Advertisement matching server, platform, and operation method thereof, using real-time IP-based multi-device audience identification

Abstract

The present invention provides an advertisement matching server, a platform, and an operation method thereof, using identification of an audience of a multi-device on the basis of a real-time IP. For example, the matching server using identification of an audience of a multi-device on the basis of a real-time internet protocol (IP) may comprise the steps of: receiving a data matching request message for a first medium from an advertisement server; generating a matching key on the basis of an IP corresponding to the data matching request message; on the basis of the data matching request message, generating or updating a matching model parameter for one or more media related to the matching key; and on the basis of the matching model parameter for the one or more media related to the matching key, selecting identification information associated with a second medium, and transmitting, to the advertisement server, a data matching response message including the identification information. Here, the first medium may correspond to an apparatus having a medium type other than a mobile device type, and the second medium may correspond to a device having a mobile device type.

Description

Advertising matching server, platform, and operating method using real-time IP-based multi-device audience identification

[0001] The present invention relates to an advertising matching server, platform, and operating method thereof that utilizes real-time IP (Internet Protocol)-based multi-device audience identification. More specifically, the present invention relates to an advertising matching server, platform, and operating method thereof that utilizes real-time IP-based audience identification of mobile devices for various media other than mobile devices.

[0002] We are moving beyond the era of monopolization of traditional media centered on TV, and with the advent of the Internet of Things (IoT) due to the advancement of communication technology and electronic devices, various media are coexisting, and advertising techniques such as audience targeting are being widely utilized.

[0003] Audience targeting, as the word itself suggests, refers to an advertising technique used to target and deliver relevant advertisements to the audience / customer. More specifically, audience (or data) targeting refers to the process of analyzing potential customers' characteristics, interests, behavioral patterns, consumption habits, and lifestyles when executing an advertisement to promote a product or service. In this regard, advertising via mobile devices, which utilizes mobile audience data (e.g., online / offline customer behavior data) to create customized segments for each (advertising) campaign and then target advertisements accordingly, is a highly targeted field.

[0004] However, unlike mobile devices, most media (devices, such as TVs and refrigerators) have their own unique ID and behavioral data, but the data formats and service environments are different until a targeted campaign is executed, and in particular, it is difficult to identify users, making targeted advertising virtually impossible. In other words, there is a problem that it is nearly impossible to provide user-targeted advertising to other media that do not utilize advertising identifiers (IDs) of mobile devices (e.g. smartphones), such as ADID (advertisement ID) for Android OS (operating system) and IDFA (ID for advertisers) for iPhone OS (e.g. iOS).

[0005] The purpose of the present invention is to solve the above-described conventional problems, and to provide an advertisement matching server, platform, and operating method thereof that utilizes real-time IP-based multi-device audience identification, capable of matching mobile audience data to various media other than mobile using a separate matching model.

[0006] According to one embodiment of the present invention, a method for operating a matching server using real-time IP (Internet Protocol)-based audience identification of multiple devices is provided, comprising the steps of: receiving a data matching request message for a first medium from an advertisement server; generating an IP-based matching key corresponding to the data matching request message; generating or updating matching model parameters for one or more mediums associated with the matching key based on the data matching request message; and selecting identification information associated with a second medium based on the matching model parameters for the one or more mediums associated with the matching key, and transmitting a data matching response message including the identification information to the advertisement server. In this case, the first medium corresponds to a device having a medium type other than a mobile device, and the second medium is a mobile device type. A mobile device type is a device having an information processing function and capable of communication, such as a smartphone, tablet, or watch, and generally used exclusively by one individual. A medium type other than a mobile device is an information processing device such as a TV or refrigerator, which, although not excluding individual use, is generally not a device for exclusive use by an individual but rather a device used jointly by a household or group.

[0007] In the present invention, the step of generating or updating the matching model parameter may include: when the first medium is determined to be a new medium with respect to the matching key, generating a matching model parameter for the first medium and setting it to an initial value, and updating a matching model parameter for a medium other than the first medium among the one or more media based on the data matching request message; and when the first medium is determined to be an existing medium with respect to the matching key, updating a matching model parameter for the one or more media including the first medium based on the data matching request message.

[0008] At this time, the matching model parameter may include a weight parameter that is updated based on a time difference between the data matching request message and the most recently received previous data matching request message related to the matching key prior to the data matching request message.

[0009] At this time, if the first medium is determined to be an existing medium in relation to the matching key:

[0010] Based on the above data matching request message, the weight parameter for the first medium is calculated according to the following mathematical expression 1:

[0011] [Mathematical Formula 1]

[0012]

[0013] Based on the above data matching request message, the weight parameters for the one or more media other than the first medium are calculated according to the following mathematical expression 2:

[0014] [Equation 2]

[0015]

[0016] The previous weight may correspond to a previous weight parameter value for each medium, and the elapsed time may correspond to a time difference between the data matching request message and the most recently received previous data matching request message related to the matching key prior to the data matching request message.

[0017] Alternatively, the matching model parameters may be configured to include a weight parameter that is updated based on a time difference between the data matching request message and the most recently received previous data matching request message related to the matching key prior to the data matching request message; and a matching counter parameter that is counted based on the number of connections for each medium.

[0018] In the present invention, selecting identification information associated with the second medium may include selecting identification information of a specific medium based on matching model parameters for all media excluding the first medium among the media associated with the matching key. In this case, the identification information associated with the second medium may correspond to identification information of the second medium or identification information of a third medium associated with the second medium and not of a mobile device type.

[0019] For example, the matching model parameter may include a weight parameter that is updated based on a time difference between the data matching request message and the most recently received previous data matching request message related to the matching key prior to the data matching request message. In this case, selecting the identification information of the specific medium may include selecting the identification information of the medium having the highest weight parameter, excluding the first medium, among the media related to the matching key, as the identification information of the specific medium.

[0020] As another example, the data matching request message may include a weight parameter that is updated based on a time difference between the data matching request message and the most recently received previous data matching request message related to the matching key before the data matching request message; and a matching counter parameter that is counted according to the number of connections for each medium. In this case, selecting the identification information of the specific medium may include selecting the identification information of the medium having the highest matching counter parameter among the media related to the matching key, excluding the first medium, as the identification information of the specific medium, and if there are a plurality of media having the highest matching counter parameters excluding the first medium, selecting the identification information of the medium having the highest weight parameter among the plurality of media as the identification information of the specific medium.

[0021] According to another embodiment for solving the object of the present invention, a matching server using audience identification of multiple devices based on real-time IP (Internet Protocol) includes: a data matching request message receiving unit for receiving a data matching request message for a first medium from an advertisement server; a matching key generating unit for generating a matching key based on an IP corresponding to the data matching request message; a matching model parameter management unit for generating or updating matching model parameters for one or more mediums related to the matching key based on the data matching request message; and a data matching response message transmitting unit for selecting identification information associated with a second medium based on the matching model parameters for the one or more mediums related to the matching key, and transmitting a data matching response message including the identification information to the advertisement server. At this time, the first medium corresponds to a device having a medium type other than a mobile device type, and the second medium corresponds to a device having a mobile device type.

[0022] According to another embodiment for solving the object of the present invention, a method for operating an advertisement matching platform using audience identification of multiple devices based on real-time IP (Internet Protocol) comprises the steps of: receiving an advertisement request message from a first medium; obtaining a data matching request message for the first medium based on the advertisement request message; generating an IP-based matching key corresponding to the data matching request message; generating or updating a matching model parameter for one or more mediums related to the matching key based on the data matching request message; and selecting identification information associated with a second medium based on the matching model parameter for the one or more mediums related to the matching key, and transmitting an advertisement response message including an advertisement corresponding to the identification information to the first medium. In this case, the first medium may correspond to a device having a medium type other than a mobile device type, and the second medium may correspond to a device having a mobile device type.

[0023] According to the present invention, targeting can be performed using mobile audience data for various media other than mobile (e.g., non-personalized media), and targeted advertising can be performed based on the same.

[0024] Figure 1 is a drawing that simply shows the overall system structure to which the present invention can be applied.

[0025] Figure 2 is a drawing briefly showing a matching model applicable to the present invention.

[0026] Figure 3 is a flowchart briefly illustrating an example of the operation of an advertisement matching server or advertisement matching platform applicable to the present invention.

[0027] Figures 4 to 6 are diagrams showing the operations of a process according to an example of the present invention.

[0028] Figure 7 is a drawing illustrating the process of Figure 6.

[0029] Figure 8 is a drawing briefly showing an example of the operation of the entire system according to an example of the present invention.

[0030] Figure 9 is a flowchart of the operation of a matching server according to an example of the present invention.

[0031]

[0032] Hereinafter, with reference to the attached drawings, various embodiments applicable to an advertising matching server, platform, and operating method thereof utilizing real-time IP-based multi-device audience identification according to the present invention will be described. The present invention is not limited or restricted by these embodiments. Furthermore, in describing the present invention, detailed descriptions of known functions or components may be omitted to clarify the gist of the present invention.

[0033] For a clear understanding of the present invention, programmatic advertising will first be explained.

[0034] Programmatic advertising is a type of intermediary technology that connects specific users (or customers) with the advertisers (or brands) most willing to pay to show them their ads. It is implemented through an automated intermediary program using wired and wireless communication between information processing devices. Below, we will explain in detail an example of the process for programmatic advertising.

[0035] Advertisers (or brands) determine the target audience for their ads. A Demand Side Platform (DSP) and a Data Management Platform (DMP) classify customers based on the target audience determined by the advertiser and deliver the ads to them. More specifically, a DMP is a data management platform that manages parameters for the target audience determined by the advertiser. Based on campaign information from the DSP, it can distinguish and analyze whether a specific consumer is a valid target and relay this information back to the DSP. A DSP is an advertising site purchasing platform that uses the information provided by the DMP to deliver ads to individuals within the target audience. These DSPs can be utilized to ensure that advertisers' ads (or campaigns) meet both investment efficiency and volume. They can also focus on targeting to deliver relevant ads to consumers, maximizing advertisers' profits. Furthermore, they can facilitate the purchase of highly efficient space through ad exchanges.

[0036] Next, customized advertisements can be provided to targeted customers using audience data through an advertisement exchange (ADX), an AD network, a supply-side platform (SSP), and media (or publisher). At this time, the AD network can be provided to advertisers as a whole network unit by grouping multiple media based on an AD server. ADX, as a collection of AD networks, can be configured to enable DSPs to send out a large volume of advertisements with precise targeting through various transaction methods and system integration. Furthermore, the SSP, as a supplier (or media) platform, can help media find and connect the most profitable advertisements (or campaigns). At this time, the media can expose advertising content to consumers and, for this purpose, provide advertising space where advertisers can place advertisements.

[0037] As mentioned above, the audience required for such programmatic advertising refers to defining the characteristics, interests, behavioral patterns, consumption habits, lifestyles, etc. of potential customers who will see the advertisement when the advertiser executes the advertisement to promote the company's product or service. Looking at a specific application example, data targeting collects related data from various DMPs. Related data may include, for example, SK Planet (SK Planet service information), Deep AD (Lotte affiliate service information), TG360 (online / offline behavioral data information), Loplat X (Wi-Fi indoor location visit history information), Hyper Local (GPS (Global Positioning System) signal information), App Profile (Android App installation history information), etc. Based on this related data, it can be composed of a step to build on / offline customer behavior data such as search, payment history, online / offline visit history, app installation history, and shopping cart history, and create customized segments for advertisements (or campaigns) to perform audience targeting.

[0038] However, while this audience targeting technique is very effective for advertising through mobile devices, it is difficult to target advertising on other media (e.g. IPTV, CTV, PC, refrigerator, clothes manager, etc.) because the data format and service environment are different and, unlike mobile devices, it is difficult to specify users.

[0039] For example, let's say an advertiser for a health functional food product wants to target women in their 20s and 30s who are interested in dieting and exercise and have recently purchased health functional foods. For mobile devices, targeting is possible through audience data. However, for media other than mobile, devices are typically shared by multiple people, making audience analysis impossible to identify specific users, making targeted ad requests and responses extremely challenging.

[0040] In order to solve such problems, the present invention discloses a method and device for performing audience analysis to identify users in media other than mobile (e.g., refrigerator, TV, cTV, PC, clothes manager, etc.) and enabling targeted advertising requests and responses.

[0041] Figure 1 is a drawing that simply shows the overall system structure to which the present invention can be applied.

[0042] As illustrated in FIG. 1, the entire system to which the present invention is applicable may include components such as various media (e.g., media A, media B, etc.), an advertising server (AD server), a matching server, a data management server (Data Management, or platform), and each component is an information processing device including a processor, a memory, and a communication module, and may perform data communication with each other through a wired or wireless communication network.

[0043] Here, the various media may include multiple mobile devices and one or more other media (e.g., non-mobile media such as a refrigerator, TV, or CTV). Depending on the embodiment, the various media may provide information such as user behavior and media operation elements to the advertising server and request advertisements for each medium. Furthermore, the various media may receive corresponding advertisements in response to the advertisement request and display them on separate display devices.

[0044] As described above, when the ad server receives an ad request message from each medium, it can select (or decide) an advertisement for each medium and provide it to each medium. To this end, when an ad request message is received from a medium corresponding to a mobile device, the ad server can select one of one or more advertisements for the mobile device based on user prediction data of the mobile device in conjunction with a data management server (or platform) that stores audience data (e.g., behavioral and interest data, etc.) for each of multiple mobile devices. Alternatively, when an ad request message is received from a medium other than a mobile device, the ad server can determine information about another medium corresponding to the medium (e.g., a specific mobile device, or a medium related to the specific mobile device, etc.) in conjunction with a matching server to select one of one or more advertisements for the corresponding mobile device. Subsequently, the ad server can provide the selected advertisement to the medium that transmitted the ad request message.

[0045] Unlike the example of FIG. 1, depending on the embodiment, the advertising server and matching server may be configured as a single system (e.g., an advertising matching platform) rather than as separate entities. Furthermore, depending on the embodiment, the data management server (or platform) may have a separate entity from the advertising server, as illustrated in FIG. 1, or may be configured as a single system with the advertising server.

[0046] Through this overall system configuration, the overall system according to the present invention can perform audience analysis to identify users for media other than mobile devices (users' personal information is de-identified and only information such as ADID is identified), and can provide targeted advertisements to customers based on the audience analysis results.

[0047] Figure 2 is a drawing briefly showing a matching model applicable to the present invention.

[0048] In the present invention, a matching model refers to a technology that matches and analyzes operational data from various network-connected media based on the same IP address and estimates the same audience. This technology can be implemented as a software program. Using such a matching model, audience targeting across various media can be achieved.

[0049] For example, using the matching model, advertisers can display desired advertisements on media such as CTV and refrigerators based on audience data for mobile users (or mobile devices). To this end, the ad server can apply a hashing function (or one-way encryption function) to information on mobile devices, other media, etc., and transmit it to the matching server. The information transmitted and received between the ad server and the matching server may include at least the following information:

[0050] (1) For Android mobile and iOS mobile: Includes at least one of the following: OS (Operating System) information, device information, log date, advertising ID, and IP.

[0051] (2) For CTV: Includes at least one of the following: OS information, device information, log date, channel ID, and IP.

[0052] (3) For PC (Personal Computer): Includes at least one of the following: OS information, device information, log date, content ID, and IP.

[0053] The matching server, which receives the above-described information from the advertising server, can compare this information based on IP addresses to identify mobile devices that are matched or associated with specific other media and provide matching information about them to the advertising server. Accordingly, the advertising server can provide targeted advertising by providing advertisements based on audience data of the identified mobile devices to the specific other media.

[0054] Figure 3 is a flowchart briefly illustrating an example of the operation of an advertisement matching server or advertisement matching platform applicable to the present invention.

[0055] FIG. 3 illustrates in detail the operation of an advertising matching server or advertising matching platform for advertising matching for three different devices (e.g., a CTV, a smartphone, and a smart refrigerator). In this case, the advertising matching server or advertising matching platform may correspond to a concept that includes both the advertising server and the matching server of FIG. 3. In other words, depending on the embodiment, the advertising server and the matching server may be implemented as a single server (e.g., an advertising matching server), or may be implemented as separate entities included in a single platform (e.g., an advertising matching platform).

[0056] As illustrated in FIG. 3, when the advertising server receives an advertising request message including information about each device (IP information, uid) from each device, the advertising server can generate a matching key (or anonymization key) based on the IP information included in the advertising request message and transmit the same to the matching server.

[0057] In response, the matching server may update the matching table (or matching model) for one or more uids having the same IP (and thus the same matching key) based on the received event value (e.g., calculate a matching count value and / or a weight), and transmit the corresponding (or determined) matching ID to the advertising server using the updated matching table (or matching model). At this time, the matching ID may be a value encrypted using a hash function. Here, the uid is the ID of each medium (device), and although the format may be different for each medium, it defines an individual device.

[0058] An advertising server that has received a matching ID can send a matched advertisement to the relevant medium using an audience segment determined (or selected) based on the matching ID.

[0059] In the above-described matching table (or matching model), the matching server can identify mobile audience data corresponding to a specific medium by considering at least one of the following two parameters, thereby increasing the efficiency (accuracy) of advertising targeting:

[0060] - Match count (match_count): This is an indicator based on the number of IP connections to the media. It is a parameter corresponding to the number of connections to a specific media in the same IP band. The larger the value, the more likely it is that users using the media are staying in the location.

[0061] - Weight: This is an indicator based on the connection time through the corresponding IP of the medium. It is a parameter corresponding to the average connection time difference of users using a specific medium. For example, the larger the value, the shorter the average connection time difference of users using the medium. This can be inferred as a high probability that the user frequently uses the medium and is more sensitive to advertisements. For example, if the value (weight) confirms that a specific TV ID has the shortest average connection time difference, this may mean that the user is very active in using TV.

[0062] Analysis results utilizing these parameters can identify individual users and enable personalized ad targeting. In particular, recent access experience is a crucial factor in programmatic advertising transactions, which require real-time ad transactions. Therefore, audience data with a high match count (match_count), indicating access frequency, and a high weighting value (e.g., indicating recent access to a media outlet) can be utilized as the most effective data for advertising on specific devices.

[0063] The matching count (match_count) and weight, which are parameter values ​​of the above-described matching model, can each have the following characteristics.

[0064] [graph]

[0065]

[0066] In the present invention, the above-described weight values ​​can be calculated according to various embodiments. As a specific example, the weight values ​​according to the present invention can be calculated as follows.

[0067] In the present invention, data can be defined based on a certain time period (e.g., from a minimum of 1 day (24 hours) to 1 month (30 days), etc.) and can be assumed to disappear after the time period elapses without any updates. For convenience of explanation, the data is defined based on 1 day (24 hours) below. However, as described above, the time period may be set differently depending on the embodiment, and for example, it may be set to 48 hours, 7 days, 30 days, etc.

[0068] First, for the first data occurrence, the matching count value may be set to 1, and the weight value may be set to 1 (i.e., the matching count and weight values ​​are each set to their initial values). Thereafter, the weight value may be calculated / updated according to the following mathematical formula based on the occurrence of an event. Here, an event may refer to the transmission of an advertisement request message.

[0069] [Mathematical formula]

[0070]

[0071] In the above mathematical formula The term part corresponds to the weight in the case where there is no matching uid among the ad request uids, and the entire mathematical expression including +1 corresponds to the case where there is a value identical to the uid held by the matching key among the ad request uids.

[0072] In the present invention, the weight value is defined in such a way that a higher value is given to the latest data, and thus can be configured in such a way that the weight decreases over time.

[0073] First, in the mathematical formula, "previous weight" may refer to the weight value of the same uid at the most recent point in time (i.e., the most recent weight value of the same uid). Furthermore, in the mathematical formula, "elapsed time" may refer to the elapsed time from the previous access time for each (ad request) uid to the present (e.g., the calculation time or update time). Whether a device is connected for this purpose may be determined by an event, i.e., whether an ad request message is transmitted.

[0074] According to the above mathematical formula, the weight value can be determined differently depending on whether there is a matching uid among the ad request uids. If there is an identical value among the uids of the ad request medium and the uids of the matching key, the corresponding weight value is determined by the basic formula ( ) can be determined by adding a weight of 1. On the other hand, if there is no identical uid in the matching key, the corresponding weight value is determined by the basic formula ( ) can be calculated. For example, if media A, B, and C are matched within 24 hours, when media B is matched again, the weight value corresponding to media B is calculated as 'basic formula + 1', while each weight value corresponding to the remaining media A and C can be calculated only by 'basic formula'.

[0075] Below, we describe in detail the operations of the entire process utilizing the matching model described above. Before proceeding, let's take a closer look at each parameter or name.

[0076] - Event: Device actions and data generated by the user (e.g. behavioral element information such as IP, OS, device, gaid, date, time, etc.)

[0077] - De-identification matching key: Generate a matching key using hash (one-way encryption) to de-identify personal information from Internet IP addresses.

[0078] - uid_type: The type of media. For example, you can identify the media type, such as gaid for Google phones, ifa for iOS phones, or tvid for TVs.

[0079] - uid: Unique identification ID for each medium

[0080] - last match_date: The time of the last match. If a matching key or a specific UID is matched, the date and time information is updated.

[0081] Here, when the matching server receives the matching key of a random medium where an event occurred, it updates the weights for all existing media (mobile devices, TVs, refrigerators, PCs, etc.) with the same IP address as that medium. Therefore, media with different IP addresses are not associated with each other. Furthermore, while the matching count is only calculated for the medium where the event occurred, the weights are calculated in real time for all media belonging to the same IP address each time an event occurs.

[0082] Hereinafter, the operations of the process according to the present invention will be described in detail based on the above-described technical definitions and FIGS. 4 to 7.

[0083] Figures 4 to 6 are diagrams showing the operations of a process according to an example of the present invention, and Figure 7 shows an explanation of Figure 6. Figures 4, 5, and 6 are divided and expressed as situations that follow each other in chronological order.

[0084] First, as illustrated in Figure 4, assume that the first ad request (event) value is generated by an Android mobile device at 10:00 AM on April 13th (Step 1). Accordingly, the mobile device can transmit an ad request message containing {ip:106.249.255.82, os:Android, device:Mobile, gaid:abcd-efgh-ijkl-mnop} to the ad server.

[0085] The ad server can generate an anonymization (matching) key based on the IP information contained in the ad request message (step 2).

[0086] As this is the first event to occur, the matching server may initialize (or count) the matching count and weight values ​​(e.g., both values ​​are set to 1) and generate a new UID (step 3). The UID value may be set to the media identification information (e.g., the value corresponding to gaid) in the ad request message received from the ad server.

[0087] Subsequently, it is assumed that an advertisement request is generated again from the same Android mobile device at 6:00 PM on April 13. Accordingly, the mobile device can send an advertisement request message including {ip:106.249.255.82, os:Android, device:Mobile, gaid:abcd-efgh-ijkl-mnop} to the advertisement server, and based on this, the advertisement server can calculate the time difference between events for weight calculation (step 4).

[0088] The matching server can determine the last update criterion for the matching key based on the most recent update date / time of the uid (step 5). Accordingly, in the case of Figure 4, since the last match_date of the uid is 6:00 PM on April 13, the last match date of the matching key can also be set to 6:00 PM on April 13.

[0089] The matching server can identify that an ad request has occurred with the same uid and increase the match count (match_count) value for the corresponding uid to 2 (step 6). In addition, the matching server can update the weight value for the corresponding uid to 1*((24-8) / 24)+1=1.66667 based on the time difference from the previous event being 8 hours (step 7).

[0090] Next, as illustrated in Fig. 5, events may be generated by other devices. In this case, the matching server may create a new matching table if a matching key with the same uid value is not found based on the uid, and if a corresponding matching key is found, the matching server may update data including the matching count and weight for the corresponding key.

[0091] More specifically, assume that an event is triggered by another mobile device with a different gaid at 8:00 PM on April 13th (step 8). Accordingly, the mobile device can send an advertisement request message containing {ip:106.249.255.82, os:Android, device:Mobile, gaid:uudf-3ndf-4mv9-d8sfj} to the advertisement server.

[0092] Based on this, the matching server can update the weight value for uid=abcd-efgh-ijkl-mnop to reflect the 2-hour time difference from the previous event with the same matching key (step 9). More specifically, the matching server can update the weight value for the above-mentioned uid to 1.66667*((24-2) / 24)=1.52778.

[0093] Additionally, the matching server can create a table for the new uid=uudf-3ndf-4mv9-d8sfj and update the matching count, weight, last match_date, etc. for it (step 10).

[0094] Next, let's assume that an event (advertisement request) occurs from a TV medium at 6:00 on April 14th (step 11). Accordingly, the TV medium can send an advertisement request message containing {ip:106.249.255.82, os:Tizen, device:Tv, tvid:tvtvtv-ididididid-tvtv} to the advertisement server.

[0095] Based on this, the matching server can update the weight values ​​for uid=abcd-efgh-ijkl-mnop and uid=uudf-3ndf-4mv9-d8sfj respectively, reflecting the 10-hour time difference from the previous event with the same matching key (step 12):

[0096] - uid=abcd-efgh-ijkl-mnop: 1.52778*((24-10) / 24)=0.8912

[0097] - uid=uudf-3ndf-4mv9-d8sfj: 1*((24-10) / 24)=0.5833.

[0098] Next, the matching server may update the last match date of the corresponding matching key to the date / time at 6:00 on April 14th, as a new UID is generated (step 13).

[0099] Then, as illustrated in FIG. 6, when an additional advertisement request (event) occurs, the matching server can more reliably identify a user (or a user of a mobile device) corresponding to the medium using the matching points and weights.

[0100] Specifically, it is assumed that an advertisement request (event) occurs again from the aforementioned TV medium at 11:00 on April 14 (Fig. 6). At this time, the matching server can update the weight values ​​for uid=abcd-efgh-ijkl-mnop, uid=uudf-3ndf-4mv9-d8sfj, and uid= tvtvtv-ididididid-tvtv respectively, reflecting the 5-hour time difference from the previous event with the same matching key, and can increase the matching count value for uid= tvtvtv-ididididid-tvtv (step 14).

[0101] Next, let's assume that an advertisement request (event) occurs from the refrigerator medium at 13:00 on April 14th. Accordingly, the refrigerator medium can send an advertisement request message including {ip:106.249.255.82, os:Tizen, device:Refrigerator, tvid:refrig-484ri-c123-girfe} to the advertisement server. In response, the matching server can create a new matching table (e.g., set a matching count value, a weight, and a last match_date value) corresponding to uid=refrig-484ri-c123-girfe (step 15).

[0102] Additionally, the matching server can update the weight values ​​for uid=abcd-efgh-ijkl-mnop, uid=uudf-3ndf-4mv9-d8sfj, uid= tvtvtv-ididididid-tvtv, respectively. At this time, the weight values ​​can be updated based on the 3-hour time difference from the previous event with the same matching key.

[0103] Next, the matching server can determine a matching response ID for the refrigerator medium and provide it to the advertising server. More specifically, the matching server can provide the information uid= tvtvtv-ididididid-tvtv to the advertising server based on the fact that the matching count values ​​of uid=abcd-efgh-ijkl-mnop and uid=uudf-3ndf-4mv9-d8sfj are the same, but the weight value for uid=uudf-3ndf-4mv9-d8sfj is greater.

[0104] At this time, each data (or matching table) for each device may be deleted or the corresponding matching count value and weight value may be changed to 0 / terminated after 24 hours (or a preset time length value) from the previous connection time. This reflects the nature of advertising that prioritizes real-time, as the data is no longer meaningful after 24 hours (or 48 hours, depending on the setting) from the last connection.

[0105] According to this matching process, the entire system according to the present invention can operate as follows.

[0106] Figure 8 is a drawing briefly showing an example of the operation of the entire system according to an example of the present invention.

[0107] First, advertising requests may occur for mobile types A, B, and C in the same IP band (1 in Figure 8).

[0108] In response, the advertising server can respond with an advertisement corresponding to each medium (e.g., mobile A, B, C) based on the mobile data it already has (Fig. 8, 2).

[0109] Assume that an advertisement request is made by a refrigerator medium in the same IP zone. In this case, the refrigerator medium can transmit an advertisement request message to the advertisement server (3 in Figure 8).

[0110] In response, the advertising server can transmit a data matching request to the matching server (4 in FIG. 8).

[0111] The matching server can update the matching counter and weight parameters of mobiles A, B, and C using the matching key generated based on the IP, and compare them (5 in Fig. 8). More specifically, the matching server can first identify the mobile medium with the largest matching counter value. At this time, if there are multiple mediums with the largest matching counter values ​​as in Fig. 8 (e.g., match_count is the same as 300 for mobiles B and C), one medium among the multiple mediums can be selected by additionally considering the weight value (e.g., since mobile C has a higher weight in real time, mobile C is estimated to be the same user as the refrigerator).

[0112] Next, the matching server can provide information related to mobile C to the advertising server as a matching data response corresponding to the refrigerator medium (6 in FIG. 8).

[0113] In response, the advertising server can respond (send) an advertisement to a refrigerator medium that cannot identify an individual based on the behavioral data of mobile C (7 in Fig. 8).

[0114] This method can be applied not only to refrigerators, but also to other media such as PCs and IPTV.

[0115] Leveraging this matching model allows for audience targeting on non-personalized media, enabling targeted advertising using mobile audience data across a variety of non-mobile platforms. Targeting and campaign performance data can then be used as upgraded audience data for the respective media.

[0116] However, the operation example of the entire system described above is only an example applicable to the present invention, and according to an embodiment, the matching server may select / identify a mobile medium corresponding to a refrigerator medium by considering only a weight value without considering a matching counter.

[0117] To summarize the contents of the invention described above, the operation of the advertisement matching server (or matching server) according to the present invention can be generalized as follows.

[0118] FIG. 9 is a diagram showing an operation flow chart of a matching server according to an example of the present invention.

[0119] As illustrated in FIGS. 1, 3, and 8, the matching server according to the present invention can receive a data matching request message for a first medium from an advertising server through a data matching request message receiving unit (S810). The data matching request message can be configured in various forms. For example, the data matching request message can directly include IP information connected to the specific medium, or can include a matching key generated based on the IP information. The data matching request message further includes the uid of the medium.

[0120] As described above, if the data matching request message includes the IP information connected to the first medium, the matching server can generate a matching key based on the IP corresponding to the data matching request message through the matching key generation unit (S820). However, the matching key generation operation may be performed selectively, and if the data matching request message includes a matching key generated based on the IP information, step 820 may be omitted.

[0121] The matching server, through the matching model parameter management unit, can generate or update matching model parameters for one or more media associated with an IP-based matching key based on a data matching request message (S830). In this case, as a preferred example, the matching server may already be managing matching model parameters for one or more media (particularly, media including one or more mobile devices) associated with the matching key, as illustrated in FIGS. 5 to 8. Accordingly, the matching model parameters for the one or more media can be updated based on the data matching request message.

[0122] More specifically, the step in which the matching server generates or updates the matching model parameters through step S830 may include the following two steps:

[0123] When the first medium is determined to be a new medium in relation to the matching key (e.g., when the first medium is determined to be a medium that has newly generated an advertisement request event in relation to the matching key), a step of generating a matching model parameter for the first medium and setting it to an initial value (e.g., 1 or a value set to a fixed value according to a setting), and updating a matching model parameter for a medium other than the first medium among one or more media related to the matching key based on a data matching request message; and

[0124] A step of updating matching model parameters for one or more media, including the first medium, based on a data matching request message, when the first medium is determined to be an existing medium with respect to the matching key (e.g., when the first medium is determined to be a medium that has already generated an advertisement request event with respect to the matching key).

[0125] In other words, when the first medium is determined to be a new medium with respect to the matching key, as in step 3 of FIG. 4, steps 10 and 13 of FIG. 5, and step 15 of FIG. 6, the matching server may generate matching model parameters for the first medium and set them to initial values. Conversely, when the first medium is determined to be an existing medium with respect to the matching key, as in steps 5 to 7 of FIG. 4, steps 9, 12, and 13 of FIG. 5, and steps 14 and 16 of FIG. 6, the matching server may update matching model parameters for one or more media including the first medium based on a data matching request message.

[0126] As an example applicable to the present invention, the above-described matching model parameters may be configured with weight parameters (parameters updated based on the time difference between a data matching request message and the most recently received previous data matching request message related to the matching key prior to the data matching request message). In this case, the matching server may select identification information related to one medium (e.g., a second medium) as a matching response ID for the first medium based on the weight parameters. The selection of the matching response ID will be described in detail later.

[0127] When the matching model parameters are configured as weight parameters, as in the example described above, the weight parameters can be updated as follows. More specifically, when the first medium is determined to be an existing medium with respect to the matching key, the matching server can update the weight parameters as follows.

[0128] The matching server may calculate and update a weight parameter for the first medium according to the following mathematical formula 1 based on the data matching request message. In addition, the matching server may calculate and update a weight parameter for one or more media other than the first medium according to the following mathematical formula 2 based on the data matching request message.

[0129] [Mathematical Formula 1]

[0130]

[0131] [Equation 2]

[0132]

[0133] In the above mathematical formulas, the previous weight corresponds to the previous weight parameter value for each medium, and the elapsed time may correspond to the time difference between the data matching request message and the most recently received previous data matching request message related to the matching key before the data matching request message (see FIGS. 4 to 6).

[0134] As another example applicable to the present invention, the above-described matching model parameters may be configured to include a matching counter parameter that counts the number of connections for each medium, together with the above-described weight parameter. In this case, the matching server may utilize both the weight parameter and the matching counter parameter to select identification information related to one medium (e.g., a second medium) as a matching response ID for the first medium. The selection of the matching response ID will be described in detail later.

[0135] When the matching model parameters are composed of weight parameters and matching counter parameters as in the example described above, the matching counter parameter corresponding to a specific medium associated with the matching key can be increased / updated each time a data matching request message for the specific medium associated with the matching key is received (see FIGS. 4 to 6).

[0136] Next, the matching server may, through a data matching response message transmission unit, select identification information associated with the second medium as a matching response ID associated with the first medium based on matching model parameters for the one or more media associated with the matching key, and transmit a data matching response message including the identification information to the advertising server (S840).

[0137] In the present invention, the matching server selecting identification information associated with the second medium may include selecting identification information of a specific medium based on matching model parameters for all media except the first medium (associated with the data matching request message) among the media associated with the matching key generated based on IP. Accordingly, the selected identification information associated with the second medium may correspond to identification information of the second medium (e.g., a specific mobile device ID or an advertisement ID corresponding to the specific mobile device itself, etc.) or identification information of a third medium associated with the second medium but not of a mobile device type (e.g., when the second medium and the third medium are already associated, the matching server and the advertisement server can infer that the first medium is associated with the second medium of a mobile device type through the identification information of the third medium) (e.g., step 16 of FIG. 6).

[0138] Here, selecting the identification information of the specific medium described above may be implemented differently depending on the configuration example of the matching model parameter. For example, when the matching model parameter is composed only of the weight parameter described above, the matching server may select the identification information of the medium having the highest weight parameter among the media related to the matching key, excluding the first medium, as the identification information of the specific medium. As another example, when the matching model parameter is composed of the weight parameter and the matching counter parameter, the matching server may select the identification information of the medium having the highest matching counter parameter among the media related to the matching key, excluding the first medium, as the identification information of the specific medium. However, when there are multiple media having the highest matching counter parameters, excluding the first medium, the identification information of the medium having the highest weight parameter among the multiple media may be selected as the identification information of the specific medium.

[0139] In the above description, the first medium may correspond to a device having a media type other than a mobile device type. In other words, the first medium described above may correspond to a medium other than a mobile device type (e.g., a refrigerator, IPTV, CTV, PC, etc.). Furthermore, the second medium may correspond to a device having a mobile device type. Therefore, according to the above-described invention, the matching server may, in response to a data matching request message for a media type other than a mobile device type, provide identification information of a mobile device determined to have high relevance (or identification information related to the mobile device) to the advertising server. Through this, the advertising server may utilize mobile audience data to perform targeted advertising for various media other than mobile.

[0140] According to the present invention, the advertising server and matching server may be configured as distinct objects, as in the example described above, or as a single object (e.g., an advertising matching server or an advertising matching platform). If the advertising server and matching server are configured as a single object (e.g., an advertising matching platform), the advertising matching platform may operate as follows:

[0141] 1. Receive an advertising request message from the first medium.

[0142] 2. Based on the above advertisement request message, obtain a data matching request message for the first medium.

[0143] 3. Generate a matching key based on the IP corresponding to the above data matching request message.

[0144] 4. Based on the above data matching request message, create or update matching model parameters for one or more media associated with the above matching key.

[0145] 5. Selecting identification information associated with a second medium based on matching model parameters for the one or more media associated with the matching key, and transmitting an advertisement response message including an advertisement corresponding to the identification information to the first medium.

[0146] At this time, the first medium corresponds to a device having a medium type other than a mobile device type, and the second medium corresponds to a device having a mobile device type.

[0147] Although the preferred embodiments of the present invention have been described with reference to the drawings as described above, those skilled in the art can modify or change the present invention in various ways without departing from the spirit and scope of the present invention as set forth in the claims below.

[0148] The present invention can be used in various industries, including the advertising industry.

Claims

1. In the operation method of a matching server using real-time IP (Internet Protocol)-based multi-device audience identification, A step of receiving a data matching request message for a first medium from an advertising server; A step of generating a matching key based on an IP corresponding to the above data matching request message; A step of generating or updating matching model parameters for one or more media associated with the matching key based on the data matching request message; and A step of selecting identification information associated with a second medium based on a matching model parameter for the one or more media associated with the matching key, and transmitting a data matching response message including the identification information to the advertising server, The above first medium corresponds to a device having a medium type other than a mobile device type, A method of operating a matching server, characterized in that the second medium corresponds to a device having a mobile device type.

2. In paragraph 1, The step of creating or updating the above matching model parameters is: In a case where the first medium is determined to be a new medium in relation to the matching key, a step of generating a matching model parameter for the first medium and setting it to an initial value, and updating a matching model parameter for a medium other than the first medium among the one or more media based on the data matching request message; and A method of operating a matching server, characterized in that it comprises a step of updating matching model parameters for one or more media including the first medium based on the data matching request message when the first medium is determined to be an existing medium in relation to the matching key.

3. In paragraph 2, The above matching model parameters are: A method of operating a matching server, characterized in that the data matching request message includes a weight parameter that is updated based on a time difference between the most recently received previous data matching request message in relation to the matching key prior to the data matching request message.

4. In paragraph 3, If the first medium is determined to be an existing medium in relation to the above matching key: Based on the above data matching request message, the weight parameter for the first medium is calculated according to the following mathematical expression 1: [Mathematical Formula 1] Based on the above data matching request message, the weight parameters for the one or more media other than the first medium are calculated according to the following mathematical expression 2: [Equation 2] A method of operating a matching server, characterized in that the previous weight corresponds to a previous weight parameter value for each medium, and the elapsed time corresponds to a time difference between the data matching request message and the most recently received previous data matching request message in relation to the matching key before the data matching request message.

5. In paragraph 2, The above matching model parameters are: A weight parameter updated based on the time difference between the data matching request message and the most recently received previous data matching request message related to the matching key prior to the data matching request message; and A method of operating a matching server, characterized in that it includes a matching counter parameter that is counted according to the number of connections for each medium.

6. In paragraph 1, Selecting the identification information associated with the second medium, Including selecting identification information of a specific medium based on matching model parameters for all media except the first medium among the media related to the above matching key, A method of operating a matching server, characterized in that the identification information associated with the second medium corresponds to the identification information of the second medium or the identification information of a third medium associated with the second medium and not a mobile device type.

7. In paragraph 6, The above matching model parameter includes a weight parameter that is updated based on the time difference between the data matching request message and the most recently received previous data matching request message related to the matching key prior to the data matching request message, Selecting the identification information of the above specific medium, A method of operating a matching server, characterized in that it comprises selecting identification information of a medium having the highest weight parameter, excluding the first medium, among the media related to the matching key, as identification information of the specific medium.

8. In paragraph 6, A weight parameter updated based on the time difference between the data matching request message and the most recently received previous data matching request message related to the matching key prior to the data matching request message; and Includes a matching counter parameter that counts the number of connections for each medium, Selecting the identification information of the above specific medium, A method of operating a matching server, characterized in that it comprises selecting identification information of a medium having the highest matching counter parameter among the media related to the matching key, excluding the first medium, as identification information of the specific medium, and, if there are multiple media having the highest matching counter parameter, excluding the first medium, selecting identification information of a medium having the highest weight parameter among the multiple media as identification information of the specific medium.

9. In a matching server using real-time IP (Internet Protocol)-based multi-device audience identification, A data matching request message receiving unit that receives a data matching request message for a first medium from an advertising server; A matching key generation unit that generates a matching key based on an IP corresponding to the above data matching request message; A matching model parameter management unit that creates or updates matching model parameters for one or more media related to the matching key based on the data matching request message; and A data matching response message transmitting unit is included, which selects identification information associated with a second medium based on matching model parameters for the one or more media associated with the matching key, and transmits a data matching response message including the identification information to the advertising server. The above first medium corresponds to a device having a medium type other than a mobile device type, A matching server, characterized in that the second medium corresponds to a device having a mobile device type.

10. In the operation method of an advertising matching platform using real-time IP (Internet Protocol)-based multi-device audience identification, A step of receiving an advertisement request message from a first medium; A step of obtaining a data matching request message for the first medium based on the advertisement request message; A step of generating a matching key based on an IP corresponding to the above data matching request message; A step of generating or updating matching model parameters for one or more media associated with the matching key based on the data matching request message; and A step of selecting identification information associated with a second medium based on a matching model parameter for the one or more media associated with the matching key, and transmitting an advertisement response message including an advertisement corresponding to the identification information to the first medium, The above first medium corresponds to a device having a medium type other than a mobile device type, A method of operating an advertising matching platform, characterized in that the second medium corresponds to a device having a mobile device type.

Images