How artificial intelligence is making a real difference to contextual advertising

Let’s say you are considering buying a car. Now typically you may start off with certain features that you want in the car and you start searching for those features. Then you start looking for reviews of cars that are good in those specific features. But when you do that, you come across something new. Maybe you come across a new feature that you'd not even thought of and then suddenly you start researching it.

This is the example Preetham Venkky, Chief Digital Officer at DDB Mudra Group, gives when he talks about the startling way the advent of AI and Machine Learning into the advertising ecosystem has given a new lease of life to contextual advertising, especially in a cookieless future, as discussed in yesterday’s companion piece.

“What happens is that in behaviour targeting, you have the ability to look at what are the previous few actions. But now in a cookieless world, you don't have that. In contextual advertising, the issue is that you have to deliver specific to what you are looking for. So, the advertising that I create, all the content that I create as a brand has to be very, very specific to that for it to be successful. And that doesn't scale.”

Just imagine a brand creating content or advertising for every piece of a feature of any product or brand. It could be from the lens of culture, consumer, the product itself, the brand itself. That becomes extremely complicated because you can't scale that beyond a particular point in time. “Now, thanks to GPT, you have the ability to do that, which is why it's just contextualization at scale, where you don't have to be limited only to the standard 16 buckets, you could actually bring it up to 1600 buckets if you want. This is possible, only largely because in the last few months we've had GPT and AI having the ability to create content at scale and on the fly while creating the parameters or the framework or the limits of that framework by using prompts.”

“But that is not to say that behaviour did not meet that level of granularity but people were just happy to use that shorthand for a while. But contextual means granularity for it to succeed, because everyone is very clear that without that level of granularity, the content that you create as a brand seems to be native to the content the user is consuming. If it is not native to them it looks extremely jarring. Just imagine you want to watch the trailer for a movie. And what you get is the mileage of the car.”

Whereas if you were to bring in AI and then you were able to connect the two, then that makes a lot more sense. So that's exactly why contextual advertising, because of AI to a large extent, has had a palpable resurgence.

Indeed, Oindrila Roy, Managing Director, Publicis Worldwide India, sees advanced NLP and AI being pressed into service of the ad industry for more sophisticated ad targeting. “These technologies will scan through the nuancing, sentiment and intent of a certain content to make it more relevant for a brand to be seen in that environment. This will eventually enable more precise targeting while respecting user privacy.”

“The other place where I see contextual ads playing a significant role are environments that serve real-time content,” she notes, adding that dynamic contextual ads will leverage the changing content or user behaviour to serve relevant ads.

Initially, relying on basic keyword matching, contextual advertising has since evolved through the integration of advanced algorithms and machine learning. Karan Anand, SVP – Strategy, Interactive Avenues (the digital arm of IPG Mediabrands India), says “This evolution considers user demographics and behaviour, expanding into richer media and video formats. Today, it's gaining even more relevance with the advent of Programmatic advertising, which introduces automation and retargeting while addressing privacy concerns and ad-blocking challenges. Importantly, it respects user privacy by targeting ads based on content, aligning with privacy regulations, and enhancing the overall user experience by reducing disruption and increasing engagement.”

Venkky points out that cookies were already doing content segmentation. But what they were doing is they were limiting content segmentation. “But now with AI what happens is within those topic categories I can be very specific, that means I can actually read what the content is about on a webpage. And I can craft my advertising at scale to deliver what content is getting delivered. And that can only be done with AI. So that's why they are now saying that with AI, you're able to have the ability to amplify the efficiency and the results that you can get from contextual advertising. That's exactly why and I would argue that it will actually probably be better than behavioral targeting if we get it right.”

Pointing out that techniques include analysing the keywords, topics, and sentiment of the content, as well as considering the user's current session behavior, Rohan Chincholi, Managing Partner – Digital, Havas Media India, says, “Machine learning and AI play a crucial role in understanding the context and selecting relevant ads. As third-party cookies face obsolescence, advertisers are recalibrating their strategies, and one prominent solution is a return to contextual advertising.”

However, this isn't the sole answer. Alternatives include Identity Resolution (e.g., with ID5), Hashed Emails (HEMs), leveraging First-party Data, embracing Cohort-based or FLOC Solutions, and harnessing the power of Artificial Intelligence (AI).

“In this evolving landscape, we are anticipating shifts and challenges in the precision of targeting, frequency control, Look-A-Like audience creation, measurement practices, and the dynamic creative optimization (DCO) of ad campaigns,” says Chincholi.

Stay tuned for more on that.