With growing concerns around the third-party cookie, contextual targeting is evolving, becoming more advanced, and as a result, more relevant and useful to marketer’s campaign strategies.
Contextual targeting, which enables you to target individual webpages, is based on contextual categories or keywords. Technology providers like Peer39 leverage a combination of natural language processing (NLP) and machine learning (ML) to identify and analyze web content and understand the category, keywords, and phrases of the page to enable marketers to reach consumers based on the content and context of the page.
To better understand the value of contextual and keyword targeting, we chatted with Mario Diez of Peer39, the pioneer in bringing a holistic approach to contextual ad targeting. They provide technology that leverages a combination of natural language processing (NLP) and machine learning (ML) to analyze every element of a web page to deliver its full suite of page-level intelligence products available across every DSP.
To start, can you explain how contextual works with Peer39?
Sure thing. At our core, we are a massive contextual classification engine. Globally, we see inventory as it is being made available to target, and we’re in real-time returning a broad array of signals that marketers use for targeting or avoidance strategies across display, mobile app, and CTV.
These signals are based on the contextual environment where our advanced semantic classification and custom keyword toolsets are applied to our page signals, which help buyers target different levels of quality environments such as ad clutter and content richness.
Can you talk a little bit about why semantic analysis is important for contextual targeting?
When you’re reading an article, we as humans can put into context what the content is about because of the surrounding elements. You read an article, and you’ll know it’s talking about amazing photography shots vs. Covid-19 shots. Or you shop for a new pitcher and are reading a review on a home design site, and you know it’s not an article about the Rockies’ new pitcher.
A semantic classification engine reads content the same way as a human, providing a contextual understanding and a sentiment classification of whether it was a positive, neutral, or negative piece of content. This is a stark difference from keyword-only systems that do not pick up on semantics or sentiment and run the risk of targeting keywords out of context, resulting in poor performance for marketers.
What are some of the first steps marketers can take to set up a contextual strategy?
For those just getting started, I’d suggest beginning a brief for your campaign or quarterly planning. When looking at your target audience, start thinking about the type of content the audience reads or spends time with. This will shape how to build a contextual strategy.
Then also, think about the type of campaigns you’re going to run. If branding-oriented, you’d want to consider page signals targeting a low number of ads on a page or maybe a content-rich environment. Once you have that foundation, you can use some of the tools to look at whether you should layer semantic categories with keywords, and if the campaign is running across connected TV (CTV) and mobile, you can match contextual targets there as well. The best part about getting started is it’s very easy to do from a setup and execution perspective with the Choozle team.
What are some common mistakes marketers make when starting with contextual keyword targeting?
The most common mistake we see is that marketers will look at the keyword tactic as a standalone one when there are now more modern approaches. The challenge for contextual as a standalone tactic is you risk targeting a keyword running on a page that is out of context. Modern marketers are now leveraging advanced contextual settings that sharpen their targeting controls layering on contextual and keywords.
From your perspective, why is contextual targeting now making a comeback?
The regulatory and browser changes certainly play a role in putting identify-free solutions like contextual in the spotlight. The other reason that often goes unpublished is that it simply works. Buyers who have scaled testing in preparation for the coming changes have surprisingly been surprised by the performance they’ve seen.
Additionally, there has been a ton of investment into modernizing toolsets. Broadening the tools set to include advanced custom category creation, contextual CTV, and modernizing brand safety to suitability controls have all contributed to the spotlight focusing more on contextual.
How do you view the future for contextual targeting?
We will see an explosion of identity-free data sources, resulting in a wider, richer data marketplace than we’ve ever had before. As the appetite for compliant, privacy-friendly data grows, advertisers will start to look beyond context and keywords. They will look towards other circumstance-based contextual data.
In addition to things like sentiment and emotion, they can think about their users’ emotions surrounding a sports score, the stock market, or their target audience’s mood in the cold, snow, or rain.