JP6165679B2 - Advertisement distribution management device and advertisement distribution management method - Google Patents

Description

  The present invention relates to an advertisement distribution management device and an advertisement distribution management method.

  In recent years, with the rapid spread of networks such as the Internet, advertisement distribution via networks has been actively performed. As one form of such advertisement distribution, a display advertisement that displays an advertisement using an image, video, or the like in a certain space within a page (hereinafter referred to as an advertisement distribution target page) that is an advertisement distribution target is known. .

  In display advertisements, effective advertisement distribution is performed by distributing advertisements to users having user attributes (gender, age, interest, etc.) designated by the advertiser (see, for example, Patent Document 1). ).

JP 2011-238020 A

  In the advertisement distribution described above, an advertisement is distributed when the user attribute of the user who has accessed the advertisement distribution target page is the user attribute specified by the advertiser. For this reason, it is desirable to predict a distribution target user as a future advertisement distribution destination including a user attribute as a future log, and accept an order from an advertiser based on the future log.

  However, when a distribution target user who becomes a future advertisement distribution destination has a plurality of user attributes, it may be difficult to accept a new order from an advertiser depending on how the distribution target user is assigned to an advertisement distribution order. .

  The present application has been made in view of the above, and an object thereof is to provide an advertisement distribution management device and an advertisement distribution management method capable of appropriately assigning distribution target users to advertisement distribution orders.

The advertisement distribution management device according to the present application acquires prediction information storage means for storing information including user attributes of a plurality of distribution target users predicted as future advertisement distribution destinations, and a plurality of order requests specifying the user attributes. Each time a new order request is acquired by the acquisition means and the acquisition means, the distribution target for the plurality of order requests including the new order request and the order requests acquired in the past by the acquisition means And allocating means for performing a process of allocating users.

  According to one aspect of the embodiment, it is possible to provide an advertisement distribution management device and an advertisement distribution management method that can appropriately allocate a distribution target user to an advertisement distribution order.

FIG. 1A is an explanatory diagram (part 1) of an advertisement distribution management process according to the embodiment. FIG. 1B is an explanatory diagram (part 2) of the advertisement distribution management process according to the embodiment. FIG. 2 is a diagram illustrating a configuration of the advertisement distribution management device according to the embodiment. FIG. 3 is a diagram illustrating an example of the advertisement information table. FIG. 4 is a diagram illustrating an example of the user information table. FIG. 5 is a diagram illustrating an example of the access log table. FIG. 6 is a diagram illustrating an example of a future log table. FIG. 7 is a diagram for explaining how to obtain the total number of user attributes. FIG. 8 is a diagram illustrating an example of the relationship between the attribute content of the user attribute pattern and the number of users. FIG. 9 is a diagram illustrating an example of assignment of the number of distribution target users belonging to the user attribute pattern assigned to the order request. FIG. 10 is a diagram illustrating an advertisement distribution management process flow of the advertisement distribution management apparatus according to the embodiment. FIG. 11 is a diagram illustrating an example of an allocation process flow of the advertisement distribution management device according to the embodiment.

  Hereinafter, a mode for carrying out an advertisement distribution management device and an advertisement distribution management method according to the present application (hereinafter referred to as “embodiment”) will be described in detail with reference to the drawings. Note that the advertisement distribution management device and the advertisement distribution management method according to the present application are not limited by this embodiment. In the following, “database” is described as “DB”.

[1. (Ad delivery management processing)
First, the advertisement distribution management process according to the embodiment will be described. FIG. 1A and FIG. 1B are explanatory diagrams of advertisement distribution management processing according to the embodiment, and distribution target users are appropriately assigned to order requests from advertisers by this advertisement distribution management processing.

  As illustrated in FIG. 1A, the advertisement distribution management system according to the embodiment includes an advertisement distribution management device 1, an information distribution device 2, an advertiser terminal 3, and a user terminal 4. The advertisement distribution management device 1, the information distribution device 2, the advertiser terminal 3, and the user terminal 4 are connected to each other via a communication network 5. The communication network 5 is, for example, an IP (Internet Protocol) network such as the Internet. Although there are a plurality of advertiser terminals 3 and user terminals 4 respectively, they are omitted for convenience of explanation in FIG. 1A.

  The advertiser terminal 3 is a terminal device of an advertiser who desires advertisement distribution, and sends an order request to the advertisement distribution management apparatus 1 to order an advertisement to be distributed to the user terminal 4 of a user having a predetermined user attribute. Do it. Specifically, the advertiser terminal 3 receives the advertisement information that the orderer desires to be posted on the advertisement distribution target page, the user attribute information designated as the distribution destination of the advertisement information, and the information on the number of advertisement information distributions. The order request including the request is transmitted to the advertisement distribution management device 1. The advertisement information includes, for example, two-dimensional contents such as banner advertisements, text advertisements, and video advertisements, and three-dimensional contents.

  The user terminal 4 is a user terminal device that accesses the advertisement distribution management device 1 and the information distribution device 2. When there is an access from the user terminal 4, the information distribution device 2 transmits an advertisement distribution target page such as a web page to the user terminal 4, for example. The advertisement distribution target page includes a URL (Uniform Resource Locator) of advertisement information, and the user terminal 4 transmits an advertisement request to the advertisement distribution management device 1 based on the URL.

  Note that the URL of the advertisement information can be the URL of the information distribution apparatus 2 instead of the URL of the advertisement distribution management apparatus 1. In this case, the user terminal 4 requests the advertisement information from the information distribution device 2 based on the URL of the advertisement information. When receiving such a request, the information distribution apparatus 2 makes a transmission request for advertisement information to the advertisement distribution management apparatus 1. The advertisement distribution management device 1 transmits advertisement information corresponding to the transmission request from the information distribution device 2 to the information distribution device 2. When the information distribution device 2 receives the advertisement information from the advertisement distribution management device 1, the information distribution device 2 transmits the advertisement information to the user terminal 4.

  The advertisement distribution management apparatus 1 performs an order request acceptance process, a user attribute assignment process for the accepted order request, and a process for executing the distribution of advertisement information specified in the order request, thereby receiving the advertisement delivery order. Perform processing up to delivery.

  The advertisement distribution management device 1 stores information including user attributes of a plurality of distribution target users predicted as future advertisement distribution destinations in the DB as a future log, and receives an order request based on the future log. And user attribute assignment processing.

  In the example illustrated in FIG. 1A, information including user attributes of the distribution target users 6a to 6e is stored in the DB as a future log. The distribution target user 6a has user attributes A, B, and C, the distribution target user 6b has user attributes A, the distribution target user 6c has user attributes A and C to E, and the distribution target user 6d is a user. It has attributes B, D, and E, and the distribution target user 6e has user attributes B to D.

  The advertisement distribution management device 1 acquires an order request transmitted from the advertiser terminal 3, and assigns a distribution target user to the order request. As described above, the order request transmitted from the advertiser terminal 3 includes information on the user attribute (hereinafter referred to as “designated user attribute”) designated as the distribution destination of the advertisement information, and the number of advertisement information distribution (hereinafter referred to as “designated user attribute”). , Specified number of distributions).

  The advertisement distribution management device 1 performs distribution target user allocation processing with a plurality of order requests as one unit. Specifically, the advertisement distribution management device 1 maximizes the number of predetermined user attributes of distribution target users not assigned to the order request based on the specified user attributes and the specified number of distributions of the plurality of order requests. As described above, distribution target users are assigned to a plurality of order requests. Hereinafter, a distribution target user that is not allocated to an order request may be described as an unallocated user.

  For example, as shown in FIG. 1B, an order request 7a having a designated user attribute “attribute A” and a designated delivery number “2”, a designated user attribute “attribute D”, and a designated delivery number “ Consider a case where an order request 7b of “1” is transmitted from the advertiser terminal 3 to the advertisement distribution management apparatus 1. Further, it is assumed that the predetermined user attributes (hereinafter referred to as appropriateness determination attributes) are “attribute B”, “attribute C”, and “attribute D”.

  The distribution target users 6a and 6c having “attribute A” as the designated user attribute of the order request 7a are assigned to the order request 7a, and the distribution target user 6e having “attribute D” as the designated user attribute of the order request 7b is assigned. In this case, the numbers of the attributes A to E that the unassigned user has are “1”, “1”, “0”, “1”, and “1”, respectively. Therefore, the total number of appropriateness determination attributes (attributes B to D) possessed by unassigned users (hereinafter referred to as unassigned score) is “2”.

  On the other hand, the distribution target users 6a and 6b having “attribute A” as the designated user attribute of the order request 7a are assigned to the order request 7a, and the distribution target user 6d having “attribute D” as the designated user attribute of the order request 7b is assigned. When assigned, the numbers of the attributes A to E of the unassigned user are “1”, “1”, “2”, “2”, and “1”, respectively. Therefore, the unassigned score, which is the total number of appropriateness determination attributes (attributes B to D) possessed by the unassigned user, is “5”.

  The unassigned score when the distribution target user is allocated to the plurality of order requests 7a and 7b is maximized when the distribution target users 6a, 6b, and 6d are allocated. Accordingly, the advertisement distribution management device 1 assigns distribution target users 6a, 6b, and 6d to a plurality of order requests 7a and 7b.

  As described above, the advertisement distribution management device 1 allocates distribution target users to a plurality of order requests so that the unallocated score is maximized. Since the unassigned score is the number of attributes for determining appropriateness possessed by unassigned users, as an appropriateness determining attribute, for example, by setting a user attribute that is frequently specified in an order request, a future order The possibility of accepting requests can be increased, and loss of sales opportunities can be suppressed.

  Although the appropriateness determination attribute has been described as a part of the plurality of user attributes, all user attributes may be set as appropriateness determination attributes. For example, when the distribution target users 6a, 6c, and 6e are assigned to the order requests 7a and 7b, the unassigned score is “4”. On the other hand, when the distribution target users 6a, 6b, and 6d are assigned to the order requests 7a and 7b, the unassigned score is “7”.

  Thus, by setting all the user attributes as appropriateness determination attributes, it is possible to increase the total number of user attributes that can be allocated to the distribution target user for the order request. Therefore, even in this case, since the possibility of accepting future order requests increases, it is possible to suppress the loss of sales opportunities.

  Hereinafter, with regard to the advertisement distribution management apparatus 1 that executes the above-described advertisement distribution management process, a specific configuration and details of specific processes will be described with reference to the drawings.

[2. Configuration of Advertisement Distribution Management Device 1]
FIG. 2 is a diagram illustrating a configuration of the advertisement distribution management device 1 according to the embodiment. As illustrated in FIG. 2, the advertisement distribution management device 1 includes a communication unit 10, an advertisement information DB 11, a user information DB 12, a past log DB 13, a future log DB 14, and a control unit 15. Further, the control unit 15 includes an advertisement distribution unit 21, a future log prediction unit 22, a reception unit 23, and an allocation unit 24.

  The communication unit 10 is an interface such as a NIC (Network Interface Card). The control unit 15 transmits and receives various types of information to and from the advertiser terminal 3 and the user terminal 4 via the communication unit 10 and the communication network 5. The advertisement information DB 11, the user information DB 12, the past log DB 13, and the future log DB 14 are storage devices such as a hard disk, a RAM (Random Access Memory), a semiconductor memory element such as a flash memory, and an optical disk.

  The control unit 15 is realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA). In addition, the control unit 15 executes a program stored in an internal storage device by a CPU (Central Processing Unit) or an MPU (Micro Processing Unit) using the RAM as a work area, so that the advertisement distribution unit 21, future log prediction Functions as a unit 22, a reception unit 23, and an allocation unit 24. Note that the configuration of the control unit 15 is not limited to such a configuration, and may be another configuration as long as information processing described later is performed.

(About the advertisement distribution unit 21)
The advertisement distribution unit 21 acquires advertisement information corresponding to the user attribute of the user of the user terminal 4 from the advertisement information table of the advertisement information DB 11 when the advertisement information distribution request is received from the user terminal 4. Information is transmitted from the communication unit 10 to the user terminal 4 that is the distribution request source.

  FIG. 3 is a diagram illustrating an example of the advertisement information table stored in the advertisement information DB 11. As illustrated in FIG. 3, the advertisement information table includes information in which “user attribute”, “distribution period”, “number of distributions”, and “advertisement information” are associated with each “advertisement ID”.

  “Advertisement ID” is identification information assigned for each received order request. “User attribute” is a user attribute (specified user attribute) specified in the order request, “Distribution period” is a distribution period of the advertisement information specified in the order request, and “Number of distributions” is the order request The number of advertisement information distributions specified in (the number of specified distributions). The “advertisement information” is advertisement information specified in the order request.

  For example, in the advertisement ID “O1”, the user attribute to be distributed is “car”, the distribution period of advertisement information is “January 1 to January 31, 2013”, and the number of advertisement information distribution is “ “100 times”, indicating that the advertisement information is “banner advertisement Ya”. Note that the distribution process of the advertisement information in response to the order request ends when the number of distributions of the advertisement information exceeds “100 times” even within the advertisement information distribution period.

  When the user of the user terminal 4 that has requested the distribution of advertisement information has the user attribute “car”, the advertisement distribution unit 21, for example, displays the advertisement information “banner advertisement” of the advertisement ID “O1” from the advertisement information table of the advertisement information DB 11 Ya ”is acquired and transmitted to the user terminal 4 that is the distribution request source.

  The user attribute which the user of the user terminal 4 has is determined based on the user information table of the user information DB 12. FIG. 4 is a diagram illustrating an example of a user information table stored in the user information DB 12. As shown in FIG. 4, the user information table has information such as “user ID”, “psychographic attribute”, and “demographic attribute”.

  “User ID” is identification information for identifying a user who uses the user terminal 4. Such “user ID” corresponds to, for example, user identification information set in an HTTP cookie (HyperText Transfer Protocol Cookie) transmitted / received between the advertisement distribution management device 1 and the user terminal 4.

  The “psychographic attribute” is information indicating the user's values, lifestyle, personality, preference, and the like. Such “psychographic attributes” are classified into attribute categories such as “car”, “cosmetics”, “clothes”, and “travel”, for example. In the example illustrated in FIG. 4, “1” is stored for each attribute category when the user preference is relatively high, and “0” is stored otherwise. “Psychographic attributes” are not limited to the example of attribute classification shown in FIG. 4, but include various attribute classifications such as economy, politics, baseball, soccer, other sports, sweets, personal computers, white goods, furniture, etc. May be.

  “Demographic attribute” indicates demographic user attribute information. Such “demographic attributes” are classified into attribute categories such as “gender” and “age” of the user, for example. In “gender” shown in FIG. 4, “1” is stored when the user is female, and “2” is stored when the user is male. The “age” stores the age of the user. Note that the “demographic attribute” is not limited to the example of the attribute classification shown in FIG. 4, and may include various attribute classifications such as the user's occupation, family structure, annual income, address, hometown, and educational background.

  For example, when the user with the user ID “U1” shown in FIG. 4 makes a distribution request for advertisement information from the user terminal 4, the user attributes of the user are “car”, “travel”, “male”, “21 years old”. And so on. Further, for example, when the user with the user ID “U2” shown in FIG. 4 makes a distribution request for advertisement information from the user terminal 4, the advertisement distribution unit 21 indicates that the user attributes of the user are “clothes”, “female”, It determines that it is "44 years old" etc.

  Further, the advertisement distribution unit 21 sets the user ID of the user who has requested distribution of advertisement information in the access log table of the past log DB 13. FIG. 5 is a diagram illustrating an example of an access log table stored in the past log DB 13.

  As shown in FIG. 5, the access log table includes information (hereinafter referred to as a past log) that associates a user ID with the date and time when the user of the user terminal 4 made a distribution request for advertisement information to the advertisement distribution management device 1. This table is set in some cases. The advertisement distribution unit 21 updates the access log table every time there is a request for distribution of advertisement information from the user terminal 4.

(About the future log prediction unit 22)
The future log prediction unit 22 determines a plurality of distribution target users predicted as future advertisement distribution destinations based on the access log table stored in the past log DB 13, and stores the determined user information of the distribution target users in the future log. Set in the future log table of DB14.

  FIG. 6 is a diagram illustrating an example of a future log table stored in the future log DB 14. As shown in FIG. 6, the future log table is set with information related to a distribution target user predicted as an advertisement distribution destination in a predetermined future period (hereinafter may be described as a future log). The future log table shown in FIG. 6 includes future log information such as “distribution target ID”, “allocation state”, “psychographic attribute”, and “demographic attribute”.

  “Distribution target ID” is identification information assigned to each distribution target user. The distribution target user is a virtual user different from the user set in the user information table of the user information DB 12.

  The “assignment state” is information indicating whether or not the distribution target user is assigned to the order request. “1” is stored if the user is a distribution target user assigned to the order request, and “0” is stored if the distribution target user is not assigned to the order request.

  “Psychographic attributes” and “Demographic attributes” are the same items as those in the user information table of the user information DB 12. Note that a user set in the user information table can be estimated as a distribution target user, and the distribution target ID can be used as the user ID. In this case, information on “psychographic attributes” and “demographic attributes” can be acquired from the user information table based on the user ID.

  The future log prediction unit 22 can predict a distribution target user who will be an advertisement distribution destination in a future predetermined period from an access log of the same period of one year ago, an access log of a predetermined period immediately before, or the like. For example, the future log prediction unit 22 can estimate a user who has made a distribution request for advertisement information in the immediately preceding past month as a distribution target user for the future month.

  Further, the future log prediction unit 22 performs a statistical process on the past log set in the access log table to determine the tendency of the distribution target user who will be the future advertisement distribution destination, and based on the determination result, the future log Can also be predicted. In this case, for example, the tendency of the distribution target user may be determined in consideration of fluctuation information such as weather and temperature.

The future log prediction unit 22 can set the predicted future log in the future log table, but can also set the future log acquired from the external device via the communication unit 10 in the future log table.
(About the reception unit 23)
The accepting unit 23 determines whether or not to accept an order request from the advertiser terminal 3 based on the future log set in the future log table, and orders the advertiser terminal 3 when accepting the order request. Make a reception response.

  The order request from the advertiser terminal 3 includes designated information such as a designated user attribute and a designated number of distributions as described above. When there is an order request from the advertiser terminal 3, the reception unit 23 counts the number of unassigned users having the specified user attribute of the order request among the plurality of distribution target users set in the future log table. The receiving unit 23 receives an order request from the advertiser terminal 3 when the count result is equal to or greater than the specified number of distributions of the order request.

  For example, if the receiving unit 23 receives an order request with the designated user attribute “car” and the designated number of distributions “100”, “1” is set in the attribute category “car” in the future log table. Count the number of distribution target users. The accepting unit 23 accepts an order request from the advertiser terminal 3 when the count result is “100” or more, and does not accept an order request from the advertiser terminal 3 when the count result is less than “100”. .

  In addition, the reception unit 23 counts a combination of two or more user attributes having a frequency specified in the order request equal to or greater than a threshold value as one user attribute (hereinafter sometimes referred to as a combined user attribute). For example, when the combination of “clothes” and “travel” is frequently specified in the order request, the reception unit 23 counts the combination of “clothes” and “travel” as one user attribute.

  In the advertisement information table of the advertisement information DB 11, information on the order request received by the receiving unit 23 is set. Based on the advertisement information table, the receiving unit 23 has a frequency specified by the order request equal to or greater than a threshold value of 2 or more. The combination of user attributes can be determined.

  For example, the accepting unit 23 obtains the total number of distributions specified in the past order request as the first total value based on the advertisement information table, and the past order request for each combination of two or more user attributes. The sum of the specified number of distributions is obtained as the second total value. And the reception part 23 counts the combination of the user attribute whose ratio of the 2nd total value with respect to a 1st total value is more than a predetermined threshold value as one user attribute.

  When the combination of “car” and “travel” is frequently specified in the order request, and the future log table is in the state shown in FIG. 6, the user attribute of the distribution target user with the distribution target ID “F3” is: Instead of “car” or “travel”, it becomes “car & travel”.

  Note that a combination of two or more user attributes whose frequency specified in the order request is equal to or greater than a threshold value may be set in the user information DB 12 separately from an external device or the like instead of being determined by the receiving unit 23.

  In addition, when the number of unassigned users is equal to or less than a predetermined value or when the number of unassigned users having a combined user attribute is equal to or less than a predetermined value, the reception unit 23 sets the combined user attribute to a plurality of user attributes. Can be divided into As a result, the total number of user attributes can be increased. As a result, the possibility of accepting an order request specifying a combined user attribute is reduced, but the possibility of accepting another order request can be increased.

(Allocation unit 24)
The allocating unit 24 allocates the distribution target user set in the future log table to the order request received by the receiving unit 23.

  Specifically, the allocating unit 24 assigns distribution target users to a plurality of order requests, based on the specified user attributes and the specified number of distributions of the plurality of order requests, by distribution target users not allocated to the order request. This is done so that the number of predetermined user attributes is maximized.

  Here, the “predetermined user attribute” is the attribute for determining the appropriateness described above. For example, an administrator of the advertisement distribution management apparatus 1 (hereinafter referred to as a distribution manager) operates the operation unit of the advertisement distribution management apparatus 1. (Not shown) can be set. The distribution manager can set all user attributes as attributes for determining appropriateness, or can set some user attributes.

  In addition, the allocation unit 24 is based on the setting made by the distribution manager from the operation unit (not shown) of the advertisement distribution management device 1, and includes the first allocation mode, the second allocation mode, and the third allocation mode. , Execute any one allocation mode.

  When the first allocation mode is set, the allocation unit 24 allocates a distribution target user to all the order requests set in the advertisement information table every time a new order request is received by the reception unit 23. The process (hereinafter may be referred to as overall process) is repeated.

  In addition, when the second allocation mode is set, the allocation unit 24 allocates the distribution target user to the newly acquired order request while maintaining the allocation to the order request to which the distribution target user is allocated. Processing (hereinafter, may be referred to as partial processing) is performed.

  Further, when the third allocation mode is set, the allocation unit 24 performs the entire process at a predetermined timing (for example, at night when there are few order requests) while performing a partial process on a new order request.

  Hereinafter, the allocation process by the allocation unit 24 will be specifically described. When there are n order requests in the advertisement information table, an order request Oi (1 ≦ i ≦ n; i is an integer, see FIG. 3) is defined, and the designated number of distributions of the order request Oi is “o_i”. When m user attribute patterns are set, a user attribute pattern Pj (1 ≦ j ≦ m; j is an integer) is defined, and the number of distribution target users belonging to the user attribute pattern Pj is “p_j”. .

The “user attribute pattern” is, for example, a combination of attribute categories such as “car”, “cosmetics”, “clothes”, and “travel” when the future log table is in the state shown in FIG. For example, when the attribute pattern is defined by {car, makeup, clothes, travel,...}, Each attribute pattern is, for example, {1, 0, 0, 0,. , 0,...}, {1, 1, 1, 0,. The number of user attributes pattern, when the number of attribute classification is, for example, six, is 2 ^6. Note that items that are not likely to be specified in the order request Oi can be excluded from the user attribute pattern.

The allocation unit 24 uses the linear programming method to maximize the unallocated score KPI with the following formulas (1) and (2) as constraints, and distributes each of the order requests O1 to On. Assign users. In the following formula (2), “a_i_j” is the number of distribution target users having the user attribute pattern Pj assigned to the order request Oi, and “c_j” is the distribution target user having the user attribute pattern Pj. Is the number of

The unassigned score KPI is a score based on the number of delivery target users not assigned to the order requests O1 to On, and is the total number of appropriateness determination attributes of the delivery target users not assigned to the order requests O1 to On. . When the number of unassigned users is “p”, an unassigned user Fk (1 ≦ k ≦ p; p is an integer) is defined, and the number of appropriateness determination attributes of the unassigned user Fk is “f_k”. Then, the unassigned score KPI is obtained by the following equation (3).

  When all the user attributes are set as the appropriateness determination attributes, “user attribute” which is the sum of “number of user attributes” of the distribution target users whose “allocation state” is set to “0” in the future log table Is the unassigned score KPI.

  For example, in the future log table shown in FIG. 6, the distribution target users with distribution target IDs “F1”, “F3”, “F4”, and “F6” are not assigned to the order requests O1 to On, and other distribution targets It is assumed that the user is assigned to order requests O1 to On.

  In this case, as shown in FIG. 7, the “number of user attributes” of the distribution target users corresponding to the distribution target IDs “F1”, “F3”, “F4”, and “F6” are “3” and “4”, respectively. ”,“ 3 ”, and“ 4 ”. Therefore, the “total number of user attributes” that is the total value of the “number of user attributes” of the unassigned users Fk not assigned to the order requests O1 to On is “13 (= 3 + 4 + 3 + 4)” as shown in FIG. The unassigned score KPI is “13”.

  The allocating unit 24 calculates the “total number of user attributes” of the unallocated user Fk as the unallocated score KPI, but determines the “total number of user attribute patterns” of the unallocated user Fk as the unallocated score KPI. You can also.

For example, the assigning unit 24 can obtain the number “e_j” of unassigned users Fk having the user attribute pattern Pj for each user attribute pattern Pj, and use the sum of the numbers as the unassigned score KPI. For example, the assigning unit 24 can obtain the unassigned score KPI using the following formula (4).

  As described above, the allocating unit 24 unassigns the distribution target users to the plurality of order requests O1 to On based on the user attributes and the number of distributions specified by the plurality of order requests O1 to On, respectively. This is performed so that the number of predetermined user attributes of the distribution target users is maximized.

  As a result, it is possible to increase the options for assigning the distribution target users to the order requests O1 to On and increase the possibility of accepting future order requests O1 to On, thereby suppressing the loss of sales opportunities. .

  Here, an example will be described in which the assigning unit 24 performs the assigning process using the linear programming method so that the above formulas (1) and (2) are the constraint conditions and the unassigned score KPI is maximized. FIG. 8 is a diagram illustrating an example of the relationship between the attribute content of the user attribute pattern Pj and the number of users c_j.

  As shown in FIG. 8, it is assumed that the user attribute patterns P1 to P6 are “car”, “cosmetics”, “clothes”, “travel”, “car & travel”, and “cosmetics & clothing”, respectively. Further, it is assumed that the number of distribution target users is set as shown in FIG. As illustrated in FIG. 8, c_1 to c_6 that are the number of distribution target users having the user attribute pattern are “9020”, “5302”, “2319”, “4287”, “4030”, and “2900”. To do.

  When the user attribute pattern is defined by {car, makeup, clothes, travel,...}, The user attribute patterns P1 to P6 are {1, 0, 0, 0}, {0, 1, 0, 0}, respectively. , {0, 0, 1, 0}, {0, 0, 0, 1}, {1, 0, 0, 1}, {0, 1, 1, 0}. Therefore, for example, the distribution target user of the user ID “F8” shown in FIG. 6 is set to the user attribute pattern P1 of {1, 0, 0, 0}, and the distribution target of the user ID “F3” shown in FIG. The user is set to the user attribute pattern P5 which is {1, 0, 0, 1}.

  The allocation unit 24 uses the linear programming method so as to maximize the unallocated score KPI with the above formulas (1) and (2) as constraints, and “a_i_j” is the allocation of the distribution target user to the order request. Set. If the relationship between the attribute contents of the user attribute pattern Pj and the number of users c_j is as shown in FIG. 8, and the advertisement information table is as shown in FIG. 3, the setting of “a_i_j” is as shown in FIG. Become. FIG. 9 is a diagram illustrating an example of assignment of “a_i_j”.

  It should be noted that the above formula (1) has a constraint that the total number of distribution target users assigned to each order request Oi matches the specified number of distributions in the order request. For example, in the order request O1 illustrated in FIG. 9, a_1_1 to a_1_6 are set so that the designated delivery number o_1 is “100” and the sum of a_1_1 to a_1_6 is “100”.

  Further, the above equation (2) indicates that in each user attribute pattern Pj, the number of distribution target users assigned to the order request Oi does not exceed the number of users c_j that is the number of distribution target users having the user attribute pattern Pj. It is a constraint condition. For example, in the user attribute pattern P1, as shown in FIG. 8, a_1_1 to a_n_1 are set in a range where the number of users c_j is “9020” and the sum of a_1_1 to a_n_1 is equal to or less than “9020”.

  When the first allocation mode is set, the allocation unit 24 allocates a distribution target user to all order requests set in the advertisement information table every time a new order request is received by the reception unit 23. Repeat the whole process. In the first allocation mode, all distribution target users set in the future log table are targeted.

  For example, it is assumed that a new order request O11 is received by the receiving unit 23 after a distribution target user is assigned to the order requests O1 to O10. In this case, the allocating unit 24 maximizes the unallocated score KPI from the plurality of distribution target users set in the future log table with the above formulas (1) and (2) as constraints as to the order requests O1 to O11. The distribution target user is assigned as follows.

  As described above, in the first allocation mode, the entire process of allocating the distribution target user for all the order requests set in the advertisement information table is repeated, so that the distribution target user is appropriately allocated to the entire order request. Can do.

  In addition, when the second allocation mode is set, the allocation unit 24 maintains the allocation to the order request to which the distribution target user is allocated, and the distribution target user allocation partial process for the newly acquired order request I do. In the second allocation mode, the distribution target user allocated to the new order request is a distribution target user not allocated to the order request among the distribution target users set in the future log table.

  For example, it is assumed that a new order request O11 is received by the receiving unit 23 after a distribution target user is assigned to the order requests O1 to O10. In this case, the assigning unit 24 performs the distribution target user assignment partial process for the order request O11 while maintaining the assignment of the delivery target user to the order requests O1 to O10. That is, the allocating unit 24 allocates the distribution target user to a_11_j without changing the distribution target user allocation to a_1_j to a_10_j.

  As described above, in the second allocation mode, every time a new order request is received by the receiving unit 23, the distribution target for the new order request is maintained while maintaining the allocation to the order request to which the distribution target user is allocated. Assign users. Therefore, it is possible to reduce the load of the assignment process while appropriately assigning the distribution target user to the new order request.

  Further, when the third allocation mode is set, the allocation unit 24 performs the entire process at a predetermined timing (for example, at night when there are few order requests) while performing a partial process on a new order request.

  In the third allocation mode, the allocation unit 24 quickly performs the allocation process of the distribution target user for the new order request every time a new order request is received by the reception unit 23. Updates can be made quickly. As a result, the next new order request can be quickly received by the receiving unit 23.

  In addition, in the third allocation mode, the allocation unit 24 repeats the entire process of allocating the distribution target user for all the order requests set in the advertisement information table, as in the first allocation mode. The distribution target user can be appropriately assigned to the entire request.

  As described above, in the third allocation mode, the distribution target user is allocated by the partial process every time a new order request is received by the receiving unit 23, and further, the distribution target user is reassigned by the entire process at a predetermined timing. I do. Thereby, it is possible to achieve both the quickness and accuracy of the allocation of the distribution target user to the order request.

  In the third allocation mode, the timing at which the entire process is performed can be set in the allocation unit 24 by the distribution manager from the operation unit (not shown) of the advertisement distribution management device 1. For example, the distribution manager can set the allocation unit 24 to perform the above-described overall process every time the timing of a predetermined period arrives (for example, every night at 2 o'clock). Also, the distribution manager performs the above-described overall processing every time the number of new order requests reaches a predetermined number or every time the total number of users of new order requests reaches a predetermined number. Can be set to

  In the above-described embodiment, distribution target users are allocated to a plurality of order requests using linear programming so that the number of attributes for determining appropriateness of distribution target users not allocated to order requests is maximized. However, methods other than linear programming can be used. That is, the number of appropriateness determination attributes of the distribution target users not allocated to the order request is maximized based on the specified user attribute and the specified number of distributions of each order request for the allocation of the distribution target users to a plurality of order requests. As long as it is performed in this way, it is not limited to linear programming.

[3. Processing flow of advertisement distribution management device 1]
Next, the procedure of the advertisement distribution management process by the advertisement distribution management apparatus 1 according to the embodiment will be described. FIG. 10 is a diagram illustrating an example of an advertisement distribution management process flow of the advertisement distribution management apparatus 1 according to the embodiment.

  As shown in FIG. 10, when receiving the order request from the advertiser terminal 3, the control unit 15 of the advertisement distribution management device 1 places an order from the advertiser terminal 3 based on the future log set in the future log table. It is determined whether or not the request is accepted (step S11). For example, the control unit 15 counts the number of user attributes specified in the order request among the user attributes of unassigned users set in the future log table, and the count result is equal to or greater than the number of distributions specified in the order request. In some cases, it is determined that an order request is accepted.

  When it determines with receiving the order request from the advertiser terminal 3 (step S11; Yes), the control part 15 sets the information contained in an order request to an advertisement information table, and updates an advertisement information table (step S12). . Note that the information included in the order request includes a user attribute, a distribution period, the number of distributions, advertisement information, and the like, and these pieces of information are set in the advertisement information table.

  Next, the control unit 15 performs distribution target user assignment processing (step S13). This allocation process is a process for allocating a distribution target user having a user attribute specified in the order request to the order request, and is, for example, the process of steps S21 to S27 shown in FIG.

  Next, the control unit 15 updates the future log table based on the result of the allocation process in step S13 (step S14). Specifically, the control unit 15 sets the “allocation state” of the distribution target user assigned to the order request among the distribution target users set in the future log table to “1”, and The “allocation state” of the distribution target user who is not allocated is set to “0”.

  Thereafter, the control unit 15 transmits an order acceptance response to the advertiser terminal 3 via the communication unit 10 (step S15), and ends the advertisement distribution management process. On the other hand, when it determines with not accepting the order request from the advertiser terminal 3 in step S11 (step S11; No), the control part 15 transmits an order rejection response to the advertiser terminal 3 via the communication part 10. (Step S16), the advertisement distribution management process is terminated.

  Next, an example of the assignment process in step S13 will be specifically described with reference to FIG. FIG. 11 is a diagram illustrating an example of an allocation process flow of the advertisement distribution management device 1 according to the embodiment.

  As illustrated in FIG. 11, when starting the distribution target user allocation process, the control unit 15 of the advertisement distribution management device 1 determines whether or not the first allocation mode is set (step S21). If it determines with having set to the 1st allocation mode (step S21; Yes), the control part 15 will allocate the delivery object user by the whole process mentioned above (step S22).

  When it is determined in step S21 that the first allocation mode is not set (step S21; No), the control unit 15 determines whether or not the second allocation mode is set (step S23). When it is determined that the second allocation mode is set (step S23; Yes), the control unit 15 allocates a distribution target user by the partial processing described above to a new order request (step S24).

  If it is determined in step S23 that the second allocation mode is not set (step S23; No), the control unit 15 determines that the third allocation mode is set, and the new order request is described above. The distribution target user is assigned by the partial processing (step S25).

  When the process of step S25 ends, the control unit 15 determines whether or not the number of order requests received after the previous overall process has reached a predetermined number (step S26). When it is determined that the number of order requests has reached a predetermined number (step S26; Yes), the control unit 15 assigns a distribution target user by the above-described overall processing (step S27).

  When it is determined that the number of order requests is not the predetermined number (step S26; No), or when the processes of steps S22, S24, and S27 are finished, the control unit 15 finishes the distribution target user assignment process.

[4. (Modification)
In the above-described embodiment, the accepting unit 23 executes the accepting process every time there is an order request from the advertiser terminal 3. For example, when there is no duplication of user attributes specified in the order request, a plurality of order requests Acceptance processing can also be performed in units. In addition, for example, when there is no duplication of user attributes specified in the order request, the reception unit 23 can perform reception processing when the user attributes specified in the order request no longer overlap.

  Further, in the above-described embodiment, in order to make the explanation easy to understand, the target period for predicting the future advertisement distribution destination and the distribution period of the order request are described as the same. However, the target period for predicting the future advertisement distribution destination may not be the same as the distribution period for the order request. For example, the target period for predicting the future advertisement distribution destination is set to one day, and the distribution period for the order request. May be set on a daily basis. In this case, for example, the assigning unit 24 can perform the assigning process with a distribution target user on a daily basis.

  In the above-described embodiment, the processing for the order request has been described. However, the control unit 15 can also receive a cancel request from the advertiser terminal 3. In this case, the allocating unit 24 executes the entire process so that the unallocated score KPI is maximized in response to the new advertisement delivery cancellation request received by the receiving unit 23.

  In the above-described embodiment, the distribution manager sets the first to third allocation modes. However, the first to third allocation modes can be changed according to the load state in the advertisement distribution management device 1. For example, the control unit 15 executes the second allocation mode when the processing load of the advertisement distribution management device 1 is high, and executes the first allocation mode when the processing load of the advertisement distribution management device 1 is low.

  In the above-described embodiment, the control unit 15 uses the entire processing in the first and third allocation modes as a single processing unit for all order requests. The above-described overall processing may be performed using one as a processing unit.

  Further, the control unit 15 assigns the distribution target users by another allocation method until the number of unassigned users becomes equal to or less than the predetermined number, and when the number of unassigned users becomes equal to or less than the predetermined number, First to third allocation modes can be executed. As “another allocation method”, the control unit 15 creates a new advertisement distribution so that a reduction in the total number of user attributes of distribution target users not allocated to an advertisement distribution order request among a plurality of distribution target users is minimized. A distribution target user is assigned to the order request. With this “other allocation method”, it is possible to reduce the processing load as compared with the first to third allocation modes.

  In the above-described embodiment, the control unit 15 assigns the distribution target users to the plurality of order requests so that the unassigned score obtained based on the unassigned users is maximized, but the unassigned score is the maximum. The allocation process may be performed so that For example, the control unit 15 may be able to assign delivery target users to a plurality of order requests so that the assigned score obtained based on the delivery target users assigned to the order request is minimized. The assigned score is a total number of distribution target users having a predetermined user attribute and assigned to the order request. Such an assigned score can be obtained, for example, by subtracting the assigned score from the total number of distribution target users having predetermined user attributes.

  In the above-described embodiment, the distribution target user is assigned to all order requests set in the advertisement information table as the overall processing. However, the distribution target user is assigned to some order requests. Also good. For example, each time a new order request is accepted by the accepting unit 23, the assigning unit 24 assigns a distribution target user to a predetermined number of order requests with a new acceptance time among the order requests set in the advertisement information table. In this case, the assigning unit 24 assigns the distribution target user to all order requests set in the advertisement information table at a predetermined timing (for example, every night at 2 o'clock).

[5. effect〕
In the advertisement distribution management device 1 according to the embodiment, the future log DB 14 (corresponding to an example of a prediction information storage unit) stores information including user attributes of a plurality of distribution target users predicted as future advertisement distribution destinations, The receiving unit 23 (corresponding to an example of an acquisition unit) acquires a plurality of order requests specifying user attributes and the number of distributions, and the allocation unit 24 (corresponding to an example of an allocation unit) is a distribution target user for a plurality of order requests. The number of appropriateness determination attributes (predetermined user attributes) of distribution target users not allocated to the order request is maximized based on the user attributes and the number of distributions specified in each of the plurality of order requests. Do as follows.

  With this configuration, distribution target users can be appropriately assigned to advertisement distribution orders, and a decrease in the possibility of accepting future advertisement distribution order requests can be suppressed, so that loss of sales opportunities can be suppressed. .

  In addition, in the advertisement distribution management device 1 according to the embodiment, the allocating unit 24 creates a new one based on the user attribute and the number of distributions specified in each of a plurality of order requests including the new order request acquired by the receiving unit 23. The distribution target user is allocated to the order request while maintaining the allocation to the order request to which the distribution target user is allocated.

  With this configuration, it is possible to reduce the load of the allocation process while appropriately allocating the distribution target user to a new order request.

  Further, in the advertisement distribution management device 1 according to the embodiment, every time the allocation unit 24 acquires a new order request by the receiving unit 23, the distribution target user for a plurality of order requests including the new order request. Make assignments.

  With such a configuration, it is possible to appropriately assign a distribution target user to the entire order request, and it is possible to accurately suppress a decrease in acceptability of a future advertisement distribution order request.

  In the advertisement distribution management device 1 according to the embodiment, the allocation unit 24 reallocates the distribution target users at a predetermined timing with respect to a plurality of order requests to which the distribution target users are allocated.

  With such a configuration, it is possible to achieve both the quickness and accuracy of allocation of distribution target users to order requests.

  Further, in the advertisement distribution management device 1 according to the embodiment, the allocating unit 24 allocates a distribution target user with a combination of two or more user attributes having a frequency specified by an advertisement distribution order request equal to or greater than a threshold as one user attribute. .

  With this configuration, it is possible to assign user attributes to the order request by a combination of two or more user attributes whose frequency specified in the order request is equal to or greater than a threshold value, thereby increasing the possibility of accepting future advertisement distribution orders. Increase sales and reduce lost sales opportunities.

[6. Others]
As described above, some of the embodiments of the present application have been described in detail with reference to the drawings. However, these are merely examples, and various modifications, including the aspects described in the disclosure section of the invention, based on the knowledge of those skilled in the art, It is possible to implement the present invention in other forms with improvements.

  The above-described advertisement distribution management device 1 may be realized by a plurality of server computers, and depending on the function, an external platform or the like may be realized by calling an API (Application Programming Interface) or network computing. Can be changed flexibly. Further, the advertisement distribution management device 1 and the information distribution device 2 may be realized by a single server computer.

DESCRIPTION OF SYMBOLS 1 Advertisement delivery management apparatus 2 Information delivery apparatus 3 Advertiser terminal 4 User terminal 5 Communication network 10 Communication part 11 Advertisement information DB
12 User information DB
13 Past log DB
14 Future log DB
21 Advertising Distribution Unit 22 Future Log Prediction Unit 23 Reception Unit 24 Allocation Unit

Claims (5)

Prediction information storage means for storing information including user attributes of a plurality of distribution target users predicted as future advertisement distribution destinations;
An acquisition means for acquiring a plurality of order requests specifying user attributes;
A process of allocating the distribution target user to the plurality of order requests including the new order request and the order request acquired in the past by the acquiring unit every time a new order request is acquired by the acquiring unit. Assigning means for performing
An advertisement distribution management device comprising: The acquisition means includes
Get multiple order requests that specify user attributes and number of deliveries,
The assigning means includes
Based on the user attributes and the number of distributions specified in each of the plurality of order requests, the number of predetermined user attributes of distribution target users not allocated to the order request is maximized with respect to the plurality of order requests. The advertisement distribution management apparatus according to claim 1, wherein the distribution target user is allocated. The assigning means includes
The advertisement distribution management device according to claim 1, wherein the distribution target user is allocated to the plurality of order requests using a linear programming method. The assigning means includes
The distribution target user is assigned as one user attribute by combining a combination of two or more user attributes whose frequency specified in an advertisement distribution order from an advertiser is a threshold value or more. The advertisement distribution management device according to one. An advertisement delivery management method executed by a computer,
Obtaining multiple order requests with user attributes;
Each time a new order request is acquired, based on the user attribute information stored in a storage unit that stores user attribute information of a plurality of distribution target users predicted as future advertisement distribution destinations, Performing a process of assigning the distribution target user to the plurality of order requests including the new order request and the order request acquired in the past from among the distribution target users ;
An advertisement delivery management method comprising:

Images