Online Political Market Research

Market research is the process of understanding the market, what consumers want, what they have, what they can afford, and what will make them want your product or service. I’m oversimplfying here of course, but bear with me. Political research is the same, it begins with understanding the voters, their wants and dislikes and the triggers that will drive them to make the actions you want them to make (usually casting a vote in your candidate’s favor).
The first place we start when doing research for a search marketing effort is in the keywords and competition, so lets assume we’d do the same when planning an online political marketing campaign. If you’ve got deep pockets, hitwise data could come in handy here, telling you which search terms drive traffic to political websites, and which websites in fact have the most traffic, but this won’t come cheap. We’ll have to rely on lower cost tools, Wordtracker, keyworddiscovery, alexa, etc. Scouring location specific keyword searches you should be able to locate the issues and candidates that are searched on most often, and you should also be able to get a rough idea of which sites get the most traffic. That will be a rough sketch of what the online political “market” in your region looks like.
Finding voters’ triggers is more akin to the persuasion and conversion analysis that we often do on commercial websites. Multivariate testing seems to be the most direct and possibly fruitful application of these concepts to the political arena. For instance, we could construct an article or essay on a topic and specific some sort of success metric, a survey or email subscription form. Then we’d test different keywords, themes, framings, etc to see which resonated most with readers and produced the highest levels of success (however we’re defining that). Of course, as is the case in commercial testing but more so in political testing, we’d have to pay close attention to the source of the traffic, perhaps pre-screening those who are most likely to already agree with the concepts were testing and those most likely to disagree.

, , , ,