The Ozone Project hands contextual content classification brief to AdmantX
- Wednesday, August 7th, 2019
- Share this article:
The Ozone Project has selected ADmantX to power the contextual classification of its content, helping it to understand its publishers’ content at page-level in order to serve contextually relevant advertising, irrespective of each individual publisher’s content structure.
The Ozone Project offers advertisers direct access to an audience of more than 44.1m people in premium environments across publications from Guardian News & Media, News UK, The Telegraph and Reach PLC.
ADmantX’s Natural Language Processing (NLP) technology reads content in the same way a human does, detecting the subtle changes in context that can change the meaning of words on a page. This, the company says, delivers more accurate contextual targeting, compared to relying solely on keywords, which can leave too much room for error, and results in large swathes of publishers’ inventory being misclassified.
Before selecting AdmantX, The Ozone Project ran a rigorous pitch, which used human-review to score the accuracy of each supplier’s contextual classification across a large sample of publisher content. In the test, ADmantX’s technology was found to be 42 per cent more accurate than the next best provider.
ADmantX’s technology will become a key component in The Ozone Project’s wider audience infrastructure, with ADmantX delivering enrichment across semantic categories, industry verticals categories, entities (people, places, products and organisations), emotions and sentiment.
“Today advertisers face increasing pressure to compete for consumer attention,” said The Ozone Project commercial director, Danny Spears. “At The Ozone Project we understand the importance of harnessing the power of first-party audience data to help advertisers build effective, targeted digital campaigns that deliver incredible ROI. After extensive head-to-head testing, ADmantX was selected based on the quality of its contextual understanding. Their quality, range of data features and ability to support our need for customisation lends itself incredibly well to our business objectives.”