Google "Project Pegasus: giving context to Google Nest"
Personalisation in online media is fraught with privacy risks, but how do you create personalised ads that don’t rely on users’ data? Project Pegasus, first piloted in the UK with The Guardian, is a breakthrough in solving this puzzle as it found a new way to target specific pages with contextual ads using Google Marketing Platform.
It works by reading publisher data in a new, more comprehensive way alongside a tool (using Google Cloud) that automates the production of creative tailored to individual articles on publisher’s site. Then there is Essence’s Vision Machine Learning model that uses Google Cloud’s Natural Language API to read article context using text and images.
An inaugural campaign for Google Nest smart speakers served 300,000 dynamic impressions that were adapted to the web page’s content and yielded strong results: 5.2% increase in purchase consideration and a 5.1% lift in category understanding.
Compare The Market "Meerkat Meals" by A Million Ads
Diageo "What’s Your Whisky Selector" by Unit9
Twitter "Magpie: a 100% AI-driven multi-channel marketing campaign" by Merkle
Bloom & Wild "Thoughtful marketing movement set in motion, thanks to Bloom & Wild" by Braze