With the topic of brand safety being front of mind for most marketers in 2020, our senior product manager Franz Etten outlines Ozone’s approach to creating safer, premium environments for advertisers.
Brand safety. Two words that have been firmly on the advertiser agenda for some time now, and as a result of 2020’s craziness have been brought even more to the digital fore. Addressing brand safety might seem like a very simplistic challenge, but the solutions available in-market - often using blunt keyword logic - can impact beyond the ways they are intended to.
Every sensible buyer understands the problem to be solved. No brand wants to appear against unsafe content - that’s a given - yet the irony is that many of the unsophisticated tools can actually drive digital spend away from text-heavy, premium environments into the long-tail of the internet. These are the places where editorial governance and regulation go out the window in a digital free-for-all. It is clear that the one-size fits all approach to brand safety doesn’t work.
That’s why at Ozone we’ve been developing our own tools that work specifically for advertisers in premium, known environments - the kind of popular, trusted websites that are more brand safe by design.
Introducing Ozone Brand Safety
Ozone Brand Safety has naturally been designed to protect advertisers from appearing against unsafe content. By ensuring we are not unnecessarily blocking safe content, we increase the marketers’ opportunity to engage with readers in the most premium environments. And with over 99% reach of the UK’s online digital population, our role is to help our advertisers reach these consumers as effectively as possible.
Our semantic contextual classification allows us to create brand safety segments that are tailored to the types of content created across the Ozone portfolio and we do this in three key functions:
Classification - We use semantic content analysis to determine whether the content of a given page is unsafe across a range of standard segment criteria (e.g. terrorism). For pages that are at risk of changing frequently (e.g. section pages), this classification is refreshed at regular intervals.
Activation - Through exclusionary and negative targeting, we can prevent campaigns from delivering against URLs that are deemed unsafe or that have not yet been classified. A similar logic is applied to our deals and across Ozone Marketplace meaning that the inventory opportunities we show them can be are pre-filtered to their individual requirements.
Reporting - our operations team is constantly analysing how different brand safety categories are impacting their campaign health, performance and delivery scores in order to optimise outcomes for their clients.
Applying Ozone Brand Safety
The first step in the application of our own tool has been to calibrate our core Ozone Brand Safety segments with those of major brand safety vendors. These have been extensively tested and optimised to improve accuracy, while allowing for greater addressability of safe impressions for our customers:
We are currently working on more customised segments to provide advertisers and agencies with solutions to their specific challenges. For example, we have developed a number of segment variations specifically for coronavirus content - by leveraging sentiment analysis to unlock positive content around this topic, we are able to provide advertisers with access to great, engaging content that would otherwise be blocked.
Results and next steps for Ozone Brand Safety
We are incredibly cognisant of the balance that exists between keeping our advertisers’ campaigns in safe territory, while maximising the scaled opportunity of consumer reach that the Ozone alliance creates.
In order to offer advertisers the broadest, yet safest, audiences, we have built a framework that enables the accuracy of our automated classification to be regularly reviewed under human-supervision. And when testing our solutions alongside those employed by our customers, we have seen a high level of accuracy for Ozone Brand Safety when compared with their buy-side tools, with <1% discrepancy. We will continue this focus on calibrating for accuracy.
The impact of this approach on Brand, Buyer and Publisher KPIs is significant when compared to traditional buy-side brand safety methodology:
We have demonstrated our ability to reduce Buyer’s brand safety ‘block rate’ by 75% whilst maintaining their (independently measured) brand safety score.
For Buyers who use ‘Firewall’ technology to suppress creative delivery after they have bought the media, Ozone’s brand safety solution can help them reduce wasted ad spend, and thus increase working media, by c.20%.
To maximise addressability, we are continuously testing new segment logic with different combinations of categories and scoring logic. This, alongside minimising our volume of unclassified content across all types of pages, is key to increasing the scale of our brand safe inventory. We have a number of client tests in the pipeline too, and we look forward to updating you with our progress in the near future
If you’d like to further discuss our approach to creating the most brand safe spaces in the most brand safe environments, please do get in touch.
