Dynamic coupon generation method and system

Abstract

The invention relates to the technical field of coupon generation, and discloses a dynamic coupon generation method and system, and the method comprises the steps: collecting the historical behavior data of a customer, building a state vector, and initializing the state information of the customer; setting a coupon strategy variable based on the customer state; constructing a cost and response benefit function, and establishing an optimization target model of the coupon strategy; solving the objective function through a numerical optimization method to obtain an optimal coupon strategy path; generating coupon information according to the strategy path, and binding the coupon information to the client identity label; the system comprises a data source module, a data cleaning module, a feature engineering module, a strategy engine module, an optimization module, a coupon code generation module and a distribution module. According to the technical scheme of constructing the customer state vector model based on the customer historical behavior data, a more accurate coupon strategy generation effect is achieved, personalized coupons can be customized for each customer, and the utilization rate of the coupons and the satisfaction degree of the customers are improved.

Description

Technical Field

[0001] This invention relates to the field of coupon generation technology, specifically a method and system for dynamically generating coupons. Background Technology

[0002] In modern marketing systems, coupons are widely used as an important promotional tool to enhance customer loyalty, boost sales conversion, and strengthen brand influence. By issuing various coupons, discount coupons, and spending-reduction coupons, coupon systems can effectively guide customers to participate in activities, stimulate their willingness to consume, and thus drive overall transaction growth. With the continuous expansion of e-commerce platforms, chain retail, and the automotive consumer market, the role of coupons in marketing plans is becoming increasingly prominent, and they have become an important tool for enterprises to implement differentiated competition and precision marketing. Especially against the backdrop of increasingly diversified brands and products and intensifying market competition, the intelligence and personalization of coupon strategies have gradually become key indicators of the maturity of a marketing system.

[0003] Most existing coupon generation methods are based on preset rules and standardized templates, such as setting fixed face values, specifying expiration dates, and uniform applicable scopes. These methods often use batch generation and distribute coupons mainly based on time nodes, promotional schedules, or customer segmentation rules.

[0004] However, existing coupon code generation methods still have room for improvement in terms of uniqueness and traceability, and pose certain risks. The push methods mostly use uniform templates, failing to adequately utilize user profiles and thus struggling to effectively attract customer participation. Furthermore, the distribution of coupons rarely considers customers' purchase records, browsing habits, and interests, easily leading to content that doesn't match demand, impacting usability and increasing resource consumption. Existing mechanisms offer limited support for dynamically balancing distribution costs and customer response, with strategies largely statically set and lacking flexible adjustment capabilities. Therefore, this invention provides a dynamic coupon generation method and system to address the shortcomings of existing technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for dynamically generating coupons, which solves the problems of lack of personalized customization, insufficient balance between cost and benefit, and poor anti-counterfeiting features in existing coupon generation methods.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for dynamically generating coupons, comprising the following steps: Acquire customers' historical behavior data, construct a customer state vector model, and initialize customer state information; Define card and coupon policy control variables based on customer status information, including card and coupon type, face value, usage conditions and validity period; Based on the relationship between customer status and strategy control variables, an optimization objective function for coupon distribution costs and customer response benefits is constructed. The optimal strategy is derived based on the objective function. The objective function is minimized within a predetermined time interval using numerical optimization methods, and the optimal coupon strategy path is obtained by solving the problem. Generate coupon information based on the optimal coupon strategy path, and bind the generated coupon structure with the customer identifier.

[0007] Preferably, the step of obtaining customer historical behavior data includes: Real-time collection of customer order data, including order type, amount, and customer location information; Acquire customer behavior data, including browsing history, consultation history, and visit frequency; External information is obtained from multiple data sources, including industry data, market research reports, and government policies; Machine learning algorithms are used to model historical data and construct a customer state vector model.

[0008] Preferably, the step of defining the card and coupon strategy control variables based on customer status information includes: Define coupon types, including discount coupons, spending-reduction coupons, service coupons, and exchange coupons; Determine the face value of the coupons and adjust them dynamically according to the needs of the customer base; The validity period and applicable scope of the coupons can be set according to different usage conditions; The optimal strategy control variables for coupons are determined based on customer profiles and historical data, and adjustments are made in real time.

[0009] Preferably, the step of constructing the objective function for optimizing coupon distribution costs and customer response benefits includes: Define a coupon issuance cost function, including the coupon's face value, issuance cost, and distribution cost; Define a customer response benefit function, including customer response probability, satisfaction, and conversion rate; Construct an optimization objective function, and adjust the coefficients in the objective function based on the relationship between customer status information and policy variables.

[0010] Preferably, the step of deriving the optimal strategy based on the objective function includes: Numerical optimization methods, including genetic algorithms and particle swarm optimization, are used to minimize the optimization objective function; The optimal coupon strategy path is derived based on the objective function, including coupon type, face value, and validity period parameters; Within a predetermined time interval, the optimal strategy is adjusted based on real-time data through an iterative optimization method. Choose the most suitable coupon scheme to ensure efficient marketing results.

[0011] Preferably, the step of generating coupon information based on the optimal coupon strategy path includes: The generated coupon information is linked to the customer's identifier to ensure that each customer has a unique coupon; Set the conditions for issuing coupons based on the optimal strategy, including time and amount limits for use; The generated coupon information is stored in a database.

[0012] Preferably, the formula for calculating the coupon distribution cost function is as follows: ; in, Costs associated with issuing coupons; The face value of the coupon; Costs associated with issuing coupons; Distribution costs for coupons; The formula for calculating the customer response benefit function is as follows: ; in, The expected response benefit value for customers to coupon types; The probability of a customer's response; For customer satisfaction; For customer conversion rate.

[0013] Preferably, the optimization formula for the optimal coupon strategy path is: ; in, This represents the optimization value for optimizing the coupon strategy path; Costs associated with issuing coupons; To improve the marketing effectiveness of coupons; For the time efficiency of coupons; , , To adjust the coefficients and control the balance between distribution costs, marketing effectiveness, and time efficiency.

[0014] Preferably, the formula for setting the conditions for issuing the coupons is: ; in, Indicates the validity period of the coupon; Indicates the minimum amount that can be used with the coupon; The expected response benefit value for customers to coupon types; This indicates the conditions under which the customer can currently use the coupons.

[0015] A dynamic coupon generation system is also provided, including: The data source module is used to collect historical customer behavior data and sales data; The data cleaning module is used to clean, standardize, and fill in missing values ​​for the collected data. The feature engineering module is used to extract customer features and build feature models; The strategy engine module is used to generate coupon strategies based on customer characteristics and business rules; The optimization module is used to generate the optimal coupon strategy based on the optimization objective. The coupon code generation module is used to generate unique coupon codes; The distribution module is used to accurately distribute the generated coupons to the target customers.

[0016] This invention provides a method and system for dynamically generating coupons. It has the following beneficial effects: 1. This invention's technical solution, which constructs a customer state vector model based on historical customer behavior data, achieves a more accurate coupon strategy generation effect. Compared to existing technologies that distribute coupons according to fixed rules, this invention dynamically analyzes customers' consumption, browsing, and inquiry behaviors to tailor personalized coupons for each customer, thereby avoiding the push of invalid coupons and improving coupon usage rate and customer satisfaction.

[0017] 2. This invention introduces a numerical optimization method, deriving the optimal coupon strategy by optimizing the objective function, achieving the best balance between distribution costs and customer response benefits. Compared with fixed or coarse optimization strategies in existing technologies, this invention uses methods such as genetic algorithms and particle swarm optimization to dynamically adjust parameters such as coupon face value and validity period, maximizing marketing effectiveness while reducing resource waste and cost expenditure.

[0018] 3. This invention ensures the uniqueness and anti-counterfeiting properties of each coupon through a coupon code generation module, thereby enhancing system security. Compared to coupon codes in existing technologies that are easily duplicated or forged, this invention employs encryption technology and a random generation algorithm to guarantee the immutability of the coupons, making the system more secure and reliable, and increasing customer trust in the coupons.

[0019] 4. This invention employs a precise coupon distribution mechanism, enabling accurate coupon delivery based on customer profiles and behavioral data, achieving highly efficient marketing results. Unlike existing technologies that indiscriminately push coupons, this invention delivers coupons at the optimal time and through the best channels, ensuring that each coupon reaches the appropriate target customers, thereby effectively improving coupon conversion rates and avoiding unnecessary marketing interference and waste. Attached Figure Description

[0020] Figure 1This is a flowchart of the method steps of the present invention; Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see the appendix Figure 1 This invention provides a method for dynamically generating coupons, comprising the following steps: S1. Obtain customers' historical behavior data, construct a customer state vector model, and initialize customer state information; S2. Define card and coupon policy control variables based on customer status information, including card and coupon type, face value, usage conditions and validity period; S3. Based on the relationship between customer status and strategy control variables, construct an optimization objective function for coupon distribution costs and customer response benefits; S4. Based on the optimization objective function, derive the optimal strategy. Minimize the objective function within a predetermined time interval using numerical optimization methods to obtain the optimal coupon strategy path. S5. Generate coupon information based on the optimal coupon strategy path, and bind the generated coupon structure with the customer identifier.

[0023] For step S1, in this embodiment, historical customer behavior data is collected in real time through multiple data sources. This includes the following types of data: Order data includes customers' historical purchase records, such as order type, amount, purchase frequency, product type, and purchase time. Analyzing order data reveals customer consumption patterns and purchasing trends over different time periods. For example, some high-frequency buyers may belong to a high-value customer group.

[0024] Browsing behavior data: This records information such as products and pages that customers browse on the platform, helping the system identify customer interests and potential needs. For example, if a customer frequently browses a particular car model or accessory, it indicates that they have a high level of interest in that product.

[0025] Consultation behavior data: Customer interactions with the platform, such as online consultations and customer service call records. By analyzing customer consultation data, we can more accurately grasp customers' purchasing intentions and their potential changes in needs.

[0026] Location information: Based on the customer's location information, the system can identify the customer's regional preferences and then generate regionally targeted coupon strategies.

[0027] In some embodiments, the system can also obtain supplementary information from external data sources, including industry data, market research reports, sales rankings, and policy subsidies. This external data provides the system with broader market background support, enabling coupon strategies to better align with market demands.

[0028] After data acquisition, the system cleans and standardizes all collected customer behavior data. The cleaning process includes removing incomplete or invalid customer data, such as filtering out users who do not meet the age requirements (e.g., customers under 18 or over 80), and standardizing category data, such as unifying "new energy vehicles" to "NEV" to ensure consistency and data usability in subsequent analysis.

[0029] After acquiring cleaned and standardized historical customer behavior data, the system will construct a customer state vector model based on this data. The customer state vector model is a high-dimensional feature space designed to comprehensively reflect each customer's behavioral characteristics, needs, and potential. The construction of this model includes the following steps: Feature extraction: By analyzing order data, browsing data, inquiry data, and geolocation information, features that reflect customer behavior are extracted. Common features include total spending, browsing preferences, and purchase frequency.

[0030] Feature transformation and encoding: In order for the data to be used for model training, the system will encode categorical features (such as using StringIndexer and OneHotEncoder to encode gender, region, etc.) and normalize numerical features (such as using StandardScaler for numerical normalization).

[0031] Customer State Vector Construction: The transformed features are combined to form a multi-dimensional customer state vector. This vector will simultaneously contain customer behavioral data (such as spending amount, purchase frequency, etc.), interests and preferences (such as browsed product categories), and geographical information (such as location, city, etc.). The customer state vector will serve as the basic data input for the subsequent coupon strategy generation process.

[0032] Alternatively, the system can use the K-Means clustering algorithm to group customers during the construction of the customer state vector model. Specifically, the system can divide customers into multiple groups based on characteristics such as spending amount, browsing frequency, and purchase frequency. For example, for high-value customer groups, the system may generate more coupons with greater discounts; for low-frequency consumers, it may generate more attractive promotional coupons to stimulate their spending.

[0033] Specifically, the customer state vector model can be represented as a vector in a high-dimensional space, where each dimension represents a certain characteristic of the customer, for example: ; in, Customer state vector; These represent different features, such as spending amount, browsing frequency, and number of inquiries. In some embodiments, these features are further optimized using machine learning algorithms to improve the model's accuracy.

[0034] In one possible implementation, the system uses machine learning algorithms (such as logistic regression, random forest, or neural networks) to transform customers' historical behavior data into "preference vectors," which represent the customer's state at a given moment. By comparing the state vectors of different customers, the system can segment customer groups and develop different coupon strategies.

[0035] For example, suppose the system extracts the following feature vector from a customer's historical data: ; in, A feature vector representing a specific customer; This is the customer's total spending amount. The number of products viewed. The probability of responding to the coupon. Its access frequency, The system identifies the customer's region. Based on this data, the system will generate a coupon strategy tailored to the customer's needs.

[0036] For step S2, in this embodiment, based on the constructed customer state vector model, several control variables related to the coupon generation strategy are set. These control variables mainly include the coupon type, face value, usage conditions, and validity period, all of which are decision parameters required for subsequent optimization algorithms.

[0037] Specifically, coupon types can be predefined into several categories to suit marketing objectives based on different customer profiles and behavioral characteristics. Generally, coupon types include discount coupons, spending-reduction coupons, service coupons, exchange coupons, loan subsidy coupons, etc. Discount coupons are used to directly reduce the percentage discount on the payment amount; Discount coupons will only be valid after a specific spending amount is met. Service vouchers are linked to a specific after-sales service, such as free maintenance; Trade-in vouchers are typically used as subsidies when trading in an old car for a new one. Loan vouchers are linked to financial services, such as down payment subsidies or interest rate discounts; Full payment vouchers and new energy vehicle subsidy vouchers are exclusive to specific vehicle models and are used for specific sales strategies.

[0038] In some embodiments, the system automatically limits the range of available coupon types based on the customer segmentation tags to which the user belongs. For example, for customer groups with high historical discount usage, priority is given to allocating discount coupons or limited-time discount coupons; while for high-value customer groups, service coupons or loan coupons can be given priority to enhance customer loyalty.

[0039] Regarding card and coupon denomination settings, the system supports dynamic range configuration. Generally, the system allows setting a minimum denomination of 50 yuan and a maximum of 2000 yuan. Specific limits can be tiered based on factors such as the customer's historical transaction amount and spending level. For example, customers with an average monthly order amount exceeding 30,000 yuan will have a higher maximum card and coupon denomination limit.

[0040] As an option, the system can also set additional value weights based on the popularity of the purchased car model or brand. For example, if the purchase frequency of a certain car model in the current region exceeds a set threshold (such as 500 vehicles / month), the system can include that car model in the "promotion priority" model, and the corresponding coupon value will be increased by 5%-10%.

[0041] The usage conditions mainly include the range of applicable vehicle models, minimum order amount, sales channel requirements (such as whether it is limited to directly operated stores), and customer type restrictions (such as whether it is limited to new customers). Specifically: The minimum amount that can be used can be set at 1.5-2.0 times the face value of the coupon; The applicable vehicle models are linked to the sales system through vehicle codes, and the vouchers can only be used on vehicles of specific configurations or batches. For service vouchers, you can also set conditions such as whether they need to be activated within 30 days of purchasing the vehicle.

[0042] Regarding the validity period, the system supports two configuration methods by default: fixed date and relative time. Fixed date format, such as "May 1, 2025 to June 30, 2025"; The relative time period indicates that it is "valid for 30 days from the date of issuance"; In one possible implementation, the system dynamically sets the validity period based on customer activity. For example, monthly active users can be assigned a shorter period (e.g., 14 days), while users who haven't made a purchase in six months can be assigned a longer validity period (e.g., 60 days) to improve their return rate.

[0043] In addition, for certain special event scenarios, the system also supports setting "overlay rules" and "mutual exclusion restrictions". The former allows the same customer to use multiple coupon combinations at the same time, while the latter is used to control the mutual exclusion between high-value coupons and regular coupons to prevent excessive marketing costs from accumulating.

[0044] For step S3, in this embodiment, after establishing the association between customer status information and strategy control variables, it is necessary to construct a mathematically defined and optimizable objective function system to balance the relationship between coupon distribution costs and customer response benefits, thereby supporting the derivation of subsequent strategy paths. This step plays a crucial role in connecting the preceding and following steps and is the core computational foundation for realizing personalized coupon distribution strategies. By clearly describing the distribution costs and customer benefits through mathematical modeling, scientific evaluation and dynamic adjustment at the strategy level can be achieved.

[0045] In this embodiment, the following calculation function is defined for the cost of issuing coupons: ; in, Costs associated with issuing coupons; It indicates the face value of the coupon, reflecting the direct benefits the customer receives when redeeming it; This indicates the cost of issuing coupons, which mainly covers the basic operating costs incurred by the platform in the process of creating and configuring coupons; This indicates the distribution cost of coupons, which is usually related to distribution channels and reach methods (such as SMS, app push, email, etc.), and may also include the portion of platform system resources consumed.

[0046] In some embodiments, the face value of the coupon The distribution cost can be dynamically set based on factors such as the user's historical spending, current user level, and predicted conversion rate; This can be a fixed configuration cost or a cost that can be fine-tuned based on policy changes; while distribution costs... The cost can vary significantly depending on the channel; for example, SMS costs are higher than in-site push notifications.

[0047] Meanwhile, the customer response benefit function is defined as follows: ; in, For customers regarding coupon types The expected response benefit value; An identifier for the type of coupon or its distribution strategy; This indicates the probability of a customer responding to the coupon, reflecting their actual tendency to interact after receiving the coupon; Customer satisfaction can be calculated from signals such as feedback submitted after receiving coupons, NPS scores, and service evaluations. This represents the customer conversion rate, which is the ratio from receiving a coupon to actually using it and converting it into an order or service consumption.

[0048] Specifically, in one possible implementation, Predictive modeling can be performed by statistically analyzing responses to similar historical coupons, such as logistic regression models or behavioral models based on collaborative filtering. This allows for the integration of unstructured data such as customer service experience ratings and coupon reviews; It is usually obtained by aggregating the conversion probabilities of each node in the order placement behavior and conversion path.

[0049] In one type of implementation, in order to achieve goal-oriented strategy formulation, the following will be used: and All components are integrated into a single optimization framework to construct the overall objective function, supporting subsequent iterative optimization methods. This design effectively balances platform resource consumption and customer-perceived value, demonstrating the technical advantages of intelligent strategy adjustment and dynamic response in this invention.

[0050] As an option, the platform can set weighting factors to... and The two can be combined to form a multi-objective optimization expression, and the ratio of the two can also be used as an evaluation index to screen the cost-effectiveness level under different strategy paths, thereby assisting the strategy module in outputting the optimal solution path.

[0051] Furthermore, to avoid over-incentivizing or creating imbalances for customers, a penalty term can be introduced into the objective function to limit the excessive skewness of high-cost strategies. This approach ensures convergence of the optimization objective while also improving the business adaptability of the strategy output.

[0052] Under normal circumstances, the optimization function and its sub-items will be updated in real time with time and customer status to adapt to the dynamic adjustment needs of marketing strategies in different cycles and scenarios, ensuring that the entire system has strong online response and strategy evolution capabilities.

[0053] In a real-world deployment scenario, the function calculation process is executed periodically in the strategy engine. After each round of evaluation, the caching mechanism is used to connect with the subsequent coupon code generation logic, enabling rapid iteration and application of personalized strategies.

[0054] For step S4, in this embodiment, after constructing the objective function, it is necessary to solve the objective function to derive the optimal strategy path that can be used to guide coupon generation. The core problem addressed in this step is: how to minimize the overall objective function within the adjustable space of the strategy control variables to obtain a coupon strategy combination scheme at the customer level. This process needs to be based on a rigorous mathematical solution framework and combined with actual business constraints and time windows.

[0055] In this embodiment, to obtain the optimal coupon strategy path, a numerical optimization method is introduced to solve the above optimization function. Specifically, the objective function to be optimized is as follows: ; in, The policy evaluation function value is the objective function that needs to be minimized. This represents the total cost of issuing coupons, including face value, issuance cost, and distribution cost. The effectiveness of coupon marketing is typically measured by a weighted average of response rate and customer activity. It represents the time efficiency of coupons and measures the time bandwidth brought about by coupon usage; , , The adjustment coefficient is used to control the relative weights among the above three items, so as to achieve balanced scheduling among multiple objectives.

[0056] In one possible implementation, the optimization method employs a genetic algorithm to perform a global search of the policy control variable space. By setting the initial population, constructing the fitness function, and introducing crossover and mutation operators, the algorithm can evolve generation by generation, approximating the global optimum. Fitness function and objective function. The mapping relationship is clear: the smaller the function value, the higher the fitness value.

[0057] In other embodiments, a particle swarm optimization algorithm is employed as the search mechanism. This algorithm updates the search direction dynamically based on the global and individual optimal states through iterative flight of individual particles in the policy space, thereby achieving rapid convergence to the optimal policy path. This algorithm exhibits good convergence stability and is suitable for finding optimal paths in high-dimensional variable spaces.

[0058] Specifically, each strategy path includes the following adjustable variables: coupon type, coupon value, coupon validity period, and applicable conditions. During optimization, these variables participate in function computation as components of the decision vector. The variable space can be configured with boundary values ​​and step sizes, forming an iterable set of strategy combinations.

[0059] In some specific implementations, the optimization process is limited to a preset time interval. The process is executed within a specific time window, with each round of solution defined within that window to ensure efficient use of computing resources and consistency with real-time data changes. This time window can be dynamically adjusted through the business model, such as triggering updates daily or hourly.

[0060] Generally, the numerical optimization methods used are replaceable and can be adjusted to heuristic search methods or policy generation methods based on reinforcement learning, depending on the latency tolerance of the computing platform and the business side. However, in the specific implementation of this invention, genetic algorithms and particle swarm optimization algorithms are used as the main computing framework.

[0061] To further enhance strategy robustness, a historical decision penalty mechanism can be introduced during the optimization process to prevent frequent strategy oscillations from degrading customer experience. Adding a volatility penalty factor to the objective function can suppress drastic strategy jumps.

[0062] In some practical applications, to ensure the optimization function has good differentiability and convergence, regularization terms can be introduced to handle some nonlinear terms, ensuring that the overall function is continuously differentiable within its domain. This process is usually adjusted in conjunction with the business logic of the strategy variables.

[0063] For step S5, in this embodiment, after the optimal coupon strategy path is derived, the system needs to generate corresponding coupon information based on the specific parameters of the path and bind it with the customer's unique identifier to achieve precise coupon distribution. This step is a key link in the strategy implementation stage, and its role is to transform the calculation results into an operable coupon structure and ensure that the coupon information is verifiable and unique.

[0064] In this embodiment, the generated coupon information includes, but is not limited to, coupon type, face value, validity period, usage restrictions, unique coupon number, and matching customer ID. The values ​​assigned to each field must strictly follow the parameter combinations output by the strategy path.

[0065] Generally, the card / coupon structure will be written to the system database in the form of data records, which includes the following fields: Coupons, coupon types (such as discount coupons, coupons with minimum spending requirement), face value, minimum spending amount, validity period, customer ID, usage status ID, creation timestamp, etc.

[0066] The “usage status identifier” is used to track whether the coupon has been activated, consumed, or expired, which facilitates subsequent management and statistical analysis.

[0067] In one possible implementation, the conditions for using the coupon are defined through the following setting function: ; in, This indicates the validity period of the coupon, which can be in days, hours, or minutes, depending on the business scenario. This indicates the minimum spending amount required for a voucher, representing the minimum spending threshold a customer must reach in a single transaction. The expected response benefit value for customers to coupon types; This indicates the actual conditions under which the customer can use the coupon, and is the criterion function used by the system to determine whether the coupon is usable during verification.

[0068] Specifically, when When the card is valid, it indicates that the card can still be used under the current conditions; otherwise, if If the card or coupon no longer meets the usage requirements, the system can refuse the usage request or notify the customer of its invalidation status.

[0069] In some embodiments, the coupon binding process requires matching using a Customer Unique Identifier (CID). This identifier can be the customer's registered mobile phone number, user ID, or anonymous encrypted identification number. Simultaneously with coupon generation, the system registers the generated coupon structure under the customer's account name in a mapping manner and records it synchronously in the marketing database.

[0070] As an alternative, to enhance system performance and security, the coupon structure can be encrypted and stored using hash encoding to ensure data integrity and prevent forgery. Using a multi-table structure in the database to maintain customer information, coupon information, and policy mapping tables separately facilitates subsequent hierarchical management and efficient querying within the system.

[0071] In another technical implementation, the coupon generation module simultaneously writes to a log module to record coupon distribution behavior for tracking purposes. This log information includes the strategy version number, coupon generation time, strategy ID, and generation source (such as system batch or personalized distribution), forming a complete distribution traceability system.

[0072] In the specific generation process, unique coupon code generation rules can also be set. For example, based on the combination of timestamp, customer ID and strategy path code hash, coupon code strings of varying lengths can be constructed to ensure that each coupon has unique identifiability.

[0073] Under normal circumstances, once the coupon information is generated, it will be immediately written into the main business system and the status initialization will be completed. After that, the system's scheduled task distribution module will connect with the subsequent distribution process to ensure the effective transmission and management of coupons throughout their lifecycle.

[0074] In some practical applications, coupons can also be set with a "cooling-off period" parameter to prevent customers from repeatedly claiming similar coupons within a short period of time. This parameter can be derived by modeling the frequency of customers' past claims and dynamically setting the cooling-off threshold.

[0075] The card and coupon dynamic generation system described below and the card and coupon dynamic generation method described above can be referred to as correspondingly.

[0076] Please see the appendix Figure 2 The present invention also provides a dynamic coupon generation system, comprising: The data source module is used to collect historical customer behavior data and sales data; The data cleaning module is used to clean, standardize, and fill in missing values ​​for the collected data. The feature engineering module is used to extract customer features and build feature models; The strategy engine module is used to generate coupon strategies based on customer characteristics and business rules; The optimization module is used to generate the optimal coupon strategy based on the optimization objective. The coupon code generation module is used to generate unique coupon codes; The distribution module is used to accurately distribute the generated coupons to the target customers.

[0077] The data source module integrates with the enterprise's internal CRM system, online shopping platform, and other external data interfaces to acquire multi-dimensional data such as customer purchase history, browsing behavior, and interaction information. Simultaneously, the system can interact with external data sources, such as third-party social media platform data and geolocation data, to comprehensively understand customer behavior and preferences. Efficient data acquisition interfaces ensure the real-time nature and accuracy of the data sources.

[0078] The data cleaning module primarily improves data quality by removing invalid data, filling in missing data, and standardizing data formats. During the cleaning process, methods such as deduplication, outlier identification, and correction are employed to ensure the input raw data has good quality and consistency. Standardization operations, such as normalization, enable data from different sources and formats to have a unified standard form, thereby supporting subsequent data analysis and model training.

[0079] For the feature engineering module, in-depth analysis of customer historical behavior and personal data extracts features that effectively represent customer preferences, such as purchase frequency, product preferences, spending power, and activity participation. Through data mining techniques, the feature engineering module can build more refined customer profiles and map these features to customer behavioral patterns, providing high-quality input features for strategy generation.

[0080] The strategy engine module automatically generates personalized coupon strategies based on specific business rules and customer characteristics. This module dynamically adjusts parameters such as coupon value, validity period, and applicable conditions, taking into account business needs like promotional goals, sales cycles, and inventory levels, thereby achieving precise marketing. Furthermore, the strategy engine supports adaptive optimization based on machine learning algorithms, continuously improving the accuracy and effectiveness of the strategies.

[0081] For the optimization module, numerical optimization methods (such as genetic algorithms and particle swarm optimization algorithms) are used to optimize the coupon distribution strategy to achieve the best coupon distribution effect. Optimization objectives include, but are not limited to, maximizing customer engagement, improving customer loyalty, and minimizing coupon costs. Through iterative calculations and strategy path adjustments, the optimization module can find the optimal balance among various business objectives and generate the most effective coupon strategy.

[0082] For the coupon code generation module, a unique encoding algorithm is used to generate a unique identifier for each coupon, ensuring that each coupon can be uniquely tracked and verified. The coupon code generation module supports multiple encoding methods, such as random number generation, timestamp combined with customer ID generation, and hash algorithms, ensuring the immutability and uniqueness of the coupons. At the same time, the system can generate multiple different coupon codes for each customer to facilitate multi-scenario application and management.

[0083] For the distribution module, by integrating with the customer management system, personalized coupons are accurately pushed to target customers based on customer profiles and strategy recommendations. Distribution methods include, but are not limited to, SMS, email, app push notifications, and social media sharing. The distribution module can also optimize the timing and frequency of coupon pushes based on customer preferences and receiving behavior, ensuring that customers receive coupons at the optimal time, thereby improving coupon usage and conversion rates.

[0084] The system in this embodiment can be used to execute the above method embodiments, and its principle and technical effect are similar, so they will not be described again here.

[0085] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for dynamically generating coupons, characterized in that, Includes the following steps: Acquire customers' historical behavior data, construct a customer state vector model, and initialize customer state information; Define card and coupon policy control variables based on customer status information, including card and coupon type, face value, usage conditions and validity period; Based on the relationship between customer status and strategy control variables, an optimization objective function for coupon distribution costs and customer response benefits is constructed. The optimal strategy is derived based on the objective function. The objective function is minimized within a predetermined time interval using numerical optimization methods, and the optimal coupon strategy path is obtained by solving the problem. Generate coupon information based on the optimal coupon strategy path, and bind the generated coupon structure with the customer identifier.

2. The method for dynamically generating coupons according to claim 1, characterized in that, The steps for obtaining customer historical behavior data include: Real-time collection of customer order data, including order type, amount, and customer location information; Acquire customer behavior data, including browsing history, consultation history, and visit frequency; External information is obtained from multiple data sources, including industry data, market research reports, and government policies; Machine learning algorithms are used to model historical data and construct a customer state vector model.

3. The method for dynamically generating coupons according to claim 1, characterized in that, The steps for defining card and coupon strategy control variables based on customer status information include: Define coupon types, including discount coupons, spending-reduction coupons, service coupons, and exchange coupons; Determine the face value of the coupons and adjust them dynamically according to the needs of the customer base; The validity period and applicable scope of the coupons can be set according to different usage conditions; The optimal strategy control variables for coupons are determined based on customer profiles and historical data, and adjustments are made in real time.

4. The method for dynamically generating coupons according to claim 1, characterized in that, The steps for constructing the objective function to optimize coupon distribution costs and customer response benefits include: Define a coupon issuance cost function, including the coupon's face value, issuance cost, and distribution cost; Define a customer response benefit function, including customer response probability, satisfaction, and conversion rate; Construct an optimization objective function, and adjust the coefficients in the objective function based on the relationship between customer status information and policy variables.

5. The method for dynamically generating coupons according to claim 1, characterized in that, The steps for deriving the optimal strategy based on the objective function include: Numerical optimization methods, including genetic algorithms and particle swarm optimization, are used to minimize the optimization objective function; The optimal coupon strategy path is derived based on the objective function, including coupon type, face value, and validity period parameters; Within a predetermined time interval, the optimal strategy is adjusted based on real-time data through an iterative optimization method. Choose the most suitable coupon scheme to ensure efficient marketing results.

6. The method for dynamically generating coupons according to claim 1, characterized in that, The step of generating coupon information based on the optimal coupon strategy path includes: The generated coupon information is linked to the customer's identifier to ensure that each customer has a unique coupon; Set the conditions for issuing coupons based on the optimal strategy, including time and amount limits for use; The generated coupon information is stored in the database.

7. The method for dynamically generating coupons according to claim 4, characterized in that, The formula for calculating the coupon distribution cost function is as follows: ； in, Costs associated with issuing coupons; The face value of the coupon; Costs associated with issuing coupons; Distribution costs for coupons; An identifier for the type of coupon or its distribution strategy; The formula for calculating the customer response benefit function is as follows: ； in, The expected response benefit value for customers to coupon types; The probability of a customer's response; For customer satisfaction; For customer conversion rate.

8. The method for dynamically generating coupons according to claim 5, characterized in that, The optimization formula for the optimal coupon strategy path is: ； in, This represents the optimization value for optimizing the coupon strategy path; Costs associated with issuing coupons; To improve the marketing effectiveness of coupons; For the time efficiency of coupons; , , To adjust the coefficients and control the balance between distribution costs, marketing effectiveness, and time efficiency.

9. The method for dynamically generating coupons according to claim 6, characterized in that, The formula for setting the conditions for issuing the coupons is: ； in, Indicates the validity period of the coupon; Indicates the minimum amount that can be used with the coupon; The expected response benefit value for customers to coupon types; This indicates the conditions under which the customer can currently use the coupons.

10. A card / coupon dynamic generation system, applied to the card / coupon dynamic generation method according to any one of claims 1-9, characterized in that, include: The data source module is used to collect historical customer behavior data and sales data; The data cleaning module is used to clean, standardize, and fill in missing values ​​for the collected data. The feature engineering module is used to extract customer features and build feature models; The strategy engine module is used to generate coupon strategies based on customer characteristics and business rules; The optimization module is used to generate the optimal coupon strategy based on the optimization objective. The coupon code generation module is used to generate unique coupon codes; The distribution module is used to accurately distribute the generated coupons to the target customers.

Images