Runtime matching of computing entities

A method for identifying one or more candidate computing entities associated with a first user of a first computing entity, the method comprising:

• maintaining, by an ad system, for each known computing entity of a plurality of known computing entities, one or more associated identity fingerprints, an associated technical ID, and associated behavioral data, including, for each known computing entity:
• capturing the associated technical ID and the associated behavioral data for the known computing entity using one or more interactions initiated from the known computing entity with one or more websites associated with a publisher system;
• determining an associated identity fingerprint for the known computing entity based upon at least a portion of an IP address of the known computing entity and user agent data for the known computing entity,
• identifying a number of other known computing entities which have an associated identity fingerprint that matches the determined identity fingerprint for the known computing entity, and
• associating the determined identity fingerprint with the known computing entity when the number of other known computing entities is less than a predetermined number;
• receiving, by the publisher system, a request for application data and a first technical ID from the first computing entity, the request for application data comprising a first IP address of the first computing entity and user agent data of the first computing entity, wherein the first technical ID is stored in a cookie on the first computing entity;
• receiving, by the ad system from the publisher system, the first technical ID associated with the first computing entity, a portion of the first IP address of the first computing entity, and a device model value from the user agent data for the first computing entity, wherein the portion of the first IP address and the device model value in combination do not uniquely identify the first computing entity;
• determining, by the ad system, a first identity fingerprint for the first computing entity based on the portion of the first IP address of the first computing entity and the device model value from the user agent data for the first computing entity;
• identifying, by the ad system, one or more candidate computing entities from the plurality of known computing entities, each candidate computing entity of the one or more candidate computing entities identified based on a similarity between the first identity fingerprint and an associated identity fingerprint of the one or more associated identity fingerprints of the candidate computing entity, wherein each candidate computing entity of the one or more candidate computing entities is a different computing entity than the first computing entity;
• determining, by the ad system, first behavioral data based on the associated behavioral data of one or more of the one or more candidate computing entities;
• determining, by the ad system, a bid amount for an ad display opportunity on the first computing entity based at least in part on the first behavioral data;
• transmitting, by the ad system, a bid for the ad display opportunity to a realtime bidding platform, the bid including the bid amount;
• receiving, by the ad system, an indication from the realtime bidding platform that the bid is accepted;
• receiving, by the ad system, a request for an advertisement to be placed in the ad display opportunity from the first computing entity, the request including
• the full first IP address associated with the first computing entity and the complete first user agent data for the first computing entity;
• storing, by the ad system, the full first IP address associated with the first computing entity and the complete first user agent data for the first computing entity;
• and
• transmitting, by the ad system, the advertisement to the first computing entity for display.