Gediminas Rickevičius
Feb 14, 2024

How contextual intelligence is changing the advertising landscape

Are the days of blunt keyword blacklisting and thematic blocks over? Now that AI is capable of detecting the tiniest linguistic nuances within natural language, contextual advertising is about to become an incredibly powerful tool for marketers in the post-cookie age.

Photo: Getty Images.
Photo: Getty Images.

A major turn in online advertising was marked by the idea that “content is king,” coined by Bill Gates in 1996 in an essay that bore the phrase as its title. After almost three decades, we have perhaps the first serious challenger for the crown of digital marketing.

In the age of automation and programmatic advertising, a possibility emerges that context is the new king. Whatever the call to action or content of the ad, its background makes all the difference.

For this reason, contextual intelligence is becoming increasingly important for brands and publishers alike. AI-powered web crawlers are among the tools that help transform and advance the very field of programmatic, context-based advertising.

The transforming landscape of programmatic advertising

Programmatic advertising refers to using automated solutions for buying digital ad space. It is based on real-time bidding, which uses software solutions to take part in an instant auction for digital ads on a per-impression basis

For publishers, programmatic advertising means an efficient way to find the highest bidder for ads and, thus, greater revenue from their ad inventory. Meanwhile, advertisers can reach the target audience faster and use their advertising budget more effectively. Thus, programmatic ad spending has been on a stable climb since 2017, on track to reach $725 billion by 2026.

However, 2017 was also the year that clearly showed programmatic advertising’s risk to brand safety. A host of incidents surfaced involving ads for various well-known brands, from Mercedes Benz to Thomson Reuters, appearing next to ISIS and other extremist group videos on YouTube. This led to a broader understanding that brands might be inadvertently supporting such groups by automatically buying ads on their websites and other platforms.

Along with hurting the advertisers, such mishaps have done great damage to the brand safety of publishers. YouTube’s owner Google has drawn criticism both as a publisher and as a major tool for advertising. More recently, Applebee’s light-hearted video in the background of the grim realities of war in Ukraine on CNN has caused a moral uproar. 

This has not only put Applebee’s in a tough spot of being both criticised and ridiculed but also resulted in blame on CNN for lack of care in placing ads. And these are the issues that major publishers face. Illegal, fraudulent, or otherwise disruptive ads hurt smaller publishers even more and thus require robust methods of defense.

All this has shown the importance of context in digital advertising and ignited some of its most recent transformations.

From brand safety to suitability

Many of the aforementioned advertising accidents have been caused due to failures to adhere to brand safety tenets. These tenets require that ads are not placed in the background of extremist, violent, pornographic, or otherwise illegal or divisive content. Public backlash forced brands to start moving from focusing on merely brand safety to thinking broader to include brand suitability.

Brand safety is more or less the same for any product. Whether you sell ice cream or vintage cars, you do not wish to see your ads next to Neo-Nazi group promos or fake news articles. Brand suitability goes a step forward and takes the entire context of the surrounding content into account. As such it is more product-specific. For example, an ad for a blood sugar tracker is well-suited to go with an article about diabetes while one for a candy bar is definitely not.

Brand suitability and ad verification challenges

Increased attention to contextual brand suitability has made it clear that the old approach focusing on keywords and blocklists is no longer enough. Contextual brand safety service providers face the following issues when attempting to thoroughly verify ads against the background of various content.

  • Keywords can have many meanings. Providing the publisher with a list of banned words might turn against the advertiser. Should a family restaurant blacklist the content with the word “blow” when advertising? It depends. Is it used as a sexual reference, a slang term for a drug, or reporting a deadly blow to the head? Or is it simply about different ways to blow up a balloon? One cannot tell without the context. 

  • Content blocks overreach. The damage of completely blocking topics manifested most clearly during the Covid-19 pandemic. A news website cannot avoid extensively reporting on a pandemic in the midst of it. Yet, some brands want to avoid all associations with it. Completely removing a brand from such discourses prevents it from reaching a huge audience and significantly hurts the publisher’s revenue.

  • Ad fraud is still persistent. For contextual ad verification, it marks the extremest point where the context can be a blank webpage boosted in the digital auction by fake, bot-generated impressions. Although viewed primarily as an issue to advertisers, a case can be made that ad fraud hurts publishers even more. By injecting the market with fake supply fraudsters artificially lower the value of ad inventory. 

  • Real-time bidding needs real-time data. Context is never stable, especially in cyberspace. To ensure brand safety and suitability when bidding for ad inventory, you need to constantly track large volumes of data on the changing content across the web.

  • Third-party cookies are dying. Another reason to pay more attention to context rather than user behavior is that tracking the latter is getting a lot harder. Without the third-party cookie, companies will have to rely on voluntarily given data and contextual intelligence for well-placed advertising.

To improve upon the old way of digital advertising, contextual brand suitability needs to utilise the newest developments in AI and automation.

How AI and web scraping advance contextual intelligence

The aforementioned challenges can be overcome with contextual analysis powered by AI-based web scrapers and other software solutions designed to fetch and analyse massive amounts of web data efficiently.

AI tools based on natural language processing (NLP) models are capable of detecting the tiniest linguistic nuances after analysing the context in which words are used. This helps to mitigate the risks of overreaching keyword and thematic blocks. Rather than blindly blacklisting sites based on words, contextual AI scrapes the content to determine its overall suitability. Using these tools advertisers can optimise bidding by leveraging applicability, affective tone, and other metrics that help identify the best-suited context for the ad.

Naturally, this and similar large-scale verification of ad inventory, like checking for ad frauds, requires an extensive web scraping infrastructure, complete with AI-powered proxies. Those who are able to use such infrastructures will benefit from real-time data that allows instant reactions to advertising opportunities and brand suitability risks. Additionally, combining such data collection architectures with AI-powered solutions to analyse visual content will improve the brands’ ability to stay away from extremist or controversial videos and imagery.

Finally, marketers who can master contextual AI and scrapers the fastest will suffer the least from the abolition of third-party cookies. Contextual advertising is not only perceived as less intrusive, but also can be much more effective than user tracking-based marketing.

Summing up

This era of digital advertising is marked by the growing importance of context. Failing to track and understand the ad inventory context in real-time leads to costly public relations fiascos for both brands and publishers. AI-powered solutions and automated web data collection tools advance contextual intelligence services, allowing marketers to benefit more from well-placed ads. 

In fact, contextual advertising might prove almost too effective as AI technology develops further. Being able to correctly assess the content sentiment, tone, and engagement by analysing instantaneously scraped data, contextual AI tools could theoretically always hit the optimal moment to target users.

This might raise the challenge of balancing business goals with the ethical requirement to avoid manipulating customers—an issue that is already familiar but now comes in dynamic and unexplored contexts.

Gediminas Rickevičius is the VP for Global Partnerships at Oxylabs. This article first appeared on Performance Marketing World.

Source:
Performance Marketing World

Related Articles

Just Published

1 hour ago

Google’s US antitrust trial comes to an end as both ...

Decision on if Google would be held accountable and face consequences might not come before Q1 2025, according to Judge Leonie Brinkema.

2 hours ago

Agency holdcos face a new crossroads: Reunite media ...

Iain Jacob predicted five years ago that buying tech and data, rather than renting it, would help agency “dinosaurs” modernise. Now, he says, merging media and creative will be a key differentiator in the AI era.

19 hours ago

40 Under 40 2024: Lana Zhang, Merkle

Zhang's visionary leadership, dedication to innovation, and contributions to marketing automation have established her as a cornerstone of the industry in China and beyond.

21 hours ago

What Chrome’s potential spin-off means for browsers ...

As the Department of Justice pushes for Google to divest Chrome, the ripple effects could redefine browser competition, shake up web standards, and disrupt the advertising ecosystem as we know it.