• by Tom Hespos , July 29, 2003

I've been talking to a number of business associates, friends and candidates lately about how to best use online marketing to support political candidates.

I find it a little strange that candidates can use their websites to present detailed information on their stance on the issues, raise money, and provoke political discussion while failing to leverage online advertising to its fullest. We've all heard about how Howard Dean raised more than half of the funds for his presidential campaign at his website and leveraged Meetup.com to engage over 50,000 people. Why then isn't the web blanketed with political advertising?

Furthermore, I think there's a lot of untapped media spending that could come the web's way from state and local races. There are billions of dollars spent here, generally in newspapers, spot TV and cable and direct mail. There's no reason why we can't pitch the power of online advertising to state and local candidates.

Accurate geotargeting will certainly help us. To tap into these dollars, website publishers will need to be able to target by ZIP code with a high degree of accuracy. For the sites collecting ZIP code information during site registration, this won't be a problem. However, many website publishers are still relying on external databases to provide this information, tagged to web surfers' IP addresses. We know this to be less accurate than gathering declared data, so this needs to improve.

Really, though, I think the true opportunity lies in targeting and segmentation by survey data. Swing voters are the true target, whether we're talking about a local race or the race for president.

Wouldn't it be valuable to political advertisers to be able to target, for instance, the people in a given group of ZIP codes who are primarily concerned with environmental issues? A combination of declared and observed data can give us the insight to be able to do that. When we can give Democratic candidates the ability to target advertising such that they don't waste media weight on hardcore Republicans who can't be swayed to vote Democrat, we introduce additional cost savings and make web advertising more appealing.

Behavioral data will likely play a stronger role in the targeting of political advertising than geographic data will. Remember how politicians wouldn't waste their time on the young crowd because they didn't turn out at the polls? That's a generalization akin to "Men 35-54 are the prime target for luxury cars." The notion of tying voting behavior to age demographics is as antiquated as doing the same with product usage. We know that there are plenty of young voters out there, we just haven't had a way to target them accurately and cost-effectively with advertising, until recently.

So let's start leveraging our data assets to provide the kind of targeting the political advertisers are looking for. Perhaps we can take a nice slice of that multi-billion-dollar pie.