Lessons from 2014: The 4 Pieces of Tech You Need To Know About

Lessons from 2014: The 4 Pieces of Tech You Need To Know About

The frenetic nature of an election cycle – with a defined end-date and a zero-sum outcome – can drive a rapid pace of innovation. But as with most midterm elections, this cycle proved to be more of an extension of the progress made in 2012 rather than a reinvention of the tech stack.

If 2008 was that digital ads that returned their investment and 2012 brought the ability to reach people not places, then 2014 was about whether or not people actually saw the message.

To prove this, there were a number of critical technological improvements that BPI led, which helped our election clients maximize the impact and results of their political advertising campaigns.

Below are the four most important advancements from 2014 that should be applied to future campaigns – political or otherwise:

  • Competitive Online Ad Tracking: Due to FEC disclosure laws and mandated political media rates, TV buys are easy to track in politics. This tracking often leads to an arms race of TV ad spending by watching what the opponent buys. However, the same regulations do not apply to digital – and as a result, it has historically been virtually impossible to know how and where an opponent is buying media online. This is made dramatically more difficult by the number of targeting options and creative formats available in the digital space – you can’t see a Kentucky-targeted advertisement unless an ad server thinks you’re in that state. But by working with technology partners to set up crawlers in 52 cities and 33 states, BPI was able to track what ads our opponents were running online and how they were purchased. This served as an early warning system in some states, and let us identify what messaging we needed to counter (and how it might differ from TV) as well as an easy way to analyze the post-election results in terms of who did what in each state online.
  • Cross-Device Individual Targeting: Targeting specific voters online was first pioneered at scale in 2012 by groups like Precision Network. But it was largely limited to banner and video ads on desktop or laptop computers, with relatively restricted potential ad placements. This election cycle, we extended this capability extend to mobile devices and additional digital channels such as social media. Now, we can reach specific voters wherever they are — regardless of device — across banner, video, Facebook, Twitter, personalized news recommendations, and more.
  • Measuring Ad Viewing and Listening Behavior: For the first time in political advertising history, we were able to measure how people actually watched our ads and interacted with them online. Distinct from metrics like “Completion Rate” that only indicate when an ad finished playing, we can now measure when people stopped watching – either jumping to another window while the ad played in the background or muting the player.  This let us know what type of creative actually resonated with voters and held their attention. It also allowed us to make much more efficient buying decisions. By focusing on total time spent viewing ads rather than simply looking at impressions delivered, we were able to quickly identify the highest quality ad placements and get more bang for the buck.
  • Data Warehousing and Centralized Reporting: All of the above involved exponential growth in the amount of data produced by our ad campaigns. We spent the first half of 2014 building infrastructure and processes to push all of our advertising data into a single database and dashboard. This let us spend more time on analysis and optimization, and less time wrangling disparate sources of data from dozens of ad platforms.

Our focus in 2014 was to maximize the impact of an ad buy. The next critical question to answer is how to precisely define the impact by tying it to offline results. We moved closer to this than ever before during this election cycle to tie shifts in polling and voting behavior to advertising activity at an individual level — it is a slow and laborious process — but we are hard at work developing a way to close the measurement loop.

Contact us if you’re interested in learning more.