Advertising keyword optimization methods, devices, electronic equipment, and readable storage media

Abstract

This application discloses a method, apparatus, electronic device, and readable storage medium for keyword optimization in advertising. The method includes: acquiring advertising delivery data for a target game on different platforms, and user behavior data within the target game; the advertising delivery data includes delivery data corresponding to multiple preset keywords; based on the advertising delivery data, determining interaction event information corresponding to the same keyword on different platforms, and matching target event information in the user behavior data according to the user identifier associated with the interaction event information; based on the interaction event information and the target event information, determining the delivery effect information of the same keyword on different platforms, and adjusting the multiple keywords according to the delivery effect information. This technical solution can specifically adjust multiple preset keywords for advertising delivery, improving the efficiency of keyword optimization.

Description

Technical Field

[0001] This application relates to the field of computers, and more particularly to a method, apparatus, electronic device, and readable storage medium for optimizing keywords in advertising. Background Technology

[0002] In the field of digital advertising for games, the selection and optimization of advertising keywords are key to acquiring high-quality users (users who effectively install, register, or make purchases in the game). Game advertisers typically place game-related ads on multiple media channels (such as search engines, news feeds, video platforms, etc.) and set different keywords for each channel.

[0003] The user attributes, behavioral habits, and advertising mechanisms of different media channels often differ significantly, leading to substantial variations in the customer acquisition effectiveness and costs (such as cost per install (CPI) and cost per paying user (CPA)) for the same game keyword across different channels. However, existing keyword optimization methods primarily rely on manual analysis or simple report statistics, making it difficult to comprehensively and accurately assess the performance of the same keyword across different platforms, resulting in relatively low efficiency in keyword optimization. Summary of the Invention

[0004] This application provides a keyword optimization method, apparatus, electronic device, and readable storage medium for advertising. By matching user identifiers associated with interactive event information, the method identifies the target event information of users corresponding to interactive event information of the same keyword on different platforms. It can evaluate the differences in customer acquisition effect of the same keyword on different platforms based on the correlation between interactive event information and the target event information of the corresponding users, thereby comprehensively and accurately determining the advertising effect of the same keyword on different platforms. This allows for targeted adjustments to multiple keywords preset for advertising, improving the efficiency of keyword optimization.

[0005] In a first aspect, embodiments of this application provide a keyword optimization method for advertisements, the method comprising: Acquire advertising data for the target game on different platforms, as well as user behavior data within the target game; the advertising data includes advertising data corresponding to multiple preset keywords. Based on the advertising data, the interactive event information corresponding to the same keyword on different platforms is determined, and the target event information is matched in the user behavior data according to the user identifier associated with the interactive event information. Based on the interaction event information and the target event information, the advertising performance information of the same keyword on different platforms is determined, and the multiple keywords are adjusted according to the advertising performance information.

[0006] Optionally, the interactive event information includes keyword click events and / or keyword display events; The step of determining the interaction event information corresponding to the same keyword on different platforms based on the advertising data includes: Extract several target keywords from the advertising data of different platforms whose semantic similarity is greater than or equal to a preset threshold. Based on a preset mapping relationship, the target keywords are mapped to the same standard keyword; the mapping relationship is the mapping relationship between the standard keyword and semantically similar synonyms; Interaction event information is extracted from the delivery data corresponding to the several target keywords, and used as the interaction event information corresponding to the same standard keyword on different platforms. Accordingly, determining the advertising performance information for the same keyword on different platforms includes: Determine the performance of the same standard keywords on different platforms.

[0007] Optionally, the interaction event information includes keyword click event information; the user behavior data includes activation event information of the target game; The step of matching the target event information in the user behavior data based on the user identifier associated with the interaction event information includes: Pre-assign a click attribution primary key containing the associated user identifier to the keyword click event information, and pre-assign an activation attribution primary key containing the associated user identifier to the activation event information; The click attribution primary key and activation attribution primary key containing the same user identifier are constructed into an association index pair, and the activation event information associated with the interaction event information is matched in the user behavior data based on the association index pair as the target event information.

[0008] Optionally, determining the advertising performance information of the same keyword on different platforms based on the interaction event information and the target event information includes: Obtain the first timestamp of the keyword click event information, and determine the target backtracking period based on the first timestamp; the target backtracking period is the duration interval extending backward from the moment the keyword click event occurs; Based on the user identifier associated with the target event information, determine each event information associated with the target event information in the user behavior data, and obtain the second timestamp corresponding to each event information respectively; Based on the second timestamp, determine the event information whose occurrence time falls within the target backtracking period from the various event information; Based on the sequence of the keyword click event information, the target event information, and the event information within the target backtracking period, a user behavior path corresponding to the same keyword within the target backtracking period is constructed, and the advertising effect information of the same keyword on different platforms is determined based on the user behavior path.

[0009] Optionally, determining the ad performance information of the same keyword on different platforms based on the user behavior path includes: If the event information within the target retrospective period includes paid event information of the target game, the time difference between each event information and the paid event information is determined in the user behavior path, and the advertising effect information of the same keyword on different platforms is determined based on the time difference.

[0010] Optionally, adjusting the multiple keywords based on the campaign performance information includes: The campaign performance of the same keyword on each platform is evaluated based on the campaign performance information, and the evaluation results for each platform are obtained respectively. Platforms whose evaluation results meet the preset conditions are identified as first-type platforms, and the output value of the same keyword on the first-type platforms is increased. Platforms whose evaluation results do not meet the preset conditions are identified as second-type platforms, and the output value of the same keyword on the second-type platforms is reduced.

[0011] Optionally, adjusting the multiple keywords based on the campaign performance information includes: The campaign performance of the same keyword on each platform is evaluated based on the campaign performance information, and the evaluation results for each platform are obtained respectively. Platforms whose evaluation results meet the preset conditions are identified as first-type platforms, and platforms whose evaluation results do not meet the preset conditions are identified as second-type platforms. Search the historical search term database to identify candidate search terms with a semantic similarity greater than or equal to a preset threshold to the same keyword, and associate the candidate search terms with the same keyword and deliver them to the first type of platform; Generate a blocking instruction for the same keyword and send the blocking instruction to the second type of platform; the blocking instruction is used to trigger the same keyword to stop being displayed.

[0012] Secondly, embodiments of this application provide a keyword optimization device for advertisements, the device comprising: The acquisition unit is used to acquire advertising data of the target game on different platforms, as well as user behavior data within the target game; the advertising data includes advertising data corresponding to multiple preset keywords. The first determining unit is used to determine the interactive event information corresponding to the same keyword on different platforms based on the advertising delivery data. The matching unit is used to match target event information in the user behavior data based on the user identifier associated with the interaction event information; The second determining unit is used to determine the advertising effect information of the same keyword on different platforms based on the interaction event information and the target event information; The optimization unit is used to adjust the multiple keywords based on the campaign performance information.

[0013] Thirdly, embodiments of this application provide an electronic device, which includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the method described in the first aspect.

[0014] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the method described in the first aspect.

[0015] In this embodiment, advertising data for the target game on different platforms and user behavior data within the target game are first obtained. The advertising data includes advertising data for multiple preset keywords. Based on the advertising data for the target game on different platforms, interactive event information for the same keyword on different platforms is determined. Target event information is then matched in the user behavior data based on the user identifier associated with the interactive event information for the same keyword on different platforms. Finally, based on the interactive event information and the matched target event information, the advertising effect information for the same keyword on different platforms is determined, and the multiple preset keywords are adjusted according to the advertising effect information.

[0016] By matching the user identifiers associated with the interaction event information, the target event information of the users corresponding to the interaction event information of the same keyword on different platforms can be identified. Based on the correlation between the interaction event information and the target event information of the corresponding users, the difference in customer acquisition effect of the same keyword on different platforms can be evaluated. This allows for a comprehensive and accurate determination of the advertising effect of the same keyword on different platforms, and further targeted adjustments can be made to the multiple keywords preset for advertising, thereby improving the efficiency of keyword optimization. Attached Figure Description

[0017] Figure 1 This is a schematic diagram illustrating keyword optimization scenarios in advertising within related technologies; Figure 2 This is a flowchart illustrating an advertising keyword optimization method according to an embodiment of this application; Figure 3 A flowchart illustrating a specific implementation process for matching target event information, provided in an embodiment of this application; Figure 4 A flowchart illustrating a specific implementation process for keyword adjustment provided in this application embodiment; Figure 5 A flowchart illustrating another specific implementation process for keyword adjustment provided in this application embodiment; Figure 6 This is a schematic diagram of the structure of an advertising keyword optimization device according to an embodiment of this application; Figure 7 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0019] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0020] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0021] The keyword optimization method, apparatus, electronic device, and readable storage medium for advertisements provided in this application will be described in detail below with reference to the accompanying drawings and through specific embodiments and application scenarios.

[0022] Please see Figure 1 First, we will provide an example of keyword optimization scenarios for advertising in related technologies: like Figure 1 As shown, Figure 1 This diagram illustrates a keyword optimization scenario in advertising within the relevant technology field. In the game digital advertising sector, the selection and optimization of advertising keywords are crucial for acquiring high-quality users (users who effectively install, register, or make payments for the game). Game advertisers typically place game-related ads on multiple media channels (such as search engines, news feeds, video platforms, etc.) and set different keywords for each channel. Furthermore, the same keyword may attract different users on different media channels, for example... Figure 1 As shown, the same advertising keyword is placed on media channel A and media channel B respectively. It may acquire user 1 on media channel A and user 2 on media channel B.

[0023] The user attributes, behavioral habits, and advertising mechanisms of different media channels often differ significantly, leading to substantial variations in the customer acquisition effectiveness and costs (such as Cost Per Install (CPI) and Cost Per Activity (CPA)) for the same game keyword across different channels. However, existing keyword optimization methods primarily rely on manual analysis or simple report statistics, making it difficult to comprehensively and accurately evaluate the performance of the same keyword across different platforms. Furthermore, they cannot automatically and dynamically select specific keywords for targeted optimization, resulting in relatively low efficiency in keyword optimization.

[0024] It is worth noting that the above description of keyword optimization scenarios for advertisements in related technologies is only an illustrative example. In actual applications, there may be other types of keyword optimization scenarios for advertisements, which are not specifically limited here.

[0025] To address the aforementioned issues, this application provides a keyword optimization method for advertising, which can specifically adjust multiple preset keywords for ad placement, thereby improving the efficiency of keyword optimization. For example... Figure 2 As shown, Figure 2 This is a flowchart illustrating an advertising keyword optimization method according to an embodiment of this application. The method includes the following steps: S201. Obtain advertising data for the target game on different platforms, as well as user behavior data within the target game.

[0026] The advertising data can include the advertising data corresponding to multiple preset keywords.

[0027] Target games refer to specific online game products that advertisers are promoting and that they need to acquire new users (players). Target games are the main carriers for advertising and also the source of user behavior data.

[0028] The platform can refer to a third-party traffic media or advertising network where advertisers place digital ads. In the embodiments of this application, the platform can refer to an online channel that supports programmatic buying and API data docking. The platform is the source of advertising data and the place where keyword bidding and ad display occur. Different platforms have different user profiles, advertising mechanisms and data standards.

[0029] Advertising delivery data refers to structured logs or reports recorded and fed back by various media platforms during the advertising delivery process, reflecting the consumption and surface effects of advertising campaigns. This structured log or report data primarily reflects the performance of the advertising front end and typically includes the delivery hierarchy (account / ad group / keyword) and basic counting metrics. In some embodiments, advertising delivery data may include impressions (keyword display events), clicks (keyword click events), spending (spending amount), downloads, and other data.

[0030] User behavior data refers to in-depth behavioral log data related to the game process generated on the game server after a user clicks on an advertisement and downloads or enters the target game. This in-depth behavioral log data can reflect the conversion quality of the advertising backend and is the fundamental basis for determining whether keywords bring real "high-value users". In some embodiments, user behavior data may include activation event information (activating the game), registration event information (registering a game account), character creation event information (creating a game character), and payment event information (in-game purchases).

[0031] Keywords refer to the bidding trigger words set by advertisers when placing ads on media platforms to match users' search intent or interest tags. They are the smallest unit of granularity for the advertising system to conduct bidding auctions. In some embodiments, keywords can be ordinary search terms (such as "Douluo Continent") or long-tail keywords with a clear game intent (such as "Hunter's World Class Recommendations" or "Hunter's World Strength Ranking").

[0032] Optionally, the methods for obtaining the target game's advertising data on different platforms may include: building corresponding API adapters for different platforms; and using the API adapters to retrieve the target game's advertising data from the API interfaces opened by the corresponding platforms.

[0033] The API adapter is used to handle data interface authentication for the corresponding platform.

[0034] An Application Programming Interface (API) adapter can refer to software middleware built specifically for each platform, encapsulating the platform's unique access logic. It can be a conversion proxy layer located between the system's core business logic and the native API interfaces of each platform. In some embodiments, the API adapter can internally encapsulate an authentication token management module, a pagination request parsing module, and a frequency control avoidance module for the specific platform's API interface. The authentication token management module can handle data interface authentication for the corresponding platform, the pagination request parsing module can be used for batch data pagination retrieval, and the frequency control avoidance module can be used for access frequency limiting.

[0035] API interfaces can refer to standardized network access endpoints provided by various platforms for programmatic reading and writing of advertising account data. API interfaces can be the only legitimate channel for third-party systems to interact with the platform's official servers. They can be "gates" defined and controlled by the platform, specifying the request format, return fields, and call frequency.

[0036] In some embodiments, the method of building corresponding API adapters for different platforms may include: for different platforms that need to be accessed, building an independent API adapter software module for each platform, each independent API adapter software module is initialized by reading the configuration information of the corresponding platform (including interface address, authentication method, token refresh cycle, frequency control threshold, etc.) when the system starts, and independently maintaining the authentication token and session state required by the platform during runtime.

[0037] In some embodiments, the method of retrieving advertising data of a target game from the API interface opened by the corresponding platform through an API adapter may include: triggering a retrieval instruction for the advertising data of the target game on the target platform according to a preset retrieval period (such as an hourly interval or a fixed time every day) through a unified task scheduler. The retrieval instruction is routed to the API adapter dedicated to the platform. After receiving the retrieval instruction, the API adapter first calls its internally encapsulated authentication token management module to obtain the currently valid platform access token, and constructs a standard HTTP request that conforms to the native API interface specification of the platform based on the platform access token and pre-configured retrieval parameters (including the list of account IDs to be retrieved, the specified data time interval, and the filtering conditions of ad groups and keyword levels). Subsequently, the API adapter initiates the standard HTTP request through the platform's API interface to retrieve the advertising data of the target game from the API interface opened by the corresponding platform through the standard HTTP request.

[0038] Optionally, the methods for obtaining user behavior data within the target game may include: establishing a secure data channel with the target game's server and using a combination of proactive push (using the log reporting SDK or data return callback interface deployed on the game server to proactively push event data in the form of HTTP POST requests in real time when users perform key behavioral events such as activation, registration, character creation, and payment) and passive retrieval (using the data export API interface opened by the game operation backend or directly connecting to the read-only slave database of the game database to batch query newly added user behavior event logs within a specified time interval according to a preset period) to obtain user behavior data.

[0039] S202. Based on advertising data, determine the interaction event information corresponding to the same keyword on different platforms, and match the target event information in user behavior data according to the user identifier associated with the interaction event information.

[0040] Interactive event information refers to the behavioral records generated when a user interacts directly with the target game corresponding to an advertisement on a media platform after seeing or clicking on an advertisement triggered by a specific keyword. These records can be recorded and transmitted back by the game server.

[0041] Optionally, the method for determining the interaction event information corresponding to the same keyword on different platforms based on advertising data may include: the interaction event information includes keyword click events and / or keyword display events; extracting several target keywords from the advertising data corresponding to different platforms whose semantic similarity is greater than or equal to a preset threshold; mapping the several target keywords to the same standard keyword based on a preset mapping relationship; extracting interaction event information from the advertising data corresponding to the several target keywords respectively, as the interaction event information corresponding to the same standard keyword on different platforms; in some embodiments, the method for determining the advertising effect information of the same keyword on different platforms may include: determining the advertising effect information of the same standard keyword on different platforms.

[0042] Keyword click events refer to a single instance of user-initiated behavior where a user actively clicks on an advertisement for a target game displayed on a media platform via mouse, touchscreen, or other interactive methods, triggering a redirect to the download page or landing page of the corresponding application for the target game. Keyword click events can be key decision points where a user transforms from an advertisement viewer into a potential game user.

[0043] Keyword display events refer to a single exposure record of the target game's advertising content being displayed on the corresponding user terminal screen of a media platform due to a successful bid for a specific keyword. Keyword display events can be used to measure the traffic coverage and brand exposure effect of keywords. They can be triggered and recorded by the platform when the advertising content is rendered and meets the effective display standards defined by the platform (such as at least 50% of the advertising material being visible on the screen for more than 1 second).

[0044] Semantic similarity can refer to the degree of closeness between two text fragments at the meaning level, rather than just a literal or character-level match.

[0045] The preset threshold can refer to a pre-defined threshold value used to filter out the target keywords mentioned above. Only keywords whose semantic similarity exceeds this threshold value will be extracted as target keywords.

[0046] Several target keywords can refer to keywords that belong to different platforms, are similar in meaning but may not be completely matched in terms of literal or character level.

[0047] The preset mapping relationship can refer to the mapping relationship between preset standard keywords and semantically similar synonyms. It can uniformly map semantically similar synonyms to the preset standard keywords to the standard keywords. Semantically similar synonyms are the target keywords whose semantic similarity is greater than or equal to the preset threshold.

[0048] Standard keywords can refer to predefined benchmark terms that serve as the unified normalization target for synonyms or near-synonyms across multiple platforms. They can be used to map multiple target keywords that, despite differences in expression habits, naming conventions, or semantic variations across different media platforms, actually point to the same game concept or user intent, to the same standardized identifier, thereby achieving unified merging and comparative analysis of cross-platform data. In some embodiments, standard keywords can be built-in standardized root word entities, possessing uniqueness and standardization.

[0049] In some embodiments, the method of extracting several target keywords from advertising data corresponding to different platforms with semantic similarity greater than or equal to a preset threshold may include: firstly, extracting all keywords currently being advertised from the advertising data of each platform, and vectorizing the root text of each keyword to obtain the root vector of each keyword; simultaneously, for each predefined and stored standard keyword, calculating the cosine similarity between the root vector of the standard keyword and the root vectors of keywords currently being advertised from different platforms; when the cosine similarity of a keyword is greater than or equal to a preset cosine similarity threshold, the keyword is determined to be a target keyword with semantic similarity greater than or equal to the preset threshold, and is then extracted and retained. In some embodiments, the method of mapping several target keywords to the same standard keyword based on a preset mapping relationship may include: the preset mapping relationship may be configured with a set of root words and / or a list of synonyms corresponding to each standard keyword; for each target keyword, first query whether the target keyword already exists in the set of root words and / or the list of synonyms configured in the preset mapping relationship, and if it exists, directly read the identifier of its corresponding standard keyword to map it to the same standard keyword.

[0050] In some embodiments, if the target keyword does not exist in the preset mapping relationship, the target keyword will be matched in the following matching rules in order of priority: exact string matching, inclusion relationship matching, and semantic similarity result matching. When any matching rule is hit, the mapping relationship between the target keyword and the corresponding standard keyword will be written into the preset mapping relationship.

[0051] In some embodiments, extracting interaction event information from the advertising data corresponding to several target keywords, and using the interaction event information corresponding to the same standard keyword on different platforms, may include: for each target keyword that serves as the mapping source corresponding to the same standard keyword, retrieving all interaction event information (including keyword display events and keyword click events) associated with the target keyword from the advertising data of the corresponding platform based on the platform identifier and keyword identifier of the target keyword; after retrieving the interaction event information of each target keyword, attaching the identifier of the standard keyword mapped to its target keyword to each interaction event information, thereby logically merging the interaction event information from each platform under the same standard keyword name.

[0052] In one embodiment, a specific implementation process for matching target event information from user behavior data based on user identifiers associated with interaction event information is described. Please refer to [reference needed]. Figure 3 , Figure 3 This is a flowchart illustrating a specific implementation process for matching target event information, as provided in an embodiment of this application. Figure 3 As shown, the specific implementation steps for matching target event information are as follows: S2021. Pre-assign a click attribution primary key containing the associated user identifier for keyword click event information, and pre-assign an activation attribution primary key containing the associated user identifier for activation event information.

[0053] The interaction event information includes keyword click event information; the user behavior data includes activation event information of the target game.

[0054] A user identifier can be a globally unique identification code that can uniquely identify a natural user (physical device or real player). User identifiers can exist across systems (such as between advertising platforms and game servers), but their forms of representation may differ.

[0055] Click attribution primary key refers to a unique event-level record identifier assigned to each keyword click event generated by the advertising platform within the system; it can be composed of platform identifier, click event ID, user identifier, and timestamp (or directly use the unique click ID returned by the platform), and is used to uniquely identify a keyword click behavior within the system.

[0056] Activation event information refers to a structured behavior log recorded and reported by the game server when a user successfully opens the game and completes the initial startup for the first time after downloading and installing the target game.

[0057] An activation attribution key is an event-level record identifier that is uniquely assigned to each activation event generated in the game server. It is used for one-to-one attribution matching with the click attribution key and can be composed of the game server identifier, activation event serial number, user identifier, and activation timestamp. It is used to uniquely identify an activation event within the system.

[0058] Optionally, the method of pre-assigning a click attribution primary key containing the associated user identifier for keyword click event information may include: when pulling advertising data from various platforms through the API adapter, for each keyword click event in the returned advertising data, the click attribution primary key allocation process is triggered in real time at the data access layer: first, the unique click ID provided by the platform and the user terminal device identifier associated with the keyword click event are extracted from the returned fields of the keyword click event; then, according to the preset primary key encoding rules, the platform identifier, the unique click ID, the user terminal device identifier, and the click timestamp are concatenated into a globally unique string in a fixed order, and a check bit is added to prevent encoding conflicts, thus forming the click attribution primary key for the keyword click event.

[0059] Optionally, the method of pre-assigning an activation attribution primary key containing the associated user identifier for activation event information may include: firstly, extracting the user terminal device identifier reported by the game server from the activation event information (which can use the same identifier type as the advertising end, such as IDFA, OAID, or the initial device fingerprint assigned by the game), and simultaneously extracting the event serial number, activation timestamp, and channel source identifier assigned by the game server for this activation event information; then, following the same encoding rules as the click attribution primary key mentioned above, concatenating the game server identifier, activation event serial number, user terminal device identifier, and activation timestamp in a fixed order into a globally unique string, and adding a check bit to prevent encoding conflicts, thus forming the activation attribution primary key for this activation event information.

[0060] S2022. Construct a pair of related indexes by using the click attribution primary key and the activation attribution primary key that contain the same user identifier, and match the activation event information associated with the interaction event information in the user behavior data based on the related indexes, and use it as the target event information.

[0061] Among them, the associated index pair refers to a one-to-one mapping relationship record established in the system memory for the click attribution primary key from the ad delivery data side and the activation attribution primary key from the user behavior data side, with the same user identifier contained in both as the matching key. It is a bidirectional queryable key-value pair structure, with one end storing the click attribution primary key and its associated click timestamp, platform identifier, keyword identifier, and other information, and the other end storing the activation attribution primary key and its associated activation timestamp, game identifier, and other information. The two ends are logically bound through the same user identifier (such as terminal device identifier).

[0062] Optionally, the method of constructing a related index pair for click attribution primary keys and activation attribution primary keys containing the same user identifier can include: constructing a bidirectional index pair by associating the click attribution primary key and the activation attribution primary key with the corresponding same user identifier as the association link, and writing the user identifier as the key and the pointer references of the click attribution primary key and the activation attribution primary key as the value to a temporary related index table.

[0063] Based on the association index pair, the activation event information associated with the interaction event information is matched in the user behavior data. The method of using the target event information as the target event information can include: extracting the pre-assigned click attribution primary key and the user identifier carried by it from the keyword click event information corresponding to the interaction event information; then using the user identifier as the hash key to query the temporary association index table in memory to locate the association index pair bound to the user identifier; reading the click attribution primary key stored in the matched association index pair and comparing it with the click attribution primary key of the current interaction event information; after confirming that the two are consistent, extracting the activation attribution primary key it points to and the corresponding storage address pointer of the activation event information in the user behavior data from the association index pair; and quickly retrieving the complete activation event information through this pointer.

[0064] S203. Based on interaction event information and target event information, determine the advertising performance information of the same keyword on different platforms, and adjust multiple keywords according to the advertising performance information.

[0065] The performance information can refer to a set of multi-dimensional, comparable numerical performance metrics calculated based on the causal relationship between the interactive event information (clicks / impressions) generated by an advertising campaign for a specific keyword on a specific platform and the target event information (activation / registration / character creation / payment) obtained through attribution matching. In some embodiments, the performance information can be a set of evaluation metrics, which may include performance indicators such as click-through rate (CTR = number of clicks / number of impressions), conversion rate (CVR = number of activations / number of clicks), and cost per install (CPI = total cost / total number of activations), used to measure the efficiency and cost of keywords in attracting user clicks and driving downloads.

[0066] Optionally, the method for determining the advertising performance of the same keyword on different platforms based on interaction event information and target event information may include: obtaining the first timestamp of the keyword click event information and determining the target backtracking period based on the first timestamp; determining each event information associated with the target event information in user behavior data according to the user identifier associated with the target event information, and obtaining the second timestamp corresponding to each event information; determining the event information whose occurrence time is within the target backtracking period based on the second timestamp; constructing the user behavior path corresponding to the same keyword within the target backtracking period according to the chronological order of the keyword click event information, target event information, and event information within the target backtracking period, and determining the advertising performance of the same keyword on different platforms based on the user behavior path.

[0067] A timestamp can refer to the point in time when a specific event occurs. The target backtracking period can refer to the time interval extending backward from the moment the keyword click event occurs; User behavior path can refer to all the key event nodes experienced by the same user from the first contact with an advertisement to the completion of a deep conversion (such as payment), which are connected in chronological order to form a causal sequence chain. For example, user behavior path can be a complete conversion narrative chain formed by splicing together a single keyword click event (starting point) on the advertisement side and various target events (intermediate nodes and end points) generated by the user on the game server side, based on the same user identifier and the increasing order of timestamps.

[0068] In some embodiments, the method of obtaining the first timestamp of keyword click event information and determining the target backtracking period based on the first timestamp may include: extracting the recorded keyword click occurrence timestamp from the keyword click event information as the first timestamp, and obtaining the corresponding backtracking window length from a preset backtracking window configuration table according to the first timestamp and the business attributes of the platform to which the keyword belongs and / or the target game; then using the first timestamp as the backtracking start time, and using the time point after adding the backtracking window length to the backtracking start time as the backtracking end time, thereby determining the continuous time interval from the backtracking start time to the backtracking end time as the target backtracking period.

[0069] In some embodiments, determining the various event information associated with the target event information in user behavior data based on the user identifier associated with the target event information and obtaining the second timestamp corresponding to each event information may include: extracting the user identifier carried by the target event information obtained by attribution matching; using the user identifier as a query condition to retrieve all behavioral event records associated with the user identifier in the user behavior data, wherein the all behavioral event records include activation events, registration events, character creation events, and payment events; setting an extended search time range based on the occurrence time of the activation event in the target event information and a preset backtracking window length, and retaining only event information whose occurrence timestamp falls within the extended search time range in the search results; subsequently reading the event occurrence timestamp field recorded in each event information retained after the search, and using the value of this field as the second timestamp corresponding to each event information.

[0070] In some embodiments, the method of determining the event information whose occurrence time falls within the target backtracking period based on the second timestamp may include: reading the start time and end time of the determined target backtracking period, traversing each event information, and extracting the second timestamp corresponding to each event information one by one; comparing the value of the second timestamp with the start time and end time of the target backtracking period respectively, and only when the second timestamp is greater than or equal to the start time of the target backtracking period and less than or equal to the end time of the target backtracking period, determining that the occurrence time of the event information falls within the target backtracking period and retaining it.

[0071] In some embodiments, constructing a user behavior path corresponding to the same keyword within a target backtracking period according to the chronological order of keyword click event information, target event information, and event information within the target backtracking period may include: using keyword click event information as a fixed starting point for the user behavior path, and then using the activation event information in the matched target event information as the first intermediate node of the user behavior path, following the click event information; then traversing the event information within the target backtracking period, and appending each event information (including but not limited to registration events, character creation events, and payment events) one by one to the activation event information in ascending order of its second timestamp, forming an ordered event chain, i.e., the user behavior path, with the keyword click event as the first node, the activation event information as the second node, the intermediate event information arranged in the order of occurrence as subsequent nodes, and the last event information within the target backtracking period as the tail node.

[0072] In some embodiments, determining the advertising performance information of the same keyword on different platforms based on user behavior paths may include: when the event information within the target retrospective period includes paid event information of the target game, determining the time difference between each event information and the paid event information in the user behavior path, and determining the advertising performance information of the same keyword on different platforms based on the time difference.

[0073] Among them, the paid event information of the target game can refer to the structured behavior log recorded and reported by the game server when a user completes a real recharge or consumption behavior in the target game. It can represent the target event information with the highest commercial value in the user behavior data in the game.

[0074] In some embodiments, when the event information within the target backtracking period includes paid event information of the target game, the method for determining the time difference between each event information and the paid event information in the user behavior path, and determining the advertising effect information of the same keyword on different platforms based on the time difference, may include: traversing each node of the included event information (including click events, activation events, registration events, character creation events, and other intermediate behavior events) in the user behavior path, except for the node corresponding to the paid event information; performing a difference calculation between the event occurrence timestamp of each event node and the payment timestamp of the last paid event node in the user behavior path to obtain the time difference value of each event node relative to the final payment time; assigning a corresponding contribution weight value to each event information node according to the size of the time difference value of each event information node, according to a preset time decay weight mapping rule (e.g., assigning a higher contribution weight to event nodes closer to the payment time and a lower contribution weight to event nodes farther from the payment time); summing up the contribution weight values ​​of all event nodes in the user behavior path as the total contribution coefficient of the user behavior path, and determining the advertising effect information of the same keyword on different platforms based on the total contribution coefficient.

[0075] In one embodiment, a specific implementation process for adjusting multiple keywords based on campaign performance information is described. Please refer to [link / reference]. Figure 4 , Figure 4 This is a flowchart illustrating a specific implementation process for keyword adjustment provided in an embodiment of this application. Figure 4 As shown, the specific implementation steps for keyword adjustment are as follows: S2031. Evaluate the performance of the same keyword on each platform based on the campaign performance information, and obtain the evaluation results for each platform.

[0076] The evaluation result can refer to the qualitative label or quantitative score given to the keyword's performance on each platform after comprehensive analysis based on the same keyword's campaign performance information on various platforms, according to preset evaluation rules and judgment criteria. In some embodiments, the evaluation result can be a scoring result.

[0077] Preset conditions can refer to preset score thresholds (when the evaluation result is a score result).

[0078] Optionally, the evaluation of the same keyword's performance on different platforms based on the campaign performance information can be achieved by: extracting the click-through rate (CTR), conversion rate, and cost per install (CPM) for each keyword on each platform; retrieving a set of evaluation rules applicable to the target game and the platform from a pre-defined evaluation rule library, which includes at least CTR thresholds, conversion rate thresholds, and CPM thresholds; comparing the CTR, conversion rate, and CPM of the same keyword on each platform with the corresponding thresholds in the evaluation rule library; and generating a comprehensive judgment value corresponding to the platform based on the comparison results of each indicator with the thresholds, which serves as the corresponding evaluation result.

[0079] S2032. Platforms whose evaluation results meet the preset conditions are identified as first-type platforms, and the output value of the same keyword on first-type platforms is increased.

[0080] The first type of platform can refer to platforms that have an advantage in keyword placement, that is, platforms with better placement results.

[0081] Value offered (VOP) refers to the maximum amount an advertiser is willing to pay for a single ad click or conversion triggered by a specific keyword on a particular media platform. It is the core price parameter in keyword bidding advertising and directly determines the keyword's competitiveness and traffic acquisition ability in the platform's ad bidding ranking.

[0082] Optionally, platforms whose evaluation results meet preset conditions can be identified as first-type platforms. Methods to increase the bid value of the same keyword on first-type platforms may include: when the evaluation result is a score, selecting platforms whose score results are greater than or equal to a preset score threshold, treating these selected platforms as a set of first-type platforms, sending a request to the bid query interface of the first-type platform through its corresponding API adapter to obtain the current bid value of the keyword on that platform, determining an upward adjustment ratio based on this, and then encapsulating the target bid value, keyword identifier, and platform identifier determined according to the upward adjustment ratio into a bid modification instruction. This instruction is then called through the API adapter of the first-type platform to access the platform's bid modification interface, writing the target bid value into the keyword's bid configuration on that platform, thereby increasing the bid value of the same keyword on the first-type platform.

[0083] S2033. Platforms whose evaluation results do not meet the preset conditions are identified as second-type platforms, and the output value of the same keyword on second-type platforms is reduced.

[0084] The second type of platform can refer to platforms where keyword placement is disadvantageous, that is, platforms with poor placement results.

[0085] Optionally, platforms whose evaluation results do not meet the preset conditions can be identified as second-type platforms, and the bidding value of the same keyword on second-type platforms can be reduced by: when the evaluation result is a score result, selecting platforms whose score results are less than a preset score threshold from various platforms, and using these selected platforms as a set of second-type platforms, sending a request to the bidding query interface of the second-type platform through the API adapter corresponding to the second-type platform to obtain the current bidding value of the keyword on the platform, and determining the reduction ratio based on this, and then encapsulating the target bidding value, keyword identifier, and platform identifier determined according to the reduction ratio into a bidding modification instruction, and calling the bidding modification interface of the platform through the API adapter corresponding to the second-type platform to write the target bidding value into the bidding configuration of the keyword on the platform, thereby reducing the bidding value of the same keyword on the second-type platform.

[0086] In one embodiment, another specific implementation process for adjusting multiple keywords based on campaign performance information is described. Please refer to [link / reference]. Figure 5 , Figure 5 This is a flowchart illustrating another specific implementation process for keyword adjustment provided in an embodiment of this application. For example... Figure 5 As shown, the specific implementation steps for keyword adjustment are as follows: S2034. Evaluate the performance of the same keyword on each platform based on the campaign performance information, and obtain the evaluation results for each platform.

[0087] S2035. Platforms whose evaluation results meet the preset conditions are identified as Type I platforms, and platforms whose evaluation results do not meet the preset conditions are identified as Type II platforms.

[0088] S2036. Query the historical search term database to identify candidate search terms with a semantic similarity to the same keyword that is greater than or equal to a preset threshold, and associate the candidate search terms with the same keyword and deploy them to the first type of platform.

[0089] Among them, the historical search term database can refer to the collection of all search term texts entered by real users for the target game during the historical launch period on various media platforms, which are collected in advance and persistently stored.

[0090] Candidate search terms refer to alternative search term texts that meet a preset similarity threshold and are not covered by the current keyword package, selected from the historical search term library by using the root of a high-value keyword as the query benchmark and through semantic similarity calculation.

[0091] Optionally, the method of querying the historical search term database to identify candidate search terms with a semantic similarity greater than or equal to a preset threshold to the same keyword, and then associating these candidate search terms with the same keyword and deploying them to the first type of platform, may include: extracting the root text of the same keyword, encoding the root text into a semantic vector to convert it into a root semantic vector; then using this root semantic vector as the query vector, performing a cosine similarity retrieval in the vector index of the historical search term database, calculating the cosine similarity value between the root semantic vector and the pre-stored semantic vector of each historical search term in the historical search term database, and filtering... All historical search terms with a similarity value greater than or equal to a preset similarity threshold are selected as candidate search terms. Using the unique identifier of the same keyword as the association primary key, a many-to-one placement association relationship is established between each candidate search term and the standard keyword, and this association relationship is written into the placement keyword package mapping table, recording the creation time, source platform identifier, and semantic similarity value of each association relationship. Subsequently, through the API adapter corresponding to the first type of platform, the keyword package addition interface of that platform is called, and the candidate search terms are added in batches to the positive placement keyword package of the platform corresponding to the ad group of the same keyword based on the placement association relationship.

[0092] S2037. Generate a blocking instruction for the same keyword and send the blocking instruction to the second type of platform; the blocking instruction is used to trigger the same keyword to stop being displayed.

[0093] Among them, the blocking command can refer to the operation command generated by a platform that is judged to be a second type platform (a disadvantaged platform), which is used to notify the platform to stop the keyword from triggering the display of advertisements on the platform.

[0094] Optionally, generating a blocking instruction for the same keyword and sending the blocking instruction to the second type of platform may include: for each platform in the second type of platform set, obtaining the ad group level identifier and ad campaign level identifier of the same keyword on that platform, and assembling a standardized blocking instruction according to a preset blocking instruction data format using the keyword text, the platform identifier, and the ad group level identifier of the platform as parameters. The standardized blocking instruction includes the operation type (add negative keyword), the target platform identifier, the target ad group level identifier, the keyword text to be blocked, and the instruction generation timestamp; submitting the standardized blocking instruction to the API adapter corresponding to the platform, and having the API adapter send the standardized blocking instruction to the aforementioned second type of platform.

[0095] In this embodiment, advertising data for the target game on different platforms and user behavior data within the target game are first obtained. The advertising data includes advertising data for multiple preset keywords. Based on the advertising data for the target game on different platforms, interactive event information for the same keyword on different platforms is determined. Target event information is then matched in the user behavior data based on the user identifier associated with the interactive event information for the same keyword on different platforms. Finally, based on the interactive event information and the matched target event information, the advertising effect information for the same keyword on different platforms is determined, and the multiple preset keywords are adjusted according to the advertising effect information.

[0096] By matching the user identifiers associated with the interaction event information, the target event information of the users corresponding to the interaction event information of the same keyword on different platforms can be identified. Based on the correlation between the interaction event information and the target event information of the corresponding users, the difference in customer acquisition effect of the same keyword on different platforms can be evaluated. This allows for a comprehensive and accurate determination of the advertising effect of the same keyword on different platforms, and further targeted adjustments can be made to the multiple keywords preset for advertising, thereby improving the efficiency of keyword optimization.

[0097] Corresponding to the above method embodiments, this application also provides an advertising keyword optimization device, such as... Figure 6 As shown, Figure 6 This is a schematic diagram of the structure of an advertising keyword optimization device according to an embodiment of this application. The device includes: The acquisition unit 601 is used to acquire advertising data of the target game on different platforms, as well as user behavior data within the target game; the advertising data includes the advertising data corresponding to multiple preset keywords; The first determining unit 602 is used to determine the interactive event information corresponding to the same keyword on different platforms based on advertising data; The matching unit 603 is used to match the target event information in the user behavior data based on the user identifier associated with the interaction event information; The second determining unit 604 is used to determine the advertising performance information of the same keyword on different platforms based on interactive event information and target event information; Optimization unit 605 is used to adjust multiple keywords based on campaign performance information.

[0098] Optionally, the interaction event information includes keyword click events and / or keyword display events; the first determining unit 602 is specifically used to extract several target keywords from the advertising delivery data corresponding to different platforms, whose semantic similarity is greater than or equal to a preset threshold; based on a preset mapping relationship, map the several target keywords to the same standard keyword; the mapping relationship is the mapping relationship between the standard keyword and semantically similar synonyms; extract the interaction event information from the delivery data corresponding to the several target keywords respectively, as the interaction event information corresponding to the same standard keyword on different platforms respectively; correspondingly, the second determining unit 604 is specifically used to determine the delivery effect information of the same standard keyword on different platforms.

[0099] Optionally, the interaction event information includes keyword click event information; the user behavior data includes activation event information of the target game; the matching unit 603 is specifically used to pre-assign a click attribution primary key containing the associated user identifier for the keyword click event information, and to pre-assign an activation attribution primary key containing the associated user identifier for the activation event information; construct an association index pair by the click attribution primary key and the activation attribution primary key containing the same user identifier, and match the activation event information associated with the interaction event information in the user behavior data based on the association index pair, as the target event information.

[0100] Optionally, the second determining unit 604 is specifically used to obtain the first timestamp of the keyword click event information and determine the target backtracking period based on the first timestamp; the target backtracking period is the duration interval extending backward from the moment the keyword click event occurs; based on the user identifier associated with the target event information, determine each event information associated with the target event information in the user behavior data, and obtain the second timestamp corresponding to each event information; based on the second timestamp, determine the event information whose occurrence time is within the target backtracking period in each event information; construct the user behavior path corresponding to the same keyword within the target backtracking period according to the chronological order of the keyword click event information, the target event information, and the event information within the target backtracking period, and determine the advertising effect information of the same keyword on different platforms based on the user behavior path.

[0101] Optionally, the second determining unit 604 is specifically used to determine the time difference between each event information and the paid event information in the user behavior path when the event information within the target backtracking period includes the paid event information of the target game, and to determine the advertising effect information of the same keyword on different platforms based on the time difference.

[0102] Optionally, the optimization unit 605 is specifically used to evaluate the performance of the same keyword on each platform based on the performance information, and obtain the evaluation results corresponding to each platform; determine the platform whose evaluation results meet the preset conditions as the first type of platform, and increase the output value of the same keyword on the first type of platform; determine the platform whose evaluation results do not meet the preset conditions as the second type of platform, and decrease the output value of the same keyword on the second type of platform.

[0103] Optionally, the optimization unit 605 is specifically used to evaluate the performance of the same keyword on each platform based on the performance information, and obtain the evaluation results for each platform; determine the platforms whose evaluation results meet the preset conditions as first-type platforms, and determine the platforms whose evaluation results do not meet the preset conditions as second-type platforms; query the historical search term library to determine candidate search terms with semantic similarity greater than or equal to the preset threshold with the same keyword, and associate the candidate search terms with the same keyword and deliver them to the first-type platforms; generate a blocking instruction for the same keyword, and send the blocking instruction to the second-type platforms; the blocking instruction is used to trigger the same keyword to stop being displayed.

[0104] like Figure 7 As shown, this application embodiment also provides an electronic device, including a processor 701, a memory 702, and a program or instructions stored in the memory 702 and executable on the processor 701. When the program or instructions are executed by the processor 701, they implement the various processes of the above-mentioned keyword optimization method embodiment for advertising and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0105] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described keyword optimization method for advertising and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0106] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0107] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0108] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0109] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

[0110] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of this application, the scope of which is determined by the scope of the claims.

Claims

1. A keyword optimization method for advertising, characterized in that, The method includes: Acquire advertising data for the target game on different platforms, as well as user behavior data within the target game; the advertising data includes advertising data corresponding to multiple preset keywords. Based on the advertising data, the interactive event information corresponding to the same keyword on different platforms is determined, and the target event information is matched in the user behavior data according to the user identifier associated with the interactive event information. Based on the interaction event information and the target event information, the advertising performance information of the same keyword on different platforms is determined, and the multiple keywords are adjusted according to the advertising performance information.

2. The method according to claim 1, characterized in that, The interactive event information includes keyword click events and / or keyword display events; The step of determining the interaction event information corresponding to the same keyword on different platforms based on the advertising data includes: Extract several target keywords from the advertising data of different platforms whose semantic similarity is greater than or equal to a preset threshold. Based on a preset mapping relationship, the target keywords are mapped to the same standard keyword; the mapping relationship is the mapping relationship between the standard keyword and semantically similar synonyms; Interaction event information is extracted from the delivery data corresponding to the several target keywords, and used as the interaction event information corresponding to the same standard keyword on different platforms. Accordingly, determining the advertising performance information for the same keyword on different platforms includes: Determine the performance of the same standard keywords on different platforms.

3. The method according to claim 1, characterized in that, The interactive event information includes keyword click event information; the user behavior data includes activation event information of the target game; The step of matching the target event information in the user behavior data based on the user identifier associated with the interaction event information includes: Pre-assign a click attribution primary key containing the associated user identifier to the keyword click event information, and pre-assign an activation attribution primary key containing the associated user identifier to the activation event information; The click attribution primary key and activation attribution primary key containing the same user identifier are constructed into an association index pair, and the activation event information associated with the interaction event information is matched in the user behavior data based on the association index pair as the target event information.

4. The method according to claim 3, characterized in that, The step of determining the advertising performance information of the same keyword on different platforms based on the interaction event information and the target event information includes: Obtain the first timestamp of the keyword click event information, and determine the target backtracking period based on the first timestamp; the target backtracking period is the duration interval extending backward from the moment the keyword click event occurs; Based on the user identifier associated with the target event information, determine each event information associated with the target event information in the user behavior data, and obtain the second timestamp corresponding to each event information respectively; Based on the second timestamp, determine the event information whose occurrence time falls within the target backtracking period from the various event information; Based on the sequence of the keyword click event information, the target event information, and the event information within the target backtracking period, a user behavior path corresponding to the same keyword within the target backtracking period is constructed, and the advertising effect information of the same keyword on different platforms is determined based on the user behavior path.

5. The method according to claim 4, characterized in that, The process of determining the advertising performance information of the same keyword on different platforms based on the user behavior path includes: If the event information within the target retrospective period includes paid event information of the target game, the time difference between each event information and the paid event information is determined in the user behavior path, and the advertising effect information of the same keyword on different platforms is determined based on the time difference.

6. The method according to claim 1, characterized in that, The adjustment of the multiple keywords based on the campaign performance information includes: The campaign performance of the same keyword on each platform is evaluated based on the campaign performance information, and the evaluation results for each platform are obtained respectively. Platforms whose evaluation results meet the preset conditions are identified as first-type platforms, and the output value of the same keyword on the first-type platforms is increased. Platforms whose evaluation results do not meet the preset conditions are identified as second-type platforms, and the output value of the same keyword on the second-type platforms is reduced.

7. The method according to claim 1, characterized in that, The adjustment of the multiple keywords based on the campaign performance information includes: The campaign performance of the same keyword on each platform is evaluated based on the campaign performance information, and the evaluation results for each platform are obtained respectively. Platforms whose evaluation results meet the preset conditions are identified as first-type platforms, and platforms whose evaluation results do not meet the preset conditions are identified as second-type platforms. Search the historical search term database to identify candidate search terms with a semantic similarity greater than or equal to a preset threshold to the same keyword, and associate the candidate search terms with the same keyword and deliver them to the first type of platform; Generate a blocking instruction for the same keyword and send the blocking instruction to the second type of platform; the blocking instruction is used to trigger the same keyword to stop being displayed.

8. A keyword optimization device for advertising, characterized in that, The device includes: The acquisition unit is used to acquire advertising data of the target game on different platforms, as well as user behavior data within the target game; the advertising data includes advertising data corresponding to multiple preset keywords. The first determining unit is used to determine the interactive event information corresponding to the same keyword on different platforms based on the advertising delivery data. The matching unit is used to match target event information in the user behavior data based on the user identifier associated with the interaction event information; The second determining unit is used to determine the advertising effect information of the same keyword on different platforms based on the interaction event information and the target event information; The optimization unit is used to adjust the multiple keywords based on the campaign performance information.

9. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein when the program or instructions are executed by the processor, they implement the keyword optimization method for the advertisement as described in any one of claims 1-7.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the keyword optimization method for an advertisement as described in any one of claims 1-7.

Images