Real-Time
Advances in AI technology over the past few years have made it possible to monitor the online activity of virtually every consumer in the US across billions of websites.
Our intent data platform tracks the online activity of 280 million US consumers across 300 billion websites on a weekly basis. This allows us to identify consumers researching our keywords and add them to our Buyer Intent Audience so we can target them with display and video ads on their mobile phones, desktops, laptops, tablets and home connected televisions (CTV).
Building the Buyer Intent Audience
There are two different methods of determining buyer intent – Predictive Intent and Real-time Intent. Predictive intent, which has been around for a couple of decades utilizes offline purchase activity to make algorithmically calculated predictions about which households are likely to purchase a particular product or service.
Real-time intent is based on online browsing and shopping activity. The types of signals utilized to determine real-time intent are search queries, website visits, online content consumption, ad clicks, and in-app activities combined with contextual information like location and time of day.
Keyword-based Natural Language Processing (NLP) AI models are utilized to measure online intent. Recent advances have built upon the original NLP keyword-based models by measuring how closely a page matches the target topic (scored 0–100). If it barely mentions the topic, it is given a low score. If it’s deeply relevant, like a product comparison page, it is given a high score.
These more advanced models also measure the strength of the intent by establishing a seven-day baseline and when someone’s activity around a topic triples vs the baseline, they’re deemed to be in-market and they are added to our dealer’s Buyer Intent Audience so we can serve the appropriate ads to their mobile phones, tablets, desktops and connected home televisions (CTV).
