Technology: Matching Engine

RELEVANCE-DRIVEN MATCHING ENGINE

The purpose of our matching engine is to match in real-time users with commercial content with a high degree of relevance. The technical challenge here is to discern a consumer’s demographical context, their psychographic context and their commercial intent with sufficient confidence level and then to match it with advertiser information.

When the match succeeds, a true dialog is established between the constituents. We do this with an array of content and context analysis techniques at run-time. Our engine also incorporates analysis of collective user behavior across web sites, time zones and task patterns to infer a user’s context.

What does the engine have?

It is composed of machine learning classifiers that discern user context from the information available for each impression, which in a probabilistic sense hypothesize this user’s intent and attempt to match with appropriate information from our back-end content engines. These hypotheses are tested, and learned from constantly to evolve our understanding of this user.