BT: Ready for Prime-Time Politics

YouTube debates and the power of the blogosphere are two signs that online is becoming integral to the American political process. Yet until now political advertising strategy has remained largely TV-centric. In 2008, behavioral targeting will begin to change that, predicts Dakota Sullivan, CMO of Blue Lithium below.

Behavioral Insider: Re-targeting was a big focus in the behavioral targeting space when we spoke last year. What has been at the forefront of development for Blue Lithium since then?

Dakota Sullivan: One thing that's become more pronounced is our use of customized segmentation and moving away from pre-packaged segments. As we evolve our approach, working with pre-packaged segments seems like just doing contextual ads by proxy. We've found that the more we're able to customize segments by overlaying demographic and geographic targeting data, the stronger those campaigns perform.

BI: How does that relate to the new political Voter Network you just launched?

Sullivan: When you look at political campaigns -- whether it's for the presidency, senate or a ballot initiative or state office -- there's no way you can just rely on pre-packaged behavioral segments. Every campaign has its own unique issue mix.

Political campaigners are interested in behavioral targeting. This makes sense if you look at the recent history of political advertising. Campaigns have taken strategies and tactics from consumer marketing and adapted them to political applications.

For example, the use of direct mail was huge in the ‘80s and ‘90s. Nearly every successful national campaign in the past two decades has also identified key demographic niches such as ‘Soccer Moms' and ‘NASCAR Dads' and used contextual media to narrowcast. In the 2004 [presidential] campaign, Karl Rove used Zip codes along with other kinds of data to identify local swing voters in the most contested ‘battleground states,' for instance, and then used direct mail and cable to reach them with very targeted messages.

BI: What are the unique advantages online targeting can provide political campaigns?

Sullivan: The rise of behavioral targeting allows campaigns to go beyond demographics or Zip codes to connect with voters based on highly specific interests and passions. One example is, say, the John Edwards campaign, which has made poverty a core issue. Using online behavior, it becomes possible to identify people who are most engaged in and motivated by the issue based on sites they've visited, searches they've made, offers and ads they've been responsive to and communities of interest. In the past, campaigns were limited to looking at demographic markers like education level, age, income and race as a proxy for who might be interested in an issue.

BI: What kinds of behaviors are leveraged?

Sullivan: In the past, if a campaign wanted to identify women in particular sections of New England who are between 21 and 35 and are conservative, with particular passion about schools and foreign policy, it was difficult to target those interests in a very direct and granular way. Online there's a far richer pool of data to work with, including sites they visit, petitions, polls, or types of publications -- and within [those pubs], specific articles they've read.

In addition to browsing, search and previous call-to-action response behavior, we're finding that new approaches such as online polling on particular issues or special interests are a great way to identify activists. Questions like ‘Do you think the price of gas is too high?' And ‘What should the government do about it?' are ways of identifying preferences and profiling attitudes, and can be highly predictive of wider political orientations.

Another incredible use of political behavioral segmentation is that once you've identified a behavioral group within a larger demographic, you can test messages and creatives -- especially video -- very quickly and effectively. You can test three different video creatives online and know, not in days or weeks but a few hours, what's working -- and then apply it to broadcast buys.

 
BI: What has the response of campaigns been so far?

Sullivan: Politicians tend to be cautious in their media strategy, for obvious reasons. There's more risk to them even than to a brand if an ad shows up in an inappropriate place. For a brand, bad ad placement is an embarrassment, but for a political campaign it could easily be lethal. The challenge is to select a wide enough range of sites for real scale, but to insure relevancy. We've screened out sites that don't want political ads or that are obviously not relevant for them. We also have in place a Total Quality Initiative that safeguards how and where clients' ads run. It includes the toughest publisher requirements in the industry, real time URL site-source reporting, and third-party & technology monitoring.

BI: What are your goals as far as adoption of behavioral targeting in political campaigns? 

Sullivan: The 2008 cycle will be about making the concept of online political targeting viable. The share of spend online will be small compared to the $3 billion projected advertising spend total, but we believe we can lay the groundwork this year to get a much higher share in 2011 and 2012. Our goal is to have a rich literature of case studies in every single type of political campaigning, to prove the case for online advertising and behavioral targeting. We think that, as with Zip code targeting in swing states and districts in 2004, behavioral innovations forged in 2008 will become the new template for future campaigns.