Yield management system and method for advertising inventory

• US 8,386,315 B1
• Filed: 12/17/2001
• Issued: 02/26/2013
• Est. Priority Date: 05/30/2001
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.), IQ Acquisition Corporation
• Original Assignee: Carl Meyer
• Inventors: Bala, Aravind, Jones, Brian B., Ranka, Sanjay, Chatwin, Richard E., Nalcacioglu, Ismail Ahmet
• Primary Examiner(s): Myhre, James W
• Application Number:
  US10/024,630
• Time in Patent Office: 4,089 Days
• Field of Search: 705/14.41, 705/14.43
• US Class Current: 705/14.43
• CPC Class Codes: G06Q 30/0241 Advertisements
• CPC Class Codes: G06Q 30/0242 Determining effectiveness of advertisements

Abstract

System, method, computer program and computer program product automate and optimize process of placing advertising campaigns on available inventory. When visitor loads one of publisher'"'"'s web pages, system will be queried for advertisement that should be shown. System periodically analyzes performance data to determine optimal allocations. When queried, identifies advertiser campaign and advertisement to be shown. Results of that visitor encounter are logged or recorded for future analysis. Determination of campaign placements may advantageously be a two-phase process. First, campaign is efficiently tested on the inventory most likely to produce good results. The testing module dynamically adapts, adjusting the amount of testing on different channels until a sufficient determination of the creative performance on different channels can be made. The second phase is the resource allocation. Campaign performance data and contractual constraints are formulated as a network problem. Efficient methods for solving network problems are well known. These two phases may be further subdivided if necessary.

Claims

1. An article comprising a tangible machine readable medium that stores a program, the program being executable by a machine to automate online advertising campaign placement using historical performance data, an available publisher inventory of locations on web pages and impressions on those web pages, a particular location on a web page being a publisher channel, and contract information that includes a set of one or more contract parameters, wherein an advertiser has associated therewith a set of one or more campaigns, and a set of creatives for each campaign, a combination of an advertiser campaign and a publisher channel defining an campaign/channel arc, the program comprising:

a pretesting code module adapted to identify, from a set of campaign/channel arcs associated with the available publisher inventory and the set of one or more campaigns, a subset of campaign/channel arcs that are to be tested, wherein the subset of arcs are identified either randomly or using performance data from a prior advertiser campaign with respect to the channels;

a testing code module adapted to use the subset of arcs identified by the pretesting code module and a first portion of the inventory to identify a further subset of campaign/channel arcs that are expected to perform well while minimizing a likelihood of excluding an effective campaign/channel arc and an amount of the inventory used by the testing code module for testing; and

an allocation optimization module adapted to use the further subset of campaign/channel arcs identified by the testing code module and a second, remaining portion of the inventory to generate a set of allocations, wherein a given allocation identifies, for each publisher channel, a fraction of the set of impressions that should be allocated to each creative of each campaign in the set of campaigns, wherein the allocations are generated by solving an objective campaign placement function as a piecewise linear function having at least one segment whose slope is defined by values associated with the one or more contract parameters.

2. The article as described in claim 1 wherein the objective campaign placement function is derived from a non-linear objective function.

3. The article as described in claim 2 wherein the piecewise linear function is associated with one of:

advertising network profitability, advertising network revenue, inventory cost, and campaign performance variance.

4. The article as described in claim 1 wherein an estimate of campaign performance is formulated into the objective campaign placement function.

5. The article as described in claim 1 wherein the piecewise linear function is a monotonically increasing function.

6. The article as described in claim 1 wherein the piecewise linear function has a first segment, a second segment, and a third segment, the first segment having a slope equal to a difference between the price per action value and the penalty per action value, the second segment having a slope equal to the price per action value, and the third segment having a slope equal to the premium per action value, wherein the first and second segments are connected at a minimum associated with the number of actions promised to the campaign, and wherein the second and third segments are connected at a maximum associated with the number of actions promised to the campaign.

7. The article as described in claim 6 wherein the campaign is a cost per click (CPC) campaign, a cost per action (CPA) campaign, or a cost per thousand (CPM) campaign.