US Patent 11562298 Predictive analytics using first-party data of long-term conversion entities

Disclosed herein are systems and method for generating synthetic events. A method may include: monitoring user activity on a website of a conversion entity; collecting, for a monitored user session, first-party data including clickstream information on the website; calculating, using a machine learning algorithm, a prediction score indicative of a likelihood that a user associated with the monitored user session will request an asset depicted on the website after a threshold period of time, wherein the machine learning algorithm is trained using training vectors that map at least website clickstream information from various user sessions to offline asset conversion information collected after the threshold period of time from the various user sessions; in response to determining that the prediction score is greater than a threshold prediction score, generating a synthetic event indicative of the request being made by the user; and transmitting the synthetic event to a content platform.