Browse Prior Art Database

Method and System for Determining an Optimum Threshold Value for Behavioral Targeting Advertisements

IP.com Disclosure Number: IPCOM000200480D
Publication Date: 2010-Oct-15
Document File: 5 page(s) / 54K

Publishing Venue

The IP.com Prior Art Database

Related People

Kevin L Chang: INVENTOR [+4]

Abstract

A method and system for determining an optimum threshold value for Behavioral Targeting (BT) advertisements is disclosed. Users with a score above the optimum threshold value have a higher probability to respond to an advertisement targeted at the users based on their behavior.

This text was extracted from a Microsoft Word document.
This is the abbreviated version, containing approximately 25% of the total text.

Method and System for Determining an Optimum Threshold Value for Behavioral Targeting Advertisements

Abstract

A method and system for determining an optimum threshold value for Behavioral Targeting (BT) advertisements is disclosed.  Users with a score above the optimum threshold value have a higher probability to respond to an advertisement targeted at the users based on their behavior.

Description

Disclosed is a method and system for determining an optimum threshold value for Behavioral Targeting (BT) advertisements.  A user is assigned a score, based on his online activity; users with a higher score have a higher probability of responding favorably to a BT targeted advertisement.  If a user’s score is above a given threshold, the user may be displayed BT targeted advertisements.  Users with a score above the optimum threshold value have a higher probability to respond to an advertisement targeted at the users based on their behavior.  The optimum threshold value is generally determined manually using a “trial and error” method.  This method is error prone and causes advertisements to not reach a set of users who might respond favorably to the advertisements.

The method and system disclosed herein analyze historical demand and supply for BT targeted advertisements.  The historical demand is related to statistics pertaining to impressions and number of users targeted by advertisers in the past.  The historical supply for BT advertisements is the number of impressions and users that have scores higher than the optimum threshold value.  The users having their score higher than the optimum threshold value are qualified to receive the BT advertisements.  If historic supply is found to be larger than historic demand, then the optimum threshold value is increased in order to reduce supply of impressions.  This results in improvement of model performance, while simultaneously ensuring that there is ample supply remaining to meet current and future demand.  Conversely, when historic supply barely satisfies or fails to meet historic demand, the optimum threshold value is decreased in order to increase the supply of impressions.  However, the optimum threshold value is not decreased by an excessive amount to ensure that model performance is not impacted. Hence, the method and system disclosed herein enables a user to select an optimum operating point for fixing the threshold.

For maintaining balance between demand and supply, the method and system disclosed herein computes four statistical metrics called Supply-Demand-Balance (SDB) metrics for a category of advertisements.  Examples of categories include but not limited to automotive, credit cards, mobile phone service.  For instance, advertisements in the category automotive are captured for search queries and web browsing activities related to cars and articles related to cars.  The four statistical metrics are, SDB_impressions, SDB_cookies, SDB_well_monetized and Views per Cook...