Advertisement attribution method and system and electronic device

Abstract

The invention provides an advertisement attribution method and system. The method comprises the following steps: S1, carrying out unique identification processing on equipment of each media terminal; s2, performing data cleaning on the data returned by each media terminal, and storing the data in corresponding databases according to different data types; s3, performing attribution processing on the cleaned and stored data; and S4, according to an attribution processing result, analyzing and displaying the result. Multi-source equipment is subjected to normalization processing and data cleaning, meanwhile, an attribution method is innovatively improved, data are redundant in different storage engines, dynamic routing is performed according to the characteristics of a strategy, query matching is performed by a proper storage engine, and in order to improve the attribution efficiency and accuracy, the attribution efficiency and accuracy are improved. In addition, the attribution strategy configuration and the final attribution effect are visualized, so that the time required for strategy effect analysis can be greatly reduced.

Description

Technical Field

[0001] The present application relates to the technical field of advertising analysis, and in particular to an advertising attribution method and system, and an electronic device. Background Art

[0002] Advertising attribution methods are used to evaluate and analyze the contribution of different advertising channels (such as which channels are used to download apps, open landing pages, mini programs, etc.) and activities to customer conversion and sales. Through these methods, companies can understand the impact of various marketing activities on the customer decision-making process, which can be used to optimize and improve advertising expenditures and effects.

[0003] The data results of advertising attribution are an important basis for measuring advertising effectiveness and evaluating channel quality. They can help advertisers optimize advertising materials and efficiently carry out new customer acquisition and marketing promotion. Therefore, advertising attribution technology is a key link in measuring advertising effectiveness, but the existing technology has the following technical defects and problems:

[0004] 1. Difficulty in cross-platform attribution:

[0005] Due to restrictions in privacy policies (such as GDPR) and the gradual disappearance of third-party cookies, cross-platform sharing of user data is limited. However, users usually access ads on multiple devices and platforms. Therefore, the interaction between users and ads lacks a unified user ID, making cross-platform attribution difficult. As a result, existing technologies find it difficult to obtain complete data in such situations, and the data sources are different and difficult to integrate.

[0006] 2. Data Accuracy and Consistency:

[0007] Due to the cross-device and multi-platform nature, the data sources are diverse and of varying quality. In addition, human errors occur during data collection and processing. Advertising attribution relies on a large amount of data, which leads to inaccurate or inconsistent data and biased attribution results.

[0008] 3. Model complexity and interpretability:

[0009] In order to improve the accuracy of analysis, many existing models, such as the multi-touch attribution model, use complex machine learning algorithms. However, the interpretability problem of machine learning algorithms makes it difficult to explain their decision-making process. In practice, advertising placement operations require subjective selectivity on the part of advertisers, and it is necessary to find a balance between accuracy and interpretability.

[0010] In addition, there are other issues that need to be considered, such as balancing the amount of data and computing power, and the fact that different advertising campaigns and user behavior patterns require different attribution windows. Summary of the invention

[0011] To solve the above problem, the present invention provides an advertising attribution method, comprising the following steps:

[0012] S1: uniquely identify each media terminal device;

[0013] S2: Clean the data sent back by each media terminal and store it in the corresponding database according to different data types;

[0014] S3: Perform attribution processing on the cleaned and stored data;

[0015] S4: Analyze and display the results based on the attribution processing results.

[0016] Furthermore, in step S2, storing in a corresponding database according to different data types includes:

[0017] Use clickhosue for complex analytical data;

[0018] Simple equal value query data uses redis;

[0019] Use Hbase for primary key query data in the returned table;

[0020] Use MYSQL for dictionary-type query data.

[0021] Further, step S3: attribution processing is performed on the cleaned and stored data as follows:

[0022] For the triggered event type, the corresponding attribution group is selected for processing. Each attribution group contains several attribution methods, which are executed sequentially. If the attribution method hits the target, the branch is pruned and returned to obtain the attribution result.

[0023] Events that do not match any rules are considered natural traffic and are routed to different destinations based on the configuration.

[0024] Furthermore, the attribution methods included in the attribution group are in the following order: clipboard attribution, asa attribution, precise click, fuzzy click, and deep attribution.

[0025] Furthermore, step S1: performing unique identification processing on the device of each media terminal includes:

[0026] S11: Build your own device ID, integrate the legal terminal device information acquisition function in the SDK, and generate a unified device ID based on the device information;

[0027] S12: Data cleaning and integration: De-duplicate and integrate the collected device IDs, aggregate multiple device IDs of the same user, and identify device associations;

[0028] S13: Database collision processing: The data sent back by the media side is compared with the device ID library in its own device library by comparing the device ID. If the same ID is found, it can be mapped to the self-built device ID.

[0029] Furthermore, in step S11, if the terminal device information is a PC device, its device information includes:

[0030] Device network information: network access mode, wireless router identification and IP address, surrounding WIFI list, network operator information, network base station information, network connection status;

[0031] Device application information: SDK host application package name and version number, and running process information;

[0032] Basic device information: motherboard serial number, CPU ID, systemuuid, HDSerialNumber, systemMachinePC, productIDPC, memoryType, cpuHardware, macAddress.

[0033] Furthermore, in step S11, if the terminal device information is Web, its device information includes: Canvas fingerprint, cookies information, and ua client browser information.

[0034] Furthermore, in step S11, if the terminal device information is a mobile terminal operating system, the device information is:

[0035] Basic information of Android device:

[0036] Device brand, device manufacturer, device model, device name, device system type and version information, device basic configuration, device basic settings, device environment, sensor information, SD card information;

[0037] Device identification information: AndroidID, GAID, IMEI, OAID;

[0038] Device network information: network access mode, wireless router identification and IP address, surrounding WIFI list, network operator information, network base station information, network connection status;

[0039] Device application information: SDK host application package name and version number, and running process information;

[0040] Basic information of IOS device:

[0041] Device brand, device manufacturer, device model, device name, device system type and version information, device basic configuration, device basic settings, device environment, sensor information, SD card information;

[0042] Device identification information: IDFA, IDFV, CAID, OAID;

[0043] Device network information: network access mode, wireless router identification and IP address, surrounding WIFI list, network operator information, network base station information, network connection status;

[0044] Device application information: SDK host application package name and version number, and running process information.

[0045] The present application also provides an advertising attribution system, including an identification processing unit, a data cleaning and storage unit, a data attribution analysis unit, and a result display unit, wherein:

[0046] The identification processing unit is used to perform unique identification processing on the equipment of each media terminal;

[0047] The data cleaning and storage unit is used to clean the data sent back by each media terminal and store it in the corresponding database according to different data types;

[0048] The data attribution analysis unit is used to perform attribution processing on the cleaned and stored data;

[0049] The result display unit is used to analyze and display the results according to the attribution processing results.

[0050] The present application also provides an electronic device, which includes a processor and a memory, wherein the memory is used to store an executable program, and the processor is used to execute the advertising attribution method.

[0051] In order to have a clearer and more comprehensive understanding of the present invention, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and those skilled in the art can obtain other drawings based on these drawings without creative work.

[0053] Figure 1 This is a flowchart of the advertising attribution method for this application.

[0054] Figure 2This is a structural diagram of the application of the advertising attribution method of this application in the advertising system.

[0055] Figure 3 This is a schematic diagram of storing different data types in the embodiments of this application.

[0056] Figure 4 This is a flow chart of the attribution method according to an embodiment of the present application. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0058] In ad attribution, an “event” generally refers to a specific behavior or action that occurs when a user interacts with an ad or app. These events can include:

[0059] Click event: The user clicks an ad link or button.

[0060] View Event: An event in which a user viewed a video ad or specific content.

[0061] Conversion event: A user completes a specific target behavior, such as registration, purchase, download, etc.

[0062] Viewing event: A user visited a specific web page or a specific part of an app.

[0063] Interaction events: other interactive behaviors between users and ads or content, such as sharing, commenting, etc.

[0064] For an advertising system, each event is an important basis for evaluating advertising effectiveness. The advertising system analyzes the user's behavior path based on these events, and then conducts attribution analysis to evaluate the effectiveness and value of advertising. Different attribution strategies can be applied to different events to better understand and optimize advertising effectiveness.

[0065] See also Figure 1 , Figure 1 The following is a flowchart of the advertising attribution method of the present application. In view of the defects of the prior art, the present application proposes an advertising attribution method, which includes the following steps:

[0066] S1: uniquely identify each media terminal device;

[0067] S2: Clean the data sent back by each media terminal and store it in the corresponding database according to different data types;

[0068] S3: Perform attribution processing on the cleaned and stored data;

[0069] S4: Analyze and display the results based on the attribution processing results.

[0070] See also Figure 2 , Figure 2 This is a schematic diagram of the structure of the application of the advertising attribution method in the advertising system. Figure 2 As shown, event attribution processing on the advertiser side can help advertising media optimize their delivery strategies, thereby improving advertisers' return on investment.

[0071] First, the advertiser's advertising pitcher places ads on the media side and gives an initial strategy. The media pushes ads to appropriate users, and the users generate events such as exposure and clicks on the ads. The media SDK forwards these events on the user's terminal device to the advertiser's monitoring service, and the user may download or open the game. The advertiser monitors the ads and processes data on events such as exposure and clicks from different media, cleans them into unified events, stores one copy in the database, and forwards one copy to Kafka.

[0072] Events such as users installing games, registering characters, and making payments are important indicators for evaluating advertising revenue. These messages are also cleaned and packaged to produce unified events, one copy is stored in the database, and one copy is forwarded to Kafka. Based on these two types of data, data attribution can be performed, and reports can be generated based on the attribution results and sent back to the media for advertising optimization.

[0073] The technical solution of the present application is described in detail below in conjunction with various specific or preferred embodiments.

[0074] S1: Perform unique identification processing on the device of each media terminal.

[0075] The data returned by different media and the data returned by different businesses on the advertiser side will have their own characteristics and some incompatibilities. We need to process these details in advance and unify them into a unified event. As a preferred implementation, the method includes the following S11-S13:

[0076] S11: Build your own device ID, integrate the legal terminal device information acquisition function in the SDK, and generate a unified device ID based on the device information;

[0077] Due to the difficulty of cross-platform device identification, this application integrates the relevant device IDs that may be obtained, and builds a self-built device ID. For the data generated by its own business, the relevant legal device information is integrated in the SDK and passed to the service side for device uniqueness marking. The self-built device ID is based on the following information of various different devices, and the obtained weak features are comprehensively processed.

[0078] Regarding the analysis of device sources, this application mainly analyzes the most commonly used terminal device types, including PC, web, adndroid, and ios. The specific information of each type is as follows:

[0079] The device information included on the PC side is:

[0080] Device network information: network access mode, wireless router identification (BSSID, SSID) and IP address, surrounding WIFI list, network operator information, network base station information, network connection status.

[0081] Device application information: SDK host application package name and version number, and running process information.

[0082] Basic device information: motherboard serial number, CPU ID, systemuuid, HDSerialNumber, systemMachinePC, productIDPC, memoryType, cpuHardware, macAddress, etc. The device information included on the Web client is as follows:

[0083] Canvas fingerprint, cookies information, UA client browser information.

[0084] The device information included on the Android side is:

[0085] Basic device information: device brand, device manufacturer, device model, device name, device system type and version information, device basic configuration, device basic settings, device environment, sensor information (magnetic field, gyroscope, acceleration sensor), and SD card information.

[0086] Device identification information: AndroidID, GAID, IMEI, OAID (anonymous device identifier).

[0087] Device network information: network access mode, wireless router identification (BSSID, SSID) and IP address, surrounding WIFI list, network operator information, network base station information, network connection status.

[0088] Device application information: SDK host application package name and version number, running process information (only obtain the current process name).

[0089] The device information included on the iOS side is:

[0090] Basic device information: device brand, device manufacturer, device model, device name, device system type and version information, device basic configuration, device basic settings, device environment, sensor information (magnetic field, gyroscope, acceleration sensor), and SD card information.

[0091] Device identification information: IDFV (application developer identifier), IDFA (advertising identifier), OAID (anonymous device identifier), CAID (China anonymous device identifier).

[0092] Device network information: network access mode, wireless router identification (BSSID, SSID) and IP address, surrounding WIFI list, network operator information, network base station information, network connection status.

[0093] Device application information: SDK host application package name and version number, running process information (only obtain the current process name).

[0094] S12: Data cleaning and integration: De-duplicate and integrate the collected device IDs, aggregate multiple device IDs of the same user, and identify device associations. After data cleaning and integration in this step, different device IDs of the same device can be identified, so that when some device IDs change, it can still be identified that this is the same device.

[0095] S13: Database collision processing: The data sent back by the media side is compared with the device ID library in its own device library. If the same ID is found, it can be mapped to the self-built device ID. After the data is aggregated into the system, the OLAP database is used for data processing to improve the recognition accuracy of device uniqueness.

[0096] S2: Clean the data sent back by each media terminal and store it in the corresponding database according to different data types.

[0097] See also Figure 3 As shown, Figure 3 This is a schematic diagram of storing different data types in the embodiment of this application. Step S2 cleans the data sent back by each media terminal, including filtering outliers, unifying uppercase and lowercase letters, and converting special values, so as to ensure the rationality and stability of the data. After processing and cleaning, these data will be dropped into the table according to the pre-configured routing rules to adapt to subsequent different types of queries, so as to meet the high performance, high availability and high stability of the entire service.

[0098] The prior art mainly uses one storage method to perform data analysis. However, due to the complexity of data analysis for advertising attribution, in order to meet the high performance, high availability and high stability of the service, this application uses different storage engines for optimization, and stores different data types in corresponding databases, including the following:

[0099] For complex analytical data use clickhosue:

[0100] As a columnar database, ClickHouse performs well in processing large-scale data and can quickly execute complex queries such as aggregation and statistics. In addition, it supports high-concurrency reading and is suitable for real-time data analysis. In addition, query performance can be further improved through compression, partitioning, and indexing.

[0101] Simple equal value query data uses redis:

[0102] Redis is an in-memory database with extremely high query speed compared to other systems. It is suitable for scenarios that require extremely low latency, such as user session storage, real-time statistics, etc. Its key-value pair storage model makes it possible to quickly access specific data. Therefore, this application will use Redis for simple equivalent query data.

[0103] Use Hbase to query the data of the primary key of the table:

[0104] HBase is a distributed column storage database suitable for random read and write operations, and can efficiently support the storage and access of large-scale data. Therefore, this application is based on the characteristics of Hbase, which is particularly suitable for processing massive data sets and using it for primary key queries in tables. In addition, Hbase can be seamlessly integrated with other tools in the Hadoop ecosystem (such as MapReduce) to facilitate data processing and analysis, thereby expanding data analysis capabilities.

[0105] Use MYSQL for dictionary-type query data, and utilize MySQL's index mechanism to quickly retrieve the required data.

[0106] S3: Perform attribution processing on the cleaned and stored data;

[0107] The attribution processing process first reads the configured policy orchestration, orchestrates according to the latest configured policy, obtains the policy group, forms the attribution logic topology, that is, the attribution method group, and then executes the specific policy according to the topological order and outputs the result.

[0108] This application stores different data types in different storage engine databases, configures different strategies based on different triggered event types, and selects corresponding attribution groups for processing, so that data can be quickly queried for attribution analysis.

[0109] As a preferred implementation, the trigger configuration rule is: select two dimensions of the event (tenant and type) as routing keys, set different scenarios, such as media and self-marketing, etc. First, the event is differentiated and routed to different strategy groups by routing keys, and attribution processing is performed according to the set scenario strategy.

[0110] For example, if the tenant of event A is 0001, the type is activation, and two policy groups, media and self-marketing, are configured, then this event will be attributed to the media policy group and marketing policy group in the activation policy group of tenant 0001, and two attribution results will be generated. Each attribution policy group contains several attribution methods, which are executed sequentially. If the attribution method hits the target, the method will be pruned and returned to obtain the attribution result.

[0111] In addition, as a more preferred implementation, each attribution group contains several attribution methods, and they are executed sequentially. If they are hit by the attribution method, they are pruned and returned to obtain the attribution result. During the execution process, if any attribution method is hit, it is pruned and returned, which can improve efficiency. In addition, in practice, the sorting configuration can be performed according to the characteristics of each attribution method. For example, the clipboard attribution is generally greater than the click attribution, so the clipboard attribution can be tried before the click attribution. Each attribution strategy outputs a result.

[0112] As a preferred implementation, the attribution methods included in each attribution group are: clipboard attribution, asa attribution, precise click, fuzzy click, and deep attribution.

[0113] Events that do not match any rules are considered natural traffic and are routed to different destinations based on the configuration. For example, events attributed to media need to be sent back to the corresponding media, while natural traffic only needs to be stored in the database.

[0114] See also Figure 4 As shown in the attribution method flow chart, this application has selected different storage engines for data operation types, so that different storage engines can be selected according to the characteristics of the event. When attributing, different attribution methods are selected:

[0115] Fuzzy click (media click fuzzy matching):

[0116] For media clicks, because this operation has a configurable matching factor and an uncertain score threshold, a more complex query statement is required to implement it. In addition, the concurrency of general installation events is not too high. Clickhouse is selected as the query engine to meet the query requirements of this scenario.

[0117] ASA Attribution:

[0118] ASA (Apple Search Ads) is an advertising attribution method in the Apple ecosystem. It queries Apple to see if there is a corresponding attribution result for the triggered event. Here, it is necessary to match based on multiple IDs of the device.

[0119] Exact click (media click exact match):

[0120] This type of event is all about equal value queries, but there are several configurable query fields. This application maps all configurable fields and click event IDs to KV, uses zset's redis data format for storage, and uses the timestamp of the click event as the score. This type of key is cleaned regularly, and only the events with the largest possible attribution window need to be saved for click data.

[0121] The clicks corresponding to several fields are attributed in the configured attribution method of first, last, and second, and then all the information of the entire click is queried in Hbase based on the rowKey composed of event ID and event time.

[0122] Redis provides fast KV storage and efficient query capabilities, which is suitable for processing click event data that needs to be quickly accessed. The score is the timestamp of the click event, and data cleaning is performed regularly to ensure that only events within the maximum possible attribution window are saved. Combined with HBase for full information query, complex attribution logic (such as first and last attribution) can be implemented.

[0123] Deep attribution (deep business matching):

[0124] The deep attribution method is based on the equivalent query of the account ID. You can directly query Hbase based on the account ID to obtain all the information. HBase is suitable for storing large-scale structured data and can efficiently query based on the account ID to obtain all relevant information.

[0125] Clipboard attribution (clipboard matching): This part of the events does not need to be queried. The results can be directly constructed according to the relevant fields in the event. This type of event does not need to be queried, so a specific storage engine is not required, reducing the dependence on the storage engine.

[0126] From the above analysis, we can see that this application configures different attribution strategy methods for different events, and abstracts the attribution strategy. By adopting different strategy groups, we can generate different strategies and compare the effects, obtain the results of A / B testing, and ensure the online effect.

[0127] In addition, when cleaning data, big data risk control can be used to prevent data from being cheated, ensuring as much as possible that the data attribution is normal user data, and performing redundant storage in multiple storage engines based on data characteristics and subsequent usage scenarios.

[0128] S4: Analyze and display the results based on the attribution processing results.

[0129] Result analysis shows that this application uses a data storage engine based on clickhouse. In addition, a low-code platform can be used to build visual reports. Different businesses can build different dashboards based on their own business characteristics, which are mapped to the server as different query statements, which is simple and fast.

[0130] In addition, this application can also visualize the attribution strategy configuration and the final attribution effect, and display the executed topological logic, which can greatly reduce the time required for strategy effect analysis. It can also use OLAP online analysis, combined with fixed chart analysis, to upgrade the special data analysis capabilities. Through the low-code platform, special data analysis can be solidified into charts, which will take effect in their respective application scenarios, isolate and do not interfere with other scenarios, and form persistence within the scenario.

[0131] It can be seen that this application normalizes devices from multiple sources and performs data cleaning. At the same time, it makes innovative improvements to the attribution method, and redundantly stores data in different storage engines. Dynamic routing is performed according to the characteristics of the strategy, and query matching is performed by a suitable storage engine. In order to improve the attribution efficiency and accuracy, the attribution strategy configuration and the final attribution effect are visualized, which can greatly reduce the time required for strategy effect analysis.

[0132] Based on the above embodiments, the present application further provides an advertising attribution system, including an identification processing unit, a data cleaning and storage unit, a data attribution analysis unit, and a result display unit, wherein:

[0133] The identification processing unit is used to perform unique identification processing on the equipment of each media terminal;

[0134] The data cleaning and storage unit is used to clean the data sent back by each media terminal and store it in the corresponding database according to different data types;

[0135] The data attribution analysis unit is used to perform attribution processing on the cleaned and stored data;

[0136] The result display unit is used to analyze and display the results according to the attribution processing results.

[0137] The present application also provides a computer storage medium, which can store an executable program. When the executable program runs on a computer, the computer executes the advertising attribution method described in any of the above embodiments.

[0138] The present application also provides an electronic device, which includes a processor and a memory, wherein the memory is used to store an executable program, and the processor is used to execute the executable program to implement the above-mentioned advertising attribution method.

[0139] It should be noted that, a person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and the storage medium can include but is not limited to: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0140] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An advertising attribution method, characterized in that: The steps include: S1: uniquely identify each media terminal device; S2: Clean the data sent back by each media terminal and store it in the corresponding database according to different data types; S3: Perform attribution processing on the cleaned and stored data; S4: Analyze and display the results based on the attribution processing results.

2. The advertising attribution method according to claim 1, characterized in that: In step S2, storing the data in the corresponding database according to different data types includes: Use clickhosue for complex analytical data; Simple equal value query data uses redis; Use Hbase for primary key query data in the returned table; Use MYSQL for dictionary-type query data.

3. The advertising attribution method according to claim 1, characterized in that: Step S3: The cleaned and stored data is attributed as follows: For the triggered event type, the corresponding attribution group is selected for processing. Each attribution group contains several attribution methods, which are executed sequentially. If the attribution method hits the target, the branch is pruned and returned to obtain the attribution result. Events that do not match any rules are considered natural traffic and are routed to different destinations based on the configuration.

4. The advertising attribution method according to claim 3, characterized in that: The attribution methods included in the attribution group are: clipboard attribution, asa attribution, exact click, fuzzy click, and deep attribution.

5. The advertising attribution method according to claim 1, characterized in that: Step S1: uniquely identifying each media terminal device includes: S11: Build your own device ID, integrate the legal terminal device information acquisition function in the SDK, and generate a unified device ID based on the device information; S12: Data cleaning and integration: De-duplicate and integrate the collected device IDs, aggregate multiple device IDs of the same user, and identify device associations; S13: Database collision processing: The data sent back by the media side is compared with the device ID library in its own device library by comparing the device ID. If the same ID is found, it can be mapped to the self-built device ID.

6. The advertising attribution method according to claim 5, characterized in that: In step S11, if the terminal device information is a PC device, its device information includes: Device network information: network access mode, wireless router identification and IP address, surrounding WIFI list, network operator information, network base station information, network connection status; Device application information: SDK host application package name and version number, and running process information; Basic device information: motherboard serial number, CPU ID, systemuuid, HDSerialNumber, systemMachinePC, productIDPC, memoryType, cpuHardware, macAddress.

7. The advertising attribution method according to claim 5, characterized in that: In step S11, if the terminal device information is Web, its device information includes: Canvas fingerprint, cookies information, and ua client browser information.

8. The advertising attribution method according to claim 5, characterized in that: In step S11, if the terminal device information is a mobile terminal operating system, the device information is: Basic information of Android device: Device brand, device manufacturer, device model, device name, device system type and version information, device basic configuration, device basic settings, device environment, sensor information, SD card information; Device identification information: AndroidID, GAID, IMEI, OAID; Device network information: network access mode, wireless router identification and IP address, surrounding WIFI list, network operator information, network base station information, network connection status; Device application information: SDK host application package name and version number, and running process information; Basic information of IOS device: Device brand, device manufacturer, device model, device name, device system type and version information, device basic configuration, device basic settings, device environment, sensor information, SD card information; Device identification information: IDFA, IDFV, CAID, OAID; Device network information: network access mode, wireless router identification and IP address, surrounding WIFI list, network operator information, network base station information, network connection status; Device application information: SDK host application package name and version number, and running process information.

9. An advertising attribution system, characterized in that: It includes an identification processing unit, a data cleaning and storage unit, a data attribution analysis unit, and a result display unit, wherein: The identification processing unit is used to perform unique identification processing on the equipment of each media terminal; The data cleaning and storage unit is used to clean the data sent back by each media terminal and store it in the corresponding database according to different data types; The data attribution analysis unit is used to perform attribution processing on the cleaned and stored data; The result display unit is used to analyze and display the results according to the attribution processing results.

10. An electronic device, characterized in that: The electronic device includes a processor and a memory, the memory is used to store an executable program, and the processor is used to execute the executable program to implement the advertising attribution method described in claims 1-8.

Images