CN117132328A

CN117132328A - Advertisement putting control method and device, equipment and medium thereof

Info

Publication number: CN117132328A
Application number: CN202311178621.4A
Authority: CN
Inventors: 刘锟
Original assignee: Guangzhou Shangyan Network Technology Co ltd
Current assignee: Guangzhou Shangyan Network Technology Co ltd
Priority date: 2023-09-12
Filing date: 2023-09-12
Publication date: 2023-11-28

Abstract

The application relates to an advertisement putting control method and a device, equipment, medium and product thereof, wherein the method comprises the following steps: acquiring advertisement putting information corresponding to commodity advertisement putting of an online store, wherein the advertisement putting information comprises commodity activity information, seed scale and a plurality of attention marketing features; performing rule matching from a marketing feature library according to a plurality of concerned marketing features, and determining a plurality of advertisement audiences with the quantity larger than the seed scale as candidate audiences; determining advertisement values of each candidate audience associated with a plurality of preset value dimensions by adopting a preset audience scoring algorithm, and screening a plurality of candidate audiences corresponding to the seed scale according to the advertisement values to serve as seed audiences; and determining target audiences by taking each seed audience as a seed crowd, and delivering commodity activity information to the target audiences. The application considers the close correspondence with the concerned marketing characteristics and the potential value generated by each seed audience to the advertisement delivery, thereby improving the advertisement delivery effect.

Description

Advertisement putting control method and device, equipment and medium thereof

Technical Field

The application relates to the technical field of electronic commerce information, in particular to an advertisement putting control method, an advertisement putting control device, advertisement putting control equipment and advertisement putting control medium.

Background

The advertisement system of the electronic commerce platform can provide commodity popularization service, and the online store of the electronic commerce platform can submit advertisement putting information to the advertisement system to put advertisements on commodities to be popularized. The advertising system may determine the corresponding target audience for the advertisement placement by a variety of techniques to maximize the assurance of the advertising effectiveness.

In the traditional technology, the mode of determining the target audience for the advertisement putting information by the advertisement system mainly comprises a label orientation technology and a seed crowd technology, wherein the label orientation technology is mainly determined by searching a keyword given by a merchant user as a label, the operation is simple, convenient and flexible, but the positioning accuracy of the target audience is lower, and the obtained advertisement putting return rate is often lower because the probability of the keyword being the same as a bid is higher, and the competition is fierce. The seed crowd technology selects a part of the target audience as the seed crowd through portrait determination on the target audience according to the attribute, behavior, interests and other information of the user and through specific screening rules and conditions, and expands the target audience on the basis of the seed crowd, and the seed crowd technology has the characteristics of simplicity, convenience and flexibility, but has some defects, and is mainly expressed as follows:

Firstly, the related technology is mainly based on feature matching implementation, various deep learning models for implementing matching often have the condition of over fitting or under fitting on certain features, so that the matching degree of a circled seed crowd and a target audience of the current advertisement delivery is difficult to grasp, the advertisement delivery return rate is unstable, and users led by the advertisement delivery do not necessarily meet the expectations of advertisement delivery operators.

Secondly, the related technology is mainly used for matching seed crowds through user portrait characteristics, but the estimation of the actual value of the seed users is lacking, and even if the influence of the potential value of the users on the advertisement putting effect is considered, the influence is quite general, so that good advertisement benefits cannot be generated.

Therefore, in the seed crowd technology, the seed crowd directly influences the advertisement putting effect, but the traditional seed crowd putting technology is poor in advertisement putting effect due to excessive dependence on model capacity and insufficient prediction capacity on actual value of users, and has larger benefit improving space, so that further exploration on the seed crowd technology is necessary.

Disclosure of Invention

The application aims to provide an advertisement putting control method, a corresponding device and equipment thereof and a nonvolatile readable storage medium.

According to one aspect of the present application, there is provided an advertisement delivery control method including the steps of:

acquiring advertisement putting information corresponding to commodity advertisement putting of an online store, wherein the advertisement putting information comprises commodity activity information, seed scale and a plurality of attention marketing features;

performing rule matching from a marketing feature library according to the plurality of concerned marketing features, and determining a plurality of advertisement audiences with the quantity larger than the seed scale as candidate audiences;

determining advertisement values of each candidate audience associated with a plurality of preset value dimensions by adopting a preset audience scoring algorithm, and screening a plurality of candidate audiences corresponding to the seed scale according to the advertisement values to serve as seed audiences;

and determining a target audience by using a seed crowd formed by the seed audiences, and throwing the commodity activity information to the target audience.

According to another aspect of the present application, there is provided an advertisement putting control device including:

the information acquisition module is used for acquiring advertisement putting information corresponding to commodity advertisement putting of an online store, wherein the advertisement putting information comprises commodity activity information, seed scale and a plurality of attention marketing features;

The audience candidate module is used for carrying out rule matching from a marketing feature library according to the plurality of concerned marketing features, and determining a plurality of advertisement audiences with the quantity larger than the seed scale as candidate audiences;

the audience selection module is used for determining advertisement values of each candidate audience, which are associated with a plurality of preset value dimensions, by adopting a preset audience scoring algorithm, and screening a plurality of candidate audiences corresponding to the seed scale according to the advertisement values to serve as seed audiences;

and the advertisement pushing module is used for determining target audiences by using seed groups formed by the seed audiences and throwing the commodity activity information to the target audiences.

According to another aspect of the present application, there is provided an advertisement delivery control device comprising a central processor and a memory, the central processor being adapted to invoke the execution of a computer program stored in the memory to perform the steps of the advertisement delivery control method of the present application.

According to another aspect of the present application, there is provided a non-volatile readable storage medium storing a computer program implemented according to the advertisement delivery control method in the form of computer readable instructions, the computer program executing the steps included in the method when being invoked by a computer to run.

Compared with the prior art, after the advertisement delivery information provided by the online store is obtained, the specified concerned marketing features are utilized to carry out rule matching in the preset marketing feature library, a plurality of candidate audiences larger than the preset seed scale in the advertisement delivery information are screened out, the rough screening of the advertisement audiences is realized, and the advertisement audiences are ensured to be checked by utilizing the marketing features; and further combining an audience scoring algorithm, determining the advertisement value of each candidate audience by utilizing a plurality of preset value dimensions, screening a plurality of seed audiences corresponding to the designated seed scale according to the advertisement value, realizing the fine screening of the advertisement audiences, and ensuring that the advertisement value determined according to the multivalent value dimensions is utilized to realize the accuracy of the advertisement audiences. The method has the advantages that the seed crowd corresponding to the current advertisement delivery is obtained rapidly and efficiently, the close correspondence with the concerned marketing features is fully considered, the potential value of each seed audience for advertisement delivery is fully considered, the expected seed scale is matched, the target audience is expanded by the seed crowd on the basis, the method has higher reliability, when the advertisement is pushed to the target audience determined according to the seed crowd, the good advertisement investment return rate can be obtained, the advertisement delivery technology is comprehensively and systematically optimized, and the quality improvement and synergy of an advertisement system are realized.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of a network architecture of an exemplary e-commerce platform of the present application;

FIG. 2 is a flow chart of an embodiment of an advertisement delivery control method of the present application;

FIG. 3 is a schematic flow chart of a method for roughly screening candidate audience members from advertisement audience members according to an embodiment of the present application;

FIG. 4 is a flow chart of a method for constructing a marketing feature library according to an embodiment of the present application;

FIG. 5 is a schematic flow chart of a method for finely screening a seed audience from a candidate audience according to an embodiment of the present application;

FIG. 6 is a flow chart of a machine learning model for preparing to determine advertisement value according to an embodiment of the present application;

FIG. 7 is a schematic diagram of a process for advertising according to a seed audience in an embodiment of the present application;

FIG. 8 is a flow chart illustrating an extreme user cleaning of a marketing feature library in accordance with an embodiment of the present application;

FIG. 9 is a schematic block diagram of an advertisement delivery control device of the present application;

fig. 10 is a schematic structural diagram of an advertisement delivery control device according to the present application.

Detailed Description

In the network architecture shown in fig. 1, the e-commerce platform 82 is deployed in the internet to provide corresponding services to its users, and the merchant user's device 80 and the consumer user's device 81 of the e-commerce platform 82 are similarly connected to the internet to use the services provided by the e-commerce platform. For example, the e-commerce platform can open advertisement delivery service for merchant users of online stores in the e-commerce platform by configuring an advertisement system, and under the condition that the merchant users submit advertisement campaign information, target audiences corresponding to advertisement delivery are determined for the merchant users, and commodity campaign information in the advertisement campaign information is delivered to the corresponding target audiences.

The exemplary e-commerce platform 82 provides matching of supply and demand for products and/or services to the public by means of an internet infrastructure, in the e-commerce platform 82, the products and/or services are provided as merchandise information, and for simplicity of description, the concept of merchandise, products, etc. is used in the present application to refer to the products and/or services in the e-commerce platform 82, and specifically may be physical products, digital products, tickets, service subscriptions, other off-line fulfillment services, etc.

In reality, each entity of the parties can access the identity of the user to the e-commerce platform 82, and the purpose of participating in the business activity realized by the e-commerce platform 82 is realized by using various online services provided by the e-commerce platform 82. These entities may be natural persons, legal persons, social organizations, etc. The e-commerce platform 82 corresponds to both merchant and consumer entities in commerce, and there are two broad categories of merchant users and consumer users, respectively. The online service can be used in the e-commerce platform 82 by the identity of the merchant user, while the online service can be used in the e-commerce platform 82 by the identity of the consumer, including the real or potential consumer, of the merchant user. In actual business activities, the same entity can perform activities on the identity of a merchant user and the identity of a consumer user, so that the user can flexibly understand the activities.

The infrastructure for deploying the e-commerce platform 82 mainly comprises a background architecture and front-end equipment, wherein the background architecture runs various online services through a service cluster, and the service functions of the background architecture are enriched and perfected by middleware or front-end services facing a platform side, services facing a consumer, services facing a merchant and the like; the head-end equipment primarily encompasses the terminal equipment that the user uses to access the e-commerce platform 82 as a client, including but not limited to various mobile terminals, personal computers, point-of-sale devices, and the like. For example, a merchant user may enter merchandise information for his online store through his terminal device 80, or generate his merchandise information using an open interface of the e-commerce platform; the consumer user can access the web page of the online store realized by the electronic commerce platform 82 through the terminal device 81 thereof, and trigger the shopping flow through the shopping keys provided on the web page, and call various online services provided by the electronic commerce platform 82 in the shopping flow, thereby realizing the purpose of purchasing orders.

In some embodiments, the e-commerce platform 82 may be implemented by a processing facility including a processor and memory that stores a set of instructions that, when executed, cause the e-commerce platform 82 to perform e-commerce and support functions in accordance with the present application. The processing facility may be part of a server, client, network infrastructure, mobile computing platform, cloud computing platform, fixed computing platform, or other computing platform, and provide electronic components of the merchant platform 82, merchant devices, payment gateways, application developers, marketing channels, transport providers, client devices, point-of-sale devices, and the like.

The e-commerce platform 82 may be implemented as online services such as cloud computing services, software as a service (SaaS), infrastructure as a service (I aaS), platform as a service (PaaS), desktop as a service (DaaS), hosted software as a service, mobile back end as a service (MBaaS), information technology management as a service (I TMaaS), and the like. In some embodiments, the various features of the e-commerce platform 82 may be implemented to be adapted to operate on a variety of platforms and operating systems, e.g., for an online store, the administrator user may enjoy the same or similar functionality, whether in the various embodiments iOS, android, homonyOS, web page, etc.

The e-commerce platform 82 may implement its respective independent station for each merchant to run its respective online store, providing the merchant with a respective instance of the commerce management engine for the merchant to establish, maintain, and run one or more of its online stores in one or more independent stations. The business management engine instance can be used for content management, task automation and data management of one or more online stores, and various specific business processes of the online stores can be configured through interfaces or built-in components and the like to support the realization of business activities. The independent station is an infrastructure of the e-commerce platform 82 with cross-border service functionality, and merchants can maintain their online stores more centrally and autonomously based on the independent station. The stand-alone stations typically have merchant-specific domain names and memory space, with relative independence between the different stand-alone stations, and the e-commerce platform 82 may provide standardized or personalized technical support for a vast array of stand-alone stations, so that merchant users may customize their own adaptive commerce management engine instances and use such commerce management engine instances to maintain one or more online stores owned by them.

The online store may implement background configuration and maintenance by the merchant user logging in his business management engine instance with an administrator identity, which, in support of various online services provided by the infrastructure of the e-commerce platform 82, may configure various functions in his online store, review various data, etc., e.g., the merchant user may manage various aspects of his online store, such as viewing recent activities of the online store, updating online store inventory, managing orders, recent access activities, total order activities, etc.; the merchant user may also view more detailed information about the business and visitors to the merchant's online store by retrieving reports or metrics, such as sales summaries showing the merchant's overall business, specific sales and participation data for the active sales marketing channel, etc.

The e-commerce platform 82 may provide a communications facility and associated merchant interface for providing electronic communications and marketing, such as utilizing an electronic message aggregation facility to collect and analyze communications interactions between merchants, consumers, merchant devices, customer devices, point-of-sale devices, etc., to aggregate and analyze communications, such as for increasing the potential to provide product sales, etc. For example, a consumer may have problems with the product, which may create a dialogue between the consumer and the merchant (or an automated processor-based proxy on behalf of the merchant), where the communication facility is responsible for interacting and providing the merchant with an analysis of how to increase sales probabilities.

In some embodiments, an application program suitable for being installed to a terminal device may be provided to serve access requirements of different users, so that various users can access the e-commerce platform 82 in the terminal device through running the application program, for example, a merchant background module of an online store in the e-commerce platform 82, and in the process of implementing the business activity through the functions, the e-commerce platform 82 may implement various functions related to supporting implementation of the business activity as middleware or online service and open corresponding interfaces, and then implant a tool kit corresponding to the interface access function into the application program to implement function expansion and task implementation. The commerce management engine may include a series of basic functions and expose those functions through APIs to online service and/or application calls that use the corresponding functions by remotely calling the corresponding APIs.

Under the support of the various components of the commerce management engine instance, the e-commerce platform 82 may provide online shopping functionality, enabling merchants to establish contact with customers in a flexible and transparent manner, consumer users may purchase items online, create merchandise orders, provide delivery addresses for the items in the merchandise orders, and complete payment confirmation of the merchandise orders. The merchant may then review and fulfill or cancel the order. The audit component carried by the business management engine instance may enable compliance use of the business process to ensure that the order is suitable for fulfillment prior to actual fulfillment. Orders can sometimes be fraudulent, requiring verification (e.g., identification card checking), a payment method that requires the merchant to wait to ensure funds are received can act to prevent such risk, and so on. The order risk may be generated by fraud detection tools submitted by third parties through an order risk API or the like. Before fulfillment, the merchant may need to acquire payment information or wait to receive payment information in order to mark the order as paid before the merchant can prepare to deliver the product. Such as this, a corresponding examination can be made. The audit flow may be implemented by a fulfillment component. Merchants can review, adjust the job, and trigger related fulfillment services by way of fulfillment components, such as: through manual fulfillment services, use when a merchant picks and packages a product in a box, purchases a shipping label and enters its tracking number, or simply marks an item as fulfilled; a custom fulfillment service that may define sending emails for notification; an API fulfillment service that may trigger a third party application to create a fulfillment record at a third party; a legacy fulfillment service that may trigger custom API calls from a business management engine to a third party; the gift card fulfills the service. Generating a number and activating the gift card may be provided. Merchants may print shipping slips using an order printer application. The fulfillment process may be performed when the items are packaged in boxes and ready for shipment, tracking, delivery, verification by the consumer, etc.

It can be seen that the service provided by the e-commerce platform is based on the fact that products are expanded as cores, corresponding commodity data are basic data of the e-commerce platform, commodity information is provided through the commodity data, mining and utilization of the commodity data are bases for realizing various technical services, and the service for providing the basis for operation of the advertisement system by utilizing user transaction data in the commodity data of the e-commerce platform is included. Therefore, the advertisement system can be operated in any one or more servers of the cluster of the e-commerce platform, so that various functions can be realized by utilizing various commodity data provided by the e-commerce platform.

Referring to fig. 2, the method for controlling advertisement delivery according to the present application includes the following steps:

step S5100, advertisement delivery information corresponding to commodity advertisements delivered by an online store is obtained, wherein the advertisement delivery information comprises commodity activity information, seed scale and a plurality of attention marketing features;

based on the e-commerce platform of the independent station, each online store operates in the corresponding independent station, and the online store always needs to promote the in-store commodity by putting advertisements so as to achieve the effect of attracting the user flow, and the transaction total of the online store is promoted by the advertisements. The online stores are typically managed by respective merchant users who may post merchandise advertisements corresponding to the merchandise within their online stores using an advertising system. Correspondingly, the e-commerce platform can open a corresponding advertisement putting page for a merchant user through the advertisement system thereof, so that the merchant user submits advertisement putting information therein, and the advertisement putting process is started.

Advertisement placement information submitted by merchant users to effect advertisement placement typically includes merchandise campaign information, seed size, and attention marketing features.

The commodity activity information mainly plays a role of designating one or more commodities in an online store, can be provided by designating commodity identifications corresponding to the commodities, and the advertisement system automatically invokes corresponding commodity data by utilizing the commodity identifications to construct corresponding formatted data to be pushed to advertisement audiences. The commodity activity information can also comprise advertisement text, and related promotion information, advertisement slogans and the like are displayed through the advertisement text, so that the effect of attracting the attention of users more easily is achieved.

The seed size may be expressed as a number or number area that is used to constrain the size of the seed population that the advertising system determines for the current advertisement placement. The seed scale can be preset by a merchant user or can be set by default by an advertisement system, and the reasonably set seed scale is favorable for controlling the accuracy of the seed population determined by the advertisement system so as to prevent the determined seed population from being excessively generalized to cover undesired seed audiences or excessively narrow to miss suitable seed audiences.

The attention marketing features may be used to determine audience marketing features to which the merchant user is interested in advertising. The audience marketing feature is a marketing feature abstracted based on the consumer transaction data of the e-commerce platform consumer user, i.e. the advertising audience. The marketing feature is determined from a marketing perspective by abstracting user transaction data generated in each online store of the e-commerce platform based on transaction performance properties. The marketing feature is different from the traditional user portrait feature in that the traditional user portrait feature is abstractly determined around the operation habit of the user behavior, the obtained feature generally represents the operation habit feature of the user on the transaction behavior, the marketing feature concentrates on the transaction effect of the user on the transaction behavior, the corresponding feature is extracted from the transaction data of the user according to the transaction effect, and the historical value of the corresponding user corresponding to advertisement delivery is comprehensively reflected through the features. For example, the same order amount, from the perspective of the user's image feature, is characterized by abstracting the fact that the user confirms the behavior of purchasing a commodity, while from the perspective of the marketing feature, is characterized by abstracting a portion of the user's shopping potential, and thus determining the corresponding marketing feature from the perspective of the transaction total that is aggregated over a period of time for all the order amounts. Therefore, the marketing features and the traditional portrait features of the application are basically homologous with each other according to the data relied by the feature engineering, but the data properties according to which the feature abstraction is carried out are different, and the specific data are necessarily different, but the marketing features and the traditional portrait features are obtained according to different ideas, so the marketing features and the traditional portrait features are not directly equivalent.

When a merchant user needs to put advertisements on certain commodity of an online store, different requirements can be met, and advertisement putting information can be flexibly processed, for example:

in one embodiment, after a merchant user designates a commodity identifier, the advertisement system automatically calls commodity title and/or commodity attribute data corresponding to the commodity identifier from a commodity database of an online store of the merchant user according to the commodity identifier, generates an advertisement file corresponding to the commodity title and/or commodity attribute data by means of an advertisement file generation model trained in advance until convergence, and constructs the advertisement file and the commodity identifier together as commodity activity information. Similarly, the commodity activity information may also include other information provided by the merchant user, such as remark information, activity description information, and the like.

In some embodiments, the attention marketing feature may be represented as option text with clear semantic expression effect, and after the option text is selected by the merchant user, the advertisement system may convert the option text selected by each user into a plurality of attention marketing features based on a preset mapping rule.

In one embodiment, a merchant user may set the seed size corresponding to the current advertisement placement by himself, given a corresponding number representing such size, so that the advertising system determines the population of seeds based on the seed size. In an alternate embodiment, the merchant user may not need to specify the seed size, but instead the advertisement system may call the value set by default as the seed size.

Step S5200, performing rule matching from a marketing feature library according to the plurality of concerned marketing features, and determining a plurality of advertisement audiences with the quantity larger than the seed scale as candidate audiences;

after obtaining the plurality of attention marketing features submitted by the merchant user, each attention marketing feature can be utilized to carry out rule matching in a preset marketing feature library so as to determine a plurality of candidate audiences. And the candidate audience is determined based on rule matching, so that the operation amount is low, and the method is more efficient and accurate.

The marketing feature library stores feature scores corresponding to a large number of advertisement audiences in different audience marketing features, and can be prepared in advance. The marketing features of the audiences in the marketing feature library can be abstracted and determined by utilizing user transaction data corresponding to transaction events triggered by the consumer users in online stores in the electronic commerce platform, the marketing features of the audiences are abstracted first, and then the data of the marketing features of the consumer users in the audiences are quantified by utilizing a preset rule, so that feature scores corresponding to the data are determined, and the association relationship data are stored in the marketing feature library.

The advertisement audience in the marketing feature library belongs to consumer users of the electronic commerce platform, and the user transaction data of the consumer users can be maintained in a centralized way by the electronic commerce platform or distributed in each online store for maintenance.

According to the analysis, the feature scores of the advertisement audiences corresponding to the audience marketing features are shown in the marketing feature library, and accordingly, the feature scores of the audience marketing features, which are matched with the advertisement audiences and the attention marketing features, can be identified according to the attention marketing features specified in the advertisement putting information, feature total scores of the advertisement audiences corresponding to all the attention marketing features are summarized and determined, and then the advertisement audiences are further screened by utilizing the feature total scores so as to determine candidate audiences. When the matching relation between the concerned marketing feature and the audience marketing feature is determined, the matching relation can be fuzzy matching or precise matching, and can be flexibly set.

When the advertisement audience is screened according to the feature total score, the advertisement audience can be determined according to the seed scale. In order to achieve the effect of data searching, the target quantity is usually determined according to a plurality of times of a preset seed scale, and candidate audiences corresponding to the target quantity are selected. Specifically, after each advertisement audience in the marketing feature library is subjected to reverse sequencing according to the feature total score, each advertisement audience which is ranked ahead is intercepted according to the target quantity and is used as a candidate audience, so that the target of primarily screening the candidate audience according to a plurality of concerned marketing features in the marketing feature library is completed.

In some embodiments, when determining the feature total score of each advertisement audience, the feature score corresponding to each marketing feature of different audiences may be weighted to obtain a corresponding feature total score, so as to provide technical support for adjusting the importance of each marketing feature, and by adjusting the weights of different marketing features, the flexibility of determining the seed population may be obtained, so as to determine the seed population with moderate inclination according to different influences corresponding to different marketing features, thereby improving the reliability of the determined seed population.

Step S5300, determining advertisement values of each candidate audience associated with a plurality of preset value dimensions by adopting a preset audience scoring algorithm, and screening a plurality of candidate audiences corresponding to the seed scale according to the advertisement values to serve as seed audiences;

based on the seed scale and the attention marketing characteristic in the advertisement putting information, a plurality of candidate audiences are determined to form a candidate audience set, and the number of the candidate audiences is huge, and the candidate audience set is comprehensive but has insufficient concentration degree, so that the candidate audience set can be finely screened by means of a preset audience scoring algorithm so as to obtain a more concentrated seed audience set.

The audience scoring algorithm can be set in advance, and is combined with the idea of utilizing marketing features to examine the advertising effect corresponding to the advertising audience, when the audience scoring algorithm is designed, the audience scoring algorithm is set to determine the advertising value corresponding to each candidate audience through association calculation by two or more preset value dimensions, so that the advertising value of each advertising audience is determined based on two or more value dimensions, and the advertising value is effectively measured by referring to each other by a plurality of value dimensions.

The value dimensions preset by the audience scoring algorithm are generally designed according to the idea of targeting advertising effectiveness, and the indexes of more than two of the activity degree, loyalty degree and importance degree of the advertising audience users relative to the e-commerce platform are respectively represented by a plurality of preset value dimensions in one embodiment. Accordingly, in one embodiment, the plurality of preset value dimensions may include an activity value dimension, a loyalty value dimension, an importance value dimension, where data corresponding to each value dimension corresponds to a number of days of a stop purchase interval corresponding to a time of a last time of a transaction event triggered by the advertisement audience, a number of repeated transactions within a preset period adjacent to the day, and a transaction total corresponding to the preset period, respectively. The number of days between the purchase stopping intervals can measure the active state of the corresponding advertisement audience in the electronic commerce platform, the number of repeated transactions can measure the loyalty degree of the corresponding advertisement audience in the electronic commerce platform, and the total transaction amount can measure the contribution value of the corresponding advertisement audience in the electronic commerce platform so as to reflect the importance degree of the electronic commerce platform.

The audience scoring algorithm determines the advertisement value of the advertisement audience by associating two or more value dimensions, so that the actual economic value of the advertisement audience in the E-commerce platform is estimated, and the determined advertisement value can more accurately measure the expected advertisement effect of each advertisement audience on advertisement delivery. Accordingly, after the advertisement value corresponding to each candidate audience in the candidate audience set is determined according to the audience scoring algorithm, a part of candidate audience with relatively high advertisement value can be screened out from the candidate audience set according to the advertisement value, the total quantity of the part of candidate audience is matched with the quantity of the preset seed scale, the part of candidate audience is used as the seed audience to be constructed as the seed audience set, and the seed population matched with the advertisement putting information can be represented. The specific screening algorithm may be a preferred manner of sorting, or may be a preferred manner of sorting according to a threshold, and in any manner, the screening algorithm is kept matched with a preset seed size. The degree of matching can be flexibly set, for example, 1 to 2 times the seed size, or the number of the matching can be completely equal to the number of the seed sizes, and the like.

According to the design thought of the audience scoring algorithm, the audience scoring algorithm does not determine the seed audience in a rule matching mode, but comprehensively measures a plurality of value dimensions of each advertisement audience corresponding to advertisement effects to obtain advertisement values, and then determines according to the advertisement values, wherein each value dimension effectively characterizes potential values of the advertisement audience from different aspects of the advertisement effects, and the correlation calculation is realized among each value dimension to determine the advertisement values, so that organic fusion among the plurality of value dimensions is realized, the effect of the advertisement values on measuring potential economic values of the advertisement audience to an electronic commerce platform is enhanced, and the fine screening of the advertisement audience can be realized more effectively.

In some embodiments, when determining the advertisement value of the candidate audience according to the multiple value dimension association, corresponding weights can be respectively matched for each value dimension, so that the influence of different value dimensions on determining the advertisement value can be adjusted through the weights, thereby flexibly adjusting the value demands of different properties, opening a richer adjusting mode for screening seed populations for the advertisement system, and strengthening the service function of the advertisement system.

In some embodiments, the audience scoring algorithm may be mathematically modeled in advance, implemented as a corresponding mathematical model, implemented as a corresponding machine learning model, to promote its ability to normalize to service a large number of requests.

And S5400, determining a target audience by using the seed population formed by the seed audiences, and delivering the commodity activity information to the target audience.

After the seed audience set is determined, the seed audience set determines the corresponding seed population, so that the corresponding target audience can be determined according to the seed population, and commodity activity information in the advertisement delivery information is delivered to each target audience.

In one embodiment, each seed audience in the set of seed audience is directly taken as a target audience, and commodity activity information is delivered to the target audience, so that the advertising effect is excellent because each seed audience is highly selected.

In another embodiment, in order to amplify the drainage effect of advertisement popularization, the similarity matching between the user image features can be performed among the massive users of the e-commerce platform according to the seed audience concentrated by the seed audience, so that more consumer users are expanded on the basis of the seed audience, the seed audience and all the expanded consumer users are used as target audiences, commodity activity information is put into the target audiences, and a larger drainage effect can be expected.

In some embodiments, before the commodity activity information is pushed to the target audience, the advertisement system may obtain the corresponding commodity information according to the commodity identifier provided by the commodity activity information, format the commodity information, and push the formatted commodity activity information to the target audience, so that the target audience can obtain more standardized advertisement information, promote the purchase desire of the target audience, and promote the transaction amount of the platform.

In some embodiments, the seed audience set and the commodity activity information may be submitted to a third party platform, such as an advertisement interface provided by an instant messaging system, a mail system, etc., and the advertisement interface may determine the target audience from the seed audience set and push advertisement information corresponding to the commodity activity information to the target audience.

According to the embodiment, after the advertisement delivery information provided by the online store is obtained, the specified concerned marketing features are utilized to carry out rule matching in the preset marketing feature library, a plurality of candidate audiences larger than the preset seed scale in the advertisement delivery information are screened out, the advertisement audiences are coarsely screened, and the advertisement audiences are ensured to be searched by utilizing the marketing features; and further combining an audience scoring algorithm, determining the advertisement value of each candidate audience by utilizing a plurality of preset value dimensions, screening a plurality of seed audiences corresponding to the designated seed scale according to the advertisement value, realizing the fine screening of the advertisement audiences, and ensuring that the advertisement value determined according to the multivalent value dimensions is utilized to realize the accuracy of the advertisement audiences. The method has the advantages that the seed crowd corresponding to the current advertisement delivery is obtained rapidly and efficiently, the close correspondence with the concerned marketing features is fully considered, the potential value of each seed audience for advertisement delivery is fully considered, the expected seed scale is matched, the target audience is expanded by the seed crowd on the basis, the method has higher reliability, when the advertisement is pushed to the target audience determined according to the seed crowd, the good advertisement investment return rate can be obtained, the advertisement delivery technology is comprehensively and systematically optimized, and the quality improvement and synergy of an advertisement system are realized.

On the basis of any embodiment of the method of the present application, referring to fig. 3, rule matching is performed from a marketing feature library according to the plurality of attention marketing features, and a plurality of advertisement audiences with the number greater than the seed size are determined as candidate audiences, including:

step S5210, obtaining a preset marketing feature library, wherein the marketing feature library comprises mapping relation data between a plurality of audience marketing features of advertisement audiences and corresponding feature scores thereof;

as described above, the marketing feature library may be pre-built by feature engineering, stored in a memory device of the e-commerce platform, and invoked accordingly when needed.

In one embodiment, the marketing feature library stores various data in the form of a two-dimensional table, specifically, a corresponding data record is set for each advertisement audience, and a plurality of preset audience marketing features are set as different fields in each data record, and feature scores of the corresponding audience marketing features are correspondingly stored in the respective fields, so that the marketing feature library contains mapping relation data between the plurality of audience marketing features of the advertisement audience and the corresponding feature scores thereof.

Step S5220, summarizing feature scores of a plurality of audience marketing features, which are matched with the plurality of attention marketing features, of each advertisement audience into feature total scores corresponding to the advertisement audience;

Determining a feature score for each advertising audience may be based on a library of marketing features, and in particular, for each given marketing feature of interest, determining a feature score for each advertising audience for an audience marketing feature that is the same as the marketing feature of interest, and summing the feature scores for each advertising audience for a plurality of marketing features of interest to obtain the feature score for each advertising audience.

In some embodiments, as described above, when determining the feature total score for each advertisement audience, the feature scores corresponding to the respective interest marketing features may also be weighted and summed to obtain the feature total score according to the preset weights corresponding to the respective audience marketing features.

A two-dimensional relationship table is obtained by calculating feature total scores corresponding to advertisement campaign information given by the advertisement delivery for each advertisement audience in a marketing feature library, wherein mapping relationship data between each advertisement audience and the feature total scores obtained under a given plurality of concerned marketing features is stored.

And step S5230, screening out partial advertisement audiences with relatively large feature total scores as candidate audiences according to the preset multiple of the seed scale.

In order to improve the recall rate, a preset multiple, for example, any value from 3 to 10 times is preset, the preset multiple is multiplied by the seed scale given by the advertisement campaign information to obtain a target quantity, on the basis, the two-dimensional relation table is firstly used for carrying out reverse ordering on each advertisement audience according to the feature total score, then a plurality of advertisement audiences which are ranked in front and correspond to the target quantity are intercepted to serve as candidate audiences, and a candidate audience set is formed, so that the corresponding candidate audience group is matched for the advertisement campaign information.

According to the embodiment, the feature total score of each advertisement audience is determined through the combination of the plurality of concerned marketing features, then the seed scale is expanded according to the preset multiple, the candidate audience set is screened out according to the feature total score, a large number of candidate audiences with a certain marketing feature association degree can be obtained, and reasonable searching of the advertisement audiences according to the plurality of concerned marketing features is achieved.

On the basis of any embodiment of the method of the present application, referring to fig. 4, before acquiring advertisement delivery information corresponding to a commodity advertisement delivered by an online store, the method includes:

step S4100, obtaining user transaction data generated by each transaction user of the e-commerce platform in a certain period, wherein the user transaction data describes order data corresponding to transaction events triggered by each online store of the corresponding user in the e-commerce platform;

When a consumer user in the e-commerce platform accesses any one online store of the e-commerce platform, the consumer user becomes a transaction user when the consumer user accesses any one online store of the e-commerce platform to conduct transaction actions, particularly actions of ordering and paying, and corresponding transaction events are triggered and corresponding user transaction data are generated in an independent station of the online store or other servers of the e-commerce platform. Whether the user transaction data is stored in a corresponding independent station of each online store or is stored in a specific server of the e-commerce platform in a centralized manner, the user transaction data can be called by the e-commerce platform.

Each user transaction data describes, in a discrete or centralized manner, order data for the consumer user during the current transaction event, including, but not limited to, shopping time for the purchased merchandise, merchandise identification, shopping amount, shopping quantity, user enjoyment of discount information, user source channel, user network address, etc., to name a few.

When the same consumer user makes multiple shopping in a certain period, user transaction data corresponding to the multiple shopping can be naturally generated, and the re-shopping condition of the consumer user can be inspected by inspecting the number of transaction events triggered by the user in a certain period; the activity condition of the transaction user in the E-commerce platform can be reflected by examining the time of the last shopping distance advertisement putting day of the transaction user in the period; the purchase force of the transaction user on the e-commerce platform can be reflected by examining the total amount of the orders of the transaction user in a certain period, and the importance of the transaction user on the e-commerce platform is also reflected.

The period for obtaining the user transaction data corresponding to the setting can be set as a history period nearest to the advertisement putting day, for example, the last 7 days, 15 days, 30 days and the like before the day, and can be flexibly set according to actual demands.

Step S4200, extracting features of user transaction data of each transaction user according to a plurality of preset audience marketing features to obtain corresponding feature data, taking the transaction user as an advertisement audience, and storing the data of each audience marketing feature into a feature database;

feature extraction may be performed on the user transaction data for each transaction user based on obtaining the user transaction data for a period of time to extract data corresponding to the plurality of audience marketing features for each transaction user.

In an exemplary embodiment, the audience marketing features may employ any of the following as desired: the method comprises the steps of user source channel characteristics, user re-purchase behavior characteristics, user shopping total amount characteristics, user shopping times characteristics, advertisement conversion rate characteristics corresponding to visiting areas of users, user commodity number transaction characteristics and user discount enjoyment characteristics. The properties and the functions of the marketing features of the audiences are analyzed one by one:

The user source channel characteristics represent the source channels of the corresponding consumer users, namely, the online stores of the E-commerce platform are accessed by the consumer users through which advertisement promotion channels, and the domain name analysis can be carried out from the webpage addresses when the consumer users visit. Consumer users in different areas often have different consumer strengths with marketing traits.

The characteristic of the user's repurchase behavior indicates whether the consumer user has the fact behavior of shopping for many times within a set period of time, that is, whether the corresponding order number is more than or equal to 2. Consumer users with the fact of repurchase typically have greater loyalty to the e-commerce platform, and thus this feature has marketing properties.

The user shopping total characteristic represents the shopping total corresponding to the transaction event triggered by the consumer user in the electronic commerce platform in a set certain period, and the shopping amount of the order generated by the consumer user in the time can be statistically determined. Shopping totals can measure the purchasing power of consumer users and thus have marketing traits.

The user shopping times feature indicates the total number of orders corresponding to the transaction event triggered by the consumer user on the E-commerce platform in a set period, and the total number of orders can be determined by counting the number of orders. Shopping times can measure the purchasing power of consumer users and thus have marketing traits.

The advertisement conversion rate characteristic represents any one or any plurality of similar indexes such as advertisement conversion rate, click and the like corresponding to the area where the consumer user is visiting in the certain period, the area where the consumer user is visiting can be determined through the IP address of the terminal equipment, and then the index corresponding to the area which is analyzed in advance is obtained from the advertisement system to determine the area. The advertising conversion rate of the region implies the effect of advertising in the corresponding region, and thus the feature also has marketing characteristics.

The user transacts the commodity quantity feature, which represents the total number of commodities purchased by the consumer user during the period of time, may be determined by counting the total amount of all commodities in the order purchased by the consumer user during the period of time. The number of the commodity in trade can reflect the requirement coverage of the consumer user, so the index also has marketing characteristics.

The user enjoys the discount feature, which represents the total discount amount enjoyed by the consumer user when purchasing goods in the certain period, and generally, the higher the discount amount is, the higher the grade representing the consumer user is or the higher the total purchased amount is, thus having certain marketing characteristics.

According to the analysis of the above, when the feature extraction is performed based on the user transaction data of each consumer user, the used audience marketing features are designed and planned from the marketing characteristics, so that the feature data corresponding to each consumer user under each audience marketing feature is obtained, the feature data comprehensively reflects the marketing characteristics of the corresponding consumer user corresponding to the advertising effect, and compared with the portrait technology purely reflecting the user behavior features, the feature data has better matching performance with the advertising effect.

After feature data of each transaction user under each audience marketing feature is obtained, the feature data can be associated with the corresponding transaction user, and the transaction user identity is converted into an advertisement audience and is stored in a feature database.

Step S4300, each advertisement audience in the feature database is independently ranked based on the feature data of each audience marketing feature, and the ranking order is converted into the feature score corresponding to the audience marketing feature of the advertisement audience and is stored in the marketing feature library.

After the feature database storing the feature data of the advertisement audiences in the marketing features of each audience is obtained, the feature data corresponding to the marketing features of each audience in the feature database is quantized, and accordingly, the feature scores obtained by each advertisement audience under the marketing features of each audience can be determined according to each marketing feature of each audience.

In one embodiment, each advertisement audience in the feature database may be independently ranked according to each audience marketing feature to obtain a ranking order corresponding to each advertisement audience, and the ranking order may be converted into a corresponding feature score. Accordingly, feature scores obtained by the same advertising audience under different audience marketing features are typically different.

In order to realize comprehensive reflection of feature scores obtained by advertisement audiences under different audience marketing features so as to improve the efficiency of subsequent operation, a marketing feature library for representing two-dimensional data can be created, wherein the advertisement audiences are taken as lines, the different audience marketing features are taken as fields, and the corresponding feature scores of the advertisement audiences under the different audience marketing features are stored.

In some embodiments, in determining the corresponding feature scores according to the ranking order, the ranking order may be normalized, for example, the feature score of the advertisement audience with the highest value of the corresponding feature data is determined to be N, where N is the total number of advertisement audiences, the feature score of the advertisement audience with the lowest value of the feature data is determined to be 1, and so on, the feature score is given from high to low according to the feature data, and then the feature score is normalized to a specific numerical space, for example, normalized to the [0,1] numerical space, so that the method is more intuitive.

According to the embodiment, the feature extraction is carried out on the user transaction data generated by the electronic commerce platform in a certain period according to the plurality of preset audience marketing features to obtain the corresponding feature data, then the corresponding feature score is determined according to the feature data on the ordering of the advertisement audience under each audience marketing feature, the standard value quantization is carried out on the feature data of the mass audience marketing features, the extracted features meet the requirements of marketing characteristics, the standardization degree of the operation process is high, the marketing feature library can be obtained for subsequent efficient and frequent calling, and the overall efficiency of determining seed crowds can be improved.

On the basis of any embodiment of the method of the present application, referring to fig. 5, determining, by using a preset audience scoring algorithm, an advertisement value of each candidate audience associated with a plurality of preset value dimensions, and screening, according to the advertisement value, a plurality of candidate audiences corresponding to the seed scale as seed audiences, including:

step S5310, obtaining data corresponding to a plurality of preset value dimensions of each candidate audience from user transaction data in a certain period of the e-commerce platform, including: the method comprises the steps of stopping purchase interval days corresponding to the ordering time of a last triggering transaction event, repeated transaction times in a preset period adjacent to the day, and transaction total corresponding to the preset period;

For the candidate audience set, which contains a large number of candidate audiences exceeding the preset seed scale, fine screening is needed, and for this purpose, the advertisement value corresponding to each candidate audience needs to be further determined. Before determining the advertisement value, a plurality of preset value dimensions can be utilized to obtain data corresponding to each preset value dimension.

In this embodiment, from the marketing perspective, the set value dimensions include an activity value dimension, a loyalty value dimension, and an importance value dimension, so that the number of days of the stop-purchase interval corresponding to the last time of the ordering of the transaction event triggered by each candidate audience, the number of repeated transactions in a preset period adjacent to the current day, and the transaction total corresponding to the preset period are obtained from the user transaction data of each candidate audience within a certain period. The period specified herein may be a historical period of time closest to the day, such as the first 7 days, the first 15 days, the first 30 days, the first 180 days, and so on.

Step S5320, adopting a preset audience scoring algorithm to integrate and quantify the number of days of purchase stopping intervals, the number of repeated transactions and the total transaction amount of each candidate audience into the advertisement value corresponding to the candidate audience;

In this embodiment, the audience scoring algorithm may be set according to the following formula:

the LTV is advertisement value, R is the number of times of purchase stopping interval, F is the number of times of repeated transactions, M is transaction total, wherein the transaction total moderately weakens the influence of the transaction total on the advertisement value through square root opening, regularization treatment is realized by adding 1 to the number of times of purchase stopping interval, the denominator is 0 is avoided, and the denominator can serve as the denominator to quantitatively stop the purchase for a long time, so that the advertisement value of candidate audiences is comprehensively determined, and the method is more scientific, reasonable, practical and effective.

Therefore, the number of the stop purchase interval days, the number of repeated transactions and the total transaction amount of the same candidate audience are fused through a specific formula, so that the advertisement value is determined, the comprehensive quantification of the activity degree, the loyalty degree and the importance degree of the candidate audience in an e-commerce platform is effectively realized, and the estimated value corresponding to the contribution of each candidate audience to the advertisement effect corresponding to the current advertisement delivery can be effectively measured.

And step S5330, optimizing each candidate audience according to the advertisement value, and obtaining a plurality of candidate audiences with the same quantity as the seed scale as the seed audiences.

After determining the corresponding advertisement value for each candidate audience in the candidate audience set, all candidate audiences in the candidate audience set can be optimized according to the advertisement value so as to obtain a plurality of candidate audiences with the same number as the preset seed scale as seed audiences to form a seed audience set.

Specifically, each candidate audience in the candidate audience set can be subjected to reverse ranking according to the advertisement value, and then, a corresponding number of candidate audiences with the front ranking are obtained as seed audiences according to a preset seed scale to be constructed as seed audience sets.

The embodiment shows that the advertisement value corresponding to each candidate audience is determined by fusing the number of days of purchase stopping intervals, the number of repeated transactions and the total transaction amount as a plurality of preset value dimensions, so that potential contribution of each candidate audience to advertisement effect of current advertisement delivery can be effectively measured, selection of the candidate audience can be realized, the seed audience can be accurately determined, the accuracy of determining the seed audience is improved, and accurate advertisement delivery is realized.

Based on any embodiment of the method of the present application, referring to fig. 6, before adopting a preset audience scoring algorithm to aggregate the number of days of outage intervals, the number of repeated transactions and the total transaction amount of each candidate audience into the advertisement value corresponding to the candidate audience, the method includes:

Step S3100, carrying out mathematical modeling on the preset audience scoring algorithm to obtain a corresponding machine learning model;

after the audience scoring algorithm is designed, mathematical modeling can be performed on the basis of a corresponding formula of the audience scoring algorithm, so as to obtain a corresponding machine learning model. The machine learning model may be trained in advance to obtain the ability to determine advertisement value for candidate audience, and the training goal is to learn the ability to determine its corresponding advertisement value according to a given plurality of value dimensions, taking the formula of the previous embodiment as an example, where the machine learning model may directly infer and determine advertisement value using its learned ability when the number of days of purchase outage intervals, the number of repeated transactions, and the total amount of transactions are given. As a generalized example of an application to this embodiment, a machine learning model may also be constructed by means of an algorithm such as Xgboost, lightGBM to determine the corresponding advertisement value for candidate audience, and a model such as Pareto, nbd mode l may also be used to determine the advertisement value.

Step S3200, obtaining data of a plurality of preset value dimensions of a target audience put by advertisement put information in historical put data of an advertisement system as a training sample, and taking advertisement achievement indexes corresponding to the advertisement put information as corresponding supervision labels of the training sample, wherein the data of the plurality of preset value dimensions comprises: the number of days of purchase stopping intervals, the number of repeated transactions and the total transaction amount are used as training samples;

After modeling the machine learning model, the machine learning model may be trained using historical impression data generated by the advertising system. The historical delivery data generated by the advertisement system comprises data corresponding to the target audience and the target audience in a plurality of preset value dimensions when each advertisement delivery exists, and advertisement effect indexes obtained by the current advertisement delivery, wherein the advertisement effect indexes can be fusion results of any one or more of click rate, conversion rate and advertisement input return rate.

And taking data corresponding to each target audience in a plurality of preset value dimensions as training samples, and taking advertisement effect indexes corresponding to the current advertisement delivery of the target audience as supervision labels to obtain a training data set.

Corresponding to the previous embodiment, the data corresponding to the plurality of value dimensions used to construct the training samples herein may include the number of days between purchases, the number of repeat transactions, and the total amount of transactions. Of course, other data of value dimensions that can serve a similar function can be added as desired.

And step S3300, performing iterative training on the machine learning model by adopting the training sample and the corresponding supervision labels thereof, so that the machine learning model is put into use after convergence.

After the training data set is obtained, the training sample can be used for carrying out iterative training on the machine learning model, the loss value of the current training is calculated by using the supervision label corresponding to the training sample during each training, and then the weight updating is carried out on the machine learning model according to the loss value, so that the iteration is carried out until the loss value of the model reaches a preset range, or the iterative training of set times is completed, thereby completing the training on the machine learning model, and enabling the machine learning model to correspondingly determine the advertisement value according to the given data of a plurality of value dimensions.

The above embodiment shows that the mapping relationship between the preset value dimensions and the advertisement value is mathematically modeled, and the mapping relationship can be realized as a machine learning model, and the machine learning model can determine the advertisement value corresponding to the data of each preset value dimension for the advertisement audience by using the prior knowledge through training the machine learning model by using the historical delivery data of the advertisement system, so that the accuracy of determining the advertisement value can be further improved, and the accuracy of seed population is further improved, and the advertisement delivery is more accurate.

On the basis of any embodiment of the method of the present application, referring to fig. 7, determining a target audience by using a seed crowd formed by the seed audience, and delivering the commodity activity information to the target audience, including:

Step S5410, according to the commodity identification specified in the commodity activity information, commodity information of a corresponding target commodity is retrieved from the online store, the commodity data is formatted into commodity activity information in a preset format, and the commodity information comprises commodity titles, commodity pictures and commodity page links;

when advertising is started to the target audience, for given advertising campaign information, a specified commodity identifier can be obtained from commodity campaign information in the advertising campaign information, then commodity information of a target commodity corresponding to the commodity identifier is called in a commodity database of an online store according to the commodity identifier, and the specific commodity information to be obtained can be flexibly set, for example, the commodity information can comprise a commodity title, a commodity picture, a commodity page link and the like.

After the required commodity information is obtained, the obtained commodity information is subjected to data encapsulation according to a preset format to obtain commodity activity information with the preset format, so that the commodity activity information can be analyzed by terminal equipment of a target audience and displayed as information content with a specific format.

Step S5420, constructing each seed audience into a seed crowd pack, and calling a preset crowd expansion interface to determine similar audiences as target audiences according to the seed audiences in the seed crowd pack;

The crowd expansion interface is preset in the advertisement system, so that other advertisement audiences with similar user portrait characteristics can be expanded based on the seed audiences in the given seed crowd pack by utilizing the user portrait characteristics on the basis of the seed audiences.

Therefore, all the seed audiences determined through fine screening can be packaged into a seed crowd pack according to the format defined by the crowd expansion interface, then the crowd expansion interface is called, the seed crowd pack is transferred to the crowd expansion interface, and the target audiences are determined by the seed crowd pack.

When the crowd expansion interface expands the target audience according to the seed audience, user portrait features of various seed audiences can be obtained, the user portrait features of the various seed audiences are utilized to carry out feature similarity matching with the user portrait features of the total consumer users in the user database of the electronic commerce platform, and the user portrait features are optimized according to the similarity of the user portrait features, so that similar consumer users similar to the various seed audiences are determined, the similar consumer users and the various seed audiences are taken as target audiences, and a target audience set is obtained.

Step S5430, pushing the commodity activity information in the predetermined format to the target audience through a preset advertisement delivery interface.

Similarly, the advertisement system is also preset with an advertisement delivery interface, which is responsible for delivering advertisements to a designated target audience. And in the previous step, after the target audience is determined through the crowd expansion interface, the commodity activity information in a preset format is further transmitted to the advertisement delivery interface. The advertisement delivery interface further obtains the target audience determined in the previous step, and sends the commodity activity information packaged in the preset format to each target audience.

In some embodiments, the crowd-extension interface and the advertisement delivery interface may be open interfaces of an instant messaging system or a mailbox system so that merchandise activity information may be sent to various target audience through instant messaging software or mailboxes.

In some embodiments, the crowd-expanding interface and the advertisement delivery interface may be implemented as the same preset interface, and thus, instead of step S5420 and step S5430, the following steps may be performed, that is: and sending commodity activity information in a preset format and seed crowd packages formed by the seed audiences to a preset interface, matching target audiences by the preset interface by utilizing the user portrait features of the seed audiences in the seed crowd packages, and pushing the commodity activity information to the target audiences to realize advertisement delivery.

It will be appreciated from the above embodiments that in particular processes for conducting advertising, merchandise campaign information may be formatted and seed populations comprising seed populations may be used to expand a wider range of target recipients, since seed populations are selected based on marketing characteristics, the target recipients determined based on seed populations are also generally targeted consumer users that are desirable, thereby ensuring good advertising results.

On the basis of any embodiment of the method of the present application, referring to fig. 8, determining a target audience by using a seed crowd formed by the seed audience, and after delivering the commodity activity information to the target audience, the method includes:

step S6100, counting advertisement achievement indexes after the commodity activity information is sent to each target audience, wherein the advertisement achievement indexes comprise any one or fusion values of a plurality of advertisement conversion rate, advertisement click rate and advertisement input return rate;

the advertisement system can track the advertisement effect corresponding to the advertisement putting activity realized according to the application, and optimize the stock advertisement audiences in the marketing feature library by utilizing the tracking result so as to exclude a small amount of advertisement audiences with obviously abnormal value data, prevent the abnormal advertisement audiences from entering the seed crowd of the subsequent advertisement putting activity and provide more accurate audience data for the subsequent advertisement putting.

Accordingly, the advertisement system can count advertisement achievement indexes after commodity activity information in advertisement delivery information is sent to each target audience in each advertisement delivery activity, the advertisement achievement indexes can be any one or a fusion value of any plurality of advertisement conversion rate, advertisement click rate and advertisement input return rate, the calculation mode is known to a person skilled in the art, the calculation mode is flexibly set by the advertisement system, and unnecessary description is avoided.

It can be understood that by tracking statistics, corresponding advertisement delivery information can be obtained for each advertisement delivery.

Step S6200, screening out all advertisement delivery information with the advertisement effectiveness index lower than a preset threshold, and merging the seed audience corresponding to each advertisement delivery information as an abnormal audience to determine an abnormal audience set;

a threshold value is preset corresponding to the advertisement effectiveness index, namely a preset threshold value, is adapted to the needs of the abnormal audience, and then all advertisement putting information with the advertisement effectiveness index lower than the preset threshold value is screened out.

Each advertisement delivery message has a corresponding seed audience, and the seed audience basically influences the determination of the target audience, and further influences the advertisement effectiveness index. When the advertisement putting information can not obtain reasonable advertisement effect indexes, the seed audiences corresponding to the advertisement putting information can be temporarily regarded as abnormal audiences and combined together to form an abnormal audience set.

S6300, averaging advertisement values determined by the abnormal audience in the abnormal audience set corresponding to each advertisement putting information to obtain an average advertisement value;

Whether or not the abnormal audience in the abnormal audience set should be eliminated, and the advertisement value determined by each abnormal audience corresponding to each advertisement delivery information is usually required to be checked for determination. Specifically, the advertisement value determined by each abnormal audience corresponding to each advertisement delivery information can be calculated to be arithmetically averaged to obtain the average advertisement value of each abnormal audience, and the performance of each abnormal audience in all advertisement delivery activities participated by the average advertisement value is measured.

Step S6400, removing abnormal audiences with average advertisement values lower than a preset threshold value from the marketing feature library as extreme audiences.

After the average advertisement value of each abnormal audience is obtained, the abnormal audience set can be filtered by utilizing a preset threshold value which is set correspondingly, in particular, all abnormal audiences with average advertisement value lower than the preset threshold value are determined, the abnormal audiences are taken as extreme audiences, the data of the abnormal audiences are removed from a marketing feature library, and the extreme audiences can not be used as seed audiences any more, so that the influence of the extreme audiences on the determination of the seed population is avoided.

According to the embodiment, the abnormal audience with poor advertising effect is detected according to the advertising effect index by tracking the advertising effect index corresponding to the advertising information, the extreme audience with lower average advertising value is determined according to the average advertising value comprehensively represented by the abnormal audience in each advertising activity, the extreme audience is removed from the marketing feature library, the extreme audience is ensured not to be used for defining the seed population, the accuracy of the advertising system for determining the seed audience for the advertising activity later can be further improved, and the advertising effect is better.

The applicant implemented comparative tests not constituting prior disclosures of the present application with respect to the technical solutions of the present application, and reported the test results as follows:

1. comparing the seed crowd pack determined by the application with the seed crowd pack determined by the traditional technology, examining the performance of the seed crowd pack and the seed crowd pack by taking the advertisement input return rate as an index, and displaying data obtained by the applicant: compared with the seed crowd pack determined by the traditional technology, the advertising investment return rate of the seed crowd pack obtained by the application is improved by 20.16% on average, and the benefit is obvious.

2. Comparing the target audience corresponding to the seed crowd pack determined by the application with the target audience determined by the traditional label orientation technology, and displaying data, the advertisement input return rate obtained by the application is improved by 99.20 percent relative to the label orientation technology, and the benefit is remarkable.

The applicant also delivers the computer program product realized according to the technical scheme of the application to limited merchant user tests of the same merchant platform in a manner which does not constitute the prior disclosure of the application, and test data show that when the seed crowd pack of the application is not used for advertising, the advertising input return rate is about 1.70; when the marketing feature is simply used to determine that the advertisement audience is used as the seed crowd pack, the advertisement input return rate is about 1.79; when the application is adopted to carefully select the determined seed crowd pack for advertisement delivery, the advertisement delivery return rate is about 2.42, and the gradient presented by the advertisement delivery return rate corresponding to each condition is very obvious.

According to the test data, the application has remarkable progress in advertising, and can obviously improve the capability of an advertising system for accurately determining advertising audiences.

Referring to fig. 9, an advertisement delivery control device provided according to an aspect of the present application includes an information obtaining module 5100, an audience candidate module 5200, an audience selecting module 5300, and an advertisement pushing module 5400, where the information obtaining module 5100 is configured to obtain advertisement delivery information corresponding to a commodity advertisement delivered by an online store, and the advertisement delivery information includes commodity activity information, a seed scale, and a plurality of attention marketing features; the audience candidate module 5200 is configured to perform rule matching from the marketing feature library according to the plurality of attention marketing features, and determine a plurality of advertisement audiences with the number greater than the seed size as candidate audiences; the audience selection module 5300 is configured to determine advertisement values of each candidate audience associated with a plurality of preset value dimensions by adopting a preset audience scoring algorithm, and screen a plurality of candidate audiences corresponding to the seed scale according to the advertisement values to serve as seed audiences; the advertisement pushing module 5400 is configured to determine a target audience by using a seed crowd formed by the seed audiences, and to deliver the commodity activity information to the target audience.

On the basis of any embodiment of the apparatus of the present application, the audience candidate module 5200 includes: the feature library calling unit is used for obtaining a preset marketing feature library, wherein the marketing feature library comprises mapping relation data between a plurality of audience marketing features of advertisement audiences and corresponding feature scores of the audience marketing features; a total score determining unit, configured to aggregate feature scores of a plurality of audience marketing features, in which each advertisement audience matches the plurality of attention marketing features, into a feature total score corresponding to the advertisement audience; and the audience checking unit is used for screening out partial advertisement audiences with relatively large feature total scores as candidate audiences according to the preset multiple of the seed scale.

On the basis of any embodiment of the device, the advertisement putting control device of the application further comprises: the data acquisition module is used for acquiring user transaction data generated by each transaction user of the e-commerce platform in a certain period, wherein the user transaction data describe order data corresponding to transaction events triggered by each online store of the corresponding user in the e-commerce platform; the feature extraction module is configured to extract features of user transaction data of each transaction user according to a plurality of preset audience marketing features to obtain corresponding feature data, and store the data of each audience marketing feature of the transaction user serving as an advertisement audience in the feature database, wherein the audience marketing features comprise any of the following: the method comprises the steps of user source channel characteristics, user re-purchase behavior characteristics, user shopping total amount characteristics, user shopping times characteristics, advertisement conversion rate characteristics corresponding to visiting areas of users, user commodity number transaction characteristics and user discount enjoyment characteristics; and the feature scoring module is used for independently sequencing each advertisement audience in the feature database based on the feature data of each audience marketing feature, converting the sequencing order into feature scores corresponding to the audience marketing features of the advertisement audience, and storing the feature scores into the marketing feature library.

On the basis of any embodiment of the apparatus of the present application, the audience selection module 5300 includes: a value data obtaining unit configured to obtain data corresponding to a plurality of preset value dimensions of each candidate audience from user transaction data in a certain period of the e-commerce platform, including: the method comprises the steps of stopping purchase interval days corresponding to the ordering time of a last triggering transaction event, repeated transaction times in a preset period adjacent to the day, and transaction total corresponding to the preset period; the advertisement value determining unit is configured to combine and quantify the number of days of the stop-purchase interval, the number of repeated transactions and the total transaction amount of each candidate audience into the advertisement value corresponding to the candidate audience by adopting a preset audience scoring algorithm; and the audience sequencing and carefully selecting unit is used for optimizing each candidate audience according to the advertisement value, and obtaining a plurality of candidate audiences with the same size as the seed audiences.

On the basis of any embodiment of the device, the advertisement putting control device of the application further comprises: the model construction module is used for carrying out mathematical modeling on the preset audience scoring algorithm to obtain a corresponding machine learning model; the sample processing module is configured to obtain data of a plurality of preset value dimensions of a target audience put by advertisement put information in historical put data of the advertisement system as a training sample, and take advertisement achievement indexes corresponding to the advertisement put information as corresponding supervision labels of the training sample, wherein the data of the plurality of preset value dimensions comprises: the number of days of purchase stopping intervals, the number of repeated transactions and the total transaction amount are used as training samples; and the model training module is used for carrying out iterative training on the machine learning model by adopting the training sample and the corresponding supervision labels thereof, so that the machine learning model is put into use after convergence.

On the basis of any embodiment of the apparatus of the present application, the advertisement pushing module 5400 includes: the information extraction module is used for calling commodity information of a corresponding target commodity in the online store according to the commodity identification specified in the commodity activity information, and formatting the commodity data into commodity activity information in a preset format, wherein the commodity information comprises a commodity title, a commodity picture and a commodity page link; the crowd expansion module is used for constructing each seed audience into a seed crowd pack, calling a preset crowd expansion interface and determining similar audiences as target audiences according to the seed audiences in the seed crowd pack; and the delivery execution module is used for pushing the commodity activity information in the preset format to the target audience through a preset advertisement delivery interface.

On the basis of any embodiment of the device of the application, the advertisement putting control device of the application comprises: the achievement statistics module is used for counting advertisement achievement indexes after the commodity activity information is sent to each target audience, wherein the advertisement achievement indexes comprise fusion values of any one or more of advertisement conversion rate, advertisement click rate and advertisement input return rate; the abnormal investigation module is used for screening all advertisement delivery information with the advertisement effectiveness index lower than a preset threshold value, merging the seed audience corresponding to each advertisement delivery information as an abnormal audience, and determining an abnormal audience set; the average value operation module is used for averaging the advertisement values determined by the abnormal audience corresponding to each advertisement putting information in the abnormal audience set to obtain average advertisement values; and the abnormal cleaning module is used for removing abnormal audiences with average advertisement values lower than a preset threshold value from the marketing feature library as extreme audiences.

The application further provides an advertisement putting control device. As shown in fig. 10, the internal structure of the advertisement delivery control device is schematically shown. The advertisement delivery control device includes a processor, a computer readable storage medium, a memory, and a network interface connected by a system bus. The non-volatile readable storage medium of the advertisement delivery control device is stored with an operating system, a database and computer readable instructions, the database can store information sequences, and the computer readable instructions can enable the processor to realize an advertisement delivery control method when the computer readable instructions are executed by the processor.

The processor of the advertisement delivery control device is configured to provide computing and control capabilities that support the operation of the entire advertisement delivery control device. The advertisement delivery control device may have stored in a memory computer readable instructions that, when executed by a processor, cause the processor to perform the advertisement delivery control method of the present application. The network interface of the advertisement delivery control device is used for connecting and communicating with the terminal.

It will be appreciated by those skilled in the art that the structure shown in FIG. 10 is merely a block diagram of a portion of the structure associated with the present inventive arrangements and is not limiting of the advertising control device to which the present inventive arrangements are applied, and that a particular advertising control device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.

The processor in this embodiment is configured to perform specific functions of each module in fig. 9, and the memory stores program codes and various types of data required for executing the above-described modules or sub-modules. The network interface is used for realizing data transmission between the user terminals or the servers. The nonvolatile readable storage medium in this embodiment stores therein program codes and data necessary for executing all modules in the advertisement delivery control device of the present application, and the server can call the program codes and data of the server to execute the functions of all modules.

The present application also provides a non-transitory readable storage medium storing computer readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the advertisement placement control method of any of the embodiments of the present application.

The application also provides a computer program product comprising computer programs/instructions which when executed by one or more processors implement the steps of the method of any of the embodiments of the application.

It will be appreciated by those skilled in the art that implementing all or part of the above-described methods according to the embodiments of the present application may be implemented by a computer program for instructing relevant hardware, where the computer program may be stored in a non-volatile readable storage medium, and where the program, when executed, may include the steps of the embodiments of the methods described above. The storage medium may be a computer readable storage medium such as a magnetic disk, an optical disk, a Read-only Memory (ROM), or a random access Memory (Random Access Memory, RAM).

In summary, the application can efficiently and accurately determine the seed population corresponding to advertisement delivery according to advertisement delivery information, fully considers the close correspondence with concerned marketing features, fully considers the potential value of each seed audience for advertisement delivery, matches the expected seed scale, expands the target audience by using the seed population on the basis of the potential value, has higher reliability, can obtain good advertisement investment return rate after pushing advertisements to the target audience determined according to the seed population, comprehensively and systematically optimizes the advertisement delivery technology, and realizes quality improvement and synergy of an advertisement system.

Claims

1. An advertisement delivery control method, comprising:

2. The advertisement placement control method according to claim 1, wherein determining a number of advertisement audiences greater than the seed size as candidate audiences by performing rule matching from a marketing feature library according to the plurality of attention marketing features comprises:

obtaining a preset marketing feature library, wherein the marketing feature library comprises mapping relation data between a plurality of audience marketing features of advertisement audiences and corresponding feature scores of the audience marketing features;

summarizing feature scores of a plurality of audience marketing features, which are matched with the plurality of attention marketing features, of each advertisement audience into feature total scores corresponding to the advertisement audience;

and screening out partial advertisement audiences with relatively large feature total scores as candidate audiences according to the preset multiple of the seed scale.

3. The advertisement delivery control method according to claim 2, wherein before acquiring advertisement delivery information corresponding to a commodity advertisement delivered by an online store, comprising:

user transaction data generated by each transaction user of the e-commerce platform in a certain period are obtained, wherein the user transaction data describe order data corresponding to transaction events triggered by the corresponding user in each online store of the e-commerce platform;

According to a plurality of preset audience marketing features, feature extraction is carried out on user transaction data of each transaction user to obtain corresponding feature data, the transaction user is taken as an advertisement audience, and the data of each audience marketing feature are stored in a feature database;

and sequencing each advertisement audience in the feature database based on the feature data of each audience marketing feature independently, converting the sequencing order into feature scores corresponding to the audience marketing features of the advertisement audience, and storing the feature scores into the marketing feature library.

4. The method of claim 1, wherein determining an advertisement value of each of the candidate audience associated with a plurality of preset value dimensions using a preset audience scoring algorithm, and selecting a plurality of candidate audience corresponding to the seed size as a seed audience based on the advertisement value, comprises:

acquiring data corresponding to a plurality of preset value dimensions of each candidate audience from user transaction data in a certain period of an e-commerce platform, wherein the data comprises: the method comprises the steps of stopping purchase interval days corresponding to the ordering time of a last triggering transaction event, repeated transaction times in a preset period adjacent to the day, and transaction total corresponding to the preset period;

Adopting a preset audience scoring algorithm to integrate and quantify the stop purchase interval days, repeated transaction times and transaction total amount of each candidate audience into the advertisement value corresponding to the candidate audience;

and optimizing each candidate audience according to the advertisement value, and obtaining a plurality of candidate audiences with the same quantity as the seed scale as the seed audiences.

5. The method of claim 4, wherein the step of adding the number of days of the outage interval, the number of repeated transactions, and the total amount of transactions for each candidate audience to the advertisement value corresponding to the candidate audience using a predetermined audience scoring algorithm comprises:

carrying out mathematical modeling on the preset audience scoring algorithm to obtain a corresponding machine learning model;

acquiring data of a plurality of preset value dimensions of a target audience put by advertisement put information in historical put data of an advertisement system as a training sample, and taking advertisement achievement indexes corresponding to the advertisement put information as corresponding supervision labels of the training sample, wherein the data of the plurality of preset value dimensions comprises: the number of days of purchase stopping intervals, the number of repeated transactions and the total transaction amount are used as training samples;

And performing iterative training on the machine learning model by adopting the training sample and the corresponding supervision label thereof, so that the machine learning model is put into use after convergence.

6. The advertisement delivery control method according to any one of claims 1 to 5, wherein determining a target audience with a seed population of the seed audience, delivering the commodity activity information to the target audience, comprises:

according to the commodity identification appointed in the commodity activity information, commodity information of a corresponding target commodity is called in the online store, the commodity data are formatted into commodity activity information in a preset format, and the commodity information comprises commodity titles, commodity pictures and commodity page links;

constructing each seed audience into a seed crowd pack, and calling a preset crowd expansion interface to determine similar audiences as target audiences according to the seed audiences in the seed crowd pack;

and pushing the commodity activity information in the preset format to the target audience through a preset advertisement putting interface.

7. The advertisement delivery control method according to any one of claims 1 to 5, wherein determining a target audience with a group of seeds comprising the seed audience, and after delivering the commodity activity information to the target audience, comprises:

Counting advertisement achievement indexes after the commodity activity information is sent to each target audience, wherein the advertisement achievement indexes comprise fusion values of any one or more of advertisement conversion rate, advertisement click rate and advertisement input return rate;

screening all advertisement delivery information with the advertisement effectiveness index lower than a preset threshold value, and combining the seed audience corresponding to each advertisement delivery information as an abnormal audience to determine an abnormal audience set;

averaging advertisement values determined by the abnormal audience in the abnormal audience set corresponding to each advertisement putting information to obtain an average advertisement value;

abnormal audiences with average advertisement value lower than a preset threshold are removed from the marketing feature library as extreme audiences.

8. An advertisement delivery control device, characterized by comprising:

9. An advertising control device comprising a central processor and a memory, characterized in that the central processor is adapted to invoke a computer program stored in the memory for performing the steps of the method according to any of claims 1 to 7.

10. A non-transitory readable storage medium, characterized in that it stores in form of computer readable instructions a computer program implemented according to the method of any one of claims 1 to 7, which when invoked by a computer, performs the steps comprised by the corresponding method.