Information processing device, information processing method, and information processing program
The information processing device calculates advertisement scores by analyzing presentation order and user attributes, addressing the challenge of attributing conversion impact, thereby improving advertising efficiency.
Patent Information
- Application Number
- JP2023083109
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-19
- Publication Date
- 2024-11-29
Smart Images

Figure 2024166773000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to an information processing device, an information processing method, and an information processing program. [Background technology]
[0002] In recent years, advertisement distribution via the Internet has become popular. For example, a technique has been provided that calculates the contribution of an advertisement provided (distributed) to a user to a conversion by the user. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Special Publication No. 2017-529587 Summary of the Invention [Problem to be solved by the invention]
[0004] However, the above conventional technology does not necessarily allow for proper calculation of a score related to an advertisement. For example, the above conventional technology assigns a contribution amount to an interaction for a conversion associated with an advertisement. Therefore, when there are multiple combinations of advertisements provided, it is difficult to properly estimate the contribution amount of each advertisement to a conversion. Therefore, it is desired to properly calculate information regarding the contribution of each advertisement even when there is a combination of advertisements provided.
[0005] The present application has been made in consideration of the above, and has an object to provide an information processing device, an information processing method, and an information processing program capable of appropriately calculating a score related to an advertisement. [Means for solving the problem]
[0006] The information processing device according to the present disclosure includes an acquisition unit and a calculation unit. The acquisition unit acquires provided advertisement information including information on a plurality of advertisements provided to a user until the user makes a conversion, the information including information on an order in which the plurality of advertisements are provided to the user. The calculation unit calculates a score indicating a relative contribution of each of the plurality of advertisements to the conversion based on the provided advertisement information acquired by the acquisition unit. Effect of the Invention
[0007] According to one aspect of the embodiment, it is possible to provide an effect that a score related to an advertisement can be appropriately calculated. [Brief description of the drawings]
[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of information processing according to an embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an example of the configuration of an information processing system according to the embodiment. [Diagram 3] FIG. 3 is a diagram illustrating an example of the configuration of the information processing device according to the embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of a user information table stored in the user information storage unit of the information processing device according to the embodiment. [Diagram 5] FIG. 5 is a diagram illustrating an example of an advertisement information table stored in the advertisement information storage unit of the information processing device according to the embodiment. [Figure 6] FIG. 6 is a diagram illustrating an example of an advertisement provision history table stored in the advertisement provision history storage unit of the information processing device according to the embodiment. [Figure 7] FIG. 7 is a flowchart showing an example of information processing by the processing unit of the information processing device according to the embodiment. [Figure 8] FIG. 8 is a hardware configuration diagram showing an example of a computer that realizes the functions of the information processing device according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0009] Hereinafter, the information processing device, the information processing method, and the information processing program according to the present application (hereinafter, referred to as "embodiments") will be described in detail with reference to the drawings. Note that the information processing device, the information processing method, and the information processing program according to the present application are not limited to these embodiments. In addition, each embodiment can be appropriately combined as long as the processing contents are not contradictory. In addition, the same parts in each of the following embodiments are given the same reference numerals, and duplicated explanations will be omitted.
[0010] [1. An example of information processing] First, an example of information processing according to the embodiment will be described with reference to Fig. 1. Fig. 1 is a diagram showing an example of information processing according to the embodiment, which is executed by an information processing device 1.
[0011] The information processing device 1 calculates a score indicating the relative contribution of each advertisement provided to the user who has reached the conversion, based on provided advertisement information including information on a plurality of advertisements provided to the user before the user reaches the conversion and information on the order in which the plurality of advertisements are provided to the user. The conversion is, for example, a user behavior targeted by an advertisement, and may be any user behavior.
[0012] For example, a conversion may be, but is not limited to, a purchase of a transaction object that is the subject of an advertisement (hereinafter, sometimes referred to as an "advertised object"), use of a transaction object that is the subject of an advertisement, registration for a membership service that is the subject of an advertisement, or viewing (displaying) of a site that is the subject of an advertisement. A transaction object may be, for example, a product or a service.
[0013] For example, a conversion may be signing up for an advertised email distribution, registering an account for an advertised online service, requesting materials for which advertising is being made, visiting a advertised physical store, or sharing on an advertised social networking service (SNS), or it may be some other action.
[0014] What kind of user behavior is regarded as a conversion can be arbitrarily set. For example, what kind of user behavior is regarded as a conversion for each advertisement is individually set by an administrator of the information processing device 1, an advertiser, or the like.
[0015] An advertisement is, for example, content for advertising an advertising target, and is also called advertising content. Examples of such advertisements include, but are not limited to, banner advertisements, listing advertisements, native advertisements, social media advertisements, and video advertisements.
[0016] An example of information processing will be described below with reference to Fig. 1. In Fig. 1, a user's purchase of product X is taken as an example of a conversion, and the information processing device 1 calculates a score indicating the relative contribution of each advertisement by taking the number of purchases of product X as the number of conversions.
[0017] In the following, a user who has achieved conversion may be referred to as a "relevant user." For simplicity, FIG. 1 will be described using an example in which three advertisements, Ads A, B, and C, are provided. For example, Ads A, B, and C are advertisements that advertise Product X and are included in one advertising campaign, but are not limited to this example.
[0018] 1 illustrates the calculation of the score when at most three advertisements, Ads A, B, and C, are provided to the user, but the number of advertisements provided to the user may be four or more, or may be two. Ad A is, for example, a broad distribution advertisement, Ad B is an interest targeting advertisement, and Ad C is, for example, a retargeting advertisement, but is not limited to such examples.
[0019] Broadband advertising is, for example, advertising provided without narrowing down to a specific target group. Interest-targeted advertising is, for example, advertising provided based on a user's interests and concerns. Retargeting advertising is, for example, advertising provided to users who have visited a website or web page that is the subject of an advertisement. Providing an advertisement means, for example, presenting the advertisement to a user, but may also include transitioning to a link destination of the advertisement by the user selecting (for example, clicking) the advertisement.
[0020] First, the information processing device 1 acquires from an internal storage unit or the like advertisement information including multiple pieces of time series information, which is information on the order in which advertisements provided to a user who has achieved conversion are provided to the user (step S1).
[0021] In step S1, the information processing device 1 acquires advertisement information for each conversion, for example, time-series information that is information on the order of advertisements provided to the user from the time the user purchased the product X until a predetermined period TA ago, and information on the user. The information on the order of advertisements is, for example, information on the order of advertisements A, B, and C provided to the user.
[0022] The predetermined period TA is, for example, one week, but is not limited to this example, and can be set appropriately depending on, for example, the attributes of the user, the category of the product X, the content of the conversion, and the like.
[0023] Each piece of time-series information included in the provided advertisement information is, for example, information indicating the order in which one or more advertisements are provided to a corresponding user. Hereinafter, the order in which advertisements are provided may be referred to as an advertisement providing order.
[0024] For example, if advertisements X, Y, and Z are provided to the user in this order from the time the user purchases product X until a certain period TA, the advertisement provision order is advertisement X, advertisement Y, and advertisement Z, and may be described below as CV(X→Y→Z). Advertisements X, Y, and Z are, for example, advertisements A, B, and C, but some or all of advertisements X, Y, and Z may be the same advertisement.
[0025] Furthermore, if one or more advertisements provided to a user in a specified period TA are in the order of advertisement X, Y, the advertisement provision order is advertisement X, advertisement Y, which may be hereinafter referred to as CV(X→Y). If one or more advertisements provided to a user in a specified period TA are only advertisement X (advertisement X is any of advertisements A, B, and C), the advertisement provision order is only advertisement X, which may be hereinafter referred to as CV(X).
[0026] The information on the user included in the provided advertisement information includes, for example, information indicating the attributes of the user. The attributes of the user are, for example, demographic attributes of the user, but may include psychographic attributes of the user instead of or in addition to the demographic attributes.
[0027] The information processing device 1 can convert, for example, the advertisement provision history stored in the internal storage unit into time-series information in a predetermined format for each conversion. The time-series information in the predetermined format includes, for example, information indicating the order in which advertisements were provided.
[0028] Here, it is assumed that the prediction model generated in step S2 is a recurrent neural network (RNN) such as a long short-term memory (LSTM) or a gated recurrent unit (GRU). In this case, the time-series information in a predetermined format is information including, for example, user attribute information and unit information including information indicating whether each advertisement is provided, in chronological order.
[0029] For example, the unit information is information such as [user age, user gender, whether advertisement A is provided, whether advertisement B is provided, whether advertisement C is provided], but the user attributes may include things other than age and gender. "User age" is, for example, a normalized value of the user's age, and "user gender" is "0" for female and "1" for male. "Whether advertisement is provided" is "1" if advertisement is provided, and "0" if advertisement is not provided.
[0030] For example, suppose that the maximum number of advertisements provided to each relevant user is 3, and the order of advertisements provided to the one relevant user is advertisement A, advertisement B, and advertisement C. Also suppose that the normalized value of the one relevant user's age is 0.31, and the one relevant user's gender is female. In this case, the time series information of the one relevant user includes information [0.31, 0, 1, 0, 0], information [0.31, 0, 0, 1, 0], and information [0.31, 0, 0, 0, 1] in that order (forward or reverse).
[0031] The time-series information of the user for which the number of advertisements provided is less than 3 is, for example, information that has been subjected to padding processing, but is not limited to such an example. The time-series information of the user is not limited to the above-mentioned example, and may be, for example, information indicating the presence or absence of the user by age group, and the user's gender may be information indicating the presence or absence of the user by gender.
[0032] In addition, a case will be described in which the prediction model generated in step S2 is a linear regression model, a GBDT (Gradient Boosting Decision Tree), etc. In this case, the time-series information in a predetermined format is information including, for example, user attribute information and information indicating the order in which each advertisement is presented.
[0033] The information indicating the order of presentation of each advertisement may be, for example, information such as [Presentation order 1 (Advertisement A, Advertisement B, Advertisement C),...,Presentation order n (Advertisement A, Advertisement B, Advertisement C)], where n is, for example, the maximum number of advertisements presented to each corresponding user.
[0034] "Presentation order 1 (Ad A, Ad B, Ad C)" is information indicating whether the last advertisement presented before the conversion by the corresponding user was Ad A, B, or C. "Presentation order n (Ad A, Ad B, Ad C)" is information indicating whether the first advertisement presented before the conversion by the corresponding user was Ad A, B, or C.
[0035] For example, in the case of advertisement A that was presented last before the conversion by the user, "presentation order 1 (ad A, ad B, ad C)" is information such as [1, 0, 0]. Also, in the case of advertisement C that was presented first before the conversion by the user, "presentation order n (ad A, ad B, ad C)" is information such as [0, 0, 1].
[0036] Next, the information processing device 1 generates a prediction model for predicting the number of conversions according to the order in which advertisements are provided to the user, based on the provided advertisement information acquired in step S1 (step S2).
[0037] The prediction model generated in step S2 is expressed by, for example, the following formula (1). In the following formula (1), the feature amount is the above-mentioned time-series information. In the following formula (1), pCV is the predicted number of conversions. f(feature) = pCV (1)
[0038] In the process of step S2, the information processing device 1 calculates the number of conversions for each advertisement presentation order for each user attribute based on the provided advertisement information acquired in step S1. For example, the number of conversions achieved when the advertisement presentation order is advertisement A, advertisement B, and advertisement C is calculated for each user attribute by aggregating the number of conversions achieved when the advertisement presentation order is advertisement A, advertisement B, and advertisement C.
[0039] For example, when the user's attribute information is information on the user's gender and age, each user attribute is represented by a combination of multiple attribute items such as gender and age. In this case, the combination of attribute items is, for example, male in his teens, female in his teens, male in his twenties, female in his twenties, etc., but is not limited to such an example, and the attribute items included in the user's attribute information may include region, occupation, annual income, etc., and may also include other attribute items.
[0040] Then, the information processing device 1 generates learning information including time-series information and information indicating the number of conversions as feature quantities for each advertisement presentation order, and generates a prediction model using the learning information. The feature quantities used in the prediction model are not limited to the above-mentioned examples, and may include, for example, advertisement display environment information for advertisements A, B, and C. The advertisement display environment information is, for example, information such as the category of the website on which the advertisement is provided, the type of application on which the advertisement is provided, and the display position of the provided advertisement.
[0041] When the prediction model is a linear regression model, GBDT, etc., the time-series information may include, instead of or in addition to the above examples, statistics such as count values, average values, minimum values, and maximum values as feature quantities. For example, in a given period TA, advertisement A is provided twice, advertisement B is provided once, advertisement A is provided x days ago (x is an integer), advertisement B is provided y days ago (y is an integer), etc. can each be a feature quantity.
[0042] Next, the information processing device 1 uses the prediction model generated in step S2 to calculate attribution scores that are scores indicating the relative contributions of the advertisements A, B, and C (step S3). The information processing device 1 calculates the attribution scores of the advertisements A, B, and C for each combination of the advertisement presentation order and the user attribute, for example, using attribute information for each user attribute and information on the advertisement presentation order.
[0043] For example, the information processing device 1 inputs user attribute information and information on the advertisement presentation order into a prediction model, and performs a calculation process to calculate the attribution score of each advertisement A, B, and C for each combination of the advertisement presentation order and user attribute, using a process such as obtaining the predicted number of conversions output from the prediction model.
[0044] For example, assume that the prediction model generated in step S2 is an RNN such as LSTM or GRU. In this case, the information processing device 1 inputs time-series information including unit information, in chronological order, including user attribute information and information indicating whether each advertisement is provided, to the prediction model, and calculates output values of a softmax function in the prediction model at this time as the attribution scores of advertisements A, B, and C.
[0045] The output value of the softmax function is obtained, for example, by applying the softmax function to the output of hierarchical attention, and is represented as a probability distribution of advertisements A, B, and C. Hierarchical attention is an example of an attention mechanism. In the example shown in Figure 1, the attribution score of advertisement A is "0.2", the attribution score of advertisement B is "0.3", and the attribution score of advertisement C is "0.5".
[0046] Also, it is assumed that the prediction model generated in step S2 is a classification model such as a linear regression model or GBDT. In this case, the information processing device 1 inputs time-series information including user attribute information and information on the advertisement presentation order to the prediction model, and performs a process of acquiring pCV, which is a predicted number of conversions for the advertisement presentation order, from the prediction model for each combination of the advertisement presentation order and the user attribute.
[0047] Then, the information processing device 1 performs a process of calculating an attribution score of each of the advertisements A, B, and C for each combination of the advertisement presentation order and the user attribute, using the pCV for each combination of the advertisement presentation order and the user attribute.
[0048] For example, assume that the advertisements are presented in the order of advertisement A, advertisement B, and advertisement C. In this case, the information processing device 1 performs the calculations of the following formulas (2) to (4). v(A) is a function related to the contribution degree of advertisement A, v(B) is a function related to the contribution degree of advertisement B, and v(C) is a function related to the contribution degree of advertisement C. v(A)={f(φ)+(f(A)-f(φ)) ···(2) v(B)={f(φ)+(f(A→B)-f(A)) ···(3) v(C)={f(φ)+(f(A→B→C)-f(A→B)) ···(4)
[0049] In the above formulas (2) and (3), f(φ) is the pCV when none of advertisements A, B, and C are provided to the user, and f(A) is the pCV when only advertisement A is provided to the user. In the above formulas (3) and (4), f(A→B) is the pCV when advertisement A and advertisement B are provided to the user in that order. In the above formula (4), f(A→B→C) is the pCV when advertisement A, advertisement B, and advertisement C are provided to the user in that order. Note that "φ" in f(φ) means the empty set.
[0050] In this way, the functions v(A), v(B), and v(C) are expressed in a form that includes an equation for subtracting the pCV in an advertisement presentation sequence that does not include one advertisement from the pCV in an advertisement combination in which one advertisement is added to an advertisement presentation sequence that does not include one advertisement.
[0051] The information processing device 1 performs the calculations of the following formulas (5) to (8) based on the calculation results of the above formulas (2) to (4). Sc(A) is the attribution score of advertisement A, Sc(B) is the attribution score of advertisement B, and Sc(C) is the attribution score of advertisement C. v(A,B,C)=v(A)+v(B)+v(C) ···(5) Sc(A)=v(A) / v(A,B,C) (6) Sc(B)=v(B) / v(A,B,C) (7) Sc(C)=v(C) / v(A,B,C) (8)
[0052] The information processing device 1 performs the above processes (2) to (8) for each user attribute. The information processing device 1 also calculates the attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C for advertisement presentation orders other than the advertisement presentation order of advertisement A, advertisement B, and advertisement C in the same manner. In this manner, the information processing device 1 performs a calculation process to calculate the attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C for each combination of the advertisement presentation order and the user attribute. Hereinafter, when multiple attribution scores including the attribution scores Sc(A), Sc(B), and Sc(C) are referred to without distinction, they may be referred to as the attribution score Sc.
[0053] Since the number of conversions for each ad serving sequence includes organic conversions, it may be desirable to calculate the attribution scores Sc(A), Sc(B), and Sc(C) using the number of conversions excluding organic conversions.
[0054] An organic conversion is a conversion that occurs as a result of a user’s voluntary action even when no advertising is provided. In the example above, it is a conversion that occurs as a result of a user’s voluntary action even when no advertising A, B, or C is provided.
[0055] The information processing device 1 can also calculate attribution scores Sc(A), Sc(B), and Sc(C) based on conversions from which organic conversions have been removed. Conversions from which organic conversions have been removed can be called lift conversions.
[0056] The information processing device 1 can calculate lift conversions and calculate attribution scores Sc(A), Sc(B), and Sc(C) based on the lift conversions. For example, the information processing device 1 calculates lift conversions for each combination of an advertisement presentation order and a user attribute, and calculates attribution scores Sc(A), Sc(B), and Sc(C) for each combination of an advertisement presentation order and a user attribute based on the lift conversions.
[0057] The information processing device 1 can also estimate the control conversion as the organic conversion and calculate the lift conversion from the difference between the control conversion and the treat conversion.
[0058] A control conversion is, for example, a conversion in which advertisements A, B, and C are not included in the target of provision. A treat conversion is, for example, a conversion in which advertisements A, B, and C are included in the target of delivery. Below, the number of control conversions, which are an example of organic conversions, may be referred to as CtrlCV, and the number of treat conversions may be referred to as TreatCV.
[0059] The information processing device 1 can treat, for example, the total number of users to whom none of advertisements A, B, and C has been provided in the provided advertisement information acquired in step S1 during a predetermined period TA before conversion as CtrlCV.
[0060] In addition, the information processing device 1 can estimate the CtrlCV for each advertisement presentation order by controlling the delivery of advertisements A, B, and C using, for example, an advertisement evaluation method (for example, A / B testing) that is a method for evaluating the performance of an advertisement.
[0061] In the advertising evaluation method, for example, users who send advertising requests are divided into Treat users and Ctrl users, and delivery control is performed in which one of advertisements A, B, or C is delivered to Treat users, and none of advertisements A, B, or C is delivered to Ctrl users.
[0062] For example, when the prediction model is an RNN, the information processing device 1 calculates the weight w shown in the following formula (9). w=(1-TreatCV / CtrlCV) ···(9)
[0063] The information processing device 1 can adjust the attribution scores Sc(A), Sc(B), and Sc(C) by weighting the hierarchical attention using the weight w. For example, the information processing device 1 can calculate the attribution scores Sc(A), Sc(B), and Sc(C) excluding organic conversions by multiplying the output of the hierarchical attention or the attribution scores Sc(A), Sc(B), and Sc(C) by the weight w.
[0064] In this way, the information processing device 1 acquires provided advertisement information including information on a plurality of advertisements provided to a user before the user achieves conversion, and information on the order in which the plurality of advertisements are provided to the user. Then, the information processing device 1 calculates an attribution score Sc, which is a score indicating the relative contribution of each of the plurality of advertisements to the conversion, based on the acquired provided advertisement information. This allows the information processing device 1 to appropriately calculate the score related to the advertisements, reflecting the relative contribution of each of the plurality of advertisements.
[0065] The information processing device 1 can provide, for example, information on the calculated attribution score Sc. The information processing device 1 provides information on the attribution score Sc calculated for the advertiser's advertisement to an advertiser terminal used by the advertiser of advertisements A, B, and C. For example, the information processing device 1 transmits information indicating the attribution score Sc of the advertiser's advertisements A, B, and C to the advertiser terminal used by the advertiser. For example, the information processing device 1 transmits content indicating the attribution score Sc of advertisements A, B, and C to the advertiser terminal used by the advertiser of advertisements A, B, and C.
[0066] Furthermore, the information processing device 1 may make various decisions based on the calculated attribution score Sc. For example, the information processing device 1 may determine the allocation of an advertising budget as a method of using the attribution analysis.
[0067] For example, the information processing device 1 determines the budget allocation for advertisements A, B, and C based on the calculated attribution score Sc. For example, the information processing device 1 allocates a larger budget to an advertisement with a larger calculated attribution score Sc. For example, the information processing device 1 determines the ratio of the attribution score Sc of each advertisement to the total value of the attribution scores Sc of all advertisements subject to budget allocation as the ratio of the budget to be allocated to each advertisement.
[0068] Here, assume that the percentage of the attribution score Sc(A) is 20%, the percentage of the attribution score Sc(B) is 30%, and the percentage of the attribution score Sc(C) is 50%. In this case, the information processing device 1 can determine to allocate half (50%) of the total budget to advertisement C, 30% to advertisement B, and 20% to advertisement B.
[0069] The configuration of an information processing system including the information processing device 1 and user terminal 2 that perform such processing will be described in detail below.
[0070] 2. Configuration of the Information Processing System 2 is a diagram showing an example of a configuration of an information processing system according to an embodiment. As shown in FIG. 2, the information processing system 100 according to the embodiment includes an information processing device 1, a plurality of user terminals 2, and a plurality of advertiser terminals 3.
[0071] The multiple user terminals 2 are used by different users U. The multiple advertiser terminals 3 are used by different advertisers O. Each of the user terminals 2 and the advertiser terminals 3 is, for example, a notebook PC (Personal Computer), a desktop PC, a smartphone, a tablet PC, or a wearable device. The wearable device is, for example, smart glasses or a smart watch, but is not limited to such examples. The user U is a user who receives advertisements provided from an information processing device 1 or the like.
[0072] The information processing device 1, the user terminal 2, and the advertiser terminal 3 are connected to each other so as to be able to communicate with each other by wire or wirelessly via a network N. Note that the information processing system 100 shown in FIG. 2 may include a plurality of information processing devices 1 and the like.
[0073] The network N includes, for example, a Wide Area Network (WAN) such as the Internet and mobile communication networks such as Long Term Evolution (LTE), 4th Generation (4G), and 5th Generation (5G: fifth generation mobile communication system).
[0074] The user terminal 2 and the advertiser terminal 3 can connect to the network N via short-range wireless communication such as a mobile communication network, Bluetooth (registered trademark), or a wireless LAN (Local Area Network), and communicate with the information processing device 1.
[0075] 3. Configuration of information processing device 1 3 is a diagram showing an example of the configuration of the information processing device 1 according to the embodiment. As shown in FIG. 3, the information processing device 1 includes a communication unit 10, a storage unit 11, and a processing unit 12.
[0076] [3.1. Communication unit 10] The communication unit 10 is realized by, for example, a communication module or a NIC (Network Interface Card). The communication unit 10 is connected to a network N by wire or wirelessly, and transmits and receives information to and from various other devices. For example, the communication unit 10 transmits and receives information to and from the user terminal 2 and the advertiser terminal 3 via the network N.
[0077] [3.2. Storage section 11] The storage unit 11 is realized by, for example, a semiconductor memory element such as a random access memory (RAM) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 11 has a user information storage unit 20, an advertisement information storage unit 21, and an advertisement provision history storage unit 22.
[0078] 3.2.1. User information storage unit 20 The user information storage unit 20 stores various pieces of information related to the user U. Fig. 4 is a diagram showing an example of a user information table stored in the user information storage unit 20 of the information processing device 1 according to the embodiment.
[0079] 4, the user information table stored in the user information storage unit 20 includes information items such as "User ID (Identifier)," "Attribute Information," and "History Information." The "User ID" is an identifier for identifying a user U, and is information assigned to each user U.
[0080] The "attribute information" is attribute information indicating attributes of the user U associated with the "user ID." The attributes of the user U are, for example, demographic attributes, psychographic attributes, etc. The demographic attributes are demographic attributes, and include a plurality of attribute items such as age, sex, occupation, place of residence, annual income, and family structure.
[0081] The psychographic attributes are psychological attributes, and include, for example, a plurality of attribute items related to lifestyle, values, interests, etc. For example, each of the plurality of attribute items in the psychographic attributes is an object of interest of the user U, such as a car, clothes, travel, games, camping, motorcycle, train, home appliance, or computer.
[0082] The "history information" includes information on the behavior history of the user U associated with the "user ID." The behavior history of the user U includes, for example, information on the behavior history of the user U in the online service. The behavior history information of the user U in the online service includes, for example, search history information, browsing history information, delivery history information, and transaction history information in the online service of the user U.
[0083] The search history information of the user U includes, for example, information on the search history by the user U in a web search service. The browsing history information of the user U includes, for example, information on the browsing history of content by the user U in an online service. The distribution history information includes information on the distribution history of content distributed to the user U's user terminal 2. The transaction history information includes information on the transaction history of products by the user U in an online service.
[0084] 3.2.2. Advertising information storage unit 21 The advertisement information storage unit 21 stores various pieces of information related to advertisements received from the advertiser terminal 3. Fig. 5 is a diagram showing an example of an advertisement information table stored in the advertisement information storage unit 21 of the information processing device 1 according to the embodiment.
[0085] 5, the advertisement information table stored in the advertisement information storage unit 21 includes information items such as "advertisement ID", "advertisement type", "advertisement", "distribution period", and "budget". The "advertisement ID" is an identifier that identifies an advertisement, and is information that is assigned to each advertisement.
[0086] The "advertising type" is information indicating the type of advertisement associated with the "advertising ID." The types of advertisements include, but are not limited to, broadband advertisements, interest targeting advertisements, and retargeting advertisements.
[0087] An "advertisement" is an advertisement associated with an "advertising ID," and is, for example, a banner advertisement, a listing advertisement, a native advertisement, a social media advertisement, a video advertisement, etc., but is not limited to these examples.
[0088] "Distribution period" is information on the distribution period of the advertisement associated with the "advertising ID". "Budget" is information on the budget of the advertisement associated with the "advertising ID". Note that the information included in the advertisement information table is not limited to the above-mentioned information, and may also include, for example, advertiser identification information, advertising target information, campaign information, and the like.
[0089] The advertiser identification information is information that identifies the advertiser O of the advertisement associated with the "advertising ID", for example, an advertiser ID. The advertising target information is information on the advertising target that is the subject of the advertisement associated with the "advertising ID", for example, information on the transaction target that is the subject of the advertisement. The campaign information is information on the advertising campaign associated with the "advertising ID". Note that the advertising information storage unit 21 is not limited to the above and can store various types of information depending on the purpose.
[0090] [3.2.3. Advertisement provision history storage unit 22] The advertisement provision history storage unit 22 stores various information related to the distribution history of advertisements. Fig. 6 is a diagram showing an example of an advertisement provision history table stored in the advertisement provision history storage unit 22 of the information processing device 1 according to the embodiment.
[0091] 6, the advertisement provision history table stored in the advertisement provision history storage unit 22 includes information items such as "User ID", "Provided Advertisement", and "Conversion". The "User ID" is the same as the user ID shown in FIG. 4, and is information on the identifier of the user U to whom the advertisement is provided.
[0092] The "provided advertisement" includes, for example, an advertisement ID of an advertisement provided (distributed) to a user U associated with the "user ID", and information indicating the date and time the advertisement was provided to the user U. The "conversion" includes, for example, information indicating the content of a conversion performed by a user U associated with the "user ID", and information on the date and time of the conversion by the user U (conversion date and time).
[0093] The content of the conversion may be, for example, but is not limited to, the purchase of the transaction object that is the subject of the advertisement, the use of the transaction object that is the subject of the advertisement, registration for the membership service that is the subject of the advertisement, or the viewing (display) of the site that is the subject of the advertisement. The transaction object may be, for example, a product or a service.
[0094] The advertisement provision history storage unit 22 can also store, in association with a user U who has achieved conversion for one transaction object, advertisements provided to the user U during a predetermined period TA in association with the user U.
[0095] For example, the advertisement provision history memory unit 22 associates advertisements corresponding to the advertisement target among the advertisements provided to the user U during a specified period TA with the user U who achieved a conversion corresponding to the advertisement target, and stores the advertisements in association with the order of presentation.
[0096] For example, the advertisement provision history storage unit 22 stores, for a product X, one or more advertisements advertising the product X among the advertisements provided during a predetermined period TA (such as the most recent week) prior to the date and time when the user U purchased the product X, in association with a user U who purchased the product X, in association with the order in which the advertisements are provided.
[0097] [3.3. Processing section 12] The processing unit 12 is a controller, and is realized, for example, by a CPU (Central Processing Unit) or an MPU (Micro Processing Unit) executing various programs stored in a storage device inside the information processing device 1 using a RAM as a working area.
[0098] The processing unit 12 may be partially or entirely realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).
[0099] 3, the processing unit 12 has an acquisition unit 30, a reception unit 31, a calculation unit 32, a determination unit 33, and a provision unit 34, and realizes or executes the functions and actions of information processing described below. Note that the internal configuration of the processing unit 12 is not limited to the configuration shown in FIG. 3, and may be other configurations as long as they perform the information processing described below.
[0100] [3.3.1. Acquisition part 30] The acquisition unit 30 acquires various pieces of information from an external information processing device, the user terminal 2, the advertiser terminal 3, or the like via the communication unit 10, and stores the acquired information in the storage unit 11.
[0101] For example, the acquisition unit 30 acquires user information, which is information about the user U, from an external information processing device or the user terminal 2 via the communication unit 10, and adds the acquired user information to a user information table in the user information storage unit 20.
[0102] Furthermore, the acquisition unit 30 acquires various types of information from the storage unit 11. For example, the acquisition unit 30 acquires user information, which is information on the user U, from the user information storage unit 20 or the like. The user information acquired by the acquisition unit 30 includes, for example, a part or all of at least one of the above-mentioned attribute information and history information.
[0103] Furthermore, the acquisition unit 30 acquires an advertisement from the advertisement information storage unit 21 or the like. For example, when an advertisement request from the user terminal 2 is accepted by the acceptance unit 31, the acquisition unit 30 acquires an advertisement from the advertisement information storage unit 21 in response to the advertisement request.
[0104] Furthermore, the acquisition unit 30 acquires the advertisement provision history from the advertisement provision history storage unit 22. For example, when an advertisement effectiveness analysis request is accepted by the acceptance unit 31 from the advertiser terminal 3, the acquisition unit 30 acquires provided advertisement information including information on a plurality of advertisements provided to the user U until the user U makes a conversion, the information being the order in which the plurality of advertisements are provided to the user U.
[0105] For example, the acquisition unit 30 converts the advertisement provision history into time-series information in a predetermined format for each conversion. The time-series information in the predetermined format includes, for example, information indicating the order in which the advertisements were provided.
[0106] Here, it is assumed that the prediction model generated by the calculation unit 32 is an RNN such as LSTM or GRU. In this case, the time-series information in a predetermined format is information including, for example, attribute information of the user U and unit information including information indicating whether each advertisement is provided, in chronological order.
[0107] For example, the unit information is information such as [user U's age, user U's gender, whether advertisement A is provided, whether advertisement B is provided, whether advertisement C is provided], but user U's attributes may include things other than age and gender. "User U's age" is, for example, a normalized value of the user's age, and "user U's gender" is "0" for female and "1" for male. "Whether advertisement is provided" is "1" if advertisement is provided, and "0" if advertisement is not provided.
[0108] For example, suppose that the maximum number of advertisements provided to each relevant user is 3, and the order of advertisements provided to the one relevant user is advertisement A, advertisement B, and advertisement C. Also suppose that the normalized value of the one relevant user's age is 0.31, and the one relevant user's gender is female. In this case, the time series information of the one relevant user includes information [0.31, 0, 1, 0, 0], information [0.31, 0, 0, 1, 0], and information [0.31, 0, 0, 0, 1] in that order (forward or reverse).
[0109] The time-series information of the user for which the number of advertisements provided is less than 3 is, for example, information that has been subjected to padding processing, etc. The time-series information of the user is not limited to the above-mentioned example, and may be, for example, information indicating the presence or absence of the user by age group, and the user's gender may be information indicating the presence or absence of the user by gender.
[0110] In addition, when the prediction model generated by the calculation unit 32 is a linear regression model, GBDT, or the like, the time series information in a predetermined format is information including, for example, attribute information of the user U and information indicating the order in which each advertisement is presented.
[0111] The information indicating the order of presentation of each advertisement may be, for example, information such as [Presentation order 1 (Advertisement A, Advertisement B, Advertisement C),...,Presentation order n (Advertisement A, Advertisement B, Advertisement C)], where n is, for example, the maximum number of advertisements presented to each corresponding user.
[0112] "Presentation order 1 (Ad A, Ad B, Ad C)" is information indicating whether the last advertisement presented before the conversion by the corresponding user was Ad A, B, or C. "Presentation order n (Ad A, Ad B, Ad C)" is information indicating whether the first advertisement presented before the conversion by the corresponding user was Ad A, B, or C.
[0113] For example, in the case of advertisement A that was presented last before the conversion by the user, "presentation order 1 (ad A, ad B, ad C)" is information such as [1, 0, 0]. Also, in the case of advertisement C that was presented first before the conversion by the user, "presentation order n (ad A, ad B, ad C)" is information such as [0, 0, 1].
[0114] The plurality of advertisements whose information is included in the provided advertisement information are, for example, the plurality of advertisements indicated in the advertisement effectiveness analysis request accepted by the accepting unit 31 or the plurality of advertisements included in the campaign indicated in the advertisement effectiveness analysis request.
[0115] [3.3.2. Reception Unit 31] The reception unit 31 receives various requests, information, and the like from the user terminal 2 and the advertiser terminal 3 via the communication unit 10.
[0116] For example, the requests received by the reception unit 31 include, but are not limited to, an advertisement request transmitted from the user terminal 2 and an advertisement effectiveness analysis request transmitted from the advertiser terminal 3. In addition, the reception unit 31 receives advertisements transmitted from the advertiser terminal 3 to the information processing device 1, and adds the received advertisements to the advertisement information table of the advertisement information storage unit 21.
[0117] 3.3.3. Calculation unit 32 The calculation unit 32 calculates an attribution score Sc, which is a score indicating the relative contribution of each of a plurality of advertisements to a conversion, based on the provided advertisement information acquired by the acquisition unit 30. In the following, the plurality of advertisements for which the attribution score Sc is calculated by the calculation unit 32 will be described as advertisements A, B, and C, but the number and types of advertisements are not limited to three advertisements A, B, and C. In addition, in the following, a user U who has performed a conversion may be referred to as a corresponding user.
[0118] The calculation unit 32 includes an estimation processing unit 40 that estimates organic conversions, a generation processing unit 41 that generates a prediction model, and a calculation processing unit 42 that calculates attribution scores Sc(A), Sc(B), and Sc(C) for advertisements A, B, and C using the prediction model generated by the generation processing unit 41.
[0119] The estimation processing unit 40 can treat, for example, the total number of users U to whom none of advertisements A, B, and C was provided during a predetermined period TA before the conversion, among the provided advertisement information acquired by the acquisition unit 30, as the number of organic conversions. As described above, an organic conversion is a conversion that occurs due to the spontaneous behavior of the user even when advertisements A, B, and C are not provided.
[0120] In addition, the estimation processing unit 40 can estimate CtrlCV as the number of organic conversions for each advertisement presentation order by controlling the delivery of advertisements A, B, and C, for example, using an advertisement evaluation method (e.g., A / B testing) that is a method for evaluating the performance of an advertisement.
[0121] The generation processing unit 41 generates a prediction model for predicting the number of conversions according to the order in which advertisements are presented to a corresponding user, based on the provided advertisement information acquired by the acquisition unit 30.
[0122] The generation processing unit 41 calculates the number of conversions for each user attribute for each advertisement presentation order based on the provided advertisement information acquired by the acquisition unit 30. For example, by tallying up the number of conversions for each user attribute when the advertisement presentation order is advertisement A, advertisement B, and advertisement C, the generation processing unit 41 calculates the number of conversions for each user attribute when the advertisement presentation order is advertisement A, advertisement B, and advertisement C.
[0123] Then, the generation processing unit 41 generates learning information including time-series information and information indicating the number of conversions as feature quantities for each advertisement presentation order, and generates a prediction model using the learning information. The feature quantities used in the prediction model are not limited to the above-mentioned examples, and may include, for example, one or more pieces of information such as advertisement display environment information and campaign information for advertisements A, B, and C.
[0124] The advertisement display environment information is information such as the website category, the type of compatible application, the display position, etc. The campaign information is information such as the campaign category including advertisements A, B, and C, the distribution period, and the number of distributions.
[0125] When the prediction model is a linear regression model, GBDT, etc., the time-series information may include, instead of or in addition to the above examples, statistics such as count values, average values, minimum values, and maximum values as feature quantities. For example, in a given period TA, advertisement A is provided twice, advertisement B is provided once, advertisement A is provided x days ago (x is an integer), advertisement B is provided y days ago (y is an integer), etc. can each be a feature quantity.
[0126] When an organic conversion is estimated by the estimation processing unit 40, the generation processing unit 41 calculates a lift conversion, which is a conversion obtained by removing the organic conversion estimated by the estimation processing unit 40, for each combination of the advertisement presentation order and the user attributes. Then, the generation processing unit 41 can generate a prediction model based on the lift conversion for each combination of the advertisement presentation order and the user attributes calculated.
[0127] For example, the generation processing unit 41 can tally up the number of lift conversions for each combination of the advertisement presentation order and the user attribute as the number of conversions, and generate a prediction model using the number of conversions.
[0128] The prediction model generated by the generation processing unit 41 may be a prediction model for each user attribute. In this case, the generation processing unit 41 generates information including time-series information not including user attribute information and information on the number of conversions as learning information for each user attribute, and generates a prediction model for each user attribute using the generated learning information for each user attribute.
[0129] The calculation processing unit 42 calculates the attribution scores Sc(A), Sc(B), and Sc(C) of the advertisements A, B, and C using the prediction model generated by the generation processing unit 41. The prediction model generated by the generation processing unit 41 is a model that predicts the number of conversions based on information including information about the user U and information about the order in which advertisements are provided to the user U.
[0130] For example, the calculation processing unit 42 uses attribute information for each user attribute and information on the advertisement presentation order to calculate attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C for each combination of the advertisement presentation order and user attribute. For example, when the attribute information of user U is information on the gender and age of user U, each user attribute is represented by a combination of multiple attribute items such as gender and age. Note that the attribute items included in the attribute information of user U may include region, occupation, annual income, and the like, and may also include other attribute items.
[0131] The calculation processing unit 42 inputs attribute information of user U and information on the advertisement presentation order into a prediction model, and performs a calculation process to calculate the attribution scores Sc(A), Sc(B), and Sc(C) of each advertisement A, B, and C for each combination of advertisement presentation order and user attributes, using a process such as obtaining the predicted number of conversions output from the prediction model.
[0132] For example, the prediction model generated by the generation processing unit 41 is assumed to be an RNN such as LSTM or GRU. In this case, the calculation processing unit 42 inputs attribute information of the user U and time-series information including unit information, in chronological order, including information indicating whether each advertisement is provided, to the prediction model, and calculates output values of a softmax function in the prediction model at this time as the attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C.
[0133] The output value of the softmax function is obtained, for example, by applying the softmax function to the output of the hierarchical attention, and is represented by the probability distribution of advertisements A, B, and C. The hierarchical attention is an example of an attention mechanism. The calculation processing unit 42 can also calculate a value obtained by using the output value of the softmax function as the attribution score Sc for each of the multiple advertisements. The value obtained by using the output value of the softmax function is, for example, a value obtained by multiplying the output value of the softmax function by the above-mentioned weight w.
[0134] Also, the prediction model generated by the generation processing unit 41 is assumed to be a classification model such as a linear regression model or GBDT. In this case, the calculation processing unit 42 inputs information including the attribute information of the user U and the information on the advertisement presentation order into the prediction model, and performs a process of acquiring pCV, which is a predicted number of conversions for the advertisement presentation order, from the prediction model for each combination of the advertisement presentation order and the user attribute. Then, the calculation processing unit 42 performs a process of calculating attribution scores Sc(A), Sc(B), and Sc(C) of each advertisement A, B, and C for each combination of the advertisement presentation order and the user attribute using the pCV for each combination of the advertisement presentation order and the user attribute.
[0135] For example, assume that the advertisement presentation order is advertisement A, advertisement B, and advertisement C. In this case, calculation processing unit 42 calculates functions v(A), v(B), and v(c) (see formulas (2) to (4) above). v(A) is a function related to the contribution degree of advertisement A, v(B) is a function related to the contribution degree of advertisement B, and v(C) is a function related to the contribution degree of advertisement C.
[0136] The calculation processing unit 42 calculates the attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C by performing the calculations of the above formulas (5) and (6) based on the calculation results of the above formulas (2) and (4).
[0137] The calculation processing unit 42 performs the above processes (2) to (8) for each user attribute. In addition, the calculation processing unit 42 similarly calculates the attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C for advertisement presentation orders other than the advertisement presentation order of advertisement A, advertisement B, and advertisement C. In this way, the calculation processing unit 42 performs a calculation process to calculate the attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C for each combination of the advertisement presentation order and the user attribute.
[0138] In this way, the calculation processing unit 42 can calculate the attribution scores Sc(A), Sc(B), and Sc(C) for each of the multiple advertisements based on the conversion predicted by the prediction model when the multiple advertisements are provided to user U and the conversion predicted by the prediction model when only a portion of the multiple advertisements are provided to user U.
[0139] In addition, the calculation processing unit 42 can also calculate the attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C by excluding organic conversions, which are conversions when multiple advertisements are not provided to user U.
[0140] For example, assume that the prediction model generated by the generation processing unit 41 is a prediction model generated based on lift conversions, which are conversions from which organic conversions have been removed. In this case, the calculation processing unit 42 can use the prediction model generated by the generation processing unit 41 to calculate the attribution scores Sc(A), Sc(B), and Sc(C) of the advertisements A, B, and C by excluding the organic conversions, which are conversions.
[0141] In addition, if the prediction model generated by the generation processing unit 41 is not generated based on lift conversions, the calculation processing unit 42 can add weighting corresponding to the organic conversions estimated by the estimation processing unit 40 to the prediction model, and calculate the attribution scores Sc(A), Sc(B), and Sc(C) of advertisements A, B, and C using the prediction model with the weighting added.
[0142] For example, when the prediction model is an RNN, the calculation processing unit 42 calculates the weight w shown in the above formula (9). Then, the calculation processing unit 42 performs weighting in hierarchical attention using the weight w, and can adjust the attribution scores Sc(A), Sc(B), and Sc(C).
[0143] For example, the calculation processing unit 42 can calculate the attribution scores Sc(A), Sc(B), and Sc(C) excluding organic conversions by multiplying the output of the hierarchical attention by a weight w.
[0144] [3.3.4. Determination unit 33] The determination unit 33 can make various decisions based on the attribution score Sc calculated by the calculation unit 32. For example, the determination unit 33 can determine the allocation of an advertising budget as a method of using the attribution analysis.
[0145] For example, the determination unit 33 determines the budget allocation for advertisements A, B, and C based on the attribution score Sc calculated by the calculation unit 32. For example, the determination unit 33 allocates a larger budget to an advertisement having a larger attribution score Sc calculated by the calculation unit 32. For example, the determination unit 33 determines the ratio of the attribution score Sc of each advertisement to the total value of the attribution scores Sc of all advertisements subject to budget allocation as the ratio of the budget to be allocated to each advertisement.
[0146] Here, it is assumed that the proportion of the attribution score Sc(A) is 20%, the proportion of the attribution score Sc(B) is 30%, and the proportion of the attribution score Sc(C) is 50%. In this case, the determination unit 33 can determine to allocate half (50%) of the total budget to advertisement C, 30% to advertisement B, and 20% to advertisement A.
[0147] In addition, when an advertising request is accepted by the accepting unit 31, the determining unit 33 determines an advertisement to be delivered to the user U from among the multiple advertisements stored in the advertising information storage unit 21 based on the information of the user U identified by the advertising request.
[0148] For example, the determination unit 33 calculates a predicted conversion rate (pCVR) for each advertisement based on information of the user U specified by the advertisement request. For example, the determination unit 33 can predict pCVR using a CVR prediction model that inputs information of the user U (e.g., attribute information of the user U) and information on the advertisement presentation order as feature quantities and outputs pCVR. Such a CVR prediction model is calculated by the calculation unit 32, for example, using a conversion rate instead of the number of conversions as a learning model.
[0149] The determination unit 33 can determine the advertisement presentation order to be input to the CVR prediction model, for example, based on the information of the user U (for example, attribute information of the user U). For example, the determination unit 33 can determine the advertisement presentation order that results in the highest pCVR as the advertisement presentation order to be input to the CVR prediction model, based on the information of the user U (for example, attribute information of the user U).
[0150] [3.3.5.Providing Department 34] The providing unit 34 provides, for example, information on the attribution score Sc calculated by the calculating unit 32.
[0151] The providing unit 34 provides information on the attribution score Sc calculated for the advertisements of the advertiser O to the advertiser terminal 3 used by the advertiser of the advertisements A, B, and C. For example, the providing unit 34 transmits information indicating the attribution score Sc of the advertisements A, B, and C of the advertiser O to the advertiser terminal 3 used by the advertiser O. For example, the providing unit 34 transmits content indicating the attribution score Sc of the advertisements A, B, and C to the advertiser terminal 3 used by the advertiser O of the advertisements A, B, and C.
[0152] In addition, the providing unit 34 provides information indicating the budget ratio determined by the determining unit 33 to the advertiser terminal 3 used by the advertiser O of the advertisements A, B, and C. For example, the providing unit 34 transmits content indicating the budget ratio determined by the determining unit 33 to the advertiser terminal 3 used by the advertiser O of the advertisements A, B, and C.
[0153] Furthermore, when an advertisement request is accepted by the accepting unit 31, the providing unit 34 provides the advertisement determined by the determining unit 33 to the user U corresponding to the advertisement request. For example, the providing unit 34 transmits the advertisement determined by the determining unit 33 to the user terminal 2 of the user U corresponding to the advertisement request.
[0154] 4. Processing Procedure Next, a procedure of information processing by the processing unit 12 of the information processing device 1 according to the embodiment will be described. Fig. 7 is a flowchart showing an example of information processing by the processing unit 12 of the information processing device 1 according to the embodiment.
[0155] 7, the processing unit 12 of the information processing device 1 determines whether or not it is time to calculate the score (step S10). The score calculation timing is, for example, the timing of receiving an advertising effectiveness analysis request from the user terminal 2, or the timing of a predetermined period, but is not limited to such examples.
[0156] When the processing unit 12 determines that it is time to calculate the score (step S10: Yes), the processing unit 12 acquires provided advertisement information including information on the user U and information on the advertisement presentation order from the storage unit 11 (step S11). Then, the processing unit 12 calculates an attribution score Sc for each advertisement based on the provided advertisement information acquired in step S11 (step S12).
[0157] When the processing of step S12 is completed, or when it is determined that the timing for calculating the score has not come (step S10: No), the processing unit 12 determines whether or not an advertisement request has been accepted (step S13).
[0158] When it is determined that the advertisement request has been accepted (step S13: Yes), the processing unit 12 determines an advertisement to be provided (step S14), and provides the advertisement determined in step S14 (step S15).
[0159] When the process of step S15 is completed or when it is determined that the advertisement request has not been received (step S13: No), the processing unit 12 determines whether or not the operation end timing has arrived (step S16). The processing unit 12 determines that the operation end timing has arrived when, for example, the power supply of the information processing device 1 is turned off.
[0160] If the processing unit 12 determines that the operation end time has not arrived (step S16: No), it transitions to step S10, and if the processing unit 12 determines that the operation end time has arrived (step S16: Yes), it terminates the processing shown in FIG. 7.
[0161] 5. Modifications In addition to the above examples, the time-series information included in the provided advertisement information may include information on the presentation time of each advertisement as a feature. The presentation time of an advertisement is the time from when the advertisement is presented to when the user converts.
[0162] In this case, the time series information is expressed as, for example, {presentation order 1, presentation order 2, presentation order 3, presentation time 1, presentation time 2, presentation time 3}. Presentation time 1 is the presentation time of the advertisement provided in presentation order 1, presentation time 2 is the presentation time of the advertisement provided in presentation order 2, and presentation time 3 is the presentation time of the advertisement provided in presentation order 3. In this case, the advertisement presentation order information is expressed as {presentation order 1, presentation order 2, presentation order 3, presentation time 1, presentation time 2, presentation time 3}.
[0163] Furthermore, the predetermined format may be, for example, an array of information indicating the type of advertisement to be presented for each time period into which a predetermined period TA is divided, when the prediction model is, for example, an RNN or an LSTM. In this case, the information on the advertisement presentation order is indicated by an array of information indicating the type of advertisement to be presented for each time period into which the predetermined period TA is divided.
[0164] Furthermore, the prediction model generated by the calculation unit 32 is not limited to the above-mentioned examples, and may be, for example, a language model that predicts the number of conversions by inputting information including information about the user U and information about the order in which advertisements are presented as features.
[0165] The language model is, for example, a model such as a transformer that is trained to estimate and output the next token from an input token string. Such a language model is in the information processing device 1 and is created independently by the operator of the information processing device 1. It is preferable that the input information is trained so as not to be used as a new answer, thereby concealing the input information, such as personal information.
[0166] In the above-described example, the calculation unit 32 calculates the attribution score Sc for each advertisement using a prediction model, but is not limited to such an example, and the calculation unit 32 can also calculate the attribution score Sc for each advertisement without using a prediction model.
[0167] For example, the calculation unit 32 counts the number of conversions for each advertisement presentation order for each user attribute. Then, the calculation unit 32 can calculate the attribution score Sc of each advertisement for each user attribute based on the number of conversions for each advertisement presentation order. In this case, the calculation unit 32 uses CV(·) indicating the number of conversions instead of f(·) indicating the prediction model. For example, assume that the advertisement presentation order is advertisements A, B, and C.
[0168] In this case, the calculation unit 32 uses, for example, CV(φ), CV(A), CV(A→B), CV(A→B→C), and CV(A,B,C) in place of f(φ), f(A), f(A→B), f(A→B→C), and v(A,B,C) in the above formulas (2) to (5).Then, the calculation unit 32 calculates the attribution scores Sc(A), Sc(B), and Sc(C) of the advertisements A, B, and C by the calculation of the above formulas (2) to (8).
[0169] Furthermore, the calculation unit 32 can also count the number of conversions for each advertisement presentation order regardless of user attributes. In this case, the calculation of CV(·) does not include information on user attributes, and the calculation unit 32 calculates the attribution score Sc for each advertisement based on the number of conversions for each advertisement presentation order regardless of user attributes.
[0170] Furthermore, the calculation unit 32 can also tally up the number of conversions for each combination of user information, including a feature obtained from the history information of the user U in addition to the user attributes, and the order of presenting an advertisement. In this case, the calculation of CV(·) includes a feature obtained from the history information of the user U, and the calculation unit 32 calculates an attribution score Sc for each advertisement based on the number of conversions for each combination of the user information and the order of presenting an advertisement.
[0171] [6. Hardware Configuration] The information processing device 1 according to the embodiment described above is realized by a computer 80 having a configuration as shown in Fig. 8. Fig. 8 is a hardware configuration diagram showing an example of the computer 80 that realizes the functions of the information processing device 1 according to the embodiment. The computer 80 has a CPU 81, a RAM 82, a ROM (Read Only Memory) 83, a HDD (Hard Disk Drive) 84, a communication interface (I / F) 85, an input / output interface (I / F) 86, and a media interface (I / F) 87.
[0172] The CPU 81 operates and controls each part based on a program stored in the ROM 83 or the HDD 84. The ROM 83 stores a boot program executed by the CPU 81 when the computer 80 is started up, and programs that depend on the hardware of the computer 80.
[0173] The HDD 84 stores programs executed by the CPU 81, data used by such programs, etc. The communication interface 85 receives data from other devices via the network N (see FIG. 2) and sends it to the CPU 81, and transmits data generated by the CPU 81 to other devices via the network N.
[0174] The CPU 81 controls output devices such as a display and a printer, and input devices such as a keyboard and a mouse, via the input / output interface 86. The CPU 81 acquires data from the input devices via the input / output interface 86. The CPU 81 also outputs data generated via the input / output interface 86 to the output devices.
[0175] The media interface 87 reads a program or data stored in a recording medium 88 and provides it to the CPU 81 via the RAM 82. The CPU 81 loads the program from the recording medium 88 onto the RAM 82 via the media interface 87 and executes the loaded program. The recording medium 88 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or a PD (Phase change rewritable Disc), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory.
[0176] For example, when the computer 80 functions as the information processing device 1 according to the embodiment, the CPU 81 of the computer 80 executes a program loaded onto the RAM 82 to realize the function of the processing unit 12. In addition, the HDD 84 stores data in the storage unit 11. The CPU 81 of the computer 80 reads and executes these programs from a recording medium 88, but as another example, these programs may be obtained from another device via the network N.
[0177] [7. Other] In addition, among the processes described in the above embodiments, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically by a known method. In addition, the information including the processing procedures, specific names, various data and parameters shown in the above documents and drawings can be changed arbitrarily unless otherwise specified. For example, the various information shown in each drawing is not limited to the illustrated information.
[0178] In addition, each component of each device shown in the figure is a functional concept, and does not necessarily have to be physically configured as shown in the figure. In other words, the specific form of distribution and integration of each device is not limited to that shown in the figure, and all or part of them can be functionally or physically distributed and integrated in any unit according to various loads, usage conditions, etc.
[0179] For example, the above-mentioned information processing device 1 may be realized by a terminal device and a server computer, or may be realized by multiple server computers. Also, depending on the functions, the configuration can be flexibly changed, such as by calling an external platform using an API (Application Programming Interface) or network computing.
[0180] Furthermore, the above-described embodiments and modifications can be appropriately combined as long as the processing contents are not contradictory.
[0181] 8. Effects As described above, the information processing device 1 according to the embodiment includes an acquisition unit 30 and a calculation unit 32. The acquisition unit 30 acquires provided advertisement information including information on a plurality of advertisements provided to the user U until the user U achieves conversion, the information being information on a provision order 1 of the plurality of advertisements to the user U. The calculation unit 32 calculates a score indicating the relative contribution of each of the plurality of advertisements to the conversion, based on the provided advertisement information acquired by the acquisition unit 30. This allows the information processing device 1 to appropriately calculate the score related to the advertisement.
[0182] Furthermore, the calculation unit 32 calculates a score indicating the relative contribution of each of a plurality of advertisements by using a model for predicting a conversion based on information including the information of the user U and information of the provision order 1 to the user U. This enables the information processing device 1 to more appropriately calculate the score related to the advertisement.
[0183] Furthermore, the calculation unit 32 calculates a score indicating the relative contribution of each of the multiple advertisements A based on the conversion predicted by the model when the multiple advertisements A are provided to the user U and the conversion predicted by the model when some of the multiple advertisements A are provided to the user U. This allows the information processing device 1 to more appropriately calculate the score related to the advertisement.
[0184] In addition, the calculation unit 32 calculates the output value of the softmax function used in the model or a value obtained by using the output value as a score indicating the relative contribution of each of the multiple advertisements. This allows the information processing device 1 to more appropriately calculate the score related to the advertisements.
[0185] Furthermore, the calculation unit 32 calculates a score indicating the relative contribution of each of the multiple advertisements, excluding an organic conversion, which is a conversion when multiple advertisements are not provided to the user U. This allows the information processing device 1 to more appropriately calculate the score related to the advertisements.
[0186] The calculation unit 32 also includes an estimation processing unit 40 and a generation processing unit 41. The estimation processing unit 40 estimates organic conversions. The generation processing unit 41 generates a model based on conversions excluding the organic conversions estimated by the estimation processing unit 40. This allows the information processing device 1 to more appropriately calculate the score related to the advertisement.
[0187] Moreover, the calculation unit 32 includes an estimation processing unit 40 and a calculation processing unit 42. The estimation processing unit 40 estimates organic conversions. The calculation processing unit 42 adds weighting according to the organic conversions estimated by the estimation processing unit 40 to a model, and calculates a score indicating the relative contribution of each of a plurality of advertisements using the model to which the weighting has been added. This allows the information processing device 1 to more appropriately calculate the score related to the advertisements.
[0188] The above describes the embodiment of the present application in detail based on the drawings, but this is merely an example, and the present invention can be implemented in other forms that include various modifications and improvements based on the knowledge of those skilled in the art, including the forms described in the Disclosure of the Invention section.
[0189] Furthermore, the above-mentioned "section, module, unit" can be read as "means" or "circuit", etc. For example, an acquisition section can be read as an acquisition means or an acquisition circuit. [Explanation of symbols]
[0190] 1. Information processing device 2. User terminal 3. Advertiser terminal 10. Communications Department 11 Storage section 12 Processing section 20 User information storage unit 21 Advertising information storage unit 22 Advertisement serving history memory unit 30 Acquisition Department 31 Reception 32 Calculation section 33 Decision Section 34 Providing Department 40 Estimation processing unit 41 Generation processing section 42 Calculation processing unit 100 Information Processing Systems
Claims
1. an acquisition unit that acquires provided advertisement information, the provided advertisement information including information on a plurality of advertisements provided to the user until the user makes a conversion, and information on an order in which the plurality of advertisements are provided to the user; a calculation unit that calculates a score indicating a relative contribution of each of the plurality of advertisements to the conversion based on the provided advertisement information acquired by the acquisition unit.
23. An information processing apparatus comprising:
2. The calculation unit is Calculating a score indicating a relative contribution of each of the plurality of advertisements using a model that predicts a conversion based on information including information about the user and information about the order in which the advertisements are presented to the user.
2. The information processing apparatus according to claim 1,
3. The calculation unit is Calculating a score indicating a relative contribution of each of the plurality of advertisements based on a conversion predicted by the model when the plurality of advertisements are provided to the user and a conversion predicted by the model when a portion of the plurality of advertisements are provided to the user.
3. The information processing apparatus according to claim 2.
4. The calculation unit is An output value of a softmax function used in the model or a value obtained by using the output value is calculated as a score indicating a relative contribution of each of the plurality of advertisements.
3. The information processing apparatus according to claim 2.
5. The calculation unit is Calculating a score indicating a relative contribution of each of the plurality of advertisements, excluding organic conversions, which are conversions when the plurality of advertisements are not provided to the user.
5. The information processing device according to claim 2, wherein the information processing device is a computer.
6. The calculation unit is an estimation processing unit that estimates the organic conversion; A generation processing unit that generates the model based on the conversions estimated by the estimation processing unit, excluding the organic conversions.
6. The information processing apparatus according to claim 5,
7. The calculation unit is an estimation processing unit that estimates the organic conversion; a calculation processing unit that adds a weighting according to the organic conversion estimated by the estimation processing unit to the model, and calculates a score indicating a relative contribution of each of the plurality of advertisements using the model to which the weighting has been added.
6. The information processing apparatus according to claim 5,
8. 1. A computer-implemented information processing method, comprising: acquiring advertisement information including information on a plurality of advertisements provided to the user until the user reaches a conversion, the information including information on an order in which the plurality of advertisements are provided to the user; and calculating a score indicating a relative contribution of each of the plurality of advertisements to the conversion based on the provided advertisement information acquired by the acquiring step.
23. An information processing method comprising:
9. an acquisition step of acquiring provided advertisement information, the provided advertisement information including information on a plurality of advertisements provided to the user until the user reaches a conversion, and information on an order in which the plurality of advertisements are provided to the user; a calculation step of calculating a score indicating a relative contribution of each of the plurality of advertisements to the conversion based on the provided advertisement information acquired by the acquisition step.
2. An information processing program comprising:
Citation Information
Patent Citations
Determination of the contribution degrees of various user interactions to conversion
JP2017529587A
