Political technology is a subculture unto itself, with many consultants and vendors as partisan as their clients. "We want Democrats to win," says Jim St. George, a principal at Voter Activation Network (VAN), a company that hosts electronic voter records for state Democratic parties; the parties make that data available to the presidential campaigns. St. George says the shared purpose among companies in what he calls "the Democratic family" (including Blue State Digital) makes it easier for vendors to integrate with each other.

VAN is working with other IT service providers and vendors to build application programming interfaces that will surround the basic voter file with data from other applications–a person's history as a donor or an activist, for example–and make it all accessible in one place. The idea is classic business intelligence, connecting the right people the right way at the right time, and ultimately to turn people out on voting day. "When you go to a house in Iowa, you should know if the person has given to a campaign, so maybe you talk about volunteer or fundraising opportunities instead of their vote, or even not knock on that door," says Blue State Digital's Franklin-Hodge. "Maybe a strong supporter who promises to vote hasn't given, so you change your approach on e-mail. But you need systems integrated and talking together to do that kind of intelligent targeting."

That level of integration can't happen fast enough for either party. Whatever Internet strategy best suits a campaign, it won't be fully effective until the basic connections are in place. And yet connecting all parts of a campaign, says St. George, presents "a meaningful challenge. I think we see where the solutions are, they are largely sketched out, but we're not there in every case." Franklin-Hodge concurs: "There is nothing approaching standardization or a turnkey software solution or strategy when it comes to integrating field and campaign operations."

The failure of the Dean campaign to integrate thousands of Internet-inspired volunteers who flooded Iowa in 2004 has become a cautionary tale. More successful was the 2004 Bush campaign, which tightly integrated its voter files with an extranet used by field offices and its own Web site. "We had a very complete picture of what we needed to do and where we stood," says Turk. "Everything online was aligned with our strategic goals." The campaign used microtargeting, a software-enabled process that matched consumer data to individual voter file records and built tools that matched volunteers with prospective voters who fit similar profiles.

That second Bush campaign, though, had two years and a lot of money to build its systems, and was not especially focused on newly emerging social media. Now, campaigns are all over the map in terms of organization and focus. Some have chief technology officers, others put technology decisions in the hands of strategists who lack technology experience. But none of them can afford to waste time on the wrong back-office system. "If you put in these applications after the work flow is established and tell people to do things a different way, you have a problem," Ruffini says. "You need that emphasis at the beginning of the campaign." Says Turk, "I don't know if any of the presidential candidates this year will build something comparable to what we had in '04 by February. Next November is more likely."

And there are some problems that technology will never solve. "Just because you've built an amazing network to support your candidate, it doesn't make you immune to Kryptonite," says Trippi, who saw the "Dean Scream" remixed and repeated endlessly across the Web. "If you say something stupid the week before the caucus, your network may not support you anymore. The Net is not so sticky that it will stick with you through anything. If you live by the Web, you can die by it."