I wrote this for the GreaterDemocracyblog, but I'm also posting it here, because I can.

The biggest mistake Howard Dean made in his 2004 campaign wasn’t his attacks on Gephardt, and it wasn’t the scream. It was his software’s failure to “scale the intimacy,” to give the 1 millionth, or 10 millionth, campaign participant the same features, and the same sense of belonging, given the 10th and 100th.

Throughout the campaign, and even to this day, Dean and his Democracy for America have relied on Movable Type as their interface with supporters. MT is a good product, but its interactivity is limited. You enter an item on the blog, and comments flow from it in a straight line.

In a true community software package, like Scoop or Slash, users have a lot more power. They can start their own comment threads. They can launch their own discussions. They can create their own polls. Interaction happens, not just between the campaign and supporters, but among the supporters.

And the same software, on the same server, can allow for the creation of local groups, specialized groups, even private groups, each of which has the same powers as the main group. You get self-organization. You get cross-fertilization, on issues, on tactics. You create leaders. You find out what works.

The difference between a two-way interaction, offered by Movable Type, and a many-to-many interaction, offered by community software, is nothing less than the difference between Microsoft Windows and open source software.

To have true open source politics, you need tools that enable it.

Managing such a scaled system is a challenge, but it’s no more difficult than managing an actual campaign. The key is to to make certain the same person with power over the mini-campaign, or the group within the campaign, has power over the blog. At minimum, the campaign manager and blog manager must be very intimate, so that the concerns of the group are always transmitted to the decision maker, even if the consensus emerges that the decision maker should be replaced.

The worst decision made by the 2003 Dean campaign, and it was reflected in later (controversial) comments by Zephyr Teachout, was that the blog was an organ of the campaign. This was said and felt without malice, but it was a central point of failure. In fact, the blog (as it scales with the community it serves) is the campaign.

This happens only when the software itself can scale while maintaining an intimate relationship among all its users, and when the campaign is organized along the same lines as the software.

If Governor Dean understands this technical Clue and follows it, Democrats will have a choice next month between making him their chairman or having him as their President.

The 2008 campaign has already begun. It will be fought in 10,000 places, online and offline. Democrats need to create an infrastructure of money, of message, and of people that can compete with what Republicans do “under the Web.” They have to get back into communities across America, fighting and winning small battles now in order to know what they’re fighting for when the time comes.

That campaign starts with software. And the decisions on what software to use should be taken now.