As Covid-19 threatens digital media, the industry must move towards semantic targeting

Michael Nevins is CMO at SmartAdServer.

Despite consumers reading more online news than ever – driven by a desire for trustworthy content in these times of uncertainty – Covid-19 is posing an unparalleled threat to digital media. Advertisers utilising strict keyword blocking and adding broad coronavirus terms to their blacklists are missing out on the current traffic spike, as well as cutting off vital revenue for publishers.

Concerns about the long-term effects are rising. Following projections of £50 million in publisher losses, multiple industry bodies have launched pleas to preserve digital media sustainability: the IAB is asking advertisers to remove coronavirus-related terms from banned keywords, as is Newsworks with its ‘back don’t block’ initiative promoting support for British journalism.

A considered approach to ad placement is, of course, crucial. But if the industry is to endure, it must understand sweeping blocking measures are not the answer to weathering this crisis. So, what is? A semantic approach that looks deeper at the context of content is one effective way.

The limitations of keywords are not a new problem. While initially hailed as a streamlined way of avoiding inappropriate ad placements at speed and scale, the drawbacks of broad blocking practices have become increasingly clear. Before the Covid-19 outbreak, digital news about hit shows such as Game of Thrones saw a significant drop in advertising due to inclusion of terms featured on wide-ranging blacklists, such as “violence” and ‘murder”.

Now, the similar fate of content associated with coronavirus is re-emphasising the negative consequences of blunt security tactics. By automatically blocking all words linked to the virus, advertisers are writing off a large portion of the web and disproportionally punishing publishers producing high-quality — and often vital — content about the current situation. Plus, they are also restricting their own opportunities for audience engagement: Ofcom data shows Covid-19 news was accessed by 99% of the UK population daily in the first week of lockdown alone.

It goes without saying that addressing the keyword problem isn’t straightforward. Brands will still need to carefully manage the content ads appear beside in line with their specific values and offering. But it’s also too simplistic to assume any content referring to coronavirus will be harmful. In fact, analysis by GumGum of almost three million web pages containing Covid-19 keywords reveals 62% are brand safe; meaning there is plenty of timely and secure digital content advertisers could be harnessing.

The brand safety conversation needs refinement. Instead of focusing solely on generic no-go terms, the industry should utilise a better way of identifying which content is suitable. This is where semantic contextual targeting comes in. Thanks to AI-driven analysis that goes beyond basic keywords to establish sentiment, the revamped approach offers closer insight into content context; making it ideally suited to assessing both suitability and risk.

Its core mechanics are based on intelligent tools such as natural language processing (NLP). By scanning page-level content, algorithms evaluate the sequences of words, the probability they will appear, and their true meaning, accurately categorising content for large-scale yet secure targeting. For example, in today’s climate, identifying coronavirus-inspired recipe articles for home cooking would be relevant to place tailored ads promoting kitchen gadgets.

The main advantage of such a semantic shift for advertisers is obvious. With efficient page-by-page analysis, positive association with content can bolster their reputation, while avoiding genuinely unsafe ad space. But there is also the marked difference more nuanced assessment could make for publishers and news sites in the form of safeguarding ad revenues. Not to mention, the bonus of not relying on third-party cookies will come in handy as Google’s block comes into force and online privacy regulations continue to rise.

The way the global crisis will ultimately shakedown is unknown, but what we can be certain of is that the industry can’t continue on its current content blocking trajectory. Audiences are depending on trusted information, digital media needs investment to survive, and advertisers can’t place messages without content.

Covid-19 has sparked unprecedented challenges for the media industry. But smarter semantic evaluation that differentiates between the good and bad of online content could not only protect the industry for as long as the coronavirus storm holds, but also help it to finally move out of the brand safety rut and thrive long after the pandemic is over. However the crisis plays out, targeting ads according to the true meaning of media has to be a better way of guaranteeing relevance and driving engagement than picking out a few select keywords. It’s time the programmatic space moved towards this more balanced and blended process for targeting, not only to remain relevant at this difficult time, but to ensure it comes out stronger on the other side.

Photo by engin akyurt on Unsplash

Interested in hearing leading global brands discuss subjects like this in person?

Find out more about Digital Marketing World Forum (#DMWF) Europe, London, North America, and Singapore.  

View Comments
Leave a comment

One comment on “As Covid-19 threatens digital media, the industry must move towards semantic targeting

Leave a Reply

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