In the post-cookie world, context, sentiments and emotions can better target ads

This article first appeared in Campaign Asia Pacific

22 September By Laura Quigley

As behavioural targeting becomes harder, contextual targeting will become increasingly important in a world where user-level data has been reduced, writes Integral Ad Science’s APAC SVP, Laura Quigley.

Marketers are increasingly cautious and are rethinking how they plan, launch, manage, and protect their campaigns in the view of the recent pandemic. To add to the woes the programmatic supply chain murkiness made headlines in the industry at large.

Not surprising that Brand safety has been thrust into the spotlight for advertisers and publishers with the Covid-19 pandemic​ and ​social justice campaigns dominating the news cycles. Before we get into the nuances of how to enable advertisers to target ad placements with unmatched precision and at scale, it’s important to make sure that we understand key differences between brand safety and suitability 

Brand safety Vs Brand suitability

Brand safety is about protecting a brand’s reputation ​online by not placing their ads next to inappropriate content. Brand Safety is focused on avoidance of content that is risky to brands​ though ​Keyword blocklists (URL), Exclusion Lists, Inclusion Lists, Content categories, etc.

Brand suitability​ ​on the other hand focuses on matching a brand’s values to the context of the content that might surround their ads. In addition to protecting brands from appearing next to unsafe content, Brand Suitability provides more granular control of content adjacencies. A common example used to illustrate this point is avoiding placement for a car ad on a news report about a car crash. This placement will be unsuitable for a car ad.

Context is an important currency

With the deprecation of cookies as well as the weakening of identifier for advertisers (IDFA), behavioral targeting (ads based on what someone has previously purchased online and/or their recent browsing and search history) will become more difficult for marketers. Contextual targeting will become increasingly important in a world where user-level data has been reduced. Contextual targeting is centered on the environment in which an ad appears, rather than individuals’ inferred intent to purchase an item or click on an ad or content based on their former, online behavioral patterns. For that reason, advertisers are starting to look at ways they can push contextual targeting beyond its current capabilities. 

Are keyword technologies enough for sophisticated classification?

Most advertisers still rely on keyword lists for contextual targeting and brand safety, i.e. lists of words they want or don’t want to be linked with. Any page or URL  including a single word found in the list will be targeted or discarded.​ ​It ​does not take the meaning of the page into consideration, but rather just look for keywords and target or discard pages solely on these grounds. This strategy ​is ineffective because ​it creates ​false ​negatives and False ​positives.​ For example, a headline such as “Brand X Kills It in Retail”, which is positive in context, is not properly understood by keyword engines, and the web page is blocked, even if it is safe. On the other hand, False Negatives are pages that should have been blocked by the brand safety tech (or targeted by the contextual targeting tech) but are not because the specific keywords were not included in the list. For instance, the keyword list may include “shooter”,​ but not its synonyms “hitman”, “hired gun”, “gunman”, etc.

The solution for advertisers is not to compile the longest and most comprehensive list, adding keyword over keyword hoping everything will be covered. Instead, advertisers should rely on technology that takes the full text into account and the full understanding of​ the context.​

Matching ads to sentiments and emotions 

Let’s begin by exploring what these terms mean- the sentiment is the general mood of the page and often inferred as positive, neutral, or negative. Emotion is described as a feeling which can be caused by the situation that you are in, such as happiness, fear, disgust, sadness, surprise, etc. So why do sentiments and emotions matter when it comes to context? It matters greatly because brands sell to people, who have emotions and do make purchase decisions on how they feel about or perceive the brand or the product. Brands do use feelings to get consumers to buy so it’s critical that technology that’s responsible for discerning content and context is able to effectively discern the emotions, sentiments, cultural nuances to capture the consumer’s frame of mind. What if the technology was able to read the page as a human would and understand the mood and feelings contained in a piece of content? Would that not help advertisers keep emotions at the heart of their campaigns and ensure that the ads are well placed in the “right emotional context”? For example, the goal of a Luxury car ad is to appeal to the reader’s sense of relaxation, luxury, and comfort — and so ads might be placed entirely outside of the automobile context and next to content that evokes those feelings. This manner of semantic targeting lets the brand values resonate in the most appropriate emotional context.

Role Context plays in controlling and enhancing brand reputation

Advanced technology and solutions are available today to allow for full-page contextual analysis, which means understanding the true meaning of the page by accounting for nuances in context, language, semantics, sentiment, and emotional impact of the content.

There’s a big difference between a story that’s talking about something negative like the increased risk of increased death toll if we open states too early, versus a really positive and uplifting story about frontline medical workers and the sacrifices they’re making.​ ​Our responsibility is to give our customers the ability to make that distinction and align against the content that they feel is right for their particular brand.

The always-on context control approach helps marketers identify brand-suitable environments while avoiding potentially damaging ones. With this approach, brands are reassured that if and when something goes wrong, they have the best possible protection already in place and are in front of any controversy before it unfolds; there’s never a 100% guarantee though. At IAS, we focus on precise control over adjacencies through a multi-layered approach called Context Control, which is our sophisticated new suite of brand safety & suitability solutions that is powered by the best in class, cutting-edge semantic technology out there. This means we can deliver precision and do so at scale – something that was not achievable until now.