March 30, 2022

Deep Fake Series - Part 2

The Ins and Outs of Contextually Targeted Campaigns

Deep Fake Series - Part 2

Welcome to part two of our series on in-market audiences! In-market audience segments vow to deliver ready-to-buy-shoppers. We’ve done a deep dive into these claims and want to share what we’ve discovered. In a nutshell, not all in-market audiences are created equally. In fact, we found that some are downright fake, which is why we refer to them as deep fakes. If you didn’t catch part 1 of our Deep Fake Series, Uncovering the Secrets to the Most Authentic In-Market Audiences, it provides a definition of deep fakes along with an introduction to the In-Market Audience Funnel.

We’re in the middle of defining and reviewing the In-Market Audience Funnel. We discussed modeled audiences and audiences created using credit data in part 1. Now, let’s discuss in-market audience segments built with contextual data.

Audiences built with contextual data are based on a website’s content. In the case of automotive contextual targeting, this would be anyone who is frequenting websites that have auto content you are targeting (there are myriad examples, including marketplace sites, auto review websites, etc.). Traditional contextual targeting serves the advertiser’s ad alongside the contextually targeted content. Contextual advertising is typically executed through ad platforms like Google Ads or Demand Side Platforms (DSP) like DV360 or theTradeDesk.

Another form of contextual targeting can be accomplished by creating a site whitelist or site inclusion list. A site inclusion list is a list of website domains, and ads are only served on those domains. For example, if a website inclusion list only has,,,, and on the list, ads would only be served on those sites.

An additional benefit to using contextual data, in the above contextual examples, is that it’s not reliant on third-party cookies, which means even consumers using Safari browsers can be reached. Safari browsers typically represent about 40% of the US population due to the fact that Safari is the default browser on iPhones.

A somewhat newer capability in contextual targeting enables marketers to buy contextual audience segments derived from the content a consumer has previously viewed. Once these segments are purchased, ads can be served to consumers on any website, not just auto websites. In essence, it is “contextual retargeting” because ads can follow the consumer across the internet. However, the foundation of this solution relies on third-party cookies, so consumers using a Safari browser can’t be targeted or reached. Also, the methodology and overall quality of these audience segments is questionable because the methodology used to create the audience is unknown. Thus, these types of contextual audience segments have the potential to fall into our “Deep Fake” category. Remember: understanding how these audiences are created will help you avoid falling into the Deep Fake trap.  

In theory, contextually targeted campaigns offer a super-efficient and cost-effective way to reach desired audiences via programmatic ad platforms such as theTradeDesk. The most significant drawbacks are the unknown quality of contextual audience segments (i.e. is the consumer really in-market) along with the reality of limited inventory availability for high-quality, website inclusion lists. Website publishers put these limits in place to protect their ad inventory from being cannibalized by programmatic ad platforms. Marketers can, however, buy contextually targeted ads directly from auto publishers. The biggest trade-off is a higher CPM along with the added complexity in media executions.

The key to identifying the highest-performing, contextually-targeted audiences is to test and learn by understanding which are most effective at driving your desired business outcomes. This will inevitably include a mix of lower and higher-cost CPM campaigns.

Demystifying in-market audiences is not as easy as it seems. But, you are well on your way to becoming a more astute buyer of in-market audiences! Stay tuned for our next part in the series: behavioral in-market audience segments.