Bringing up Pandora and setting up a playlist can often be a deeply personal matter for some people. While some have dozens of channels to suit the mood at any particular time, others have just one or two channels devoted to a specific interest or feeling. But Pandora (News - Alert) thinks that it can take data mining to a whole new level by using those channels to determine just how people voted in previous elections, predict how they are likely to vote in the future, and sell advertising space accordingly.

Pandora already has a system in place to find out if this belief actually holds water, and will be rolling out the service this week.  With this advertising capability in place, reports suggest that both political organizations and individual candidates will have a better chance at advertising to the desired target audience.

The system, at last report, works via a simple combination of playlist and ZIP code, and then compiles the combinations to see just how similar the contents of said playlists are to the geographic areas in which they are found. The system may notice a user who listens to a lot of country music, which is big in Republican-dominated areas, and thereby throw Republican-themed advertising at the user. A profile belonging to a woman over 30 that has a child-themed music station, meanwhile, is likely a parent.

This isn't the first time that demographic information was taken like this in a bid to drive advertising; Facebook is largely considered the first to use demographic information from user profiles to augment advertising's targeting mechanisms, with attempts going back to the 2008 elections. Even back in November, there was something of a limited trial as advertisers could target users in areas with high Hispanic concentrations who listened to salsa music or Spanish-language stations.

Given that Pandora is in a semi-constant struggle for advertising dollars with close rival Spotify (News - Alert), among other competing services like Beats, the value inherent in having an excellent targeting mechanism for advertisement is clear. Going with political advertising as a way to boost revenue seems like a good idea at the outset. There will likely always be a demand for it—every two years or so there's a nice big ad buy as the various senators, representatives, and even presidential candidates get into the mix, with various offices getting different runs based on the time of year—and candidates aren't really much different than any other product. Candidates want to get their message out where the voters are, and indeed, there are plenty of voters to be had on Pandora. 

It’s reasonable enough, but it must be noted that this isn't a perfectly reliable method. There will always be outlying elements; the red state Republican, for example, who favors jazz on the weekend or the Democrat in rural Wisconsin who likes a bit of Garth Brooks. This kind of targeting likely won't be without flaw, but it's apt to cover the waterfront sufficiently to make advertisers happy with the overall reach and the solid targeting impact.

Targeted advertising is the big draw these days; being able to put out just the right ad for just the right user makes plenty of sense, and it makes advertising more likely to be effective in the long run. Advertisers crave value, and are likely to get it in spades thanks to services like Pandora's new targeting system.