Recent A.I. research by GumGum, Inc. has shown that most of the popular Covid-19 search keywords are “brand safe.” This means companies can leverage the coronavirus-related search terms that their keyword tool comes up with and not worry about negatively affecting their brand image.
The company uses an artificial intelligence engine called Verity to leverage machine learning for analyzing content. Verity was able to find over 2.8 million unique pages that contained popular Covid-19 search keywords between the dates of 3/25/2020 and 4/6/2020 — just a little more than week’s time. According to the algorithm, over 60% of that content was deemed as “safe.”
Keyword Concerns Valid But Overblown
Phil Schraeder, the CEO of GumGum, indicated that he felt publishers were justified in their concerns about blocking coronavirus keywords. However, Schraeder elaborated on his stance, explaining that advertisers were also locking themselves out of a huge amount of keywords that are safe for their brand precisely at a time when more people are online browsing content.
In that single week, GumGum estimates that brands missed out on over 1.4 billion impressions on their network. And they point out that GumGum’s publisher network is a good sample of the web at large. Instead of beneting from additional traffic, advertisers would have missed out on all those impressions because they included words like “quarantine,” “coronavirus,” “covid-19,” and other related terms.
Machine Learning is the Secret Sauce Behind Evaluation
Verity’s method to determine brand safety is derived from natural language processing. It uses evaluations from almost a dozen machine learning models. This technology is designed to assess the level of threat that a given keyword will pose to an advertiser on the publisher’s networks and across the web in general, using different categories of threats.
There are different threat settings that GumGum uses. For this particular study, they used the sensitivity appropriate for a Fortune 100 brand. They achieved this by adjusting the threshold of statistical confidence used to evaluate the estimated keyword safety.
Ken Weiner is the company’s CTO. He claims that even under more conservative models, the majority of terms that marketers would uncover with a traditional keyword tool are safe.
CTO Suggests That Coronavirus Terms Can Be Safe for Brands
Weiner’s theory is that coronavirus has affected every aspect of society to a point where even mundane and innocent content makes a mention of it. Therefore, blocking related keywords is going a little overboard. He also believes that people have the wrong idea about brand safety in general, with most companies sacrificing reach for safety. The reality is, according to Weiner, that brands can achieve a balance of both.
Mr. Weiner added that the company’s analysis demonstrated certain categories of content to be demonstrably safer regarding COVID keywords, especially IAB (Interactive Advertising Bureau) genres.
The technology expert also went on to make a suggestion surrounding specific content niches. Pop culture, video gaming, and technology are particularly good for brand safety right now. These categories will grant brands more reach and the COVID-19 terms are mostly safe, according to Weiner. It may be time for advertisers to break out their keyword tool and use coronavirus keywords creatively.
0 Comments