Product promotion system and method based on supply chain resource data management

Abstract

This invention discloses a product promotion system and method based on supply chain resource data management. This invention relates to the field of supply chain resource data management technology and solves the technical problems of scattered underlying supply chain resource data, lack of quantifiable traceability of product advantages, single and inaccurate customer matching dimensions, and homogeneous and poorly adaptable promotional content. This invention constructs a structured database by splitting the product into four layers, integrating scattered raw material and production process underlying supply chain data. Then, combining Pearson correlation coefficient, entropy weight method, and control variable algorithm, it quantifies the correlation degree, feature contribution degree, and comprehensive advantage score, transforming implicit supply chain resource advantages into explicit, traceable, and quantifiable product advantages, completely solving the problems of vague advantages and lack of data support in traditional promotion.

Description

Technical Field

[0001] This invention relates to the field of supply chain resource data management technology, specifically to a product promotion system and method based on supply chain resource data management. Background Technology

[0002] In today's increasingly competitive business environment, the core pain point of product promotion has shifted from "information reach" to "precise matching." Enterprises often possess abundant supply chain resources (such as raw materials, production processes, and component characteristics), but the data on these underlying resources is scattered and lacks systematic integration, making it impossible to deeply explore and quantify product advantages. At the same time, traditional promotion models rely heavily on experience-based judgment, failing to establish a precise mapping relationship between product advantages and customer needs, and failing to combine underlying supply chain parameters to form a differentiated promotion logic. Ultimately, this results in problems such as low promotion efficiency, insufficient customer conversion rates, and wasted resources. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a product promotion system and method based on supply chain resource data management, which solves the problems of scattered underlying supply chain resource data, lack of quantitative traceability of product advantages, single and inaccurate customer matching dimensions, and homogeneous and poor adaptability of promotion content.

[0004] To achieve the above objectives, the present invention provides a product promotion method based on supply chain resource data management, comprising the following steps: Step 1: Obtain a product P0 and break it down into four levels: assembly, parts, raw materials, and manufacturing process. Based on the hierarchical relationships, parameter sets, and complete bottom-up mapping paths between the levels of product P0, construct a mapping database. Step 2: Based on the raw material parameters and production process parameters of each component, the Pearson correlation coefficient is used to calculate the parameter correlation degree, and the feature contribution degree is calculated by the control variable method and the information entropy method. The advantage level of the component is comprehensively evaluated, and the raw material parameters and production process parameters that simultaneously meet the requirements of high comprehensive weight ranking, contribution exceeding the set threshold, and correlation reaching the strong dependence level are selected according to the correlation degree, contribution, and parameter weight to form the component advantage feature vector. Step 3: Construct a product advantage vector based on advantageous components, construct a customer demand vector, and screen customers using similarity matching. Combine geographical location, demand volume, and pre-order price to segment and prioritize customers to determine the target customers for promotion. Step 4: Use AI to generate customized promotional content based on customer type and deliver it accurately and follow up.

[0005] As a further aspect of the present invention: the product hierarchical decomposition specifically includes: splitting product P0 into multiple product assemblies A, each product assembly A into multiple parts B, each part B into corresponding raw materials C and production processes D, and constructing a database based on the hierarchical mapping relationship, assigning a unique code, name and technical parameter attributes to each hierarchical object, wherein the database is named with the unique ID of product P0, and stores the association information of assemblies, parts, raw materials and production processes in layers.

[0006] As a further aspect of the present invention: the parameter correlation calculation adopts the Pearson correlation coefficient, specifically including: for all raw material C and production process D parameters under each component B, the original value of the correlation degree of each parameter is calculated by the control variable method and mapped to the [0,1] interval. According to the correlation degree value, it is divided into four levels: complete dependence, strong dependence, weak dependence and no dependence, and assigned corresponding weights. Then, the total correlation degree R of component B is calculated by weighted average. B .

[0007] As a further aspect of the present invention: the evaluation of the component advantage level includes: based on the total correlation degree R B and feature contribution C total B Through the weighted formula S B =R B ×0.4+Ctotal B Calculate the overall advantage score S by multiplying by 0.6. B And according to S B The scores categorize components into three levels: core advantages, potential advantages, or no advantages, providing a basis for subsequent quantitative analysis of features.

[0008] As a further aspect of the present invention: the parameter screening includes: for each component B, screening for parameters that satisfy W C or W D The top 1-2 weighted parameters, with a single parameter contribution of ≥0.15 (C parameter) or ≥0.1 (D parameter), and a correlation level M>0.7 for raw material C and production process D parameters, can retain a maximum of 2 C parameters and 2 D parameters for each B parameter; The component advantage vector includes: B-code, S-code, etc. B Scores, key C parameters, key D parameters, and core feature descriptions are entered into the database in the format of B code, parameter type, data name, and advantage indicator.

[0009] As a further aspect of the present invention: the construction of the product advantage vector includes: constructing a B advantage vector with core advantage B and potential advantage B as independent units; taking product assembly A as a unit, aggregating the advantage vectors of all its subordinate B to construct an A-level advantage vector, and calculating the score SA of product assembly A, the calculation formula of which is: S A=Core B's S B Score × 0.7 + Potential B's S B Score × 0.3; Select product assembly A that meets SA≥0.7 and whose key performance indicators exceed those of competing products by ≥10%, and construct the core advantage vector of product P0; The dimensions of the advantage vector at each level include the corresponding level's encoding, score, key parameters, and core feature description.

[0010] As a further aspect of the present invention: the customer matching adopts a cosine similarity algorithm, specifically including: constructing a customer demand vector whose dimension is aligned with the core advantage vector of product P0; after calculating the similarity, only retaining highly matched customers with a similarity ≥ 0.8 and moderately matched customers with a similarity ≤ 0.6 and < 0.8, and labeling the customer ID and core demand; The customer segmentation is based on three dimensions: geographic location, demand volume, and pre-order price. Matched customers are divided into six core segment types, including high-end scale type, high-end precision type, mid-range scale cost-effective type, mid-range balanced type, mid-range basic fit type, and potential growth type. Each type has a clear corresponding three-dimensional characteristics, core demands, and value orientation.

[0011] As a further aspect of the present invention: the AI ​​personalized promotion in step four is implemented as follows: AI-powered promotion generates customized promotional content based on the product's P0 core advantage vector, A and B advantage vector data, the customer's three-dimensional information, segmentation types, core demand data, and suitable promotional scenarios, according to the customer segmentation type. After distribution through the corresponding channels, the system tracks customer clicks, dwell time, and inquiry interaction data in real time, triggers follow-up reminders, and pushes relevant customer information to sales personnel. It also generates standardized and personalized response scripts for common inquiry questions.

[0012] A product promotion system based on supply chain resource data management includes: Data acquisition and database construction module: used to split product P0 into four layers, collect information such as codes, names, core functions, and key parameters of each layer, and build and store database documents of ABC / D mapping relationship; B. Advantage Analysis Module: Used to calculate the correlation degree R of D. B Feature contribution C total B Comprehensive advantage score S B Determine the level of advantage of B, screen key C / D parameters and summarize core features; Advantage Vector and Customer Matching Module: Used to construct BA-P0 hierarchical advantage vectors and customer demand vectors, calculate vector similarity to filter high / medium matching customers, supplement customer information and segment in three dimensions, and select Q optimal customers through weighted scoring; AI Promotion Module: Used to input product-side, customer-side, and promotion scenario data, generate customized promotion content, adapt to distribution channels, track customer interaction data and intelligently follow up, and generate response scripts that combine standardization and personalization.

[0013] This invention provides a product promotion system and method based on supply chain resource data management. Compared with existing technologies, it has the following advantages: (1) This invention constructs a structured database by splitting the product into four layers, integrates the scattered raw material and production process supply chain bottom data, and then combines Pearson correlation coefficient, entropy weight method and control variable algorithm to quantify the correlation degree, feature contribution degree and comprehensive advantage score, transforming the implicit supply chain resource advantages into explicit, traceable and quantifiable product advantages, and completely solving the problems of vague advantages and lack of data support in traditional promotion. (2) This invention establishes a precise matching logic between advantages and needs by matching the similarity between hierarchical advantage vectors and customer demand vectors, and by combining the three dimensions of geographical location, demand volume and pre-order price. This allows promotional resources to focus on high-matching, high-value customers, thus avoiding the waste of resources in traditional experience-based promotion. (3) Based on the key advantages of the underlying C / D parameters of the supply chain, this invention constructs a hierarchical advantage system of components, product assemblies and products, and adapts differentiated A / B advantage combinations and promotional content for different customer types, forming a unique promotional logic that is different from competitors, so that customers can clearly perceive that the product advantages come from the hard power of the supply chain, and enhance brand recognition and willingness to cooperate. Attached Figure Description

[0014] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0015] The technical solutions of 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. Example

[0016] Please see Figure 1 This application provides a product promotion method based on supply chain resource data management, which specifically includes the following steps: Step 1: Decompose product P0 into four levels: assembly, parts, raw materials, and manufacturing process. Construct a mapping database based on the hierarchical relationships, parameter sets, and complete bottom-up mapping paths between the levels of product P0. Step 2: Based on the raw material parameters and production process parameters of each component, the Pearson correlation coefficient is used to calculate the parameter correlation degree, and the feature contribution degree is calculated by the control variable method and the information entropy method. The advantage level of the component is comprehensively evaluated, and the raw material parameters and production process parameters that simultaneously meet the requirements of high comprehensive weight ranking, contribution exceeding the set threshold, and correlation reaching the strong dependence level are selected according to the correlation degree, contribution, and parameter weight to form the component advantage feature vector. Step 3: Construct a product advantage vector based on advantageous components, construct a customer demand vector, and screen customers using similarity matching. Combine geographical location, demand volume, and pre-order price to segment and prioritize customers to determine the target customers for promotion. Step 4: Use AI to generate customized promotional content based on customer type and deliver it accurately and follow up. Example

[0017] Please see Figure 1 As a second embodiment of this application, this embodiment is implemented based on the first embodiment, and the method provided in this embodiment includes the following steps: Step 1: S11: First, obtain a product P0. Then, break down product P0 into four levels: product assembly A, components B, raw materials C, and manufacturing process D. Construct a mapping relationship between A, B, C, and D, specifically: P0 is broken down into multiple product assemblies, denoted as: A1, A2, A3, ..., Aa, where a is the nth product assembly; Each product assembly A is broken down into corresponding components B. Taking A1 as an example, the components B are denoted as: B1, B2, B3, ..., Bb, where b is the nth component. Each component B is broken down into its corresponding raw material C. Taking B11 as an example, the raw material C is denoted as C1, C2, and Cc, where c is the nth raw material. Each component B is broken down into its corresponding production process D. Taking B11 as an example, the raw materials C are denoted as D1, D2, and Dd, where d is the nth production process.

[0018] S12: Based on the four-layer structure of product disassembly derived from P0, a database document is constructed according to its mapping relationship, specifically as follows: Assign a unique ID to product P0 as the name of this database document. With product assembly A as the top level, enter the product assembly's code / name, core functions, and a set of subordinate parts B. Component B is the middle layer, which enters its own code / name, its own core features, and the code of the所属产品总成A (Product Assembly A), for example, B1 belongs to the set A1; Raw material C is the bottom layer, which enters its own code / name, the key technical parameters of raw material C, and the code of the所属零部件B (Component B) it belongs to. For example, C1 belongs to the set B1; Generating process D is the bottom layer, which enters its own code / name, the key technical parameters of process D, and the code of the所属零部件B (Component B) it belongs to. For example, D1 belongs to the set B1.

[0019] Step Two S21: Extract the raw material C parameters and production process D parameters corresponding to all components B under each assembly A from the database document. Based on the raw material C parameters and production process D parameters at the lower level of each component B, calculate the original value m of the single parameter correlation degree of each C and D in the B set by using the Pearson correlation coefficient, and map it to the range [0, 1] after taking the absolute value to obtain the single parameter correlation degree M; S22: Divide into four levels according to the value of M, and simultaneously assign corresponding weights. Specifically: Full dependence: When 0.9 < M ≤ 1, the weight is 1; Strong dependence: When 0.7 < M ≤ 0.9, the weight is 0.8; Weak dependence: When 0.3 < M ≤ 0.7, the weight is 0.5; No dependence: When M ≤ 0.3, the weight is 0.

[0020] S23: For all C or D parameters of each B, multiply the single parameter correlation degree of each parameter by the weight of the corresponding level of the parameter to obtain the weighted correlation degree of each parameter. Then sum up the weighted correlation degrees of all C or D parameters, and finally divide the summation result by the total number of all C or D parameters corresponding to this B to obtain the total correlation degree R B ; For example, B1 includes three C parameters, and their single parameter correlation degrees are 0.8, 0.7, and 0.9 respectively, and the corresponding weights are 0.8, 0.5, and 0.8 respectively. Then R B1 = (0.8×0.8 + 0.7×0.5 + 0.8×0.9) / 3 = 0.441.

[0021] S24: Calculate the feature contribution degree C of each B 总B , specifically: S241: For all C or D parameters corresponding to each B, when testing the improvement amplitude of a single C parameter on the features of B, keep all D parameters unchanged and observe the impact brought by the change of the C parameter adjustment. For example, the D parameters in B1 are fixed. When the luminous efficiency of C1 rises from 80 lm / w to 90 lm / w, the brightness of B11 increases by 20%. Then the contribution degree C of the single parameter C1 ciThat is, 20%, with a contribution value of 0.2. Similarly, when testing the improvement of a single D parameter on the B feature, all C parameters are kept unchanged, and the impact of changes in the D parameter is observed.

[0022] S242: Based on the improvement data of all C or D parameters corresponding to B, through standardization, information entropy and information utility value, calculate the proportion of information utility value between the two groups C and D within the parameter group, objectively allocate weights, and finally obtain the comprehensive weight W of parameter group C. C Combined with the D parameter group weight W D (W) C+ W D =1).

[0023] S243: C is obtained by averaging the contributions of all C parameters. 均 Specifically, this involves summing the contribution values ​​C of all parameters related to C and dividing by the number of C parameters. The average of the contributions of all D parameters is used to obtain D. 均 Specifically, this involves summing the contribution values ​​D of all parameters related to D and dividing by the number of parameters D. C is derived 均 With D 均 Substitute the numerical values ​​into the corresponding weights W C与 W D得出 C 总B The specific formula is as follows: C 总B= W C ×C 均 +W D ×D 均 For example, the average contribution value of C1, C2, and C3 in set B1 is 0.6, corresponding to a weight W. C The average contribution of D1, D2, and D3 in set B1 is 0.6, corresponding to a weight W. D If it is 0.4, then C 总B1 =0.6×0.6+0.5×0.4=0.56.

[0024] S25: Obtain the total correlation degree R of all components B based on S23. B Contribution C of feature B of each component in S24 总B， The advantage score of each component B is comprehensively evaluated based on both the relevance and feature contribution dimensions. The specific formula is as follows: Overall advantage score S B= R B ×0.4+C 总B ×0.6, where 0.4 is the correlation weight (40%) and 0.6 is the feature contribution weight (60%).

[0025] S26: Combining comprehensive advantages score SB The calculation results are used to classify all components B into three categories: core advantages, potential advantages, and no advantages, providing a basis for subsequent quantitative analysis of features. Specifically: when S B When S ≥ 0.6, it is considered a core computational advantage; when S ≤ 0.3, it is considered a core computational advantage. B When S < 0.6, it is considered a potential technological advantage. B When the value is less than 0.3, it is considered to have no advantage.

[0026] S27: For all Bs with established dominance levels, combine with W in step S242. C With W D The contribution of individual parameters in step S241 and the correlation level in step S22 are used to screen the key parameters that have the most significant impact on the performance of B and to calibrate them. Specifically: Selecting C from the same set B requires satisfying: W C The parameters with the highest percentage (top 1-2), a single parameter contribution of ≥0.15, and a correlation level (M) >0.7 are selected, and each B parameter retains a maximum of 2 C parameters to ensure that the core support is not redundant. To select D from the same set B, it must satisfy: W D For parameters that rank in the top 1-2, have a single parameter contribution of ≥0.1, and a correlation level M>0.7, each B can retain a maximum of 2 D key processes; Enter the data into the database document in the format of B code - parameter type - parameter name - advantage index, for example, B1-key C1-luminous efficiency 90lm / w, B1-key D1-distillation temperature stability ±2°.

[0027] S28: Based on the advantages of key C and D parameters in step S31 and the core functions of B, proceed according to S... B The scoring classification summary extracts core features, specifically: Core Advantage B: Features are defined by specific performance metrics (i.e., key C and D parameters + specific indicators + industry comparisons), for example, B4 (S... B =0.68): Long life (C4 aging-resistant material, D4 sealing process, service life of 5000h, exceeding the industry average by 20%) decalibrated characteristics; Potential Advantage B: Define features based on specific performance (i.e., key C and D parameters + specific indicators + optimization directions), such as B4 (S B =0.53): Low power consumption (C4 low power material, power consumption reduced by 12%, optimized D4 power supply process can reduce it by another 3%).

[0028] Step 3 S31: Construct a B advantage vector with each core advantage B and potential advantage B as an independent unit. The vector dimensions include: B encoding, S...B Scoring, key C parameters, key D parameters, core feature description; Taking product assembly A as the unit, aggregate the advantage vectors of its subordinate B to construct A and its advantage vectors. The vector dimensions include: A code, S... A The scoring includes the core strength vector set of B, the potential strength vector set of B, a summary of A's core strengths, and a summary of A's supplementary strengths, where S... A The scoring formula is: S A =Core B's S B ×0.7+S supporting B B ×0.3; From all A-level advantage vectors, select those that satisfy: S A ≥0.7, exceeding competitors by ≥10%, construct the core advantage vector of product P0. The vector dimensions include: unique ID of P0, two sets of core A advantage vectors, summary of product core competitiveness, and key parameter indicator matrix.

[0029] S32: Based on the core advantage vector dimension of product P0, construct a demand vector for downstream customers through market research, customer questionnaires, and industry data collection. The vector dimension is aligned with the core advantage vector of P0 and includes: customer ID, customer level (high-end / mid-range / low-end), core demand type A, demand performance indicators, type of attention parameters, cost sensitivity, and procurement priority.

[0030] S331: Based on the core advantage vector of P0 and the construction demand vector, the cosine similarity algorithm is used to calculate the similarity between the customer demand vector and the product P0 vector. Only high-matching customers with similarity ≥ 0.8 and medium-matching customers with similarity ≤ 0.6 and < 0.8 are retained, and the customer ID and core demand are labeled. Low-matching customers with similarity < 0.6 are removed.

[0031] S332: Supplement the screening of high-match and low-match customers with three key pieces of information, specifically: In terms of geographical location: it is divided into two categories: core market areas and non-core market areas; In terms of demand: it is divided into three categories: large-scale, medium-scale, and small-scale. Regarding pre-order prices: they are divided into high-price range, mid-price range, and low-price range.

[0032] S34: Based on a combination of geographical location, demand, and pre-order price, highly matched and moderately matched customers are divided into 6 core sub-categories. Each sub-category has clearly defined three-dimensional characteristics, core demands, and value orientations, as detailed below: High-end large-scale customers: meet the needs of core regions, large scale, and high price range. Their core demands focus on building technological barriers, ensuring the stability of bulk supply, and providing localized service support in core regions. Their value orientation is mainly based on technological leadership and long-term large-scale cooperation. High-end, precision-oriented customers: those who meet the needs of core or non-core regions, medium-sized enterprises, and high-price ranges, whose core demands are the ultimate performance of core A / B advantages and customized parameter adaptation, and whose value orientation focuses on the precise matching of technology and their own needs. Mid-range scale cost-effective customers: meet the needs of core regions, large scale, and mid-price range. The core demand is controllable batch cost on the basis of meeting performance standards, while also paying attention to the supply timeliness in core regions. The value orientation is based on scale cost-effectiveness and stable cooperation. Mid-range balanced customers: These customers are located in non-core areas, are of medium size, and are in the mid-price range. Their core demand is a comprehensive match between core advantages and complementary advantages, while also requiring convenient after-sales service. Their value orientation is comprehensive and balanced value. Mid-range basic adaptation customers: meet the needs of non-core areas, small scale, and mid-price range. The core demand is that the basic functions of the core A advantage meet the standards and the basic B parameter support. At the same time, they pay attention to the convenience of the procurement process. The value orientation is basic adaptation and low-cost cooperation. Potential growth-oriented clients: These clients are located in core regions, are of medium or small scale, and are in the mid-price range. Their core demand is the long-term use value and potential optimization space of their core advantages. They hope to obtain long-term cooperation and empowerment, and their value orientation is growth empowerment and long-term binding.

[0033] S351: Based on the acquired high-match and medium-match customers and supplemented by three types of key information, establish a scoring system for high-match and medium-match customers, specifically as follows: Customer priority score = matching score × matching weight + demand volume score × demand weight + pre-order price fit score × pre-order fit weight + geolocation value score × geolocation value weight, where matching weight is 40%, high-match customers get 10 points, medium-match customers get 8 points; demand weight is 30%, large-scale customers get 10 points, medium-scale customers get 7 points, small-scale customers get 4 points; pre-order fit weight is 20%, high-price range gets 10 points, low-price range gets 8 points; geolocation value weight is 10%, core areas get 10 points, non-core areas get 6 points.

[0034] S352: By obtaining the scores of all customers, sorting them in descending order of customer priority scores, and combining them with the company's actual business capabilities, select the top Q customers (Q is the threshold, ranging from 15 to 30) to form the Q optimal customer selections.

[0035] Step Four S41: Based on the Q selected promotion customers, the system automatically determines the type of promotion customer in step S34, and combines the advantages of A / B and key C / D parameters to accurately match the focus of the push content, ensuring that the push content for each promotion customer is in line with the core needs and value orientation.

[0036] S42: Input core data into the AI ​​promotion system: specifically: Product-side data includes P0 core advantage vectors, adapted A / B advantage vectors, key C / D parameter indicators, and industry comparison data; The customer-side data consists of three dimensions of information, segmentation type, core demand keywords, and matching degree tags for the Q best customers; Promotional scenarios are determined based on customer preferences and include email communication, PPT presentations at industry exhibitions, and targeted WeChat push notifications.

[0037] S43: Based on input data, the AI ​​system generates customized promotional content according to customer segmentation types and selects suitable distribution channels based on customer preferences. The AI ​​system tracks customer interaction data with promotional content in real time, including click count, dwell time, and inquiry focus. When a customer interacts, it automatically triggers follow-up reminders, pushing customer segmentation types, core needs, and suitable advantages to sales personnel to guide precise communication. For common customer inquiries (such as bulk supply cycles, pre-order price discount policies, and parameter customization processes), the AI ​​generates standardized and personalized response scripts based on the optimal customer list information to ensure consistent and targeted communication. Example

[0038] Please see Figure 2 This paper proposes a product promotion system based on supply chain resource data management, used to implement product promotion methods based on supply chain resource data management as described above, including: The data acquisition and database construction module is used to break down the target product P0 into a four-layer structure: product assembly A, components B, raw materials C, and production process D. It collects the codes, names, core functions, and key technical parameters of each layer. It constructs and stores a database document with the ABC / D mapping relationship according to the hierarchical relationship of A as the top layer, B as the middle layer, and C / D as the bottom layer. The top layer A is associated with the set information of its subordinate B, the middle layer B is associated with the code information of its A, and the bottom layer C / D is associated with the code information of its B. The B advantage analysis module is used to extract the C and D parameters corresponding to each component B from the database documents, calculate the correlation degree of the parameters through relevant algorithms, classify them into levels and assign weights, and then solve the total correlation degree. The contribution of individual C and D parameters to feature B is tested using the controlled variable method. The combined weight of C and D parameter groups is then assigned using the entropy weight method to calculate the feature contribution. The comprehensive advantage score is obtained by combining the correlation degree and feature contribution degree, and three levels are defined: core advantage, potential advantage, and no advantage. Key C / D parameters are selected based on the weight ratio of parameter groups, parameter contribution threshold, and correlation degree level. They are entered into the database in the prescribed format, and the core features of B are summarized based on the key parameters. The Advantage Vector and Customer Matching Module is used to construct B-level advantage vectors, including B-codes, advantage scores, key C / D parameters, and core features, based on core / potential advantages (B). It aggregates subordinate B advantage vectors to construct A-level advantage vectors, including A-codes, advantage scores, a set of B advantage vectors, and a summary of A advantages. It also filters A-level advantages with core competitiveness to construct P0-level core advantage vectors, including unique P0 IDs and a set of core A vectors. Furthermore, it constructs customer demand vectors aligned with P0 advantage vectors through market research, using a similarity algorithm to filter highly matched and moderately matched customers. Finally, it supplements customer information with three categories: geographic location, demand volume, and pre-order price, and further segments them in three dimensions. Based on matching degree, demand volume, pre-order price fit, and geographic location value, it constructs a weighted scoring system, sorting customers in descending order of score and considering the company's business capabilities to select several optimal customers. The AI-powered promotion module is used to input the product's core advantage vectors, advantage vectors at all levels, key parameters, and industry comparison data; the optimal customer's three-dimensional information, segmentation types, core appeal keywords, and matching tags; and promotion scenario data determined based on customer preferences. It generates customized promotional content based on customer segmentation types, adapting to corresponding distribution channels such as email communication, industry exhibition displays, and targeted push notifications. It tracks customer interaction data with promotional content in real time, automatically triggering follow-up reminders when customers interact, and pushing relevant key customer information to sales personnel. For common customer inquiries, it generates standardized yet personalized response scripts based on the optimal customer list information.

[0039] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0040] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A product promotion method based on supply chain resource data management, characterized in that: Includes the following steps: Step 1: Obtain a product P0 and break it down into four levels: assembly, parts, raw materials, and manufacturing process. Based on the hierarchical relationships, parameter sets, and complete bottom-up mapping paths between the levels of product P0, construct a mapping database. Step 2: Based on the raw material parameters and production process parameters of each component, the Pearson correlation coefficient is used to calculate the parameter correlation degree, and the feature contribution degree is calculated by the control variable method and the information entropy method. The advantage level of the component is comprehensively evaluated, and the raw material parameters and production process parameters that simultaneously meet the requirements of high comprehensive weight ranking, contribution exceeding the set threshold, and correlation reaching the strong dependence level are selected according to the correlation degree, contribution, and parameter weight to form the component advantage feature vector. Step 3: Construct a product advantage vector based on advantageous components, construct a customer demand vector, and screen customers using similarity matching. Combine geographical location, demand volume, and pre-order price to segment and prioritize customers to determine the target customers for promotion. Step 4: Use AI to generate customized promotional content based on customer type and deliver it accurately and follow up.

2. The product promotion method based on supply chain resource data management according to claim 1, characterized in that, The product hierarchical decomposition specifically includes: splitting product P0 into multiple product assemblies A, each product assembly A into multiple parts B, each part B into corresponding raw materials C and production processes D, and constructing a database based on the hierarchical mapping relationship, assigning a unique code, name and technical parameter attributes to each hierarchical object, wherein the database is named with the unique ID of product P0, and stores the association information of assemblies, parts, raw materials and production processes in a hierarchical manner.

3. The product promotion method based on supply chain resource data management according to claim 1, characterized in that, The parameter correlation calculation uses the Pearson correlation coefficient, specifically including: for all raw material C and production process D parameters under each component B, the original value of the correlation of each parameter is calculated using the control variable method and mapped to the [0,1] interval. Based on the correlation value, it is divided into four levels: complete dependence, strong dependence, weak dependence, and no dependence, and assigned corresponding weights. Then, the total correlation degree R of component B is calculated by weighted average. B .

4. The product promotion method based on supply chain resource data management according to claim 1, characterized in that, The feature contribution calculation includes: testing the improvement effect of a single C parameter or D parameter on feature B of component by fixing other parameters, and obtaining the contribution of each parameter; based on the contribution data of the C parameter group and the D parameter group, assigning objective weights W through standardization processing and information entropy calculation. C and W D Finally, based on the average contribution of parameters C and D, C 均 With D 均 And the total feature contribution of component B, C, is calculated by weighting. B The formula is C_total B =W C ×C 均 +W D ×D 均 .

5. The product promotion method based on supply chain resource data management according to claim 1, characterized in that, The evaluation of component advantage levels includes: based on the total correlation coefficient R. B and feature contribution C total B Through the weighted formula S B =R B ×0.4+Ctotal B Calculate the overall advantage score S by multiplying by 0.

6. B And according to S B The scores categorize components into three levels: core advantages, potential advantages, or no advantages, providing a basis for subsequent quantitative analysis of features.

6. The product promotion method based on supply chain resource data management according to claim 1, characterized in that, The parameter filtering includes: for each component B, filtering parameters that satisfy W. C or W D The top 1-2 weighted parameters, with a single parameter contribution of ≥0.15 (C parameter) or ≥0.1 (D parameter), and a correlation level M>0.7, are raw material C and production process D parameters. Each B parameter can retain a maximum of 2 C parameters and 2 D parameters. The component advantage vector includes: B-code, S-code, etc. B Scores, key C parameters, key D parameters, and core feature descriptions are entered into the database in the format of B code, parameter type, data name, and advantage indicator.

7. The product promotion method based on supply chain resource data management according to claim 1, characterized in that, The construction of the product advantage vector includes: constructing a B advantage vector with core advantage B and potential advantage B as independent units; taking product assembly A as a unit, aggregating the advantage vectors of all its subordinate B units to construct an A-level advantage vector, and calculating the score SA of product assembly A, the calculation formula of which is: S A =Core B's S B Rating × 0.7 + Potential B's S B Score × 0.3; Select product assembly A that meets SA≥0.7 and whose key performance indicators exceed those of competing products by ≥10%, and construct the core advantage vector of product P0; The dimensions of the advantage vector at each level include the corresponding level's encoding, score, key parameters, and core feature description.

8. The product promotion method based on supply chain resource data management according to claim 1, characterized in that, The customer matching uses a cosine similarity algorithm, which specifically includes: constructing a customer demand vector whose dimensions are aligned with the core advantage vector of product P0; calculating the similarity and retaining only highly matched customers with a similarity ≥ 0.8 and moderately matched customers with a similarity ≤ 0.6 and < 0.8; and labeling the customer ID and core demand. The customer segmentation is based on three dimensions: geographic location, demand volume, and pre-order price. Matched customers are divided into six core segment types, including high-end scale type, high-end precision type, mid-range scale cost-effective type, mid-range balanced type, mid-range basic fit type, and potential growth type. Each type has a clear corresponding three-dimensional characteristics, core demands, and value orientation.

9. The product promotion method based on supply chain resource data management according to claim 1, characterized in that, The implementation method of AI personalized promotion in step four is as follows: AI-powered promotion generates customized promotional content based on the product's P0 core advantage vector, A and B advantage vector data, the customer's three-dimensional information, segmentation types, core demand data, and suitable promotional scenarios, according to the customer segmentation type. After distribution through the corresponding channels, the system tracks customer clicks, dwell time, and inquiry interaction data in real time, triggers follow-up reminders, and pushes relevant customer information to sales personnel. It also generates standardized and personalized response scripts for common inquiry questions.

10. A product promotion system based on supply chain resource data management, capable of executing the product promotion method based on supply chain resource data management as described in any one of claims 1-9, characterized in that, include: Data acquisition and database construction module: used to split product P0 into four layers, collect the code, name, core function and key parameter information of each layer, and build and store the database document of ABC / D mapping relationship; B. Advantage Analysis Module: Used to calculate the correlation degree R of D. B Feature contribution C total B Comprehensive advantage score S B Determine the level of advantage of B, screen key C / D parameters, and summarize core features; Advantage Vector and Customer Matching Module: Used to construct BA-P0 hierarchical advantage vectors and customer demand vectors, calculate vector similarity to filter high / medium matching customers, supplement customer information and segment in three dimensions, and select Q optimal customers through weighted scoring; AI Promotion Module: Used to input product-side, customer-side, and promotion scenario data, generate customized promotion content, adapt to distribution channels, track customer interaction data and intelligently follow up, and generate response scripts that combine standardization and personalization.

Images