Emerging AI Gives Advertising Greater Contextual Relevance

Emerging AI Gives Advertising Greater Contextual Relevance

July 24, 2020

Silverpush CRO Karthik Mehta discusses how AI-powered ‘contextual relevance’ is helping brands engage customers in targeted video advertising. 

By Karthik Mehta

The brand safety crisis, that first caught the attention of the mainstream media and advertisers back in 2017, is even more real today. With millions of pieces of user-generated visual content added to video-sharing platforms daily, brand safety has taken centre stage in the advertising world.

This was driven home in a recent survey around the State of Brand Safety and Suitability, where around 90% industry professionals agreed that unsafe exposure impacts brand perception negatively, and over 60% believe brand safety risks can result in revenue loss ranging from reduced buying to complete boycott of the brand.

In order to understand the extent of the issue, Silverpush reviewed around 15 million videos across the largest video hosting and sharing platforms, and found nearly 8-9% of analysed videos to be brand unsafe, i.e. featuring one or more contexts like nudity, smoking, violence, arms and guns, and more. Therefore, 1 in 10 ad placements across video content can potentially be brand unsafe.

Like every crisis, this has also resulted in practical and workable solutions that have provided a semblance of control to advertisers in varying degrees. However, some of the most widely used brand safety measures including blocklists, whitelisted channels/pages, third-party measurement, bring along their own set of efficiencies and pitfalls.

The pitfalls of static blanket exclusion and inclusion

Whitelists and keyword-based blocklists are the most widely used brand safety measures today. However, Blocklists often fail to provide the full context and can let damaging ad placements pass through, as well as kill reach. For instance, a video featuring smoking might not be described so in its title, description or meta tags. Limiting keyword-based solutions in identifying and filtering out these damaging contexts. On the other hand, keywords like shoot, kill, crash, and guns (some of the most blocked keywords) can easily be used within perfectly safe contexts like movies and songs.

Blocklists are often developed reactively — only excluding certain sites once they’ve already been exposed, and even when used preemptively, blocklists can accidentally inoculate brands against favourable audiences.

Even as marketers continue to use these measures, they’ve become less confident that they’re reaching their desired audiences — nearly 64% respondents from the survey said that using current brand safety measures has created an inability to reach specific audiences. This gained more weight recently as coronavirus topped keyword blocklists, squeezing ad revenues and killing brand reach.

What COVID-19 taught us about over-blocking

The COVID-19 pandemic has established a new normal and news-related content has become a big part of our lives. And where consumers go advertising follows. However, fearing unsafe exposure to news stories related to morbidity and highly negative impact of the pandemic, brands are forced to follow aggressive avoidance. In our April 2020 survey on Brands Response to COVID-19, 78% said they are either strongly or moderately worried about the unsafe brand exposure due to ad adjacency across COVID-19 related content.

One of the key factors behind extreme COVID-19 related over-blocking is the inability of keyword-based measures to detect if the stories are informative versus stories that can harm brands – which has led to blanket exclusions. This has not only led to demonetization of online news at a time when the public needs reliable information but has meant that brands failed to reach their target audience.

Brand safety is not one size fits all

Each brand is different and must define its own guidelines for inappropriate and damaging context in accordance with its specific needs, values and brand image. And brand safety measures and tools should be able to provide required controls to amplify or lower restrictions to allow a highly customized approach. Blanket restrictions lack the flexibility to offer custom controls. This is where brand suitability steps in to offer advertising controls that are highly customized to the needs of individual brands.

Emerging AI-powered solutions are increasingly focusing on providing context relevance, and are fast becoming an answer to brand safety woes. AI enables processing of large volumes of data at speed, with better context, at a higher scale and improved targeting efficiencies.

Specifically, computer vision powered context detection is promising in many ways. Nearly 17% of our respondents are using this tech to ensure brand safety. It allows in-video and in-image context detection, offering a higher degree of context relevance that surpasses limitations of traditional keyword-based targeting. This offers unparalleled insight for advertisers to place context-relevant video ads and to exclude unsafe content in a highly structured manner.

In-video context detection opens a whole new set of audience to improve reach, with unparalleled brand safety. The ad has a higher probability to match its environment in terms of context and messaging. It runs on the principle that users are engaging with their interests while consuming certain content, and engaging at the right moment can augment this experience and gain moment interest and trust. With brand safety as a key consideration, an advertisement when served in a context that matches the content is more likely to achieve increased clicks, views, and completed views.

(Ed. Kartik Mehta is the Chief Revenue Officer at adtech firm Silverpush. Mehta says he is responsible for its sales, partnerships and global expansion.


Leave a Reply

Your email address will not be published. Required fields are marked *