A mix of self-aggrandizing PR and overly eager journalists willing to take the bait has given many political observers – including practitioners – the false idea that digital ad targeting is a precise science with the accuracy of a laser-guided missile. “Microtargeting” for example is often targeted for the public’s ire about what’s wrong with campaigns.
In the popular imagination, campaigns use sensitive data to send customized messages to specific voters so one group can be told exactly what they want to hear and another group can be told the opposite. This isn’t how an effective campaign uses online ad targeting.
Unfortunately, I know too many candidates and campaigners also believe some parts of this myth. When they do, it leads to ineffective online advertising and can derail campaign strategy. And while it’s true that digital advertising can be more precisely targeted than TV, radio, and print, the fundamental principle of advertising – repeating a message to a desired audience – remains the same across each medium.
To properly target your ads, you first need to know how the targeting works.
In order to use online targeting effectively, you need to have a basic technical understanding of the different methods for determining which individuals belong within an audience along with the limitations of each.
Behaviorally targeted ads rely on actions users take to determine whether they should see an ad or not. These could be visitors to your site that you remarket with ads, individuals who have engaged with conservative political content elsewhere online, or actions they’ve taken on other sites, including social networks. This targeting typically relies on “cookies” or related tracking technology that follows a user as they browse the web.
The key limitation with this form of targeting is effectively deducing intent from behavior. Perhaps the person who visited your website earlier this month was a supporter and they’re interested in donating to you, but it could just have also been staff from the opposing campaign. On the whole, however, past behavior is a reliable indicator of future action.
Data Matched Targeting
This targeting technique ingests a dataset you provide, typically a voter file, that the ad network or platform matches based on what they know about the users. The more data you provide, the higher the match rate will be. For example, providing just “John Smith” is unlikely to result in a match, but “John Smith, 1234 Main Street, Anywhere, USA” increases the success rate.
This match rate is the most significant limitation for data-matched targeting. Depending on your data and the platform, the match rate can vary between 30-60%, which means that 40-70% of your intended audience won’t receive ads targeted this way.
Location information can be gathered from a user in a few different ways. They can provide it when they’re signing up for an account, it can be based on the IP address of their computers, or it can come from their phone’s GPS.
Each method has its drawbacks and limitations. Users might not update their information when they move, their work IP address might not be where they vote, and they may only open an app when they’re traveling.
Whichever type of targeting (or combination) you employ for your campaign’s digital advertising, you will never know with certainty whether someone saw your ad or not because ad networks must protect the privacy of their users. Understanding this simple fact dispels most myths about online advertising.