As the online brand safety debate rages on, advertisers are thought to be considering alternative formats for their programmatic ads in an attempt to tackle the issue. With standard display ad units seemingly more likely to lead to placement concerns, for many, the native advertising format seems a more desirable alternative. In fact, spend on native ad placements in the UK alone is expected to reach $5.81 billion by 2020.

The reasoning behind the thinking is logical: the very nature of native is that it closely matches ads with their surroundings, ensuring greater contextual relevance and, by extension, brand safety. But the approach isn’t without its issues.

Typically, the processes used to identify appropriate native placements have been manual and can’t be applied at the scale needed for successful and efficient digital advertising. The most obvious way of boosting reach and achieving scale is adopting automated tools similar to those used in programmatic, such as keyword searches, which puts brands at risk of landing in the same, potentially unsafe, boat.

So, the question is: how can brands safely leverage native at scale?

Too slow or too broad: the analysis conundrum

The key problem facing brands when it comes to native is the long-standing issue of quantity versus quality. Either they can attempt to physically assess every piece of online content to make sure it’s appropriate; an approach that is not only limiting for campaign scope, but also unfeasible. Or they can use automated mechanisms such as keyword searches to analyse content en masse — and these come with many flaws of their own when it comes to brand safety.

Firstly, keyword assessment is reliant on probability; using statistics to assess what terms are likely to mean in context, and classifying content as safe or unsafe accordingly. Yet this doesn’t consider the shape-shifting nature of words. The exact meaning of a term can vary depending on how and where it’s used. For example, ‘house’ could be a dwelling, type of music, Australian retailer, or US TV show. Assessing words at such at a general, isolated level raises the risk that their true meaning won’t be identified and ad placements will be unsuitable.

Secondly, inflexible keyword models might take the opposite route: using lengthy blacklists that restrict the number of domains or URLs ads can be placed. Acting as generic and constantly growing ledgers of banned terms, these lists fence off content regardless of how certain words are employed. Thus, this creation of ‘false positives’ means brands are missing opportunities to connect ad messaging with positive content.

New tech for a high-speed age

None of this is to say that native isn’t an effective method of serving safe and engaging ads to a larger audience when executed well; or that humans can’t match machines in terms of deep analysis. It’s just that with huge stores of content to sift through — all of which contain potential brand reputation hazards — marketers need a better way of quickly defining what terms mean.

Enter artificially intelligent (AI) technology with expertise in semantics. Capable of efficiently assessing content without skipping over the details, AI-driven semantic analysis provides just the right mix of speed and safety for native and any other kind of online ad placement. The chief reason for this is it automates the process the human brain has perfected, such as natural language processing (NLP). By reading content in the same way humans do, technology can pick up on small but crucial variances that keyword searches overlook.

Thus, the semantic analysis offers an accurate and thorough evaluation of every article, video or social media post: including the topics they cover and the sentiment they express. And this gives marketers all the information required to determine good ad placements from bad before ads are served. Moreover, the automated production of this insight not only facilitates impactful in-the-moment ad targeting — with messages instantly tailored to blend with their environment and audience interests — but also allows physical teams to take a step back; thereby achieving the scalability necessary for native to flourish.  

Language is fluid and every day new terms and interpretations enter the lexicon. Technologies built around a fixed set of instructions aren’t fit for purpose when it comes to understanding context, regardless of the ad format being used. If marketers want to successfully navigate the ocean of potential placements out there – no matter whether it’s programmatic, native or direct – taking advantage of the best and most relevant opportunities while avoiding the unsafe ones, then they need to make semantic technology their top priority for 2018.