Method and a system for delivering messages

Information

  • Patent Grant
  • 8676682
  • Patent Number
    8,676,682
  • Date Filed
    Wednesday, June 11, 2008
    16 years ago
  • Date Issued
    Tuesday, March 18, 2014
    10 years ago
  • Inventors
  • Original Assignees
  • Examiners
    • Hamilton; Lalita M
    Agents
    • Novak Druce Connolly Bove + Quigg LLP
Abstract
A system for generating a schedule for use in triggering transmission of message content in a telecommunications network is provided. The system comprises a store for storing profiles of a plurality of recipients, and a plurality of message delivery campaigns with one or more criteria. The campaigns to be executed in relation to a given inventory, which may relate to a predetermined time period, may be arranged into one or more campaign clusters. A suitable allocation method may be used individually for each of said plurality of campaign clusters to allocate, among the plurality of recipients, target recipients for each of the campaign clusters based on campaign criteria and the profiles. The allocation process may be repeated to obtain, for example as high a number of messages as possible with the network resources available without exceeding a predetermined maximum number of messages per any given recipient. After the allocation process is completed, the campaign messages are delivered to the allocated target recipients via a communications network.
Description
FIELD OF THE INVENTION

The invention relates to communications systems, and particularly to a method and a system for delivering messages via communications systems.


BACKGROUND OF THE INVENTION

Mobile marketing is considered by advertisers as the next new channel to reach direct to the user by utilizing the core assets and characteristics of the mobile media: it being personal, “always on”, mobile and naturally forming groups of people who communicate actively with each other. These characteristics combined with social networks-based marketing approach of the Internet could form a very powerful base to execute marketing strategies.


In general, mobile marketing and advertising can be divided into the following four categories:

    • a) Mobile Marketing: The systematic planning, implementing and control of a mix of business activities intended to bring together buyers and sellers for the mutually advantageous exchange or transfer of products where the primary point of contact with the recipient is via their mobile device.
    • b) Mobile Advertising: The paid, public, non-personal announcement of a persuasive message by an identified sponsor; the non-personal presentation or promotion by a firm of its products to its existing and potential customers where such communication is delivered to a mobile phone or other mobile device. Examples of mobile advertising would include: Wireless Application Protocol (WAP) Banner ads, Web page Banner ads, mobile search advertising, mobile video bumpers, and interstitial ads in on device portals.
    • c) Mobile Direct Marketing: Sales and promotion technique in which the promotional materials are delivered individually to potential customers via the potential customer's mobile phone or other mobile device. Examples of mobile direct marketing include the sending of Short Message Service (SMS), Multimedia Message Service (MMS) or WAP push messages, Bluetooth messaging, personalized WAP and Web page banner and other advertisements and other interrupt based marketing to mobile phones or other mobile devices.
    • d) Mobile customer relation management (CRM): Combination of all the above in a manner that establishes a long-term, engaging relationship between the customer and the promoting company.


Today's mobile marketing is usually mostly based on push campaigns to opt-in recipient mobile number database, or pull campaigns that acquire mobile phone numbers from recipients. A popular method is direct advertisement done using text and picture messaging. As the mobile device is conceived very personal, it cannot be filled with junk advertisement, as people stop receiving advertisements, either by unsubscribing from the advertisement service or by blocking the advertisements with some technical mechanism. Thus, the receivers have either to opt-in to certain campaigns, or provide enough personal details for targeting the advertisements more precisely than in the traditional media. Using personal details makes the advertisements more compelling and more effective to the advertisement receiver, and may thus enable a higher number of advertisements to be sent to a single receiver without causing the receiver to consider the marketing information received via a mobile channel as a burden. Also the utilization of the capacity of the network is improved. On the other hand, the more individually the delivery of the marketing messages is targeted in the network, the more processing capacity is needed for selecting targets and scheduling the delivery and/or reserving network resources. The advertisement reservation and allocation may involve dozens or hundreds of parameters that can be freely selected on an advertisement basis, which may easily lead to a reservation method that cannot effectively reserve the advertisement receivers and achieve a full booking in the utilization of mobile network resources point of view. The reservation is also time-constrained challenge, as there is only a short period of time (e.g. few hours) available in a day to run the reservation, which imposes further requirements for the processing/computing capacity. The number of advertisement receivers (i.e. recipients) may be extremely high (e.g. hundreds of thousands or millions of recipients). If a reservation method using a limited processing capacity is optimized to provide a high accuracy and guarantee a maximum number of messages sent in a network per day, the processing is easily be too slow for such a high number of advertisement receivers and the time available for processing. On the other hand, if a reservation method using a limited processing capacity is optimized for the speed of processing, a high number of advertisement receivers can be processed with the time available for processing, but the accuracy is reduced, i.e. the customers may receive wrong type of or too many messages and/or the network resources may not be used efficiently in terms of a number of messages sent in a network per day.


SUMMARY OF THE INVENTION

In accordance with at least one embodiment of the invention, methods, systems and software are provided for supporting or implementing functionality to provide an output, as specified in the independent claims. This is achieved by a combination of features recited in each independent claim. Accordingly, dependent claims prescribe further detailed implementations of the present invention.


According to an aspect of the invention, there is provided a scheduling method for generating a schedule for use in triggering transmission of message content in a telecommunications network according to an available inventory of messaging events, the method comprising:


i) maintaining group data identifying a plurality of message groups, each of said groups being associated with one or more predefined criteria;


ii) maintaining profile information relating to potential message recipients;


iii) using a first allocation algorithm to define a first allocation of messages relating to one or more of the groups based on said one or more predefined criteria and the profile information;


iv) using a second allocation algorithm, different to the first allocation algorithm, to define a second allocation of messages relating to the one or more of the groups based on said one or more predefined criteria and the profile information;


v) selecting, on the basis of one or more selection criteria, one of said first allocation and said second allocation; and


vi) inserting scheduling data into said schedule on the basis of the selected allocation,


wherein the selected allocation defines message recipients and respective message content to be sent to the defined recipients.


Accordingly, in embodiments of the present invention, multiple algorithms are used to define inventory resource allocations, allowing a preferred allocation to be selected and a schedule thereby generated. This enables resources to be allocated more efficiently than in systems using only a single allocation algorithm. In some embodiments, the criteria used for this selection may be varied according to a scheduling condition, for example, an amount of inventory resource available and/or an amount of message content to be scheduled; this allows a flexible system able to adapt to varying conditions and thereby able to allocate resources efficiently.


In some embodiments, steps iii) to vi) are performed with respect to each of a first plurality of message groups and a second plurality of message groups, with different algorithms being used for the second plurality of message groups to the algorithms used with respect to the first plurality of message groups. This allows algorithms to be varied according to a message group, improving the flexibility and efficiency of the method.





BRIEF DESCRIPTION OF THE DRAWINGS

In the following the invention will be described in greater detail by means of exemplary embodiments with reference to the attached drawings, in which



FIG. 1 is a block diagram illustrating a high-level architecture of an exemplary telecommunications system wherein the concepts of the present invention may be implemented;



FIG. 2 is a block diagram illustrating an example of an advertisement management system in accordance with an aspect of the invention,



FIG. 3 is a flow diagram illustrating an example of a reservation process that uses a single reservation algorithm,



FIG. 4 is a flow diagram illustrating an example of planning and delivery of campaigns for several days, and



FIGS. 5 and 6 are a flow diagram illustrating an example of a reservation process that uses multiple allocation algorithms in parallel.





In the drawings, the same references symbols refer to similar structures or functions.


DETAILED DESCRIPTION OF THE INVENTION

According to an exemplary concept of the invention a mobile communications service provider to provide free or subsidized mobile communications services to users or recipients in return for receiving messages from a sending party, typically as part of a messaging campaign of the sending party. The sending party could be a company wishing to bring their goods and/or services to the attention of the recipient, to inform them of the URL of their homepage, to request feedback on or assess interest in goods and services using, for example, multiple choice questions, or to provide general information. Messages sent as part of a messaging campaign are herein referred to as “campaign messages”. The messaging campaigns can be viewed as each defining a group of messages; each of the groups may be associated with predefined criteria, as described below.


The mobile communications service provider or operator is referred to as a Commercial Virtual Operator (CVO) and the service concept is referred to as a Commercial Virtual Operator Network (CVON). The CVO is a virtual operator in a sense that while the users experience the CVO as a normal network operator, the CVO may not necessarily have any mobile communications network of its own but it may lease or hire part or all of the required mobile network resources from a mobile communications network of another network operator for business purposes of building a virtual operator which offers free or subsidized mobile communications services. In order to obtain free or subsidized communications services the recipient may be required to receive a predetermined number of campaign messages per a time period, respond to certain messages, etc. For example, a recipient may be expected to receive 4 short messages (SMS) or multimedia messages (MMS) per day 7 days per week. An example of mobile communications services offered in return to recipient may be 100 minutes of voice service and 200 text messages and some data services. Technically the mobile communication services may be provided and the campaign messages delivered via network operator's network to recipients. The free or subsidized mobile communications services offered as a reward to the users or recipients may be provided before, during and/or after said target users receive and/or react to the campaign messages. The reward may be provided to the user in a single action or in a plurality of actions. Rewards given before receiving/reacting to campaign messages may be different from rewards given after receiving/reacting to campaign messages. A first reward may be a reward for joining/completing a subscription agreement. A reward given during receiving/reacting to campaign messages may a free or subsidized communication or service relating to an advertisement, for example. The CVON may cancel (bill/charge afterwards) free or subsidized mobile communications services given beforehand if a user fails to meet the terms and conditions, e.g. receive or react to a predetermined number of campaign messages. Embodiments of the invention may be also used by non-virtual operators when they offer fully or partly subsidized services for their subscribers.


Sending parties may make agreements directly or via an agent with the CVO on delivery of campaign messages to target recipients.


Sending parties may pay for the services of message delivery to the CVO. Payment fees may include delivery cost of the actual message, profiling cost of the messages, monthly payments, profit margin of the CVO, etc. There may be a set of different types of messages offered by the CVON to third parties. These may include but are not limited to multimedia message service (MMS) based messages, short message service (SMS) based messages, message tagging, wireless application protocol (WAP) push based messages, banner advertisements, search related messages, and interactive messages. The above list describes only at a high level the functionality of exemplary messaging possibilities that may be offered with the CVON concept to third parties. There may also be, for example, video, audio, etc. type of messages provided by the CVON. Typically more than one message type is offered to third parties in order to offer a comprehensive enough portfolio of tools for the sending party to create innovative messaging campaigns.



FIG. 1 is a block diagram illustrating a high-level architecture of an exemplary telecommunications system wherein the concept according to FIG. 4 may be implemented. Reference numeral 6 denotes a mobile terminal. The mobile terminal 6 may be a mobile phone, a personal digital assistant (PDA), a multimedia computer, a personal computer, a lap top, etc., or generally any terminal capable of access in services, such as content download, web browsing, streaming, Wireless Application Protocol (WAP) browsing, voice and messaging.


The telecommunications system of FIG. 1 comprises one sending party 1 but there are typically many such sending parties. The sending party 1 may reserve, program, and/or book a campaign via a web interface of a message management system 2, for example. The sending party 1 may provide message content in the form of, for example, data, text, pictures, audio, video, links, Hypertext Markup Language (HTML), Extensible Markup Language XML, and/or Extensible Hypertext Markup Language (XHTML) to the message management system 2. The sending party 1 may also provide parameters for the messaging campaign or advertisement, such as: when to send messages, to which target group, on which format, a target price level of the advertisement, a target feedback level of the advertisement, demographics of the target audience, a duration of the advertisement campaign.


The message management system 2 manages messaging campaigns and may comprise a message database that may contain all relevant information for the messaging campaign to be run. The message management system 2 may also have tools for the sending party 1 to define a campaign so that all or some of the campaign parameters and the rules may be set, cancelled, modified, updated, or otherwise processed by the sending party 1. The message management system 2 may be used to keep and maintain rules of the sponsoring of access to web services for the users.


It should be appreciated that, as used herein, a message delivery refers to any type of triggering of a message being sent based on the results of the optimized reservation according to the invention. For example, the delivery may be direct sending of the message, programming or configuring any network element to perform the message sending, providing commands to a message delivery system of another operator to perform the message sending, etc.


In the illustrating example shown in FIG. 1, the message is delivered to users 6A and/or 6B via communication network 4. The communication network 4 can be any cellular, broadcast, wide area, local area network or the Internet. Examples of cellular network technologies include but are not limited to GSM (Global System for Mobile communication), WCDMA (Wideband CDMA), CDMA (Code Division Multiple Access), GPRS (General Packet Radio Service), UTRAN (UMTS Radio Access Network), UMTS (Universal Mobile Telecommunications System), MBMS (Multicast Broadcast Multimedia System). Examples of other network technologies include but are not limited to local area networks, such as Wireless Local area networks (WLAN), BlueTooth (BT), and other technologies, such as WiMax (Worldwide Interoperability for Microwave Access), Broadcasting over cellular, Broadcasting over DVB-H (Digital Video Broadcasting—Handhelds), ISDB-T (Terrestrial Integrated Services Digital Broadcasting), DMB (Digital Multimedia Broadcasting). The communication network 4 may also be provided by any generic Internet access using any transport methods. The communication network 4 can be also a combination of different communication network technologies.


In the example of FIG. 1, the communication network 4 comprises a cellular network and is shown with exemplary network elements, such as base stations (BSs) 43 and 44, a Short Message Service Center (SMSC) 41 and a Multimedia Message Service Center (MMSC) 42.


There may also be provided a value added service gateway (VAS GW) 3 that connects the communication network 4 or some of the elements thereof to the message management system 2. The VAS gateway 3 may also be connected to a billing system 5. The VAS gateway 3 may include a message delivery component for sending campaign messages to the customers through the communications network 4. The VAS gateway 3 may also include a database and an action-tracking component for monitoring user actions regarding the messages.


A billing system 5 represents any real-time billing system or close-to-real-time billing system that may be employed for monitoring the usage of the communication services in the communications network 4. Services, i.e. communication events, may include but are not limited to voice and/or messaging services (Short Message Service, Multimedia Message Service, Instant Message Service, Electronic mail services), video telephony services, push to talk services, data services such as Internet or Wireless Application Protocol (WAP) browsing services, content usage (television, radio, video) services, download services, software (including games, applications, tools and utilities), Flash or Quicktime animations/shows, SIM toolkit messages. The billing system 5 may also refer to any 3rd party offered service running in a server or a computer system, such as a proxy server or a web server that offers services to mobile users. The billing system 5 may receive charging records from other network elements, each charging record comprising all or some of the information required for the billing of a given communication by a user in the communications network, possibly excluding price information. A charging record may specify the content and format of the file that is delivered to the billing system 5. Charging records are often referred to as call detail records (CDRs) or charging data records, or service detail records (SDRs) in value added services. The charging records may include not only the user's calling or originating number/address or similar identity but also a destination of the communication, for example, a called telephone number of a communication, an Uniform Resource Locator (URL) or similar network address accessed via the communication network 4, a telephone number or network address number to which a message is sent, etc. The information can be delivered to the billing system 5 from any communication network element handling a given communication, such as via a short message service center (SMSC) handling SMS messages of a user, a multimedia message service center (MMSC) handling MMS messages of a user, Wireless application protocol gateway (WAP-GW) handling a WAP communications of a user, and an Internet access point (Internet AP), a serving GPRS support node (SGSN), a gateway GPRS support node (GGSN), etc.


In an embodiment of the invention, the billing system 5 may be used to meter usage of the services. The metered usage of the services may be compared with a free/subsidized balance allocated to each recipient in the communication network. In embodiments of the invention, typically no invoices are sent to recipients but the metered usage may be compared with the business rules associated with customer and metered cost may be invoiced directly or indirectly from advertisers, such the advertiser 1.


The type and format of a campaign message as well as the delivery method used may be selected among those available in the communication network 4 employed. Such message formats and delivery methods may include but are not limited to messaging services, such as short message service (SMS), multimedia message service (MMS), Instant Message Service (IMS), electronic mail, file downloading or browsing services, such as Wireless application protocol (WAP), Word Wide Web (WWW), audio streaming, video streaming or other data services, etc.


The message management system 2 manages messaging campaigns and its databases may contain all relevant information for the messaging campaign to be run. The messaging system 2 may comprise one or more databases, computer programs and/or related hardware that can be used to offer a tool and an interface, such as a web interface, for the sending parties to access the messaging management system 2 to reserve, program and define an advertisement campaign. An example of the message management system 2 will be described below with reference to FIG. 2.


The sending party 1 may access the message management system 2 via a web interface using personal computers with a web browser, or via any similar interface 201 suitable for the access. The sending party 1 may program a campaign including the preferred profiles, pricing, timing, advertisement content, desired impact levels, feedback requests, etc. to the message management system 2, or to look, modify, add, extract reports, etc. The message management system 2 may include a database or databases 202 for storing the campaigns and related content and other information. The recipients (users) of the mobile services offered by the CVON may be referred as members, subscribers, users, or recipients. The database 202 may also comprise a recipient database containing required information on the members, such as profile information of users or members. Typically the information may be collected prior to allowing the recipient to join free mobile service. The information may be updated later as the preferences of the member change and/or there is need to collect more information on the members. The information may be updated also automatically as the member or peer group of the member interacts with the messaging system directly or indirectly. The sending party 1 may use the profile information as they define a target audience for the messaging campaigns using these profiles. The member information may be presented to sending parties in such a format that no direct association can be made between any specific recipient and his/her profile, i.e. the anonymity of the recipients is conserved.


In the profile information, at least one of the profiles may comprise one or more of the following pieces of information: a sociological background of a user; age; sex; a target telephone type; an income level; status of a user; a location of a user; historical data of a user's behavior; information of sent direct messages to a user; information of content vouchers or coupons sent to a user; codes of vouchers or coupons sent to a user; lifestyle and interest related; behavior; demographics; education; marital status; zip code; preferred times for messages. Moreover, the database 202 may store campaign parameters, such as campaign messages, a user profile; preferences of when to send messages; to which target group an advertisement is sent; on which format a message is sent; a target price level of a message; a target feedback level of the message; a target audience; demographics of a target audience; a duration of a messaging campaign, cost per message; type of message; a sociological background of a target audience; age; sex; a target telephone type; an income level; status of a user; a location of a user; historical data of a user's behavior; historical data on a behavior of a profile of users; information of sent direct messages to a user; information of content vouchers or coupons sent to a user; codes of vouchers or coupons sent to a user; a time of a day or a week or a month or a date, alpha and omega scores, other calculated values from user behavior or user mass analysis, etc. The message management system 2 may also be used to maintain rules of the sponsoring communications and/or service access.


The message management system 2 may also contain an optimization unit 203 that may perform the campaign delivery optimization, analyze the effect of the optimization, store learning data of the campaign optimization and define the actual delivery rules for delivery (210) of the messages to the recipients. In addition, there may be a receive and feedback unit 204 in the message management system 2. The receive and feedback unit 204 may collect direct and/or indirect feedback and responses to the messages and feed the related information to the optimization unit 203 and/or to the database 202 and/or to the sending party 1. The direct feedback refers to a recipient (MT) 205 responding (207) to a direct question, a query etc. The indirect feedback (208) refers to, for example, the purchase of goods or the usage of a coupon, and the indirect feedback may be obtained, for example, from a purchasing system 209.


The message management system 2 uses message requests received from the sending party 1 to fill an inventory for sending message content. The inventory may comprise an amount of network resource allocated for sending messages during a predetermined time period (a “time window”). The inventory may comprise a plurality of slots for sending messages of a fixed or variable size. The message management system 2 uses allocation algorithms to allocate resources of the inventory for delivering messages in accordance with embodiments of the present invention.


The allocation process will now be explained by way of example, and in the context of allocating message recipients. In a message management system 2 there may be dozens or hundreds of criteria which may be used to define a campaign. In order to explain the invention we use as an example a case with 3 criteria (on/off criteria) A, B, and C, a total of 5 recipients 1, 2, 3, 4, 5, and a plurality of different campaigns c1, c2, c3, . . . cN. For example, a criterion could have a value of ON if a recipient is aged above 22, and OFF otherwise. An exemplary user profile database is assumed be as shown in Table 1.









TABLE 1







A profile database











A
B
C
















1
ON
ON
ON



2
ON
OFF
OFF



3
ON
ON
OFF



4
OFF
ON
ON



5
ON
OFF
OFF










Example 1

In a first example of selection algorithm there are campaigns C1, C2, and C3 Where


In campaign C1 it is defined that 2 messages are to be sent to recipients having a profile A.


In campaign C2 it is defined that 2 messages are to be sent to recipients having a profile B.


In campaign C3 it is defined that 2 messages are to be sent to recipients having a profile C


Allocation algorithm 1 is assumed to use a recipient number based allocation and order of campaigns as 1, 2, and 3. Only one message per recipient is allowed. Allocation algorithm 1 gives rise to the following delivery of the campaigns C1, C2, C3. Messages of the campaign C1 are allocated to be sent to the recipients 1 and 2. Messages of the campaign C2 are allocated to be sent to the recipients 3 and 4. No messages of the campaign C3 are allocated to be sent to the recipients, since there are no recipient with a profile C left: a message has already been allocated to be sent to recipient 1 and 3 in campaigns C1 and C2, and only one message per recipient is allowed. The only recipient 5 who has not yet received any messages does not have profile C.


Allocation algorithm 2 uses a different sending order, the order of campaigns being 3, 2, and 1, for example. Only one message per recipient is allowed. This results in a different allocation of inventory resources of the recipient to the campaigns. More specifically, one message of the campaign C3 is allocated to be sent to each of the recipient 1 and 4, one message of the campaign C2 is allocated to be sent to the recipient 3 and one message of the campaign C1 is allocated to be sent to each of the recipients 2 and 5. Thus, the allocation algorithm 2 results in sending a message to each one of the recipients 1 to 5, and all campaigns are carried out, even though one campaign only partially. Therefore, the allocation algorithm 2 leads to a better utilization of resources (a total of 5 messages) than the previous algorithm 1 (a total of 4 messages), which demonstrates that the selection of an allocation algorithm has an impact on the results.


Allocation algorithm 3 is similar to the allocation algorithm 1 but allows up to 3 messages per recipient. Again, this results in a different allocation of inventory resources. More specifically, the messages of the campaign C1 are allocated to be sent to recipients 1 and 2. The messages of the campaign C2 are allocated to be sent to the recipients 1 and 3. The messages of the campaign C3 are allocated to be sent to recipients 1 and 4. Thus, the allocation algorithm 3 results in the best utilization result (6 messages) but also in a greater load of messages to the recipient 1 and no messages to the recipient 5.


Allocation algorithm 4 selects the recipient having the least number of messages sent per period of time but not having restrictions relating to a maximum number of messages per recipient. Optimization rounds would result in the following reservation of the recipients to the campaigns. The messages of the campaign C1 are allocated to be sent to the recipients 2 and 5. The messages of the campaign C2 are allocated to be sent to the recipients 1 and 3. The messages of the campaign C3 are allocated to be sent to the recipients 1 and 4. Thus, the allocation algorithm 4 leads to a full utilization and an even distribution.


In some embodiments, the algorithms are based on fuzzy logic techniques; instead of the criteria having discrete values (for example an age of 22 in the above example), a set of discrete values (e.g 21, 22, 23) could be set and probabilities assigned to selecting recipients in dependence on their ages.


As a summary, the exemplary algorithms presented above show that different algorithms for allocating resources with respect to the same recipients and campaigns give different results. One algorithm may have an advantage of being simple and fast to be executed with a computer program, for example, while another algorithm may lead to inventory resources being allocated for all messages of all campaigns to be sent, and some other algorithms could be targeted to provide an even distribution of the messages among users. However, the challenge is that, typically in practice, no single algorithm can meet all requirements, for example there may be conflicting issues between a full utilization of resources (as high a total number of sent messages as possible) and recipient satisfaction (reasonable number of received messages, high relevance of the messages), profit optimization versus serving all sending parties etc.



FIG. 3 illustrates a prior art allocation process using a single allocation algorithm. Campaigns 310 are fed 312 in the message management system 2, the allocation algorithm 314 is used and the campaigns are delivered 316 to the allocated users. The problem has further dimensions, such as the time aspect illustrated in FIG. 4. The allocations can be typically performed up to several days or weeks (the planning phase in FIG. 3) before the actual delivery, which may be performed over several days. After the initial allocation there might be new campaigns defined by the same or another sending party, some of the campaigns might be cancelled etc., meaning that the original allocation may not be accurate at all (e.g. due to cancelled or added campaigns) at the time of actual delivery. As already discussed above, this and other factors lead to an ineffective utilization and/or to a slow allocation process. Thus, it is important to have an efficient reservation process at the time of the delivery.


In some embodiments of the present invention, a plurality (two or more) allocation algorithms are used with respect to the same campaign or cluster of campaigns, in parallel or in sequence, in order to maximize the usage of the inventory (i.e. the number of advertisement receivers (i.e. users) multiplied by the message slots per given day) in the restricted amount of a inventory resources available during a given day. In some cases, the message management system 2 may also learn from its own behavior and effectiveness, and it may apply this learned knowledge to subsequent allocation processes to maximize the amount of users covered and to minimize the time spent in the allocation process. This learning will be described in more detail below.


Clustering and multiple allocation algorithms allow both a more accurate and faster allocation process with the limited processing capacity and the limited processing time available in the message management system 2; clustering will be described in more detail below. The clustering and multiple allocation algorithms allow use of algorithms optimal to each specific cluster to enhance accuracy, while at same the time parallel processing is possible for reducing the processing time and/or making several iterations. Due to the enhanced accuracy, the message management system 2 can more effectively allocate inventory resources and improve utilization of mobile network resources and improve the service to the users, i.e. the efficiency is improved. The allocation is also time-constrained, as there is only a short period of time (e.g. few hours) available in a day to perform the allocation, which imposes further requirements for the processing/computing capacity.


An exemplary process utilizing multiple allocation algorithms in accordance with an embodiment of the invention is described below with reference to FIGS. 5 and 6. The message management system 2 contains user information, such as the user profiles in a database (step 602), as described in the exemplary embodiments above. A plurality of message delivery campaigns C1, C2, . . . , CN are created or defined with a plurality of parameters in the message management system 2 (step 604), for instance in a similar manner as described above. The sending parties 1 may access the message management system 2 via a web interface using personal computers with a web browser, or via any similar interface 201 suitable for the access to reserve, program and define messaging campaigns. The parameters may include, for example, the preferred user profiles, pricing, timing, advertisement content, desired impact levels, feedback requests, etc. Further examples of suitable parameters, such as the parameters as used in the user profiles, have been given above in the description of the advertisement management system 2, without restricting the invention to these examples. The parameters form the basis for campaign criteria for each of the campaigns C1, C2, . . . , CN. The campaign criteria are used in determining recipients for messages of the campaign and relate to profile information of target recipients, such as an age range, a location, a specified interest or preference, gender, and so on. Each campaign is individually defined to be carried out during one or more message delivery time windows or slots, e.g. during one or more days. The campaigns may be stored in the database 202.


At the time of the delivery, the message management system 2, e.g. the optimisation unit 203, may study the campaigns to be carried out in the present message delivery time window, e.g. today or during following 24 hours, and create campaign lists based on the criteria used in the campaigns. The campaigns may be clustered into a plurality of campaign clusters 502, 504 and 506 according to at least one similarity or commonality between the campaign criteria (step 606), such as one or more of a priority of an advertiser, a number of target users, strictness of campaign criteria, and requirements set for user equipment. Some of the clusters may also contain only one campaign. The order of campaigns C1, C2, . . . , CN may be optimized, albeit loosely, based on the clusters 502, 504 and 506 (step 608). It may be defined, for example that campaigns of cluster 506 are first in order, campaigns of cluster 502 are second in order, and campaigns of cluster 504 are third in order. Naturally, the campaigns may be ordered also within each cluster based on the algorithm(s) used for the specific cluster. Examples of clustering, and how it affects the allocation process, are given below.


The allocation process may be executed by selecting allocation algorithms A1, A2, . . . , AN for the clusters 502, 504 and 506 or for individual campaigns C1, C2, . . . , CN (step 610). The suitable allocation algorithms may be selected among a plurality of allocation algorithms available in the message management system 2. Such algorithms may include the algorithms 1 to 4 described above, for example, but any type of algorithm may be used depending on the desired results. As the campaigns in each specific cluster 502, 504 and 506 have similar criteria, at least regarding the criteria used for the clustering, same algorithm or algorithms can be used for all campaigns within a cluster. However, the algorithm or algorithms may typically differ from one cluster to another. A campaign that cannot be clustered with any other campaign (a one-campaign cluster) may also have a different allocation algorithm. The selected algorithms A1 . . . A4 are executed for the respective clusters 502, 504 and 506 or individual campaigns at step 612. For example, the allocation algorithms may aim to allocate inventory resources among all users having profiles in the database, target users for each of said plurality of campaigns based on the campaign criteria and the user profiles, such that as high number of messages as possible with the network resources and the user base available within the given message delivery time window can be delivered without exceeding a predetermined maximum number of message per any given user. Thereby, a first set of allocations for the campaigns are obtained.


Because the allocation is also a time-constrained process, the process may check whether the time available for the allocation process has expired (step 614). If the time has expired, the allocation process may be completed and the process may proceed to step 620. If the time has not expired, the allocation results may be analyzed (step 616), and if it appears that the allocation results may still be improved (step 618), i.e. more messages can be allocated within the delivery schedule and/or reach a broader recipient list etc, as determined by the selection criteria particular to this allocation run, the process may return to step 610 wherein one or more of the selected allocation algorithms may be changed to improve the allocation results. In some cases, the process may return to step 606 wherein the clustering is changed in order to make even more radical changes to the allocation results. One of the allocation results to be monitored may be that all campaigns are adequately allocated. Another allocation result to be monitored may be conflicts between the allocations made by different algorithms. Further, in case of conflicts between the allocation results, also special optimization algorithms, such as peep-hole and rescheduling, may be run, as is described below. The optimization or changes may be made until the reservation results cannot be further improved.


As mentioned above, the allocation results are analyzed at step 616 and a determination as to whether the allocation results can be further improved is made at step 618. Some examples of optimization criteria that can be used for making this determination are now given. One criterion is whether the optimization round could assign more messages for a period of time to which the inventory corresponds. This is especially important when the number/size of message requests received from the sending party 1 is too great for all messages to be sent within the available inventory resource.


Some optimization criteria may relate to a value or values associated with messages and recipients thereof. The operator of the message management system 2 may charge different prices for different messaging campaigns, or messages sent to different users; in some circumstances, it may be desirable to optimize the selection criteria such that the profit for the operator is maximized. Alternatively, or additionally, in some circumstances the sending party may wish to prioritize certain messages/and or message campaigns over others; this could be achieved by assigning different values to the messages/campaigns, and applying an algorithm such that a value for the sending party 1 is optimized.


Further, some optimization criteria may relate to a distribution of messages amongst users i.e. ideally every message recipient should get roughly the same amount of messages and they should also be distributed somewhat evenly in the time or according to delivery preferences specified e.g. in their user profile.


Different optimization criteria may lead to different result sets, and depending on the situation, different allocations are desirable. For example, if there is a big backlog of as yet unsent messages, it may be desirable to weight these messages at the beginning of the allocation process differently to how they are weighted towards the end of the allocation process, to ensure that a relatively greater number of these types of unsent messages are entered into the delivery schedule. Alternatively, or additionally, if the number of message requests is such that there are insufficient inventory resources to allocate to all messages, the messages may be selected based on a message value representing an importance of the message to the sending party 1; this value may be assigned by the sending party 1.


However, results obtained using these criteria may not make full use of the available inventory resources i.e. there may be some available inventory resources that are not assigned to any users. This can be mitigated by using peep-hole optimization, in which the user to which a high-value message, for example, is assigned is changed so that a previously unassigned message can be assigned to the user to which the high-value message was originally assigned. This may be continued until no message candidates for message reassignment can be found, or until messages or users run out.


Returning to FIG. 6, at step 619 the system selects which of the allocations obtained as a result of applying different allocation algorithms is to be used for generating a schedule for sending the messages. This selection is performed based on one or more selection criteria. The selection criteria are typically the same as or similar to the optimization criteria described above.


In step 620, a schedule for sending the message is generated, and the messages are delivered at step 621. The messages of the campaigns C1, C2, . . . , CN to are delivered the reserved target users via the communications network 4 in the determined order during a given message delivery time window, such as a day. Also other delivery rules may naturally be followed in the delivery.


Moreover, at step 622, the created clusters, used algorithms and obtained allocation results, i.e. the allocation strategy used, may be stored in a learning/knowledge database 508 in order for the system to learn optimization strategies. Since there may be a limited time period, and limited processing capacity, available for allocating inventory resources, it is advantageous for the system to be able to learn allocating strategies for efficient allocation of resources within these limitations. The message management system 2 may use the stored allocation strategy for using a similar approach for similar cases, i.e. the system is able to learn 510 from previous allocations. This may further reduce the number of optimization rounds in the allocation process and/or allow time to optimize the allocation results further in subsequent allocations.


Examples of “clustering” according to some embodiments of the present invention are now described. Campaigns are clustered according to which clusters produce the best results according to optimization criteria as described above. Accordingly, the clustering criteria by which campaigns are organized into clusters may vary according to the situation; the criteria may vary according to how optimization criteria are weighted, for example. Furthermore, the system may employ a learning process, as described below; consequentially, the clustering criteria may vary with time.


In some embodiments, clustering is used in conjunction with a partitioning of the database 202 with respect to potential recipients; that is, recipients may be categorized, and one or more of the resulting categories allotted to one or more of the clusters, as is described in the following example.


For the purposes of this example, it is assumed that each recipient included in the database is associated with four criteria A, B, C and D, each of which can have a value of ON or OFF. In this example, each the recipients is categorised into one of 8 categories, a-h, as shown in table 2:














TABLE 2







Category
A
B
C









A
ON
ON
ON



B
ON
ON
OFF



C
ON
OFF
ON



D
ON
OFF
OFF



E
OFF
ON
ON



F
OFF
ON
OFF



G
OFF
OFF
ON



H
OFF
OFF
OFF










In this example, the recipients are categorized according to values of criteria A, B and C; no categorization is made according to criterion D.


Further, in this example, each campaign has a preference of ON or OFF for each of the criteria A, B, C and D, or no preference may be specified for some criteria. For example, a particular campaign may specify that recipients should have an associated criterion value of ON for criterion A, a criterion value of OFF for criterion B, with no preference for criteria C (and D); potential recipients of this campaign would have associated criterion values of ON for criterion A and OFF for criterion B, the values of criterion C and criterion D being irrelevant for this campaign.


In the present example, the campaigns are clustered according to similarity with respect to criteria A, B and C; this could result in, for example, the six campaign clusters, i-vi, each of which corresponds to a particular set of criteria, as shown in table 3:














TABLE 3







Cluster
A
B
C









i
OFF
ON




ii
OFF

ON



iii

ON
ON



iv
ON



v
OFF
OFF
OFF



vi


ON










In this example, criterion D is ignored for the purposes of clustering. Although in this example clustering of campaigns is performed with respect to the same criteria with respect to which the database 202 is partitioned, in some embodiments different criteria are used.


For each cluster, categories of recipients can be identified from which potential recipients can be drawn. Table 4 shows which of categories a-f contain recipients which are potential recipients of messages associated with each of the campaign clusters i-vi.















TABLE 4











Potential



Cluster
A
B
C
recipient categories









i
OFF
ON

e, f



ii
OFF

ON
e, g



iii

ON

a, b, e, f



iv
ON


a, b, c, d



v
OFF
OFF
OFF
h



vi


ON
a, c, e, g










Thus in this example, recipients in category h can only receive messages from campaigns in cluster v; recipients in category h are also the only potential recipients of messages of campaigns in cluster v. It is therefore advantageous to assign category h to cluster v only; that is, processing for the campaigns of cluster v is only performed on recipients in category h. Since processing is performed with respect to fewer recipients, this results in a faster processing time and/or requires fewer resources than in systems in which no clustering is performed. Since no potential recipients are excluded, this is done without any change in the number or distribution of messages allocated.


Other similar assignations may also be advantageous. For example, category e could be assigned to cluster i only. This also results in faster processing time and/or lower usage of resources; however, since, for example, category ii now only has potential recipients in category g, it may result in fewer messages being allocated overall. Whether or not this partitioning is effective therefore depends on how speed of processing is weighted as compared to number of messages allocated, etc.


Another aspect with respect to which clustering is advantageous over systems which do not use clustering, is in prioritization of campaigns. Each campaign may specify different levels of strictness with respect to completion within a given “window” i.e. predetermined time period. For example, this could take the form of a further criterion, specifically criterion E, which also can take a value of ON or OFF; campaigns specifying value ON must be completed within the predetermined time period e.g. a day, whereas those with a value of OFF can be completed later than this. If clustering is performed according to this criterion, in addition to criteria A, B and C, a total of 12 clusters may result, as shown in table 5:














TABLE 5





Cluster
A
B
C
E
Potential recipient categories







i
ON
ON

ON
e, f


ii
ON

ON
ON
e, g


iii

ON

ON
a, b, e, f


iv
ON


ON
a, b, c, d


v
OFF
OFF
OFF
ON
h


vi


ON
ON
a, c, e, g


vii
OFF
ON

OFF
e, f


viii
OFF

ON
OFF
e, g


ix

ON

OFF
a, b, e, f


x
ON


OFF
a, b, c, d


xi
OFF
OFF
OFF
OFF
h


xii


ON
OFF
a, c, e, g









Processing for clusters having a value of ON for criterion E (clusters i to vi) may be completed before processing for clusters having a value of OFF for criterion E (clusters vii to xii) is started, ensuring that processing for the former clusters is completed within the predetermined time period. A number of consequences arise from this: firstly, processing for vii to xii might not be completed during the allocated time period-so messages for these campaigns might not be sent during the predetermined time period. Secondly, assuming that the algorithms used specify a maximum number of recipients, some recipients will reach this maximum with allocations from clusters i to vi, meaning that there may be a smaller number of potential recipients for vii-xii, which may result in a fewer total number of messages being sent. Thirdly, whether or not the algorithms specify a maximum number of recipients, it is more difficult to spread the messages evenly. Thus, whether clustering in this way is advantageous may depend on how optimization criteria are weighted.


In some embodiments, campaigns within campaign clusters are allocated to different recipients; this may result in an even distribution of messages amongst recipients.


Although in the above examples campaigns are arranged into clusters according to a similarity or commonality between the campaigns, in some embodiments campaigns may be clustered according to different criteria. For example, in some cases clustering may be according to a dissimilarity, or lack of commonality, between campaigns; this may allow processing to be completed with fewer optimization rounds, since there will be less overlap between potential recipients of campaigns clustered in this way. Further, although criteria having a finite number of possible values were used in the above examples, in some embodiments fuzzy logic may be used. In some embodiments, the system learns from its own behaviour to improve the criteria used for clustering. An initial set of clustering criteria may be determined by a simulation; this may involve creating a database having similar attributes to the actual recipient database, creating a set of trial campaigns, and running these using a limited combination of criteria. The combination that generates the most favourable results is then selected as an initial combination of criteria.


The learning process may be implemented as follows. The messaging system first quantifies the start situation, for example, whether there an excessive number of message requests, a large backlog of previous requests etc, and evaluates an initial allocation. Next, the system tries a given optimization technique, by for example selecting allocation algorithms and/or optimization algorithms as described above, and stores the results compared to the initial allocation. The results are evaluated with respect to a metric or metrics appropriate to the situation e.g. utilization of inventory resources, delivery of message backlog, revenue, etc. that quantify the improvement. If improvements are satisfactory (for example, if defined thresholds with respect to the optimization criteria are reached), a further optimization round is commenced. If the results are not satisfactory, another optimization technique is tried, the results stored, compared with one another and so forth.


Based on these quantifications, the system learns the effectiveness of optimization techniques for given initial situations and can start to try the optimization techniques in a more efficient order. If it is found that one optimization technique tends to be superior in a given situation, it may always be tried first. If it is found that another optimization technique produces inferior results in a given situation, it may never be used in that situation.


Similarly, the system may try various clustering techniques in a given situation and monitor the results, enabling it to ascertain which clustering techniques are effective in the given situation.


The techniques described herein may be implemented by various means. For example, these techniques may be implemented in hardware (one or more devices), firmware (one or more devices), software (one or more modules), or combinations thereof. For a firmware or software, implementation can be through modules (e.g., procedures, functions, and so on) that perform the functions described herein. The software codes may be stored in any suitable, processor/computer-readable data storage medium(s) or memory unit(s) and executed by one or more processors/computers. The data storage medium or the memory unit may be implemented within the processor/computer or external to the processor/computer, in which case it can be communicatively coupled to the processor/computer via various means as is known in the art. Additionally, components of systems described herein may be rearranged and/or complimented by additional components in order to facilitate achieving the various aspects, goals, advantages, etc., described with regard thereto, and are not limited to the precise configurations set forth in a given figure, as will be appreciated by one skilled in the art.


It will be obvious to a person skilled in the art that the inventive concept can be implemented in various ways. For example, although in the example described above with reference to FIG. 6 the process involved steps of analyzing results and determining whether they can be improved, in some cases there may be no such analysis and determination; instead, two or more algorithms could be applied to the same set of campaigns, and allocations resulting from these used to select, according to the selection criteria, a preferred allocation, with the latter being used as the basis for generating the schedule at step 620. The invention and its embodiments are not limited to the examples described above but modifications and changes can be made without departing from the scope of the attached claims.

Claims
  • 1. A computer implemented method comprising: allocating, by a processor, campaign messages of a plurality of messaging campaigns to target recipients selected from a plurality of potential recipients based upon a first allocation algorithm that dictates a first order in which messages from each of the plurality of campaigns are allocated, wherein a target recipient for each message is selected by matching campaign criteria associated with the respective messaging campaign to profile data associated with each of the plurality of potential recipients, the resulting allocation creating a first allocation of target recipients;allocating, by the processor, campaign messages of the plurality of messaging campaigns to target recipients selected from the plurality of potential recipients based upon a second allocation algorithm that dictates a second order in which messages from each of the plurality of campaigns are allocated, wherein a target recipient for each message is selected by matching campaign criteria associated with the respective messaging campaign to profile data associated with each of the plurality of potential recipients, the resulting allocation creating a second allocation of target recipients; andselecting, by the processor, either the first allocation of target recipients or the second allocation of target recipients based upon which allocation best satisfies a selection criteria indicating a favorable outcome to the plurality of messaging campaigns.
  • 2. The computer implemented method of claim 1, wherein the first allocation algorithm and the second allocation algorithm further dictate a maximum number of messages which can be delivered to each recipient.
  • 3. The computer implemented method of claim 1, wherein the selection criteria accounts for which allocation of target recipients results in more messages being delivered.
  • 4. The computer implemented method of claim 1, wherein the selection criteria accounts for which allocation of target recipients results in a larger number of target recipients receiving messages.
  • 5. The computer implemented method of claim 1, wherein the selection criteria accounts for which allocation of target recipients results in all messages of a campaign being sent.
  • 6. The computer implemented method of claim 1, wherein the selection criteria accounts for which allocation of target recipients results in an even distribution of the messages among recipients.
  • 7. The computer implemented method of claim 1, wherein the selection criteria accounts for which allocation of target recipients results in a higher profit.
  • 8. The computer implemented method of claim 1, wherein the selection criteria accounts for which allocation of target recipients results in the best balance between multiple requirements.
  • 9. The computer implemented method of claim 1, wherein the selection criteria include a time period for which to evaluate a result of the first allocation algorithm and the second allocation algorithm.
  • 10. The computer implemented method of claim 1, further comprising: transmitting the messages of the plurality of messaging campaigns based on the selected allocation of target recipients.
  • 11. The computer implemented method of claim 1, wherein the campaign criteria includes an age range of the recipient.
  • 12. The computer implemented method of claim 1, wherein the campaign criteria includes a specified interest or preference of the recipient.
  • 13. The computer implemented method of claim 1, wherein the plurality of messaging campaigns are evaluated together based on similarity or commonality between campaign criteria.
  • 14. The computer implemented method of claim 1, wherein the plurality of messaging campaigns are evaluated together based on dissimilarity or lack of commonality between campaign criteria.
  • 15. A system comprising: a processor; andmemory containing instruction that, when executed, causes the processor to:allocate campaign messages of a plurality of messaging campaigns to target recipients selected from a plurality of potential recipients based upon a first allocation algorithm that dictates a first order in which messages from each of the plurality of campaigns are allocated, wherein a target recipient for each message is selected by matching campaign criteria associated with the respective messaging campaign to profile data associated with each of the plurality of potential recipients, the resulting allocation creating a first allocation of target recipients;allocate campaign messages of the plurality of messaging campaigns to target recipients selected from the plurality of potential recipients based upon a second allocation algorithm that dictates a second order in which messages from each of the plurality of campaigns are allocated, wherein a target recipient for each message is selected by matching campaign criteria associated with the respective messaging campaign to profile data associated with each of the plurality of potential recipients, the resulting allocation creating a second allocation of target recipients; andselect either the first allocation of target recipients or the second allocation of target recipients based upon which allocation best satisfies a selection criteria indicating a favorable outcome to the plurality of messaging campaigns.
  • 16. The system of claim 15, wherein the first allocation algorithm and the second allocation algorithm further dictate a maximum number of messages which can be delivered to each recipient.
  • 17. The system of claim 15, wherein the selection criteria accounts for which allocation of target recipients results in more messages being delivered.
  • 18. The system of claim 15, wherein the selection criteria accounts for which allocation of target recipients results in a larger number of target recipients receiving messages.
  • 19. The system of claim 15, wherein the selection criteria accounts for which allocation of target recipients results in all messages of a campaign being sent.
  • 20. The system of claim 15, wherein the selection criteria accounts for which allocation of target recipients results in an even distribution of the messages among recipients.
  • 21. The system of claim 15, wherein the selection criteria accounts for which allocation of target recipients results in a higher profit.
  • 22. The system of claim 15, wherein the selection criteria accounts for which allocation of target recipients results in the best balance between multiple requirements.
  • 23. The system of claim 15, wherein the selection criteria include a time period for which to evaluate a result of the first allocation algorithm and the second allocation algorithm.
  • 24. The system of claim 15, wherein the instructions further cause the processor to: transmit the messages of the plurality of messaging campaigns based on the selected allocation of target recipients.
  • 25. The system of claim 15, wherein the campaign criteria includes an age range of the recipient.
  • 26. The system of claim 15, wherein the campaign criteria includes a specified interest or preference of the recipient.
  • 27. The system of claim 15, wherein the plurality of messaging campaigns are evaluated together based on similarity or commonality between campaign criteria.
  • 28. The system of claim 15, wherein the plurality of messaging campaigns are evaluated together based on dissimilarity or lack of commonality between campaign criteria.
  • 29. A non-transitory computer-readable medium having computer executable instructions stored thereon that, when executed, cause a computing device to: allocate campaign messages of a plurality of messaging campaigns to target recipients selected from a plurality of potential recipients based upon a first allocation algorithm that dictates a first order in which messages from each of the plurality of campaigns are allocated, wherein a target recipient for each message is selected by matching campaign criteria associated with the respective messaging campaign to profile data associated with each of the plurality of potential recipients, the resulting allocation creating a first allocation of target recipients;allocate campaign messages of the plurality of messaging campaigns to target recipients selected from the plurality of potential recipients based upon a second allocation algorithm that dictates a second order in which messages from each of the plurality of campaigns are allocated, wherein a target recipient for each message is selected by matching campaign criteria associated with the respective messaging campaign to profile data associated with each of the plurality of potential recipients, the resulting allocation creating a second allocation of target recipients; andselect either the first allocation of target recipients or the second allocation of target recipients based upon which allocation best satisfies a selection criteria indicating a favorable outcome to the plurality of messaging campaigns.
Priority Claims (1)
Number Date Country Kind
0711527.2 Jun 2007 GB national
CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit under 35 USC 119 of U.S. Provisional Patent Application Ser. No. 60/934,525 filed Jun. 14, 2007 and GB Application No. GB0711527.2 filed Jun. 14, 2007, the entire disclosure of each of which is incorporated herein by reference.

US Referenced Citations (166)
Number Name Date Kind
4811382 Sleevi Mar 1989 A
4935956 Hellwarth Jun 1990 A
5408519 Pierce et al. Apr 1995 A
5592535 Klotz Jan 1997 A
5613213 Naddell et al. Mar 1997 A
5678179 Turcotte et al. Oct 1997 A
5752238 Dedrick May 1998 A
5978775 Chen Nov 1999 A
5978833 Pashley et al. Nov 1999 A
6006197 d'Eon et al. Dec 1999 A
6009410 LeMole et al. Dec 1999 A
6023700 Owens et al. Feb 2000 A
6044259 Hentila Mar 2000 A
6097942 Laiho Aug 2000 A
6202086 Maruyama et al. Mar 2001 B1
6253189 Feezell et al. Jun 2001 B1
6286005 Cannon Sep 2001 B1
6334145 Adams et al. Dec 2001 B1
6345279 Li et al. Feb 2002 B1
6411946 Chaudhuri Jun 2002 B1
6502076 Smith Dec 2002 B1
6639977 Swope et al. Oct 2003 B1
6647269 Hendrey et al. Nov 2003 B2
6718178 Sladek et al. Apr 2004 B1
6889054 Himmel et al. May 2005 B2
6990330 Veerepalli et al. Jan 2006 B2
7013125 Henrikson et al. Mar 2006 B2
7027799 Morin et al. Apr 2006 B2
7039599 Merriman et al. May 2006 B2
7072947 Knox et al. Jul 2006 B1
7136871 Ozer et al. Nov 2006 B2
7149500 Halsell Dec 2006 B2
7155243 Baldwin et al. Dec 2006 B2
7164927 Koch et al. Jan 2007 B1
7236768 Levitan Jun 2007 B2
7251478 Cortegiano Jul 2007 B2
7280818 Clayton Oct 2007 B2
7333823 Kamphuis Feb 2008 B2
7343317 Jokinen et al. Mar 2008 B2
7346630 Eichstaedt et al. Mar 2008 B2
7496362 Wolfman et al. Feb 2009 B2
7529357 Rae et al. May 2009 B1
7730017 Nance et al. Jun 2010 B2
7734632 Wang Jun 2010 B2
7925283 Smith et al. Apr 2011 B2
7945242 Kahn May 2011 B2
7979314 Ulenas Jul 2011 B2
8046810 Srivastava Oct 2011 B2
20010008404 Naito et al. Jul 2001 A1
20020032602 Lanzillo, Jr. et al. Mar 2002 A1
20020046341 Kazaks et al. Apr 2002 A1
20020052788 Perkes et al. May 2002 A1
20020077130 Owensby Jun 2002 A1
20020120498 Gordon et al. Aug 2002 A1
20020128908 Levin et al. Sep 2002 A1
20020137507 Winkler Sep 2002 A1
20020138291 Vaidyanathan et al. Sep 2002 A1
20020161770 Shapiro et al. Oct 2002 A1
20020164977 Link, II et al. Nov 2002 A1
20020183040 Lundstrom et al. Dec 2002 A1
20030003929 Himmel et al. Jan 2003 A1
20030023489 McGuire et al. Jan 2003 A1
20030040297 Pecen et al. Feb 2003 A1
20030040300 Bodic et al. Feb 2003 A1
20030076940 Manto Apr 2003 A1
20030083931 Lang May 2003 A1
20030110171 Ozer et al. Jun 2003 A1
20030197719 Lincke et al. Oct 2003 A1
20030208754 Sridhar et al. Nov 2003 A1
20040068435 Braunzell Apr 2004 A1
20040103157 Requena et al. May 2004 A1
20040133480 Domes Jul 2004 A1
20040158612 Concannon Aug 2004 A1
20040180648 Hymel Sep 2004 A1
20040185827 Parks Sep 2004 A1
20040192359 McRaild et al. Sep 2004 A1
20040203851 Vetro et al. Oct 2004 A1
20040204133 Andrew et al. Oct 2004 A1
20040209649 Lord Oct 2004 A1
20040240649 Goel Dec 2004 A1
20040254993 Mamas Dec 2004 A1
20040259526 Goris et al. Dec 2004 A1
20050027676 Eichstaedt et al. Feb 2005 A1
20050055408 Dong Mar 2005 A1
20050075929 Wolinsky et al. Apr 2005 A1
20050123111 Philbin et al. Jun 2005 A1
20050125397 Gross et al. Jun 2005 A1
20050186940 Schatzberger et al. Aug 2005 A1
20050229209 Hildebolt et al. Oct 2005 A1
20050238154 Heaton et al. Oct 2005 A1
20050245241 Durand et al. Nov 2005 A1
20050267798 Panara Dec 2005 A1
20060025163 Smith et al. Feb 2006 A1
20060031327 Kredo Feb 2006 A1
20060041470 Filho et al. Feb 2006 A1
20060085395 Cradick et al. Apr 2006 A1
20060109969 Oh May 2006 A1
20060123014 Ng Jun 2006 A1
20060194595 Myllynen et al. Aug 2006 A1
20060200460 Meyerzon et al. Sep 2006 A1
20060200461 Lucas et al. Sep 2006 A1
20060252512 Walker Nov 2006 A1
20060253327 Morris et al. Nov 2006 A1
20060276170 Radhakrishnan et al. Dec 2006 A1
20060276213 Gottschalk et al. Dec 2006 A1
20060282328 Gerace et al. Dec 2006 A1
20060286962 Davis Dec 2006 A1
20060286963 Koskinen et al. Dec 2006 A1
20060286964 Polanski et al. Dec 2006 A1
20060288124 Kraft et al. Dec 2006 A1
20070004380 Ylikoski Jan 2007 A1
20070016488 Ulenas Jan 2007 A1
20070027760 Collins et al. Feb 2007 A1
20070027762 Collins et al. Feb 2007 A1
20070033104 Collins et al. Feb 2007 A1
20070038711 MacBeth et al. Feb 2007 A1
20070055565 Baur et al. Mar 2007 A1
20070061195 Liu et al. Mar 2007 A1
20070074262 Kikkoji et al. Mar 2007 A1
20070077911 Raman Apr 2007 A1
20070094066 Kumar et al. Apr 2007 A1
20070100805 Ramer et al. May 2007 A1
20070105536 Tingo, Jr. May 2007 A1
20070136457 Dai et al. Jun 2007 A1
20070174295 Abraham et al. Jul 2007 A1
20070192409 Kleinstern et al. Aug 2007 A1
20070232300 Wolfman Oct 2007 A1
20070237330 Srivastava Oct 2007 A1
20070260624 Chung et al. Nov 2007 A1
20080052158 Ferro et al. Feb 2008 A1
20080057920 Pettit et al. Mar 2008 A1
20080065491 Bakman Mar 2008 A1
20080071929 Motte et al. Mar 2008 A1
20080075249 Zach Mar 2008 A1
20080086369 Kiat et al. Apr 2008 A1
20080091796 Story Apr 2008 A1
20080120186 Jokinen et al. May 2008 A1
20080125080 Phillips May 2008 A1
20080140508 Anand et al. Jun 2008 A1
20080215436 Roberts Sep 2008 A1
20080228568 Williams et al. Sep 2008 A1
20080249832 Richardson et al. Oct 2008 A1
20080271068 Ou et al. Oct 2008 A1
20080319836 Aaltonen et al. Dec 2008 A1
20090006194 Sridharan et al. Jan 2009 A1
20090029721 Doraswamy Jan 2009 A1
20090049090 Shenfield et al. Feb 2009 A1
20090063249 Tomlin et al. Mar 2009 A1
20090075627 Hwang Mar 2009 A1
20090106111 Walk et al. Apr 2009 A1
20090125377 Somji et al. May 2009 A1
20090132395 Lam et al. May 2009 A1
20090141875 Demmitt et al. Jun 2009 A1
20090197619 Colligan et al. Aug 2009 A1
20090216847 Krishnaswamy et al. Aug 2009 A1
20090286520 Nielsen et al. Nov 2009 A1
20090298483 Bratu et al. Dec 2009 A1
20100082397 Blegen Apr 2010 A1
20100082423 Nag et al. Apr 2010 A1
20100088152 Bennett Apr 2010 A1
20100114654 Lukose et al. May 2010 A1
20100125505 Puttaswamy May 2010 A1
20100138271 Henkin Jun 2010 A1
20100161424 Sylvain Jun 2010 A1
20100169157 Muhonen et al. Jul 2010 A1
20100169176 Turakhia Jul 2010 A1
Foreign Referenced Citations (160)
Number Date Country
199 41 461 Mar 2001 DE
19941461 Mar 2001 DE
1 073 293 Jan 2001 EP
1073293 Jan 2001 EP
1 109 371 Jun 2001 EP
1109371 Jun 2001 EP
1 161 093 Dec 2001 EP
1161093 Dec 2001 EP
1 193 955 Apr 2002 EP
1193955 Apr 2002 EP
1 220 132 Jul 2002 EP
1220132 Jul 2002 EP
1 239 395 Sep 2002 EP
1239395 Sep 2002 EP
1 253 542 Oct 2002 EP
1249995 Oct 2002 EP
1253542 Oct 2002 EP
1 298 945 Apr 2003 EP
1298945 Apr 2003 EP
1 324 250 Jul 2003 EP
1324250 Jul 2003 EP
1 365 604 Nov 2003 EP
1365604 Nov 2003 EP
1445703 Aug 2004 EP
1 455 511 Sep 2004 EP
1455511 Sep 2004 EP
1 542 482 Jun 2005 EP
1587332 Oct 2005 EP
1587332 Oct 2005 EP
1 613 102 Jan 2006 EP
1 615 455 Jan 2006 EP
1613102 Jan 2006 EP
1615455 Jan 2006 EP
1 633 100 Mar 2006 EP
1633100 Mar 2006 EP
2 352 856 Feb 2001 GB
2352856 Feb 2001 GB
2 356 777 May 2001 GB
2356777 May 2001 GB
2 383 149 Jun 2003 GB
2383149 Jun 2003 GB
2 406 996 Apr 2005 GB
2 407 002 Apr 2005 GB
2406996 Apr 2005 GB
2407002 Apr 2005 GB
2 414 621 Nov 2005 GB
2002140272 May 2002 JP
9624213 Aug 1996 WO
WO 9624213 Aug 1996 WO
9733421 Sep 1997 WO
WO 9733421 Sep 1997 WO
9837685 Aug 1998 WO
WO 9837685 Aug 1998 WO
9853637 Nov 1998 WO
0044151 Jul 2000 WO
WO 0044151 Jul 2000 WO
0062564 Oct 2000 WO
0122748 Mar 2001 WO
WO 0122748 Mar 2001 WO
0131497 May 2001 WO
WO 0131497 May 2001 WO
0157705 Aug 2001 WO
0158178 Aug 2001 WO
0163897 Aug 2001 WO
WO 0155892 Aug 2001 WO
WO 0157705 Aug 2001 WO
WO 0158178 Aug 2001 WO
0165411 Sep 2001 WO
0169406 Sep 2001 WO
0171949 Sep 2001 WO
0172063 Sep 2001 WO
WO 0165411 Sep 2001 WO
WO 0169406 Sep 2001 WO
WO 0171949 Sep 2001 WO
WO 0172063 Sep 2001 WO
0177840 Oct 2001 WO
0178425 Oct 2001 WO
WO 0177840 Oct 2001 WO
WO 0178425 Oct 2001 WO
0191400 Nov 2001 WO
WO 0191400 Nov 2001 WO
0193551 Dec 2001 WO
0197539 Dec 2001 WO
WO 0193551 Dec 2001 WO
WO 0197539 Dec 2001 WO
0223423 Mar 2002 WO
0231624 Apr 2002 WO
WO 0231624 Apr 2002 WO
0244834 Jun 2002 WO
WO 0244834 Jun 2002 WO
02054803 Jul 2002 WO
WO 02054803 Jul 2002 WO
02059720 Aug 2002 WO
WO 02059720 Aug 2002 WO
02069585 Sep 2002 WO
02075574 Sep 2002 WO
WO 02069585 Sep 2002 WO
WO 02075574 Sep 2002 WO
02080595 Oct 2002 WO
02084895 Oct 2002 WO
WO 02080595 Oct 2002 WO
WO 02084895 Oct 2002 WO
03015430 Feb 2003 WO
WO 03015430 Feb 2003 WO
03019845 Mar 2003 WO
03024136 Mar 2003 WO
WO 03019845 Mar 2003 WO
WO 03024136 Mar 2003 WO
03038638 May 2003 WO
WO 03038638 May 2003 WO
03088690 Oct 2003 WO
WO 03088690 Oct 2003 WO
2004034671 Apr 2004 WO
WO 2004034671 Apr 2004 WO
2004054205 Jun 2004 WO
WO 2004054205 Jun 2004 WO
2004084532 Sep 2004 WO
WO 2004084532 Sep 2004 WO
2004086791 Oct 2004 WO
WO 2004086791 Oct 2004 WO
2004100521 Nov 2004 WO
WO 2004100470 Nov 2004 WO
WO 2004100521 Nov 2004 WO
2004107795 Dec 2004 WO
2004114109 Dec 2004 WO
WO 2004114109 Dec 2004 WO
2005015806 Feb 2005 WO
WO 2005015806 Feb 2005 WO
2005029769 Mar 2005 WO
WO 2005029769 Mar 2005 WO
2005062637 Jul 2005 WO
WO 2005062637 Jul 2005 WO
2005076650 Aug 2005 WO
WO 2005076650 Aug 2005 WO
2005083980 Sep 2005 WO
2005086927 Sep 2005 WO
2006002869 Jan 2006 WO
WO 2006002869 Jan 2006 WO
WO 2006005001 Jan 2006 WO
2006026505 Mar 2006 WO
2006027407 Mar 2006 WO
WO 2006026505 Mar 2006 WO
WO 2006027407 Mar 2006 WO
2006093284 Sep 2006 WO
WO 2006093284 Sep 2006 WO
2006110446 Oct 2006 WO
WO 2006105202 Oct 2006 WO
WO 2006110446 Oct 2006 WO
2006119481 Nov 2006 WO
WO 2006119481 Nov 2006 WO
2007056698 May 2007 WO
WO 2007056698 May 2007 WO
2008013437 Jan 2008 WO
WO 2008013437 Jan 2008 WO
WO 2008053062 May 2008 WO
2009009507 Jan 2009 WO
2009061914 May 2009 WO
2009077888 Jun 2009 WO
2009099876 Aug 2009 WO
2009158097 Dec 2009 WO
Non-Patent Literature Citations (28)
Entry
Notification Concerning Transmittal of International Preliminary Report on Patentability (Chapter 1 of the Patent Cooperation Treaty) mailed Dec. 30, 2009 in related International Application PCT/EP2008/057503 (1 page).
International Preliminary Report on Patentability mailed Dec. 30, 2009 in related International Application PCT/EP2008/057503 (1 page).
Written Opinion of the International Searching Authority mailed Dec. 30, 2009 in related International Application PCT/EP2008/057503 (6 pages).
International Search Report and Written Opinion of the International Searching Authority) dated Nov. 3, 2008, issued in counterpart International Application PCT/EP2008/057303.
“On the Design and Evaluation of Job Scheduling Algorithms,” Jochen Krallman, Uwe Schwiegelshohn and Ramin Yahyapour, Lecture Notes in Computer Science, vol. 1659, Jan. 1, 1999, pp. 17-42.
U.K. Search Report under Section 17 dated Oct. 11, 2007 of corresponding U. K. Patent Application No. GB0711527.2.
U.K. Search Report under Section 17 dated Jun. 27, 2008 of corresponding U. K. Patent Application No. GB0808041.8.
Communication pursuant to Article 94(3) EPC issued Jan. 14, 2011 by the European Patent Office in related European Application No. 08 761 022.6 (6 pages).
“Advertisement System, Method and Computer Program Product”, IP.com Prior Art Database Disclosure, Pub No. IPCOM000138557D, dated Jul. 24, 2006, IP.com, Amherst, NY (Available online at http://priorartdatabase.com/IPCOM/000138557, last visited Aug. 30, 2010), (39 pages), Jul. 24, 2006.
“Communication Pursuant to Article 94(3) EPC”, European Application No. 08101640.4 (4 pages), Nov. 25, 2009.
“European Search Report”, European Patent Application No. EP 08101640, (2 pages), Mar. 19, 2008.
“European Search Report”, European Application No. EP 08101331, (2 pages), Jun. 4, 2008.
“European Search Report date Jun. 26, 2008”, European Patent Application No. 08155590.6, (3 pages).
“International Preliminary Report on Patentability and Written Opinion”, International Application PCT/EP2008/005084, (5 page), Nov. 17, 2009.
“International Search Report and Written Opinion (Form PCT/ISA/210)”, International Application No. PCT/EP2008/058117, (13 pages), Feb. 13, 2009.
“International Search Report dated May 6, 2008”, International Application No. PCT/EP2008/050208.
“Office Action”, related U.S. Appl. No. 12/221,028, (14 pages) Feb. 19, 2009.
“U.K. Search Report under Section 17”, U.K. Application No. GB 0809536.6, (1 page) Jun. 9, 2008.
“U.K. Search Report Under Section 17”, U.K Application No. GB 0712291.4, (1 pages) Aug. 16, 2007.
“United Kingdom Search Report dated Apr. 13, 2007”, UK Application No. GB0700340.3, (3 pages).
“United Kingdom Search Report under Section 17”, Application No. GB0709426.1, (1 page) Sep. 10, 2007.
“Written Opinion of the International Search Authority May 6, 2008”, International Application No. PCT/EP2008/050208, (8 pages).
Hillard, Dustin et al., “Improving Ad Relevance in Sponsored Search”, Proceedings of the third ACM international conference on Web search and data mining, WSDM'10, Feb. 4-6, 2010, Session: Ads, pp. 361-369, ACM, New York, New York, USA, 2010, (9 pages).
Internet Reference, “Specific Media Behavioral Targeting Index”, Specific Media, Inc., Irvine, CA, 2010, Available online at http://www.specificmedia.com/behavioral-targeting.php, (1 page).
Langheinrich, Marc et al., “Unintrusive Customization Techniques for Web Advertising”, Computer Networks: The International Journal of Computer and Telecommunications Networking, vol. 31, No. 11, May 1999, pp. 1259-1272, Elsevier North-Holland, Inc., New York, NY, (18 pages).
Regelson, Moira et al., “Predicting Click-Through Rate Using Keyword Clusters”, Proceedings of the Second Workshop on Sponsored Search Auctions, EC'06, SSA2, Jun. 11, 2006, ACM, (6 pages).
Richardson, Matthew et al., “Predicting Clicks: Estimating the Click-Through Rate for New Ads”, Proceedings of the 16th international conference on World Wide Web, Banff, Alberta, Canada, May 8-12, 2007, Session: Advertisements & click estimates, pp. 521-529, ACM, (9 pages).
Shaikh, Baber M. et al., “Customized User Segments for Ad Targeting”, IP.com Prior Art Database Disclosure, Pub No. IPCOM000185640D, dated Jul. 29, 2009 UTC, IP.com, Amherst, NY (Available online at http://priorartdatabase.com/IPCOM/000185640, last visited Aug. 30, 2010), (15 pages).
Related Publications (1)
Number Date Country
20080312948 A1 Dec 2008 US
Provisional Applications (1)
Number Date Country
60934525 Jun 2007 US