Thirty-five years ago, when author Joe McGinniss compared a presidential campaign to selling a bar of soap, his conclusion was dead-on but still jarring. Today, the technology brought to bear on political research, strategy and marketing alarms no one, but it should. It is putting the old school to shame.

I recently had the chance to listen in on a panel of the American Association of Political Consultants, where the people who design and sell the latest techniques for targeting voters gathered to sell their wares to hungry campaign managers. What I heard confirmed for me that politics has increasingly become a mathematical exercise, designed to reach just enough people to get a candidate elected and no more, at the least possible cost.

That's not necessarily bad. The idea is still to communicate with potential voters, if in an increasingly sophisticated and manipulative way. But something tells me that the new wave of demographic and cultural targeting is not healthy for our democracy. The point is no longer to debate the issues of the day and build a mandate for your cause. It is simply to identify and move, by whatever means possible, those voters the computer tells you are most likely to side with your candidate.

Consider the banner ad. Those strips of copy atop an Internet page on the World Wide Web are the latest hot tool of political marketing. Using research about who is likely to look at what kind of page, plus actual data about individuals from online surveys, consultants can now target their message to an ever-narrower slice of the electorate.

Want to bash the president over the failure of the economy to generate enough jobs? Buy an ad on Monster.com, where only those people looking for work are likely to see it. If that's not precise enough for you, try screening delivery of the desktop message to only women who own homes and whose last job paid them between $40,000 and $50,000 a year. It's probably doable.

"Targeting doesn't get much better than that," said Hal Malchow, a consultant who cut his teeth in direct-mail marketing but is increasingly convinced that the Internet is the wave of our political future.

If you are skeptical, the old media are offering plenty of new opportunities for similar endeavors. Cable television executives can tell you exactly how many black female Democrats under 40 are likely to be watching a particular network in Seattle at any given time. If that's who your campaign wants to reach, you can buy an ad there without having to pay for a lot of viewers who aren't receptive to your candidate's ideas. And even stodgy old newspapers are making a play in the new world, offering data that will help a candidate target high-value households with flyers, post-it notes or plastic wrappers advertising their message on Election Day.

But let's back up a moment. The banner ad or the cable television commercial is really the end of the process. The beginning is research.

Cheap and fast computers have made it a snap to combine waves of census, voter registration and marketing information into a highly nuanced database that becomes the foundation for a communications strategy.

The smartest consultants have realized that they need more than just a few basic facts about voters to predict their behavior at the polls. A sloppy data cruncher, for example, might look at a voter who earns $28,000 a year, is a union member and drives a pick-up truck and conclude that he will lean Democratic. But throw in one more fact - he watches the Rev. Jerry Falwell on television - and the entire picture changes.

Similarly, it used to be that consultants guessed which voters were likely to vote by simply looking at voter history. If you voted in three of the past four elections, for example, you were considered a reliable bet to do so again. Now voter history is combined with age, income and other factors to dial that prediction down to the decimal point.

Early polling then refines the data further. It gives the consultant a statistical profile of the electorate and a ranking of voter sub-groups according to who is most persuadable. Don't waste your money talking to young single men on the west side of town when the biggest bang for your buck is among elderly women on the city's northern edge.

I don't doubt that these tactics work. But what's increasingly missing from data-driven campaigns is a broader community conversation about the issues and problems that confront us all. By targeting only those people already inclined to vote, and communicating in narrower and narrower messages, campaigns ignore or turn off everybody else, driving political participation ever lower while driving the premium on targeting still higher. It's a vicious cycle.