Many publishers are losing ad revenue as their content is wrongly blacklisted as misinformation, inflammatory, or inappropriate, making it almost impossible to monetise. However, emerging AI-driven technology is providing brand advertisers with greater confidence in media safety, offering publishers a new way to recover lost revenue and ensure their content is properly categorised.

The amount of publisher inventory wrongly flagged as misinformation or inappropriate content can vary, but it’s a significant issue. Studies consistently show that ad inventory is often inaccurately flagged due to technical errors or flaws in the ad-tech supply chain, causing high-quality content to be overlooked. This issue is amplified during sensitive times, like elections or crises, where ad filtering systems may incorrectly flag or block legitimate inventory​.

At News UK, these challenges prompted the development of a new solution, Nucleus Narr(ai)te, designed to address the growing need for more nuanced content tagging and contextual advertising.

We spoke with Charlie Celino, Head of Strategic Development at News UK (left), to learn more about the thought process behind this initiative and the benefits it brings to publishers and advertisers alike.

What were the main challenges News UK faced that led to the creation of Nucleus Narr(ai)te?

Charlie Celino: One of the biggest issues was the broad and often inaccurate way third-party vendors were tagging our content. These systems typically relied on basic keyword matching to flag what they deemed “unsafe” content. The problem was that these systems lacked the nuance to understand context. So, content that was perfectly suitable for advertisers was often blocked because certain words were flagged without considering the actual sentiment or relevance.

For instance, during the COVID-19 pandemic, some of our most important reporting was getting flagged as unsafe because the tools were focusing solely on keywords like “pandemic” or “death.” This not only limited our ability to monetise that content, but also created missed opportunities for advertisers looking to reach engaged audiences.

We also found a lack of transparency in the way these systems worked. Publishers like us had little control over how content was being classified, which made it difficult to refine or improve these models to suit our specific needs.

What did you hope to achieve by developing Nucleus Narr(ai)te?

CC: The main goal was to regain control over how our content is tagged and classified, while creating a more sophisticated, context-driven solution. Nucleus Narr(ai)te is designed to provide a deeper understanding of the content we produce—moving beyond simple keyword analysis to evaluate sentiment, context, and overall suitability for brands.

This solution allows us to unlock more inventory that might have been restricted by traditional systems. For example, sports content like boxing was often flagged because of terms like “hit” or “punch.” With Nucleus Narr(ai)te, we can now analyse the full context, ensuring advertisers don’t miss valuable placements on content that is actually safe and engaging for their audiences.

How did you end up working with Illuma?

CC: We ended up working with Illuma after understanding in greater detail their collaboration with other publishers around IP protection and realising the value of their publisher-first mindset. They clearly understood the industry’s challenges, especially the growing importance of contextual advertising with the upcoming switch-off of third-party cookies. We found Illuma’s AI offering to be superior to anything else available; it added a much-needed human layer to our human-first content, which is ultimately consumed by humans.

How have these changes impacted monetisation and content diversity at News UK?

CC: The ability to classify content with more nuance has had a significant impact on our monetisation efforts. So far, the results from our early validation testing have been very promising. We’ve seen a +10% average increase in segment accuracy, with particularly notable improvements in several key verticals. For example, on The Sun, segment accuracy improved by +25% for ‘Beauty & Skincare’, +11% for ‘Tech’, and +12% for ‘Travel’. On The Times, we saw improvements of +20% for ‘Sport’, +16% for ‘Autos’, and +13% for ‘Retail’.

The increased segment accuracy allows for more precise targeting, ensuring ads reach the right audience in the right context. Additionally, our proof of concept testing has demonstrated a +16% average increase in brand-safe inventory across the News UK portfolio.

This improvement in accuracy has also opened up more brand-safe inventory for advertisers. For us, it means we’re able to monetise a broader range of content, including articles that touch on sensitive or complex topics, which are often the most important stories to tell.

How does Nucleus Narr(ai)te improve the advertising experience for brands?

CC: One of the key benefits for advertisers is the enhanced precision of our audience segments. By analysing the full context of the content, we can create much more granular and relevant audience segments, especially around breaking news or high-interest stories. This allows advertisers to target audiences more effectively, ensuring that their campaigns are reaching the right people at the right time.

What’s particularly useful is the way Nucleus Narr(ai)te works with dynamic content. It’s always learning and adjusting in real time, which means advertisers can optimise their campaigns on the fly based on how audiences are engaging with content at that moment. This flexibility is invaluable in today’s fast-paced digital environment.

What impact do you see these developments having on the wider publishing industry?

CC: I think this shift toward more sophisticated contextual advertising solutions is going to be transformative. It’s no longer enough to rely on blunt tools that tag content based on keywords alone. Publishers need more transparency, more control, and the ability to fine-tune these systems to better align with their content and audiences.

This means more opportunities for monetisation, particularly when it comes to quality journalism. A more refined approach to content tagging allows publishers to capitalise on important stories without worrying about lost ad revenue. It’s also about creating a more sustainable model for journalism, where advertisers and publishers can work together in a way that benefits both sides.

Lastly, are there any direct benefits for readers in this new approach?

CC: Absolutely. From the reader’s perspective, this technology ensures that the ads they see are more relevant and aligned with the content they’re consuming. It makes advertising feel less intrusive and more tailored to their interests. At the same time, by opening up more inventory for monetisation, it allows publishers like us to continue investing in high-quality journalism.

Ultimately, it’s about delivering a better overall experience—both for the advertisers looking to reach engaged audiences and for the readers who benefit from more relevant, brand-safe content.

Originally published in Media Makers Meet, October 2024.