Contextual targeting will not be enough in the cookieless environment

The main challenge of digital advertising is reaching users with high purchase potential. This is related to displaying proper ads to the right users at the right time and in a placement that matters. 

Airlines ads just next to news about a plane catastrophe? That’s definitely not the context in which the brand could shine. But displaying ticket promo ads on a booking website? That's the context where the possibility of reaching interested users is definitely high.

Contextual targeting is a solution that allows targeting certain contexts on the Internet. It enables matching the right ads with specific audiences, depending on the contents of the websites they are browsing.  

Contextual targeting – evolution

Contextual advertising isn't a new approach. In the past, the vast majority of marketers  relied on keyword targeting as a preferred strategy to connect with their desired audience. We can say that keywords were the first iteration of contextual targeting. The technology has evolved over the years, and things like image recognition and semantic analysis are now a standard. This allows for a deeper understanding of context beyond just keywords. To make all of this possible, today’s contextual targeting relies heavily on advanced Machine Learning and natural language processing (NLP) algorithms. 

Contextual – what’s the purpose?

Brand safety

It takes 20 years to build a reputation and five minutes to ruin it. If you think about that, you’ll do things differently. 

– Warren Buffett

Brand safety in digital marketing refers to the strategy that advertisers and marketers utilize to ensure their brand's online presence remains free from association with harmful, offensive, or inappropriate content. The ultimate goal is to protect the brand's reputation in the digital space.

Brand image is critically important, as it plays a central role in shaping consumer perceptions and influencing their purchasing decisions. There are several examples of how negative brand perception influenced the company’s revenue drop, to name a few: 

• BP: crisis following the oil spill in 2010 in the Gulf of Mexico.
• VW: shares price drop after revealing that the company had manipulated emissions tests.
• Uber: controversies and negative public perception related to workplace culture.

For brand safety purposes, contextual targeting is a universal and best-performing marketing tool.

Targeting to wide audiences

We can all easily recall iconic marketing slogans like Nike's "Just do it" or "Red Bull gives you wings." These are examples of well-known global brands that aspire to be instantly recognizable to people everywhere on the planet.

Such a company needs to communicate as widely as possible. They are communicating to a broad audience by using very general labels, such as “sports,” “fashion,” etc. For this use case, contextual targeting seems like a perfect match. Brands want their ads to appear on pages that are contextually related (but not strictly related!) to their product or service to maximize the chances of reaching potential customers.

Precise targeting

There is only one winning strategy. It is to carefully define the target market and direct a superior offering to that target market.

– Philip Kotler

While contextual targeting is highly valuable for brand safety purposes as well as reaching a wide audience, its defects are highlighted when more precise targeting is needed. Contextual targeting doesn't take into account individual user profiles or behaviors. It treats all users on a particular web page or app as having the same interests, which can lead to missed opportunities for personalization and targeting based on user history and preferences.

Moreover, a big share of websites do not have any concrete context, especially on big, multi-topic portals. Such articles fall under the “news” category, and it’s hard to extract relevant purchase intentions from them.  In such cases, advertisers have two options: either remove such contexts, limit scale significantly, or target them, despite knowing that it’s like blindly guessing user interests.

Additional signals enriching the contextual targeting

There are some interesting signals marketers can utilize together with contextual so that its usability is improved, while still preserving the user’s privacy. The most promising ones are:

Topics API: designed for cohort-based targeting, it's a privacy-preserving approach that groups users with similar interests into cohorts. Advertisers can still use these cohorts to deliver contextually relevant ads to a broader audience without tracking individual users. As the browser assigns a Topics cohort to the user, this information will also be included in the bid request. Consequently, the combination of contextual data with the user's assigned Topics cohort will be a common practice for targeting new and potentially engaged audiences.

First-party data: Publishers and advertisers can leverage first-party data, which is data collected directly from users with their acceptance. This data can provide insights into user preferences, interests, and behaviors without relying on third-party cookies. There are two possible ways how and where that data can be utilized:

• On the publisher’s website—directly in the place where first-party data is collected. This is possible, for example, via Seller-Define Audiences (you can learn more about this tool on the IAB website). 
• On external websites that offer ad placements—enabled via Protected Audience API (learn more in our article in Advertising Week).

Alternative sources of first-party data 

There are a variety of websites where contextual targeting isn’t allowed at all—because there are no ad placements. At the same time, users going there may indicate a strong purchase intent—so potentially is a valuable data source for advertisers and marketers. What are we talking about? Comparison and click-through websites of all kinds.

Let’s imagine that the user needs to buy a new laptop and has absolutely no idea what the market is offering at the moment. The majority of consumers would devote some time to do research online, read reviews, analyze pros and cons, etc. But Internet users are clever—they know how to recognize sponsored search results, and they are conscious that frequently, a producer website is not the best place to find real-user reviews. That’s why they will end up on some independent websites with rankings e.g., “Top 10 TVs for a family” or “Best laptop for gamers,” etc. Basically, it's possible to find this type of website in any category you can imagine—so for each product sold online, there’s a potential to find a user at the very end of their customer journey.

Once a user goes to one of these websites, it appears that there are no ads displayed there. The business model is different here, relying on click-through provision. Advertisers would be super interested in reaching purchase-intent users there, but that will undermine the portal’s independence and worsen the user experience. 

So why not make use of that data outside of the portal, in the open web environment?
The users could be segmented into cohorts, assuring that one particular internet user can not be identified while assuring a personalized ad experience at the same time. 

Today, in the cookie-based world, this type of strategy is hard to implement, mainly due to a lack of appropriate tools. Publishers are not able to track what happens with the user outside of their ecosystem. So no feedback is given about what happens next—what their data is mixed with, how it is used, what the effects of applying their data are, etc.

However, with tools such as the Protected Audience API, this will be easier. Data from a specific site, such as "exampleBestTVs.com," will not be mixed with other sites, such as "exampleBestRadios.com." That will facilitate analyzing the audience behavior in an isolated environment which fosters proper user classification and market value estimation, respecting data privacy at the same time.