Leverage the Power of StackAdapt Contextual Targeting

by Abby Woodcock | Mar 28, 2023 | Programmatic Display | 0 comments

Our team here at Clix has had a lot of fun these past few months diving into Programmatic advertising. StackAdapt has been our primary Programmatic platform, and we’ve loved learning about all the possibilities for our clients. StackAdapt has several advanced targeting features but the one we’ve been most excited about is Contextual Targeting. With increasing regulations concerning the use of cookies and some browsers saying goodbye to third-party cookie tracking, advertisers need to find targeting strategies outside of Behavioral targeting to continue capturing the right audience.

How Does Contextual Targeting Work?

Contextual targeting basically uses context clues and aims to align ads with the websites they’re showing on, with the idea that if a user is on a site, they have an interest in the content within that site. Those who have used the Google Display Network, for example, would be familiar with this concept and targeting type.

As I mentioned above, there are some key differences between Behavioral and Contextual targeting. There is a great outline here showing how these targeting types differ:

What Contextual Targeting Does StackAdapt Offer?

Our reps provided a succinct breakdown of the three current contextual targeting strategies they offer:

PCAI – Reaching users in real-time reading about the content/topics you input on the platform. Our AI determines the best placements based on context.
Keyword Rule Targeting – Using Keyword matching and a precise set of rules to reach your audience.
Browsing – Reaching users who have historically read the content/topics you input in the platform. This is based on historical data, not in real time.

Page Context AI Targeting

With Page Context AI, or PCAI, targeting you set a group of keywords, or as StackAdapt calls them ‘In Context Phrases’, and the StackAdapt AI will find websites with that context on which to place your ads.

You’ll find this feature in the Targeting section of your campaign creation process:

Once you’ve put in your group of phrases, you can click on the Generate Content button and see some example sites, which is a really cool feature and a great way to see if you agree with the site matching:

You can read about some of the advantages of this type of targeting, but one callout I wanted to be sure to note here is that StackAdapt can expand the targeting to what it feels is relevant based on the phrases you’ve provided in an effort to maximize performance. I look at this as being somewhat similar to audience expansion in LinkedIn – except you don’t have the ability to opt-out.

*Pro tip: when leveraging PCAI targeting, we’d recommend this strategy with a Conversion-focused campaign instead of Awareness, as you’ll see high CPCs/CPMs with Awareness strategies.

Keyword Rule Targeting

Keyword Rule Targeting is a newer targeting strategy in StackAdapt and is helpful if you want to get even more specific than with PCAI. You’ll find this option right underneath PCAI targeting within campaign creation:

Here you can either Include or Exclude keywords, even specifying if they are Case Sensitive, if they show in just the Heading or the Heading and body of the page, as well as how frequently they show there – at least once, at least twice or 3+ times. You can create both AND as well as OR rules.

As with PCAI, you can click the Preview button to see a list of possible sites your ads would show on.

Probably the best thing about this targeting type is that StackAdapt provides keyword performance data! If you head to Reporting > Inventory Reports > Keywords you’ll see it there:

You can get insight into Impressions, Clicks, CTRs, Conversions, Cost, and more. This level of insight is extremely helpful to optimize and hone in on what is working in your campaign.

*Pro tips: StackAdapt recommends using a CPM bid with no goal and as a standalone targeting criterion in your campaign.

Browsing Targeting

The last type of Contextual targeting is Browsing Audiences, which is actually a hybrid Contextual and Behavioral strategy, and uses the same technology as PCAI targeting. Our reps sent along this explanation which I think perfectly outlines how these two targeting options work and how they differ:

“If you’re using PCAI and a user is browsing an identified page, and you also have an ad that matches one of the placements on that page, the system will place an ad right there. This makes PCAI a contextual targeting solution, as it’s just based on the page other than basic user details like geo and demo.
If you’re using a Browsing Audience and a user is browsing an identified page, the system does no additional checks and just captures their IP address and device ID. We then send that to LiveRamp to match against the user’s other devices and store all of those in the browsing audience. It’s both contextual (since it’s based on the context they browse) and behavioural (since it’s based on past browsing behavior by the time you target it).
Notably, this is why a browsing audience starts out empty and grows over time. It’s more like “PCAI retargeting” than anything else, and starts catching users to retarget as soon as you set up your “net” (the browsing parameters).”

To create a Browsing audience, head to the Audiences tab and choose Create an Audience, then choose Browsing Audience:

From there, you can establish your audience by adding both in-context and out-of-context phrases:

You can also click the Generate Content button to see examples of words and phrases related to website page topics users are viewing:

You can also leverage the Audience Size slider to help control the audience size and interest level in your established topics.

*Pro tip: once the audience has populated, you can go to Detailed View to see more information:

You’ll see how many people are in your segment, how many people were added to the audience day-over-day for the past 20 days, user geo-location, as well as the topical phrases you chose and the example topics. This is helpful for seeing if any additional example phrases should be added as out-of-context. Here, for instance, I can see that TikTok has been added to the sample topics so I’ll be adding that as an out-of-context phrase to the audience to help narrow it down and make it more relevant to this client and campaign:

Ultimately, while Contextual targeting offers a great way to help fill in the gaps with the hurdles surrounding cookies and Behavioral targeting methods, we’d recommend testing both types of targeting, and hybrid versions like Browsing Audiences, to help target a wider range of users and be able to test to see which get the best results for your clients.