win would afflict Dean by the end of the year:

It’s possible that this approach would be wildly successful, allowing Dean’s campaign to target a broad variety of voters with distinct messages…. But the decentralized approach could also allow unofficial groups to hijack the Dean campaign, dilute Dean’s message, and lead to strange arguments over who controls the “authentic” Dean message–the candidate, or the spontaneously organized groups that so far have been invaluable to his campaign?

The main problem I had with Jock’s essay was that he was too quick to denounce the “Amway-style organizational pyramid” as being “top-down”, “command and control”; to him “radical theology comes to mind”. Jock’s comments were echoed by Susan Epstein, in a post-election diary post. (I met Susan at the DNC this past summer). Susan had a litany of complaints about the coordination gaps due to technological shortcomings and legal barriers (as did I), and considered how to improve the “grassroots” efforts:

The GOP Amway-style pyramid relies on neighbors but is so top-down that I don’t see it working well for Dems. Dems like creativity and input into how they do things, in my experience.

Let’s peel away the Amway-style connotations here. Fundraising networks depend on personal connections. The Democratic Party bundled contributions in a very similar manner to the Republican Party, though with some minor differences; I generalize both as Social Network Fundraising. The $4000 I raised got me into the DNC, and gave me a conversation starter with people I met at fundraising events. If the party did nothing for my level of fundraisers (they refused to display my last name on the Kerry website, or to give me the contact information for other fundraisers), do we consider that a victory of a rejection of top-down values?

Furthermore, whether I was to go to my friends or to my neighbors, what sort of values decision does that mean? The ultimate decision that was made was for the DNC to contract out to Grassroots Campaigns to raise money door-to-door. (Slate reported that by the end of September that they raised $15m in donations and took a hefty $5m commission. The proper label for this is type of fundraising is not “top-down” but “shakedown”).

Ascribing “creative” as a positive value and “top-down” as a negative value paint such broad strokes that they are not useful at all. The volunteer management software that the Kerry campaign used was rudimentary at first, and conveyed very little vision at all. It improved over the season, adeptly shifting over to use “volunteer points,” of which fundraising was just one factor. By virtue of their points, some would become leaders, and some would not. The next step would be to make sure they’d stay connected. And here’s where the star model comes in.

The Star Model

A few civic activists have warmed to the need for more flexible network models. Zephyr Teachout, the director of Internet organizing for the Dean campaign, gave her insight her post-election analyis: Come Together, Right Now: The Internet’s Unlit Fuse:

All powerful networks have hubs. In the best case, the hub is responsive, but even in the worst case, an evangelist at the core allows for networks to grow and coordinate — and the individuals in the network to know each other and feel powerful.

I’ll call it the star model. The star metaphor is friendly, as it is to some degree recursive, celestially speaking: stars revolve around galaxies; planets around stars; moons around planets.

The proponents of mesh networks would point out that each person in a community may be connected to each other by no more than six degrees. Now rephrase this by pointing out that the six (in reality, probably four) degrees is all that’s needed to connect to the candidate. He’s the star.

The star model works for information. Example: This essay responds to Jock’s article and Zephyr’s article, both of which are posted in prominent Internet sites. Readers may choose to add responses, but they are to be no more than 2000 characters (350 words) so as not to overshadow the main piece. These are condensed in the comments section below. Other readers may not have time at all to respond, but may be able to offer an opinion through a one-click ViewPoint.

The star model breaks of course when it is streched thin so that its network resembles an asteroid belt. When several months ago I wanted to respond to a David Brooks column in an open manner, I thought to post to the Times‘s online forums. I discovered that his column had drawn an average of two hundred responses over the course of 4 months. I cannot even guess how many more letters poured in to the editor. This may be called the super-star effect, or perhaps more accurately a black hole.

Similarly, the super-star effect was employed when the Kerry campaign drafted the big stars of the Democratic party– Clinton, Carville, Cahill, Clark– to sign emails asking for money. Why so many emails? I had asked Zack Exley, the Kerry campaign’s director of Online Communication and Organizaiton, this summer. Zack pointed out that certain emails could raise $750,000– a small price for people getting multiple messages a day. The downside was that the email was a communications black hole– the typical recipient had no one to respond to.

One suggestion for avoiding the black hole was in my proposal for improving campaign emails, that I had shared with the campaign. This would identify certain local stars– top fundraisers, activists, volunteers, even local elected officials. These local stars would be one step closer to the core. Yet they would by no means be fixed into a designated hierarchy; they would compete with each other for attention and relevance.

This is how politics works in the real world. If community/campaign software adheres solely to the values exhibited by mesh networks, it risks its relevance. The values in mesh and star networks are not the exclusive province of either of the major parties; they are values exhibited by political parties in general. To suceed in building communities and campaigns, we need focus squarely on building constructive relationships no matter what the network terminology.