Browse Prior Art Database

ADVERTISEMENT PERFORMANCE PREDICTION AND SELECTION

IP.com Disclosure Number: IPCOM000236879D
Publication Date: 2014-May-20
Document File: 12 page(s) / 1M

Publishing Venue

The IP.com Prior Art Database

Abstract

The performance of an advertisement (which can be text, video, audio) on a Web site may relate closely to the content of the advertisement and the interests of a potential viewer of the advertisement. For example, a viewer who has an interest in cinema may be more receptive to a movie trailer than a viewer who does not have that interest. However the same viewer may not be interested in watching the movie trailer if the viewer is searching a music video of a favorite singer. Advertisement performance may be measured in multiple ways, including Click Through Rate, View Through Rate (for skippable video advertisements), and Conversion Rate (i.e., probability that the advertisement leads to a certain result, such as a purchase or a subscription).

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 9% of the total text.

Page 01 of 12

ADVERTISEMENT PERFORMANCE PREDICTION AND SELECTION

Background

     The performance of an advertisement (which can be text, video, audio) on a Web site may relate closely to the content of the advertisement and the interests of a potential viewer of the advertisement. For example, a viewer who has an interest in cinema may be more receptive to a movie trailer than a viewer who does not have that interest. However the same viewer may not be interested in watching the movie trailer if the viewer is searching a music video of a favorite singer. Advertisement performance may be measured in multiple ways, including Click Through Rate, View Through Rate (for skippable video advertisements), and Conversion Rate (i.e., probability that the advertisement leads to a certain result, such as a purchase or a subscription).

Description of the Drawings

     The detailed description refers to the following figures in which like numerals refer to like items, and in which:

     Figure 1 illustrates aspects of an environment that supports selection and evaluation of advertisements based on the interests and moods of potential viewers (i.e., Web site visitors) of the advertisements

Figures 2A - 2C illustrate aspects of an example advertisement selection system; and

     Figure 3 is a flowchart illustrating an example advertisement selection method executed by the system of Figures 2A - 2C.

Detailed Description

     The herein disclosed systems and methods operate to find a match between or among viewer interests/moods and advertisements in order to maximize advertisement performance. The systems may capture viewer interests and moods by using machine learning techniques that process viewer actions (e.g., search, browse, playing videos, etc.) and predict advertisement performance for a certain advertisement or advertisement vertical (e.g., music, cars, movies).

     A video delivery system provides viewers with videos. The viewer may play the videos in a video player, which is a software program executing on the viewer's media device. The video delivery system may include one or more advertisements in conjunction with the provision of the video. Some advertisements are video advertisements. Some video advertisements may play prior to the video. These video advertisements may be referred to as pre-roll video advertisements. Some video advertisements include a skip feature. Some skip features allow the video ad to be skipped aft five seconds of play. Other video advertisements cannot be skipped. If the viewer watches a skippable video advertisement in its entirety, the video delivery system operator may charge the advertiser for display of the video advertisement. Thus, the video delivery system operator has an incentive to serve advertisements in which the viewer is likely to be interested and therefore likely is to watch to conclusion or near conclusion (i.e., the viewer does not skip the video advertisement). However, other indications of viewer interest, besides not skipping, a...