Contextual advertising has been present in some form since the birth of digital advertising. But in the pre-GDPR era, user-level data was too readily available for alternative solutions to be explored with the enthusiasm they are today. Now, as third-party cookies prepare for their final descent, the industry is scrambling to find new ways to discover and target addressable audiences, leading to a surge of interest in contextual solutions.
It’s not just a strict regulatory environment that has given contextual advertising a shot in the arm. The proliferation of artificial intelligence in advertising technology—in particular, machine learning models that can process and learn from vast volumes of data—has expanded the scale, speed and precision of contextual solutions.
The biggest AI breakthrough for contextual advertising has been natural language processing, allowing AI programs to interpret language with a human-like level of understanding. Where legacy solutions might have had a hard time distinguishing between a director shooting a movie and a person shooting a gun, AI-powered contextual solutions can apply their semantic understanding to categorize the content appropriately.
Interest-based targeting is also riding the AI wave. While much of the user identification that interest-based targeting requires remains deterministic (through identifiers such as email addresses), probabilistic identification deploys AI to weave together various signals and data points to predict whether an individual falls into a particular audience segment. Though less precise, probabilistic IDs offer scale and can plug the gaps left by consumers that can’t be reached deterministically.
The most significant development in interest-based targeting is the gradual rollout of Google’s Topics API. Topics combines anonymized browser-level IDs with the user’s web activity to tag them with interests from a predefined list that can then be used for targeting.
Who’s ahead in the contextual versus interest competition?
The benefit of interest-based targeting is that a consumer’s prior behavior can be taken into account to determine the value of their impression. It can be reasonably inferred that someone who has been reading tourist guides, shopping for luggage and looking for travel insurance is planning a trip and can be targeted appropriately.
Depending on the type of content the consumer is engaging with, it may also be possible to design campaigns that serve different messaging depending on their likely position in the sales funnel. For example, someone watching a “best budget getaways” video is probably still in the consideration phase, while someone searching for flights is teetering on a conversion.
However, interest-based solutions might miss the sales window completely and continue targeting a consumer who is no longer relevant. We’ve all experienced seeing an ad for an item we’ve already purchased. Until the interest-based signals associated with the consumer expire, the advertiser is wasting spending on impressions that have zero chance of conversion.
This is an inherent limitation of interest-based targeting, and Google Topics will be no different as the user’s interests are refreshed weekly—an age in modern life. While a user irritated by out-of-date targeting could dig into their Chrome settings and delete topics they are no longer interested in, it’s safe to assume most won’t bother to maintain their advertising profiles manually.
Contextual advertising, on the other hand, is designed to work solely at the moment a bid request is sent. While this means a consumer’s prior interests can’t be considered, the immediacy of contextual suits the rapid pace of today’s sales funnel, where the wealth of information and ease of e-commerce has cut the gap between consideration and conversion.
Whether interest-based or contextual targeting is more appropriate will depend on what is being advertised and the campaign’s objectives. While contextual advertising can still infer sales funnel position through clues in the content, campaigns that require a closer eye on the journey to purchase may be better off using interest-based solutions.
But it’s not just the buy side that needs to weigh the pros and cons of contextual and interest-based targeting. Some media owners—such as premium publishers with large and loyal subscriber bases—have rich, consented first-party data that can be activated for audience segmentation and interest-based targeting. Many others, however, do not, either because they have no login requirements or they simply can’t afford the tech and skills investment required to collate and activate user data at scale. Such sites might opt to pack their pages with ad slots to compensate, undermining the user experience.
Even publishers with a large number of known, addressable users have many more unknown users who come and go without a trace. Think of all the times you visit a website without logging in; the content may have been valuable to you, but if the publisher relied solely on interest-based targeting, they would not be able to unlock the full value of your impression.
Contextual targeting is, by comparison, quite plug-and-play. Because it works from the page outward, it can operate on any site and begin improving addressability from the moment it’s activated. Publishers can raise their average revenue per user, take a quality-over-quantity approach to ad placements, and sell against a broader and more detailed range of content categories than if they rely on their CMS for tagging.
There’s no clear winner yet—but would we even want one?
Both contextual and interest-based targeting solutions continue to have their place in the privacy-first ecosystem. While the lack of a definitive winner may make investing in tech stacks and planning campaigns more complex, this diversity of solutions is a sign of health.
During the cookie era, user data was so ubiquitous and easy to acquire that most solutions followed the same approach, with innovation focused on making it easier to target individuals with more precision and across more devices. The post-cookie ecosystem is far more exciting. Contextual and interest-based solutions take entirely different approaches to increasing advertising’s relevance to target audiences, with both producing a regular stream of iterative improvements. A rising tide lifts all boats, and the flood of innovation flowing from the contextual versus interest competition is taking digital advertising to exciting new heights.
