Customer portrait-based marketing method and related device

Abstract

The embodiment of the invention relates to the field of data processing, and provides a marketing method based on a customer portrait and a related device, and the method comprises the steps: obtaining customer behavior information of a target customer during consumption, carrying out the construction of the customer portrait according to the customer behavior information, and obtaining a target customer portrait, therefore, marketing activity preference information can be generated according to the target customer portrait, target marketing activity information corresponding to the target customer is further determined according to the marketing activity preference information, and the target marketing activity information can be sent to a terminal corresponding to the target customer. Precise marketing based on customer consumption behaviors can be realized, and the matching degree of marketing activities and customer demands and the customer response rate can be improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a marketing method and related apparatus based on customer profiles. Background Technology

[0002] With the rapid development of the digital economy, enterprise marketing models have shifted from traditional broad-based promotion to refined operations, and customers' demand for personalized services is becoming increasingly prominent. On the one hand, existing technologies mostly rely on basic customer information (such as gender, age, and region) or fragmented behavioral data (such as single purchase records and page browsing history) for superficial analysis, making it difficult to truly reflect customers' potential needs and dynamic preferences. On the other hand, existing marketing campaigns are often designed based on common group characteristics, lacking precise adaptation to individual preferences, resulting in widespread indiscriminate push notifications and ineffective outreach. Therefore, how to achieve precise marketing based on customer consumption behavior has become an urgent problem to be solved. Summary of the Invention

[0003] This application provides a marketing method and related apparatus based on customer profiles, which can achieve precise marketing based on customer consumption behavior and improve the matching degree between marketing activities and customer needs and the customer response rate.

[0004] A first aspect of this application provides a marketing method based on customer profiles, the marketing method based on customer profiles including: Obtain customer behavior information related to the consumption of target customers; Customer profiles are constructed based on customer behavior information to obtain target customer profiles. Generate marketing campaign preference information based on target customer profiles; Determine the target marketing campaign information corresponding to the target customers based on marketing campaign preference information; Send targeted marketing campaign information to the corresponding terminals of the target customers.

[0005] A second aspect of this application provides a marketing device based on customer profiles, the marketing device based on customer profiles comprising: The acquisition unit is used to acquire customer behavior information of the target customer at the time of consumption. A construction unit is used to construct a customer profile based on the customer behavior information to obtain a target customer profile. A generation unit is used to generate marketing activity preference information based on the target customer profile. The determining unit is used to determine target marketing activity information corresponding to the target customer based on the marketing activity preference information. The sending unit is used to send the target marketing activity information to the terminal corresponding to the target customer.

[0006] A third aspect of this application provides a terminal including a processor, an input device, an output device, and a memory, wherein the processor, input device, output device, and memory are interconnected, wherein the memory is used to store a computer program, the computer program including program instructions, and the processor is configured to invoke the program instructions to execute the step instructions as described in the first aspect of this application.

[0007] A fourth aspect of this application provides a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform some or all of the steps described in the first aspect of this application.

[0008] A fifth aspect of this application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in the first aspect of this application. The computer program product may be a software installation package.

[0009] Implementing the embodiments of this application has the following beneficial effects: By acquiring customer behavior information of target customers during consumption, customer profiles can be constructed based on the customer behavior information to obtain target customer profiles. Marketing activity preference information can then be generated based on the target customer profiles, and further, target marketing activity information corresponding to the target customers can be determined based on the marketing activity preference information. Subsequently, the target marketing activity information can be sent to the terminal corresponding to the target customers, which is conducive to achieving precise marketing based on customer consumption behavior and can improve the matching degree between marketing activities and customer needs and the customer response rate. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This application provides a flowchart illustrating a marketing method based on customer profiles. Figure 2 This is a schematic diagram of the structure of a terminal provided in an embodiment of this application; Figure 3 This application provides a schematic diagram of the structure of a marketing device based on customer profiles. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0014] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.

[0015] To better understand the customer profiling-based marketing method provided in this application, a brief overview of current customer profiling-based marketing methods is provided below. On one hand, customer profiling lacks depth and dynamism. For example, existing technologies often rely on basic customer information (such as gender, age, and region) or fragmented behavioral data (such as single purchase records and page browsing history) to construct profiles, failing to conduct in-depth, multi-dimensional analysis of consumer behavior throughout the entire process. This includes ignoring pre-purchase decision-making paths (such as the duration of product comparison and changes in search keywords), consumption scenario characteristics (such as whether it is an immediate purchase or whether it is influenced by promotional activities), and post-purchase feedback behaviors (such as evaluation sentiment and repurchase triggers). This superficial data collection results in customer profiles exhibiting labeled and static characteristics, making it difficult to truly reflect customers' potential needs and dynamic preferences, and even leading to discrepancies between the profile and reality (such as misclassifying a customer who occasionally buys baby products as a frequent baby product user). On the other hand, marketing activities are significantly disconnected from customer needs. Due to the ambiguity of customer profiles, existing marketing activities are often designed based on common group characteristics, lacking precise adaptation to individual preferences. For example, pushing high-end membership benefits to price-sensitive customers and focusing on online coupons to customers who prefer offline shopping leads to widespread indiscriminate and ineffective outreach. This misalignment not only wastes significant marketing resources (such as advertising budgets and manpower costs) but can also cause customer resentment due to frequent pushes of irrelevant information, reducing brand trust and customer retention. Furthermore, existing marketing campaign adjustment mechanisms are lagging. Most companies rely on periodic data reviews to optimize marketing strategies, failing to dynamically adjust campaign content based on real-time customer spending behavior. For instance, when customer spending frequency declines recently, timely reminder promotions are not pushed; when customers give negative feedback on a certain type of activity, the same type of information continues to be pushed, further exacerbating the problem of insufficient marketing precision.

[0016] To address the aforementioned issues, this application provides a customer profiling-based marketing method. This method can construct a dynamic and accurate target customer profile by deeply collecting behavioral information from the entire consumption process of target customers. Furthermore, through sub-preference decomposition, association grouping, and fusion extraction, marketing activity preference information tailored to customer needs is generated from the profile. This allows for the customization of appropriate target marketing activities based on preference information and precise reach to customers. This effectively solves the problems of vague customer profiles and disconnect between activities and preferences in existing marketing, reduces indiscriminate push notifications and resource waste, and improves marketing accuracy and customer experience.

[0017] Please see Figure 1 , Figure 1 This application provides a flowchart illustrating a marketing method based on customer profiles. Figure 1 As shown, customer profiling-based marketing methods include: S10: Obtain customer behavior information of the target customer during consumption.

[0018] In this context, target customers refer to a specific customer group that a business or marketing campaign focuses on, possessing potential consumption needs or having already made purchases. Examples include credit card users of a bank or frequent buyers of mother and baby products on an e-commerce platform. Essentially, target customers can be seen as the precise reach of a marketing campaign.

[0019] Customer behavior information during consumption refers to quantifiable and recordable behavioral data generated by customers throughout the entire consumption process, covering all stages of behavior records, including pre-consumption (such as product search and page browsing), during consumption (such as adding to cart, choosing payment method, and consumption amount), and post-consumption (such as order evaluation, after-sales consultation, and feedback on repurchase intention).

[0020] Specifically, for online scenarios, if it's a company's own platform, data such as customer visit pages, dwell time, button clicks, search keywords, shopping cart additions, payment records, and reviews can be extracted from the logs of the company's own apps, mini-programs, and official websites. If integrating with third-party platforms (such as e-commerce platforms and social media platforms), customer consumption-related behaviors on third-party channels (such as clicking on brand ads or sharing product links on social media platforms) can be obtained through API interfaces. For offline scenarios, the offline spending amount, product categories purchased, and payment methods of customers can be collected through the store's POS system; or the browsing path and areas where customers linger in the store can be captured through store cameras (with customer authorization) (such as the duration of time spent in front of a particular product shelf); or customer feedback during offline consultations can be recorded by sales staff (such as interest in certain promotional activities).

[0021] Optionally, the user behavior data collected from the above-mentioned multiple channels can be cleaned (e.g., removing duplicate data and filling in missing fields), standardized (e.g., unifying data formats), and stored in an enterprise data warehouse or customer data platform (CDP) to form the above-mentioned complete customer behavior information of the target customer during consumption. This application does not impose any restrictions on this.

[0022] S20: Based on the customer behavior information, construct a customer profile to obtain a target customer profile.

[0023] Customer profiling can be understood as building a virtual customer model based on multi-dimensional customer data, reflecting the customer's characteristics, preferences, and needs. It can be seen as a precise digital portrayal of the customer. Customer profiling typically includes dimensions such as basic attributes, behavioral characteristics, preference characteristics, and consumption characteristics, and this application does not impose any limitations on this.

[0024] A target customer profile refers to a personalized profile of a target customer. Understandably, this profile focuses more on characteristics relevant to the marketing campaign (such as spending preferences, discount sensitivity, and campaign participation habits) rather than generalized customer information.

[0025] Specifically, the aforementioned target customer profile can be constructed using a process of feature extraction, dimension construction, and tagging. For example, key features can be extracted from customer behavior information, such as purchase frequency (e.g., number of purchases per month, whether it is a high-frequency purchase), purchase preferences (e.g., preferred product categories, price ranges, brand preferences), and behavioral habits (e.g., commonly used purchase channels, purchase time (weekday evenings / weekends), whether it relies on promotional activities for decision-making). Furthermore, the extracted features can be categorized into pre-defined profile dimensions. Core dimensions can include behavioral dimensions, such as purchase frequency, browsing duration, and click preferences; consumption dimensions, such as average purchase amount, proportion of purchased product categories, and payment method preferences; and demand dimensions, such as potential consumption needs (e.g., frequently browsing a certain category but not purchasing, indicating potential demand) and responsiveness to marketing activities (e.g., whether they have participated in discount promotions). It can generate quantifiable or descriptive tags for each dimension of features, such as consumption frequency - high frequency (e.g., monthly consumption ≥ 5 times), product preference - mother and baby category (e.g., mother and baby category consumption accounts for ≥ 60%), discount sensitivity - high (e.g., having placed orders 3 times due to discount activities), thereby forming a structured target customer profile. This application does not impose any restrictions on this.

[0026] S30: Generate marketing campaign preference information based on the target customer profile.

[0027] Marketing campaign preference information refers to customer preference data extracted from target customer profiles that is directly related to marketing campaigns. This information reflects customer preferences regarding the type, format, and benefits of marketing campaigns and can be considered a core basis for subsequently determining target marketing activities.

[0028] Specifically, the aforementioned marketing activity preference information can be generated through a process of profile feature mapping, preference stratification, and association verification. For example, a mapping relationship can be established between the tags of the target customer profile and marketing activity preferences. If the profile tag is "High Discount Sensitivity + Has Participated in Discount Promotions," it can be mapped to "Prefer Discount Promotions"; if the profile tag is "Offline Consumption ≥ 70% + Inclined Towards In-Store Experiences," it can be mapped to "Prefer In-Store Interactive Activities"; if the profile tag is "Membership Level - Platinum + High Frequency Repurchase," it can be mapped to "Prefer Membership Exclusive Benefits (such as Double Points, Exclusive Discounts)." Furthermore, the preliminary preferences obtained from the mapping can be prioritized. For example, if a target customer profile maps to both "Prefer Discount Promotions" and "Prefer Membership Benefits," historical behavior can be analyzed (e.g., the last 3 purchases were all triggered by discount promotions) to determine that "Discount Promotion Preference" has a higher priority than "Membership Benefit Preference." Furthermore, feedback data from customers' historical participation in marketing activities (such as whether they have clicked on a certain type of activity link and whether they placed an order after participating) can be used to verify the accuracy of the above preferences. For example, if the customer profile shows a "preference for live-streaming e-commerce activities", but historical data shows that they have never clicked on a live-streaming link, the preference can be adjusted to "potential preference for live-streaming activities, requiring low-threshold guidance", thereby generating accurate marketing activity preference information. This application does not impose any restrictions on this.

[0029] S40: Determine the target marketing activity information corresponding to the target customer based on the marketing activity preference information.

[0030] The target marketing campaign information can refer to marketing campaign content that is filtered, adjusted, or customized based on the client's marketing campaign preferences, specifically tailored to that target client. This target marketing campaign information may include campaign type (discounts, special offers, trials, membership benefits, etc.), campaign rules (discount amount, participation conditions, validity period), campaign delivery methods (APP push, SMS, offline store notifications), etc., and this application does not impose any restrictions on this.

[0031] Specifically, the aforementioned target marketing activity information can be determined through a process of matching activity pools, adjusting rules, and personalizing customization. For example, businesses can pre-establish a marketing activity pool to store various standardized marketing activities (such as ¥50 off for purchases over ¥200, 20% off for new customers' first order, member points redemption, etc.). Based on the marketing activity preferences of target customers, they can initially select matching marketing activities from the pool. For instance, if target customer A prefers discount activities, all discount-type activities (such as ¥20 off for purchases over ¥100, ¥80 off for purchases over ¥300, etc.) can be selected from the pool. Further, considering customer spending characteristics (obtainable from target customer profiles), the rules for the initially matched activities can be adjusted. For example, if target customer A's average spending is ¥180, and the threshold for the selected ¥50 off ¥200 activity is slightly higher than their spending amount, it can be adjusted to ¥45 off ¥180 to lower the participation threshold. If target customer B prefers weekend spending, the validity period of the marketing activity can be adjusted to Saturday and Sunday to align with target customer B's spending habits. Optionally, if there are no perfectly matching marketing campaigns in the marketing campaign pool, they can be customized based on the preferences of the target customers. For example, if the target customer profile of target customer C shows a preference for the maternal and infant category and high sensitivity to discounts, a marketing campaign of 100 off for every 300 spent can be customized specifically for the maternal and infant category. The above methods can generate and obtain target marketing campaign information containing the marketing campaign type, marketing campaign rules, and marketing campaign reach methods; this application does not impose any restrictions on this.

[0032] S50: Send the target marketing activity information to the terminal corresponding to the target customer.

[0033] The terminal corresponding to the target customer can refer to the device or platform that the target customer uses daily and can receive information. This terminal can include mobile terminals (such as through mobile apps and SMS), PC terminals (such as through website pop-ups), and offline terminals (such as through store displays and sales staff handheld devices). It can be understood that the terminal corresponding to the target customer can be seen as the carrier through which the target marketing campaign information reaches the target customer.

[0034] Specifically, the optimal reach terminals and methods can be selected based on customer behavior habits (such as those obtained from target customer profiles). For example, based on the behavioral habit dimensions in the target customer profile, the priority reach terminals can be determined. If the target customer prefers online consumption and frequently uses the company's app, the aforementioned target marketing activity information can be pushed through the company's app (e.g., via pop-ups or message notifications). If the target customer prefers offline consumption and has provided a mobile phone number, the aforementioned target marketing activity information (including offline store addresses) can be sent via SMS. If the target customer is a member and frequently visits the store, the aforementioned target marketing activity information can be pushed in real time when the customer visits the store using a handheld device (e.g., if you have an exclusive discount for mother and baby products, this discount can be applied to this purchase). This application does not impose any restrictions on this.

[0035] Optionally, the format of the target marketing campaign information can be optimized for different terminals. For example, app push notifications can use a combination of text and images, including campaign rules and product links, while SMS messages can use concise text, including core campaign benefits and short links, to ensure the information is clear and easy to understand. Furthermore, after the target marketing campaign information is sent, terminal feedback data, such as the click-through rate of app push notifications and the open rate of SMS messages, can be used to initially assess the reach. This application does not impose any restrictions on this.

[0036] By implementing a closed-loop process of acquiring consumer behavior data, building precise profiles, extracting activity preferences, customizing marketing campaigns, and precisely reaching customers, this approach avoids the waste of marketing resources caused by traditional indiscriminate push notifications. Through precise characterization of customer profiles and preferences, marketing campaigns directly address customer needs (e.g., pushing exclusive discounts to high-frequency mother and baby users, rather than generalized home appliance offers), significantly improving customer response rates and engagement. Dynamically building profiles and adjusting campaign rules based on customer behavior ensures the personalization and adaptability of marketing campaigns, reducing customer interference from irrelevant marketing activities, enhancing customer experience and brand affinity, and helping businesses improve marketing effectiveness and return on investment (ROI). This is particularly suitable for industries such as retail, e-commerce, and finance that require precise customer reach.

[0037] In this embodiment, by acquiring customer behavior information of target customers during consumption, a customer profile can be constructed based on the customer behavior information to obtain a target customer profile. Marketing activity preference information can then be generated based on the target customer profile, and further, target marketing activity information corresponding to the target customer can be determined based on the marketing activity preference information. The target marketing activity information can then be sent to the terminal corresponding to the target customer, which is conducive to achieving precise marketing based on customer consumption behavior and can improve the matching degree between marketing activities and customer needs and the customer response rate.

[0038] In one possible implementation, when generating marketing campaign preference information, the target customer profile can be hierarchically decomposed and fused. First, the profile is decomposed into multiple sub-preferences, the relationships between these sub-preferences are analyzed and grouped, and then the grouped sub-preferences are fused into higher-dimensional preferences. Finally, accurate marketing campaign preference information is extracted from the sub-preferences and the fused preferences, achieving a deep conversion from customer profile to marketing needs. Specifically, a method for generating marketing campaign preference information based on the target customer profile may include: A1. Perform preference analysis on the target customer profile to obtain k sub-preference information; A2. Obtain the correlation information between k sub-preference information to obtain a set of correlation information; A3. Based on the aforementioned set of related information, perform related preference grouping to obtain n sub-preference information groups, where n is a positive integer less than or equal to k; A4. Group the n sub-preference information and determine the fusion preference information separately to obtain m fusion preference information, where m is a positive integer less than or equal to n; A5. Extract marketing preferences from k sub-preference information and m integrated preference information to obtain marketing campaign preference information.

[0039] Preference analysis can be understood as the process of extracting the specific preferences of target customers across different dimensions from a profile. It is understood that this preference analysis focuses more on features related to marketing activities, such as product categories, types of offers, and activity formats; this application does not impose any restrictions on this.

[0040] The k sub-preferences refer to the k independent and detailed preferences (k being a positive integer) obtained after preference analysis. It should be noted that each sub-preference can correspond to a specific dimension. For example, sub-preference 1 can be a preference for maternal and infant products, sub-preference 2 can be a preference for discount activities, sub-preference 3 can be a preference for weekend consumption periods, etc. This application does not impose any restrictions on this.

[0041] Specifically, natural language processing or feature extraction algorithms can be used to extract sub-preferences from the tags of the target customer profile. For example, from the customer profile of high-frequency maternal and infant consumption, high sensitivity to discounts, and preference for online consumption on weekends, sub-preferences such as maternal and infant product category preference, discount preference, weekend time slot preference, and online channel preference can be extracted (understandably, in this example, k=4), and this application does not impose any restrictions on this.

[0042] Relationship information can refer to the correlation or influence between various sub-preferences, reflecting the linkage between different preferences. For example, the correlation between a preference for discounts and online channels can be higher than that between a preference for discounts and offline channels. The set of relationship information can include one or more relationship information items, which can be regarded as a summary of the relationship information between all sub-preferences. Optionally, this set of relationship information can usually be presented in matrix or list form, and this application does not impose any restrictions on this.

[0043] Specifically, the frequency or influence coefficient of co-occurrence of sub-preferences can be calculated through statistical analysis or association rule algorithms (such as the Apriori algorithm). For example, if analyzing customer historical behavior reveals that 80% of spending during discount promotions is done online, then the association between discount preference and online channel preference can be set as a strong association; if 60% of spending on maternity and baby products occurs on weekends, then the association between maternity and baby product preference and weekend time preference can be set as a medium association. This can be further summarized and integrated to form the aforementioned set of association information, and this application does not impose any restrictions on this.

[0044] Association-based preference grouping can be understood as the process of grouping highly correlated sub-preferences together based on the correlation information between them. Association-based preference grouping can improve the synergy of sub-preferences within each group. The n sub-preference groups can refer to the n preference sets (n≤k) obtained after association-based preference grouping. Optionally, each sub-preference group can contain one or more highly correlated sub-preferences; this application does not impose any restrictions on this. For example, sub-preference group 1: baby and maternity product category preference + discount preference; sub-preference group 2: weekend time slot preference + online channel preference.

[0045] Specifically, clustering algorithms (such as hierarchical clustering) can be used to group sub-preferences with a correlation strength higher than a preset threshold (e.g., a strong correlation threshold or a medium correlation threshold can be set) based on the correlation strength in the associated information. For example, if the preference for maternal and infant products is strongly correlated with the preference for discounts, and the preference for weekend time slots is strongly correlated with the preference for online channels, they can be divided into two sub-preference information groups (in this case, n=2). If a certain sub-preference has a low correlation with all other sub-preferences, it can be grouped separately, and this application does not impose any restrictions on this.

[0046] Determining fused preference information can be understood as integrating various sub-preferences within the same sub-preference information group to generate a comprehensive preference that reflects the core needs of that sub-preference information group. The m fused preference information pieces can refer to the m higher-order preferences obtained after fusion (m≤n). It can be understood that each fused preference information piece can cover the collaborative needs of a sub-preference information group; for example, fused preference 1 could be: a need for discounts on maternity and baby products, and fused preference 2 could be: a need for online consumption on weekends.

[0047] Specifically, for each sub-preference group, core needs can be extracted through semantic fusion or weighted calculation. For example, for the group of preferences for maternal and infant products + discount preferences, it can be merged into a discount preference for maternal and infant products; for the group of preferences for weekend time slots + online channels, it can be merged into a consumption preference for weekend online scenarios. It is understandable that if a group of sub-preferences overlaps with other groups after fusion, such as discount preferences and discount preferences both pointing to price-sensitive discounts after fusion, they can be merged into a single fused preference (in this case, m...). <n)。

[0048] Marketing preference extraction can be understood as filtering and extracting preferences directly related to marketing activities from k sub-preference information and m integrated preference information, while excluding preference content irrelevant to activity design. Marketing activity preference information is the set of preferences output after marketing preference extraction that can be directly used to design marketing activities. For example, preferences for discount activities in the maternal and infant category, preferences for weekend online channel pushes, and high sensitivity to discounts and offers.

[0049] Specifically, information relevant to the marketing campaign can be filtered from k sub-preference information and m integrated preference information based on the dimensions of the marketing campaign design (such as campaign type, category, channel, and time period). For example, sub-preferences that are not related to the marketing campaign, such as "gender" and "age," can be excluded, while sub-preferences and integrated preferences that are directly related to the campaign design, such as "discounts," "baby and maternity products," and "weekend online shopping," can be retained to form the aforementioned marketing campaign preference information after aggregation and integration.

[0050] In this embodiment, through hierarchical processing of sub-preference decomposition, associated grouping, and fusion extraction, the details of customer preferences (sub-preferences) are preserved, while the collaborative needs between preferences (fusion preferences) are also explored. This avoids the one-sidedness of single-dimensional preference analysis, making the generated marketing campaign preference information more comprehensive and accurate in reflecting the real needs of customers. This lays the foundation for designing highly adaptable marketing campaigns and effectively solves the problem of marketing campaigns being out of touch with customer needs caused by fragmentation and weak correlation in traditional preference analysis.

[0051] In one possible implementation, when determining the target marketing activity information corresponding to the target customer, the marketing activity preference information can be encoded and matched with the current marketing activity information to filter out reference marketing activities. Then, the reference activities can be dynamically adjusted based on the target customer's current consumption status to determine the target marketing activity information that accurately matches customer needs and the real-time scenario, thereby achieving precise conversion and scenario-based adaptation from preference to activity. Specifically, a method for determining the target marketing activity information corresponding to the target customer based on the marketing activity preference information may include: B1. Encode the marketing activity preference information to obtain a preference information encoding vector; B2. Obtain a collection of information on current marketing campaigns; B3. Decompose each current marketing activity information in the current marketing activity information set into activity items to obtain the activity item set corresponding to each current marketing activity information; B4. Encode the set of activity items corresponding to each current marketing activity information to obtain the activity item encoding vector corresponding to each current marketing activity information; B5. Based on the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information, determine the reference marketing activity information from the current marketing activity information set; B6. Obtain information on the current consumption status of target customers; B7. Adjust the reference marketing activity information based on the current consumption status information to obtain the target marketing activity information.

[0052] The encoding process can be understood as transforming unstructured preference information into a quantifiable vector form, facilitating subsequent mathematical calculations and similarity comparisons. The preference information encoding vector refers to the numerical vector corresponding to the marketing campaign preference information obtained after encoding. It should be noted that each preference dimension can correspond to a quantifiable value for a preference feature, such as a preference for maternal and infant products being 0.9, a preference for discounts exceeding a certain amount being 0.8, a preference for weekend periods being 0.7, a preference for online channels being 0.8, and a preference for discounts ≥20% being 0.9. This application does not impose any restrictions on this.

[0053] Specifically, feature quantization algorithms (such as one-hot encoding and weighted assignment) can be used to quantify each feature in the marketing activity preference information. For example, category preference, activity type, channel, time period, and discount can be used as vector dimensions, and values ​​can be assigned according to the strength of customer preference (e.g., the assignment range is between 0 and 1, where 1 can represent the most preferred and 0 can represent the least preferred), thereby generating a multi-dimensional preference information encoding vector, such as [0.9, 0.8, 0.8, 0.7, 0.9]. This application does not impose any restrictions on this.

[0054] The current marketing campaign information set can include one or more current marketing campaigns, which can refer to currently actionable marketing campaigns. This set can contain complete information on various marketing campaigns, such as a "spend 200 get 50 off" campaign applicable to all product categories, online, weekdays; or a "20% off" campaign for mother and baby products, offline, weekends, etc. In essence, the current marketing campaign information set can be considered a foundational library for selecting and referencing marketing campaigns.

[0055] Specifically, currently valid activity data can be retrieved from the marketing activity management system, which may include information such as activity type, applicable product category, channel, time period, and discount rules, thereby forming a structured collection of current marketing activity information. This application does not impose any restrictions on this.

[0056] Activity item decomposition can be understood as breaking down each current marketing activity into its core elements (i.e., activity items) to clarify the specific characteristics of each activity. An activity item set refers to the set of elements corresponding to each marketing activity obtained after activity item decomposition. For example, activity item set 1 could be: {Type = Discount, Category = Baby & Maternity, Channel = Online, Time Period = Weekend, Discount Threshold = 300, Discount Amount = 100}. Specifically, each current marketing activity can be decomposed according to preset activity element dimensions (such as type, category, channel, time period, and rule parameters).

[0057] An activity item encoding vector refers to the numerical vector corresponding to each activity item in a set of activity items after encoding them. It's understood that the dimension of this activity item encoding vector can be consistent with the dimension of the aforementioned preference information encoding vector to facilitate subsequent calculations and analysis. For example, the encoding vector for a certain activity might be [0.9, 0.8, 0.8, 0.7, 0.8], corresponding to: Mother and Baby Category = 0.9, Discount Type = 0.8, Online Channel = 0.8, Weekend Period = 0.7, and Discount Amount = 0.8.

[0058] Specifically, the same dimensions and quantification rules as those used for preference information coding can be used to assign values ​​to each element in the set of activity items. For example, if the category in an activity item is mother and baby products, the corresponding dimension is assigned a value of 0.9; if a marketing activity does not involve a certain dimension, then that dimension can be assigned a value of 0.5 (a neutral value).

[0059] Reference marketing campaign information refers to marketing campaigns selected from the current set of marketing campaign information that best match the client's marketing campaign preferences. It can be understood that this reference marketing campaign can serve as the basis for subsequent adjustments. For example, if a client prefers discounts on mother and baby products and online shopping on weekends, then the reference marketing campaign could be: 100 off for every 300 spent on mother and baby products, online shopping, weekends.

[0060] Specifically, the similarity (such as cosine similarity or Euclidean distance) between the preference information encoding vector and the encoding vector of each activity item can be calculated to select the 1-3 activities with the highest similarity as reference marketing activity information. For example, if the cosine similarity between the activity item encoding vector and the preference information encoding vector of a marketing activity is 0.92 (the highest), then that marketing activity can be identified as a reference marketing activity.

[0061] Current consumption status information refers to the real-time or recent consumption scenarios and behavioral data of target customers when a marketing campaign is determined. This current consumption status information can reflect the customer's current consumption context, such as currently browsing a page for baby and maternity products, not having made a purchase in the past 3 days, or currently being in a physical store.

[0062] Specifically, customers' dynamic status can be obtained through real-time data interfaces, including but not limited to: online scenarios (such as the current page being viewed, items in the shopping cart, login status), offline scenarios (such as whether they are in the store, and interaction records with sales staff), and recent behaviors (such as consumption records in the past 24 hours and participation in activities), in order to form the aforementioned current consumption status information.

[0063] Further adjustments based on the current consumption status can yield target marketing campaign information, which can be considered as the final personalized marketing campaign pushed to the customer. For example, if the reference marketing campaign is "¥100 off for every ¥300 spent on baby and maternity products," and considering the current shopping cart amount is ¥280, the campaign can be further adjusted to "¥90 off for every ¥280 spent" to obtain the target marketing campaign information.

[0064] Specifically, the rule parameters of the reference activity can be optimized based on the current consumption status information. For example, the discount amount can be reduced proportionally. If the amount of the product a customer is currently browsing on the page is 280 yuan, and the threshold for the reference marketing activity is 300 yuan, then the threshold can be adjusted to 280 yuan. Additional usage conditions can be added. For example, if the customer is currently in the store and the reference marketing activity is an online activity, rules for simultaneous participation in offline stores can be added. The discount can be increased. For example, if the customer has not made a purchase in the past 3 days, an additional coupon wake-up mechanism can be added on the basis of the reference marketing activity. This application does not impose any restrictions on this.

[0065] In this embodiment, the dual mechanism of encoding vector matching and real-time state adjustment ensures the basic match between marketing activities and customers' long-term preferences, solving the problem of disconnect between activities and preferences in existing solutions. By dynamically adjusting the current consumption status, the activities are adapted to the real-time scenario, avoiding the mechanical push problem in existing solutions. This allows the target marketing activities to not only meet the core needs of customers but also flexibly respond to immediate consumption situations, significantly improving the accuracy and customer acceptance of marketing activities and reducing ineffective outreach and resource waste.

[0066] In one possible implementation, when determining reference marketing campaign information, a first distance can be calculated between the customer preference information encoding vector and the activity item encoding vector of each current marketing campaign to filter out the campaign with the smallest distance as the reference marketing campaign information, thereby achieving precise campaign matching based on quantitative similarity. Specifically, a method for determining reference marketing campaign information from the set of current marketing campaign information based on the preference information encoding vector and the activity item encoding vector corresponding to each current marketing campaign information may include: C1. Calculate the first distance between the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information to obtain the first distance set; C2. Determine the current marketing activity information corresponding to the minimum distance in the first distance set as the reference marketing activity information.

[0067] The first distance can refer to a quantitative metric used to measure the similarity between the preference information encoding vector and the activity item encoding vector. It is understood that a smaller first distance indicates a higher degree of matching. This first distance can be a cosine distance or an Euclidean distance; this application does not impose any limitation on this. The first distance set can refer to the sum of the first distances corresponding to each current marketing activity. This first distance set can contain the similarity measurement results between the activity item encoding and the customer preference information encoding corresponding to each current marketing activity. For example, the first distance set could be {Marketing Activity A: 0.12, Marketing Activity B: 0.35, Marketing Activity C: 0.21}; this application does not impose any limitation on this.

[0068] Optionally, the process of calculating the first distance between the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information to obtain the first distance set can be found in the following formula: Where D(P,A) represents the first distance, indicating the first distance between the preference information encoding vector P and the activity item encoding vector A of a current marketing campaign. It is the core quantitative indicator for measuring the matching degree between the two. P can represent the preference information encoding vector, P=[p1,p2,...,pd], where d is the number of dimensions of the preference information encoding vector. A can represent the activity item encoding vector of a current marketing campaign, such as the activity item encoding vector of the current marketing campaign A, A=[a1,a2,...,ad], where d is the number of dimensions of the activity item encoding vector. The degree corresponds to the dimension of the preference information encoding vector; k can represent the number of distances contained in the effective distance set D′ (i.e., the number of dimensions that were not eliminated), and can be used as the denominator when calculating the mean to ensure that the calculation result can reflect the average matching deviation of the effective dimensions; di can represent the distance of a single dimension, di=|pi-ai|, which can be used to calculate the absolute difference between the i-th preference feature and the corresponding activity item feature, and can reflect the matching deviation between marketing activities and customer preferences under this dimension; D′ can represent the effective distance set, that is, the valuable matching deviation data retained after eliminating the dimensions with extreme deviations.

[0069] It should be noted that the above formula calculates the first distance between the preference information encoding vector P and the activity item encoding vector A of a certain current marketing activity. The above formula can be used to further calculate the first distance between the preference information encoding vector P and the activity item encoding vectors of all current marketing activities (such as current marketing activity B, current marketing activity C, current marketing activity D, etc.) until all current marketing activities have been calculated (e.g., there are y in total) to obtain y first distances. These y first distances are then organized into a set (such as a list or array), which is the first distance set mentioned above.

[0070] The minimum distance can refer to the smallest distance value in the first set of distances. Understandably, the marketing campaign corresponding to this minimum distance has the highest degree of alignment with customer preferences.

[0071] The marketing campaign corresponding to the smallest distance is the aforementioned reference marketing campaign information. This reference marketing campaign information can be understood as a selection of the most suitable marketing campaigns for a given customer, based on their preferences. For example, if marketing campaign A has the smallest calculated distance, then marketing campaign A can be the aforementioned reference marketing campaign.

[0072] In this embodiment, by quantifying the first distance between the preference vector and the activity vector and filtering the activity corresponding to the minimum distance, an objective and accurate match between marketing activities and customer preferences is achieved, avoiding the bias of traditional subjective activity selection. At the same time, using distance as a quantitative indicator ensures the interpretability and consistency of the matching process, laying a highly adaptable foundation for subsequent adjustments to activities based on real-time scenarios, and effectively improving the initial matching accuracy of marketing activities.

[0073] In one possible implementation, feedback from target customers regarding the pushed marketing campaigns can be received, and the original campaigns can be further adjusted based on the feedback to generate optimized campaign information, which is then pushed back to the customer's terminal, forming a marketing closed loop of push, feedback, adjustment, and re-push. Specifically, the customer profile-based marketing method further includes: D1. Receive feedback information sent by the target user regarding the target marketing campaign information; D2. Based on the feedback information, the target marketing campaign is corrected to obtain the corrected target marketing campaign information; D3. Send the corrected target marketing activity information to the terminal corresponding to the target customer.

[0074] Feedback information can refer to the attitudes, needs, or suggestions expressed by target users regarding the target marketing campaign. This feedback information can be divided into explicit feedback, such as clicking "not interested" or "hoping to extend the validity period," and implicit feedback, such as clicking the campaign link but not placing an order or spending too little time browsing the campaign page. This application does not impose any restrictions on this.

[0075] Specifically, feedback information can be collected through multiple channels. For example, online scenarios include setting up "like / dislike" buttons on the activity page within the app, pop-up questionnaires, customer service consultation records, and other feedback collection channels; offline scenarios include sales staff recording customers' verbal feedback on the activity; and behavioral data can be used to infer information, such as not clicking on the activity link can be considered as "implicit disinterest," and clicking but not placing an order can be considered as "unmet needs," etc. This application does not impose any restrictions on this.

[0076] Correction processing can be understood as the process of adjusting the core elements (such as rules, benefits, and format) of the original target marketing campaign based on feedback information to better align it with customer needs. The corrected target marketing campaign information refers to the optimized version of the campaign information generated after correction processing. For example, if the original marketing campaign was "spend 280 and get 90 off," and based on user A's feedback requesting a lower threshold, the campaign could be corrected to "spend 250 and get 80 off."

[0077] Specifically, feedback information can be categorized and analyzed to clarify the direction of correction. For example, if the feedback indicates disinterest (specifically due to low discount), the discount amount can be increased (e.g., adjust the discount from 280 minus 90 to 280 minus 100); if the feedback indicates a short validity period, the marketing campaign duration can be extended (e.g., extend from 3 days to 7 days); if the feedback indicates an incompatible product category, the applicable product category for the marketing campaign can be changed (e.g., change from FMCG to a dedicated mother and baby product category).

[0078] In this embodiment, by receiving feedback from target customers and making targeted adjustments to the campaign, the limitations of one-time pushes in the traditional marketing process are broken, enabling the marketing campaign to dynamically adapt to the real-time needs of customers and effectively solving the problem of discrepancies between the marketing campaign and customer needs. At the same time, the closed-loop design of feedback, adjustment, and re-push not only enhances the customer's sense of participation and being valued, but also reduces customer aversion caused by ineffective activities, ultimately improving the conversion rate and customer satisfaction of the marketing campaign.

[0079] For examples consistent with the above embodiments, please refer to... Figure 2 , Figure 2 This is a schematic diagram of the structure of a terminal provided in an embodiment of this application, such as... Figure 2 As shown, it includes a processor, an input device, an output device, and a memory, which are interconnected. The memory is used to store a computer program, which includes program instructions. The processor is configured to call the program instructions. The program includes instructions for performing the following steps. Obtain customer behavior information related to the consumption of target customers; Based on the customer behavior information, a customer profile is constructed to obtain a target customer profile. Marketing campaign preference information is generated based on the target customer profile. Based on the marketing activity preference information, determine the target marketing activity information corresponding to the target customers; The target marketing campaign information is sent to the terminal corresponding to the target customer.

[0080] The above mainly describes the solutions of the embodiments of this application from the perspective of the method execution process. It is understood that, in order to achieve the above functions, the terminal includes the corresponding hardware structure and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments provided herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0081] This application embodiment can divide the terminal into functional units according to the above method example. For example, each function can be divided into a separate functional unit, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0082] For those consistent with the above, please refer to Figure 3 , Figure 3 This application provides a schematic diagram of a marketing device based on customer profiles as an embodiment of the present application. Figure 3 As shown, the device includes: The acquisition unit 101 is used to acquire customer behavior information of the target customer at the time of consumption. Construction unit 102 is used to construct a customer profile based on the customer behavior information to obtain a target customer profile; The generation unit 103 is used to generate marketing activity preference information based on the target customer profile; Determining unit 104 is used to determine target marketing activity information corresponding to the target customer based on the marketing activity preference information; The sending unit 105 is used to send the target marketing activity information to the terminal corresponding to the target customer.

[0083] In one possible implementation, the generation unit 103 is configured to generate marketing activity preference information based on the target customer profile, specifically for: Preference analysis is performed on the target customer profile to obtain k sub-preference information; Obtain the correlation information among the k sub-preference information to obtain a set of correlation information; Based on the set of associated information, associated preferences are grouped to obtain n sub-preference information groups, where n is a positive integer less than or equal to k; The n sub-preference information groups are grouped and fused preference information is determined to obtain m fused preference information, where m is a positive integer less than or equal to n; Marketing preferences are extracted from the k sub-preference information and the m fused preference information to obtain marketing campaign preference information.

[0084] In one possible implementation, the determining unit 104 is configured to determine target marketing activity information corresponding to the target customer based on the marketing activity preference information, specifically for: The marketing activity preference information is encoded to obtain a preference information encoding vector; Obtain a collection of information on the current marketing campaign; Each current marketing activity information in the current marketing activity information set is split into activity items to obtain the activity item set corresponding to each current marketing activity information; The activity item set corresponding to each current marketing activity information is encoded to obtain the activity item encoding vector corresponding to each current marketing activity information. Reference marketing activity information is determined from the set of current marketing activity information based on the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information. Obtain the current consumption status information of the target customer; The reference marketing campaign information is adjusted based on the current consumption status information to obtain the target marketing campaign information.

[0085] In one possible implementation, the determining unit 104 is configured to determine reference marketing activity information from the current marketing activity information set based on the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information, including: Calculate the first distance between the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information to obtain a first distance set; The current marketing campaign information corresponding to the minimum distance in the first distance set is determined as the reference marketing campaign information.

[0086] In one possible implementation, the sending unit 105 is further configured to: Receive feedback information sent by the target user regarding the target marketing campaign information; The target marketing campaign is corrected based on the feedback information to obtain the corrected target marketing campaign information. The corrected target marketing campaign information is sent to the terminal corresponding to the target customer.

[0087] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the customer profile-based marketing methods described in the above method embodiments.

[0088] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program that causes a computer to perform some or all of the steps of any of the customer profile-based marketing methods described in the above method embodiments.

[0089] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0090] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0091] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0092] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0093] Furthermore, the functional units in the various embodiments of the application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software program module.

[0094] If the integrated unit is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0095] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory, a random access memory, a magnetic disk, or an optical disk, etc.

[0096] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A marketing method based on customer profiles, characterized in that, The customer profiling-based marketing methods include: Obtain customer behavior information related to the consumption of target customers; Based on the customer behavior information, a customer profile is constructed to obtain a target customer profile. Marketing campaign preference information is generated based on the target customer profile. Based on the marketing activity preference information, determine the target marketing activity information corresponding to the target customers; The target marketing campaign information is sent to the terminal corresponding to the target customer.

2. The marketing method based on customer profiles according to claim 1, characterized in that, The step of generating marketing campaign preference information based on the target customer profile includes: Preference analysis is performed on the target customer profile to obtain k sub-preference information; Obtain the correlation information among the k sub-preference information to obtain a set of correlation information; Based on the set of associated information, associated preferences are grouped to obtain n sub-preference information groups, where n is a positive integer less than or equal to k; The n sub-preference information groups are grouped and fused preference information is determined to obtain m fused preference information, where m is a positive integer less than or equal to n; Marketing preferences are extracted from the k sub-preference information and the m fused preference information to obtain marketing campaign preference information.

3. The marketing method based on customer profiles according to claim 2, characterized in that, The step of determining the target marketing activity information corresponding to the target customer based on the marketing activity preference information includes: The marketing activity preference information is encoded to obtain a preference information encoding vector; Obtain a collection of information on the current marketing campaign; Each current marketing activity information in the current marketing activity information set is split into activity items to obtain the activity item set corresponding to each current marketing activity information; The activity item set corresponding to each current marketing activity information is encoded to obtain the activity item encoding vector corresponding to each current marketing activity information. Reference marketing activity information is determined from the set of current marketing activity information based on the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information. Obtain the current consumption status information of the target customer; The reference marketing campaign information is adjusted based on the current consumption status information to obtain the target marketing campaign information.

4. The marketing method based on customer profiles according to claim 3, characterized in that, The step of determining reference marketing activity information from the set of current marketing activity information based on the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information includes: Calculate the first distance between the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information to obtain a first distance set; The current marketing campaign information corresponding to the minimum distance in the first distance set is determined as the reference marketing campaign information.

5. The marketing method based on customer profiles according to any one of claims 1-4, characterized in that, The method further includes: Receive feedback information sent by the target user regarding the target marketing campaign information; The target marketing campaign is corrected based on the feedback information to obtain the corrected target marketing campaign information. The corrected target marketing campaign information is sent to the terminal corresponding to the target customer.

6. A marketing device based on customer profiles, characterized in that, The device includes: The acquisition unit is used to acquire customer behavior information of the target customer at the time of consumption. A construction unit is used to construct a customer profile based on the customer behavior information to obtain a target customer profile. A generation unit is used to generate marketing activity preference information based on the target customer profile. The determining unit is used to determine target marketing activity information corresponding to the target customer based on the marketing activity preference information. The sending unit is used to send the target marketing activity information to the terminal corresponding to the target customer.

7. The marketing device based on customer profiles according to claim 6, characterized in that, The generation unit is used to generate marketing activity preference information based on the target customer profile, specifically for: Preference analysis is performed on the target customer profile to obtain k sub-preference information; Obtain the correlation information among the k sub-preference information to obtain a set of correlation information; Based on the set of associated information, associated preferences are grouped to obtain n sub-preference information groups, where n is a positive integer less than or equal to k; The n sub-preference information groups are grouped and fused preference information is determined to obtain m fused preference information, where m is a positive integer less than or equal to n; Marketing preferences are extracted from the k sub-preference information and the m fused preference information to obtain marketing campaign preference information.

8. The marketing device based on customer profiles according to claim 7, characterized in that, The determining unit is used to determine target marketing activity information corresponding to the target customer based on the marketing activity preference information, specifically for: The marketing activity preference information is encoded to obtain a preference information encoding vector; Obtain a collection of information on the current marketing campaign; Each current marketing activity information in the current marketing activity information set is split into activity items to obtain the activity item set corresponding to each current marketing activity information; The activity item set corresponding to each current marketing activity information is encoded to obtain the activity item encoding vector corresponding to each current marketing activity information. Reference marketing activity information is determined from the set of current marketing activity information based on the preference information encoding vector and the activity item encoding vector corresponding to each current marketing activity information. Obtain the current consumption status information of the target customer; The reference marketing campaign information is adjusted based on the current consumption status information to obtain the target marketing campaign information.

9. A terminal, characterized in that, The device includes a processor, an input device, an output device, and a memory, which are interconnected. The memory is used to store a computer program, which includes program instructions. The processor is configured to invoke the program instructions to execute the customer profile-based marketing method as described in any one of claims 1-5.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the customer profile-based marketing method as described in any one of claims 1-5.

Images