Patent Application
20110029377
The invention relates generally to computer systems, and more particularly to an improved system and method for forecasting an inventory of online advertisement impressions by sampling in a map-reduce framework.
A major problem faced by an online advertising publisher is to forecast available inventory of advertisement impressions for sale to online advertisers. Online advertisers would like to target users visiting certain web pages with certain demographics, geographies, behavioral interests, as well as many other attributes. For example, an advertiser may want to target users in a publisher's website with the following profile: male, over 25 years old, with a behavioral interest in automobiles. Thus an online advertising publisher needs to make accurate forecasts for any combination of those attributes efficiently.
One method of forecasting available inventory of advertisement impressions for sale to online advertisers that are untargeted is described in U.S. Pat. No. 6,801,945, entitled “SYSTEMS AND METHODS FOR PREDICTING TRAFFIC ON INTERNET SITES”. However, the problem of forecasting available inventory of advertisement impressions for sale to online advertisers for targeting profiles of attributes is complicated by the fact that there may be thousands of targeting attributes. Moreover, online advertising applications may need to forecast advertisement impression inventory from hundreds of gigabytes of advertisement impression data collected daily. The greater the number of advertising properties managed by the online publisher, the more complicated the problem becomes to forecast the thousands of targeting attributes across the advertising properties.
What is needed is a way for an online publisher to efficiently forecast available inventory of online advertisement impressions for targeting profiles of attributes. Such a system and method should be able to make accurate forecasts for any combination of those attributes.
Briefly, the present invention may provide a system and method for forecasting an inventory of online advertisement impressions by sampling in a map-reduce framework. In an embodiment of a map-reduce framework for sampling online advertisement impressions, one or more sample mappers each executing on one or more servers may be operably coupled to one or more sample reducers each executing on one or more reducer servers in a distributed computer system. A sample mapper may receive targeting profiles of attributes for displaying advertisements and may collect samples of profiles of visitors served impressions of advertisements from impression logs. The samples of visitor profiles collected may be matched to the targeting profiles for displaying advertisements. Each of the targeting profiles may be mapped to a reducer server for integrating the samples collected by each of the one or more sample mappers executing on each of the servers into a reduction sample set for a targeting profile. The or more sample mappers may send the collected samples of visitor profiles matched to targeting profiles to the one or more mapped reducer servers for integrating the samples collected by each of the one or more sample mappers into a reduction sample set for a targeting profile.
One or more sample reducers each executing on one or more reducer servers may receive sets of samples of visitor profiles matched to a targeting profile for integration into a reduction sample set. A count of samples of visitor profiles matched to a targeting profile may be aggregated by a sample reducer for a targeting profile, and the sets of samples of visitor profiles matched to a targeting profile may be integrated into a reduction sample set for each targeting profile. In an embodiment, a multitude of sample sets may be integrated into a uniformly distributed reduction sample set for each of the targeting profiles. The one or more sample reducers may forecast a targeting profile inventory for each targeting profile from the aggregated count of samples of visitor profiles matched to the targeting profile. And a targeting profile inventory may be output by a sample reducer with a reduction sample set of visitor profiles for the targeting profile.
Advantageously, the present invention may accurately forecast an inventory of advertisement impressions for online advertising. By providing a map-reduce framework for sampling online advertisement impressions, the present invention may be scalable for online advertising applications that may forecast advertisement impression inventory from hundreds of gigabytes of advertisement impression data collected daily. Samples of visitor profiles matched to a targeting profile may be collected in parallel across multiple machines and reduced into a sample set representing a uniform distribution in parallel by multiple reducer servers.
Other advantages will become apparent from the following detailed description when taken in conjunction with the drawings, in which:
The invention may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and so forth, which perform particular tasks or implement particular abstract data types. The invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in local and/or remote computer storage media including memory storage devices.
With reference to
The computer system 100 may include a variety of computer-readable media. Computer-readable media can be any available media that can be accessed by the computer system 100 and includes both volatile and nonvolatile media. For example, computer-readable media may include volatile and nonvolatile computer storage media implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can accessed by the computer system 100. Communication media may include computer-readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. The term “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. For instance, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared and other wireless media.
The system memory 104 includes computer storage media in the form of volatile and/or nonvolatile memory such as read only memory (ROM) 106 and random access memory (RAM) 110. A basic input/output system 108 (BIOS), containing the basic routines that help to transfer information between elements within computer system 100, such as during start-up, is typically stored in ROM 106. Additionally, RAM 110 may contain operating system 112, application programs 114, other executable code 116 and program data 118. RAM 110 typically contains data and/or program modules that are immediately accessible to and/or presently being operated on by CPU 102.
The computer system 100 may also include other removable/non-removable, volatile/nonvolatile computer storage media. By way of example only,
The drives and their associated computer storage media, discussed above and illustrated in
The computer system 100 may operate in a networked environment using a network 136 to one or more remote computers, such as a remote computer 146. The remote computer 146 may be a personal computer, a server, a router, a network PC, a peer device or other common network node, and typically includes many or all of the elements described above relative to the computer system 100. The network 136 depicted in
The present invention is generally directed towards a system and method for forecasting an inventory of online advertisement impressions for targeting profiles of attributes by sampling in a map-reduce framework. As used herein, a targeting profile of attributes means one or more attributes associated with web page properties, with web browser properties, with one or more users including demographics, online behavior, and so forth. A map-reduce framework may support an implementation of a task that may separated into a map phase and a reduce phase. Sample mappers distributed on servers may collect and match samples of visitor profiles of attributes to targeting profiles of attributes. Sets of samples of visitor profiles matched to the targeting profiles may be mapped and distributed to reducer servers to integrate the samples collected by sample mappers into a reduction sample set for each targeting profile. The sample reducers may forecast a targeting profile inventory for each targeting profile and output the targeting profile inventory with a reduction sample set of visitor profiles. In an embodiment, the targeting profile inventory may be an inventory forecast of advertisement impressions available on display advertising properties during a time period. As used herein, a display advertising property means a collection of related web pages that may have advertising space allocated for displaying advertisements. As used herein, a reduction sample set means a sample set generated from integrating a subset of samples from sets of samples.
As will be seen, a forecast of an inventory of online advertisement impressions may be generated to target many different profiles of attributes. Thus, the present invention may provide a publisher with the capability to forecast available inventories of advertisement impressions for targeting different combinations of attributes before selling them to online advertisers. As will be understood, the various block diagrams, flow charts and scenarios described herein are only examples, and there are many other scenarios to which the present invention will apply.
Turning to
In various embodiments, one or more ad log servers 202 may be operably coupled to one or more reducer servers 222 by a network 220. The ad log server 202 and the reducer server 222 may each be a computer such as computer system 100 of
The ad log server 202 may include a sample mapper 204 that receives various ad targeting profiles, collects samples of visitor profiles, matches the samples of visitor profiles to ad targeting profiles, and outputs sets of visitor profile samples for each of the various ad targeting profiles. The sample mapper 204 may include a visitor profile sample collector 206 that collects samples of profiles of visitors served impressions of advertisements that are stored in ad logs 214 on the ad log server 202. The sample mapper 204 may also include an ad targeting match engine 208 that matches the samples of visitor profiles 218 to ad targeting profiles 216. And the sample mapper 204 may additionally include a mapper stream handler 210 that maps the targeting profiles to reducer servers 222 and sends representative samples of visitor profiles that match ad targeting profiles to a mapped reducer server 222. Each of these components may be any type of executable software code that may execute on a computer such as computer system 100 of
The reducer server 222 may include functionality for receiving sets of samples of visitor profiles matched to a targeting profile, aggregating a count of samples of visitor profiles matched to a targeting profile, integrating sets of samples of visitor profiles matched to a targeting profile, and forecasting the targeting profile inventory from the aggregated count of samples of visitor profiles matched to a targeting profile. The reducer server 222 may be operably coupled to a computer-readable storage medium such as reducer storage 232 that may store one or more ad targeting profile inventories 234 that include a reduction sample set of representative visitor profile samples 236.
The reducer server 222 may include a sample reducer 224 that receives one or more sets of samples of visitor profiles matched to a targeting profile, forecasts a targeting profile inventory from the one or more sets of samples of visitor profiles matched to a targeting profile, and outputs a targeting profile inventory with a reduction sample set of visitor profiles for the targeting profile. The sample reducer 224 may include a visitor profile sample processor 226 that aggregates a count of samples of visitor profiles matched to a targeting profile and integrates sets of samples of visitor profiles matched to a targeting profile. The sample reducer 224 may also include an ad targeting forecaster 228 that forecasts the targeting profile inventory from the aggregated count of samples of visitor profiles matched to a targeting profile. The sample reducer 224 may additionally include a reducer stream handler 230 that receives one or more sets of samples of visitor profiles matched to a targeting profile. Each of these components may be any type of executable software code that may execute on a computer such as computer system 100 of
There are many applications that may use the present invention to forecast inventory for displaying advertisement impressions. For instance, the present invention may be used to determine untargeted inventory for displaying advertisement impressions to online users. Or the present invention may be used to forecast targeting profile inventory for displaying advertisement impressions. For example, an application may send a request to obtain an inventory forecast of available advertisement impressions for targeting profiles of attributes. In an embodiment, the ad logs 224 may include recorded information of advertisement impressions served. The recorded information may include a web page ID, a user ID, an advertisement ID, a timestamp, and other information such as a web browser ID. The information gathered from the logs and other lookup tables, such as page hierarchy tables and visitor attribute tables, may include web page attributes such as properties of the page and the web page position of an advertisement; visitor attributes such as age, gender, country, behavioral interests; time attributes such as date and hour of the day; and other attributes such as attributes of a browser. Samples of visitor profiles that include visitor attributes may be collected from the ad logs on a multitude of ad log servers, may be matched to a targeting profile, and may be sent to mapped reducer servers that forecast a targeting profile inventory from sets of samples of visitor profiles matched to a targeting profile. The forecast of the targeting profile may then be sent as a response to the application request for an inventory forecast of available advertisement impressions for targeting profiles of attributes.
At step 308, each of the targeting profiles may be mapped to a reducer server for integrating the samples collected on each of the multitude of servers into a reduction sample set for a targeting profile. In an embodiment, an index may be constructed of targeting profile identifiers with each having an associated address of a reducer server. This index may be sent to the multitude of servers. At step 310, the multitude of servers may send the collected samples of visitor profiles matched to targeting profiles to the mapped reducer servers for integrating the samples collected on each of the multitude of servers into a reduction sample set for a targeting profile. At step 312, each reducer server may integrate the samples received from each of the multitude of servers into a reduction sample set for a targeting profile. In various embodiments, a number of samples proportional to the number received and processed may be randomly selected for including into the reduction sample set for the targeting profile. In various other embodiments, random samples may be uniformly selected from samples received for including into the reduction sample set for the targeting profile. And step 314, a targeting profile inventory may be output with a reduction sample set of visitor profiles for each of the targeting profiles by the reducer servers.
Each targeting profile may be mapped at step 408 by a sample mapper to a reducer server for sending the matched samples of visitor profiles for integration into a reduction sample set. In an embodiment, a single reducer server may be designated for integrating sample sets of visitor profiles from a multitude of sample mappers for a targeting profile. At step 410, samples of visitor profiles matched to a targeting profile may be sent to the mapped reducer server by a sample mapper for each of the targeting profiles for integration into a reduction sample set.
At step 602, a new sample set and a count of new samples may be received. In various embodiments, a reducer server may receive the new sample set and the count of new samples from one of multiple servers. For the first sample set of a multitude of sample sets received by the reducer server for the targeting profile, the samples from the first sample set may be stored into an empty reduction sample set and an aggregate count of samples received may be set to zero. At step 604, a random number may be chosen between 1 and the sum of an aggregate count of samples received and the count of the new samples. It may then be determined at step 606 whether the random number chosen may be less than or equal to the aggregate count of samples received. If the random number chosen may be less than or equal to the aggregate count of samples received, then a count of samples to be taken from the reduction sample set may be updated at step 608 by incrementing the count by one. Otherwise, if the random number chosen may be greater than the aggregate count of samples received, then a count of samples to be taken from the new sample set may be updated at step 610 by incrementing the count by one.
At step 612, it may be determined whether the last random number for the given sample reduction set size has been chosen. If not, then processing may continue at step 604 where a random number may be chosen between 1 and the sum of an aggregate count of samples received and the count of the new samples. Otherwise, if it may be determined that the last random number for the given sample reduction set size has been chosen at step 612, then samples are randomly chosen from the reduction sample set for the count of samples to be taken from the reduction sample set at step 614. At step 616, samples are randomly chosen from the new sample set for the count of samples to be taken from the new sample set. And at step 618, the samples in the reduction sample set may be replaced by the samples randomly chosen from the reduction sample set at step 614 and the samples randomly chosen from the new sample set at step 616.
At step 620, the count of new samples may be added to the aggregate count of samples received. It may be determined at step 622 whether the last sample set has been received for integrating the sets of samples into a reduction sample set. If not, then processing may continue at step 602 where a new sample set and count of samples may be received. Otherwise, if it may be determined that the last sample set has been received, then processing may be finished for integrating the sets of samples into a reduction sample set.
Because online advertising applications may forecast advertisement impression inventory from hundreds of gigabytes of advertisement impression data collected daily, those skilled in the art will appreciate that the steps described in conjunction with
Importantly, the forecast of an inventory of online advertisement impressions may be generated to target many different profiles. For instance, web page attributes such as properties of the page and the web page position of an advertisement may be used. User attributes for online behavior and/or demographics including age, gender, and country, may be used for targeting profiles of attributes. Or profiles of attributes may be targeted by time, browser attribute or type, and so forth. The present invention may make accurate forecast for any combination of thousands of targeting attributes. Thus, the present invention may provide an online publisher with the capability to forecast available inventories of advertisement impressions for targeting different combinations of attributes before selling them to online advertisers who, for instance, would like to target users visiting certain web pages with certain demographics, geographies, behavioral interests, as well as many other attributes. Those skilled in the art will appreciate that the present invention may be used to target any impression attributes including page attributes, user attributes, browser attributes, time attributes, and so forth. For example, a content match application may use the present invention for matching an advertisement to content of a web page. Or an application on a mobile device may use the present invention for matching an advertisement to browser attributes for a browser that may process a particular type of advertisement such as video, text, so forth.
As can be seen from the foregoing detailed description, the present invention provides an improved system and method for forecasting an inventory of online advertisement impressions for targeting user profiles by sampling in a map-reduce framework. Samples of visitor profiles may be collected and matched to targeting profiles for displaying advertisements by distributed sample mappers. Sets of samples of visitor profiles matched to the targeting profiles may be mapped and distributed to reducer servers for integrating the samples collected by sample mappers into a reduction sample set for each targeting profile. Each reducer server may integrate the sets of samples of visitor profiles matched to a targeting profile into a reduction sample set for each targeting profile, and a count of samples of visitor profiles matched to a targeting profile may be aggregated. The sample reducers may forecast a targeting profile inventory for each targeting profile and output the targeting profile inventory with a reduction sample set of visitor profiles. Such a system and method may accurately and efficiently forecast availability of advertisement impressions targeting a user profiles for a particular combination of targeting attributes requested by online advertisers. As a result, the system and method provide significant advantages and benefits needed in contemporary computing and in online applications.
While the invention is susceptible to various modifications and alternative constructions, certain illustrated embodiments thereof are shown in the drawings and have been described above in detail. It should be understood, however, that there is no intention to limit the invention to the specific forms disclosed, but on the contrary, the intention is to cover all modifications, alternative constructions, and equivalents falling within the spirit and scope of the invention.
