Customer return visit management method and system for e-commerce platform

Abstract

The invention provides a customer return visit management method and system for an e-commerce platform, and belongs to the technical field of data processing, and the method specifically comprises the steps: taking the constitution data of the on-shelf commodities of different exchange visit optimization demand types as the basis, and determining the manual return visit of the commodities when the manual return visit of the commodities is needed; determining manual return visit commodities in the shelf commodities according to return visit optimization demand types of the shelf commodities and customer association conditions between the shelf commodities and other shelf commodities, and determining change conditions of customer feedback data according to return visit processing results of the manual return visit commodities; according to the return visit management method for the shelf commodities except the manual return visit commodities, the return visit management method for the shelf commodities except the manual return visit commodities is determined based on the manual return visit commodities associated with the shelf commodities and the change condition of the client feedback data, and the processing efficiency of manual return visit is improved.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, and in particular relates to a customer follow-up management method and system for e-commerce platforms. Background Technology

[0002] Regular customer follow-ups can greatly improve customer satisfaction and loyalty. Existing technical solutions often involve regular customer follow-ups to understand the problems customers encounter during product use and to make targeted product improvements. However, the above technical solutions have the following technical problems: When conducting customer follow-up, existing technical solutions often determine the proportion of customers to be followed up using AI or manual methods based on a fixed ratio. However, they neglect the correlation between customers to be followed up for different listed products. Specifically, if there is a high degree of overlap between customers to be followed up for other listed products, the proportion of manual follow-up can be adjusted to identify changes in customer demand for multiple products, and then the script can be adjusted during the AI ​​follow-up process. Therefore, determining the follow-up processing plan based on the correlation between customers to be followed up for other listed products has become an urgent technical problem to be solved.

[0003] Therefore, there is an urgent need for a customer follow-up management method and system for e-commerce platforms. Summary of the Invention

[0004] To achieve the objectives of this invention, the following technical solution is adopted: Specifically, this application provides a customer follow-up management method for e-commerce platforms, which includes: S1 uses the listing data of products on the e-commerce platform to determine the products listed on the e-commerce platform, and determines the type of follow-up optimization needs for the products based on the changes in the sales data of the listed products. Based on the composition data of the listed products with different mutual visit optimization needs, S2 determines the products that need to be manually revisited when determining the products that need to be manually revisited, by considering the revisit optimization needs of the listed products and the customer relationship between the listed products and other listed products. S3 determines the changes in customer feedback data based on the results of the manual follow-up of the products. Based on the changes in customer feedback data of the manually followed products that are related to the listed products, S3 determines the follow-up management method for the listed products other than the manually followed products.

[0005] The beneficial effects of this invention are as follows: Based on the types of follow-up optimization needs for listed products and their customer relationships with other listed products, the products for which manual follow-up is required are determined. This approach considers the differences in manual follow-up needs due to sales fluctuations within each listed product, and further combines this with the customer relationships with other listed products. This ensures that during manual follow-up, the needs of customers for other listed products can be accurately identified to the greatest extent possible, thereby improving the comprehensiveness of follow-up processing and the matching of user needs.

[0006] Based on changes in manually followed-up products and customer feedback data related to the listed products, a follow-up management method for listed products excluding those manually followed up is determined. This method considers the differences in the number of manually followed-up products related to the listed products, which leads to differences in the reliability of identification and processing when customer feedback issues change. Furthermore, by combining the changes in customer feedback issues, the follow-up management method is determined from the perspectives of the reliability of identification and processing and the risk of abnormal changes in feedback issues, thereby further improving the reliability of follow-up management.

[0007] Furthermore, the products listed on the e-commerce platform are those listed on the e-commerce platform.

[0008] Furthermore, the changes in the sales data of the listed products are determined based on the changes in the sales volume of the listed products between different time periods.

[0009] Furthermore, the method for determining the type of return visit optimization requirement for the listed products is as follows: Based on the unit time period, the time is divided into multiple time intervals. Based on the sales data of the products listed in different time intervals, a benchmark value for the sales volume in the time interval is determined. Based on the sales data of products listed in different time intervals, the time intervals in which sales are lower than the benchmark value are determined and regarded as poor sales intervals. Based on the data of the poor sales period, determine the type of follow-up optimization requirement for the listed products.

[0010] Furthermore, the method for determining the follow-up management method for listed products, excluding those requiring manual follow-up, is as follows: Products that are associated with the listed products and are manually followed up are considered as associated follow-up products. Based on the changes in customer feedback data in the associated follow-up products, determine the changes in the number of customers for the listed products across different demand types; Based on the changes in the number of customers under different demand types and the associated return visit data of the listed products, a return visit management method for the listed products is determined.

[0011] Secondly, the present invention provides a computer system comprising: a memory and a processor connected in communication, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the above-described customer follow-up management method for an e-commerce platform when running the computer program.

[0012] Other features and advantages will be set forth in the following description, and the objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0013] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0014] The above and other features and advantages of the present invention will become more apparent from a detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0015] Figure 1 This is a flowchart of a customer follow-up management method used on e-commerce platforms; Figure 2 This is a flowchart illustrating the method for determining the types of follow-up optimization needs for listed products; Figure 3 This is a flowchart for determining which products require manual follow-up visits; Figure 4 This is a flowchart illustrating the method for determining the products to be manually followed up on among the listed products. Detailed Implementation

[0016] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0017] Example 1 like Figure 1 As shown, this application provides a customer follow-up management method for e-commerce platforms, specifically including: S1 uses the listing data of products on the e-commerce platform to determine the products listed on the e-commerce platform, and determines the type of follow-up optimization needs for the products based on the changes in the sales data of the listed products. Furthermore, the products listed on the e-commerce platform are those listed on the e-commerce platform.

[0018] Furthermore, the changes in the sales data of the listed products are determined based on the changes in the sales volume of the listed products between different time periods.

[0019] Specifically, such as Figure 2 As shown, the method for determining the return visit optimization requirement type for the listed products is as follows: S11 divides the time into multiple time intervals based on the unit time duration, and determines the benchmark value of sales within the time interval based on the sales data of the products listed in different time intervals. In the above steps, time intervals are divided and sales benchmark values ​​are calculated. The total analysis time is divided into multiple intervals according to unit time, and a unified sales benchmark is calculated for the entire analysis period.

[0020] Time interval: Dividing the total duration into equal-length periods (e.g., one month) for comparative analysis. Sales benchmark (Sales_Benchmark): Average sales across all time intervals * preset multiple. This is a passing grade for determining whether sales are "poor".

[0021] This example calculates (product P): time intervals and sales volume: Interval 1 (first month): sales volume = 120 units, Interval 2 (second month): sales volume = 80 units, Interval 3 (third month): sales volume = 50 units, average sales volume = (120 + 80 + 50) / 3 ≈ 83.3 units, sales volume baseline = 83.3 * 0.7 ≈ 58.3 units.

[0022] S12 determines the time intervals in which sales are less than the benchmark value based on the sales data of the products listed in different time intervals, and regards these as the poor sales intervals. Identify periods of poor sales performance by comparing the actual sales volume of each time interval with the baseline value to find the periods of poor performance. The period of poor sales performance (Underperformance_Interval) is the time interval in which the sales volume is less than the sales baseline value.

[0023] In this example, the conclusion for product P is: poor sales range = {range 3}.

[0024] S13 determines the type of follow-up optimization requirement for the listed products based on the data of poor sales in the poor sales range.

[0025] It should be noted that the unit duration is divided based on half a month to one month.

[0026] Specifically, the benchmark value of the sales volume is a preset multiple of the average sales volume of the listed products in different time intervals, wherein the preset multiple ranges from 0.5 to 0.8.

[0027] Furthermore, based on the data regarding the poor sales period, the types of follow-up optimization needs for the listed products are determined, specifically including: S131 Based on the poor sales range data in the most recent preset time period, determine whether the listed product has a poor sales range in the most recent preset time period. If yes, proceed to the next step. If no, determine that the return visit optimization requirement type of the listed product is no optimization processing required. S131: Does a poor sales range exist? Judgment: Product P has a poor sales range (range 3) within the most recent preset period (one quarter). Decision: Yes, proceed to S132.

[0028] S132 obtains the number of poor sales intervals of the listed products in the most recent preset time period, and determines whether the number of poor sales intervals of the listed products in the most recent preset time period is greater than the preset interval number threshold. If yes, the return visit optimization demand type of the listed products is determined to be a type of demand. If no, proceed to the next step. Determining the frequency of problems (number of problematic intervals): This assesses how often a problem occurs. Frequent occurrences of problems suggest that the problem may be persistent.

[0029] Type 1 Request: High-priority optimization requests usually mean that the problem is serious and urgent. Management significance: If the performance is poor for most of the time, it means that there may be a fundamental problem with the product (such as insufficient product strength or excessive pricing), and it is necessary to focus on follow-up and handling.

[0030] Is the number of poor sales intervals (1) greater than T_bad_intervals (2)? No, the problem does not occur frequently. Proceed to S133.

[0031] S133 determines the sales deviation value by the difference between the baseline value of the poor sales range in the most recent preset time period and the sales volume, and determines whether there is a poor sales range where the sales deviation value is greater than the preset deviation threshold. If yes, proceed to the next step; otherwise, determine that the return visit optimization demand type of the listed product is a second type of demand. In the above steps, determine whether there is a significant deviation range (question depth), and determine whether there is a "sharp drop" in sales during a certain period of time, rather than just a slight failure to meet the target.

[0032] Sales Deviation: Baseline value - Actual sales in this poor range. The larger the value, the larger the sales gap. Second type of demand: Medium-priority optimization demand, usually indicating that a problem exists but is not yet critical.

[0033] Identify the points in time when significant anomalies occur, which may be caused by specific events (negative reviews, competitive shocks) and require in-depth analysis.

[0034] For the only poor sales range (range 3): sales deviation = 58.3 - 50 = 8.3 units. Determine: Is there a range where the sales deviation is >15? No (8.3 < 15), no sales plunge occurred. According to the rules, the return visit optimization requirement type for product P is determined to be a type II requirement.

[0035] S134 obtains the percentage of poor sales ranges where the sales deviation value is greater than the preset deviation threshold in the most recent preset time period, and determines the comprehensive deviation coefficient of the listed product based on the percentage of poor sales ranges in the most recent preset time period. Based on the comprehensive deviation coefficient, the type of follow-up optimization requirement for the listed product is determined.

[0036] In one possible embodiment, assume data for a product Q (analyzed for one quarter): Interval 1: Sales = 100 units, Interval 2: Sales = 25 units (assuming a sharp drop in sales due to major negative reviews), Interval 3: Sales = 95 units; Calculate the baseline and identify the poor sales range: Average sales = (100 + 25 + 95) / 3 = 73.3 units, baseline = 73.3 * 0.7 ≈ 51.3 units, poor sales range = {range 2} (because 25 < 51.3).

[0037] The comprehensive deviation coefficient is calculated by taking into account both the severity of the deviation and the scope of the problem, and a final judgment is made. The percentage of poor sales intervals (P_severe) with deviation values ​​exceeding the preset deviation threshold is calculated as: (Number of intervals with deviation values ​​> threshold / Total number of poor sales intervals). In this example, it is 1 / 1 = 100%. The percentage of poor sales intervals within the most recent preset time period is calculated as: (Number of poor sales intervals / Total number of intervals). In this example, it is 1 / 3 ≈ 33.3%.

[0038] Comprehensive Deviation Coefficient: Calculated by combining the above two ratios, it is used for final classification. The calculation formula is: Comprehensive Deviation Coefficient = 0.1 multiplied by 1 + 0.333 = 0.433. Is the Comprehensive Deviation Coefficient (0.433) > T_comprehensive_deviation (0.3)? Yes, the return visit optimization requirement type for product Q is determined to be a type of requirement.

[0039] It is understood that, based on the aforementioned comprehensive deviation coefficient, the type of return visit optimization requirement for the listed products is determined, specifically including: When the overall deviation coefficient of the listed product is greater than the preset deviation coefficient threshold, the return visit optimization requirement type of the listed product is determined to be a type 1 requirement. If the overall deviation coefficient of the listed product is not greater than the preset deviation coefficient threshold, the return visit optimization requirement type of the listed product is determined to be a type 2 requirement.

[0040] It should be noted that the need for follow-up optimization of listed products of the first type of demand is greater than that of listed products of the second type of demand.

[0041] Based on the composition data of the listed products with different mutual visit optimization needs, S2 determines the products that need to be manually revisited when determining the products that need to be manually revisited, by considering the revisit optimization needs of the listed products and the customer relationship between the listed products and other listed products. Specifically, such as Figure 3 As shown, the determination of products requiring manual follow-up includes: Based on the composition data of the listed products for different mutual visit optimization demand types, S21 determines the listed products for one type of demand and the listed products for two types of demand, and treats them as one type of demand product and the other as two type of demand product. Category 1 Demand Products: Products identified as "Category 1 Demand Products" require the highest priority for optimization intervention. Category 2 Demand Products: Products identified as "Category 2 Demand Products" require medium priority for optimization intervention. Total Number of Demand Products: Number of Category 1 Demand Products + Number of Category 2 Demand Products. This represents the total number of "problem products" requiring attention across the entire category, aggregating scattered individual product issues into category-level data that can be grasped macroscopically, providing a basis for resource allocation decisions.

[0042] In this example, the data (assumptions) is as follows: there are 100 products listed in this category, the number of products in category 1 demand (N1) = 3 (e.g., products A, B, C), the number of products in category 2 demand (N2) = 10 (e.g., products D to M), and the total number of products in demand (N_total) = 3 + 10 = 13.

[0043] S22 determines the quantity of the first-category demand products and the quantity of the second-category demand products among the listed products based on the composition data of the first-category demand products and the second-category demand products. S23 determines whether manual follow-up visits are needed based on the quantity of Category I demand products and the quantity of Category II demand products among the listed products.

[0044] It is understandable that, based on the quantity of Category 1 and Category 2 demand products among the listed products, the determination of whether manual follow-up is required for these products includes: S231 Based on the quantity of Category I demand products and Category II demand products in the listed products, determine the total quantity of demand products in the listed products, and determine whether the total quantity of demand products in the listed products is greater than a preset quantity threshold. If yes, proceed to the next step. If no, since the number of listed products with high demand is small, the difficulty of follow-up processing is small, so it is determined that no manual follow-up is required. In the above steps, it is determined whether the total number of "problem products" that need to be handled has reached a scale that requires special attention. If the total number of problem products is small, the manager or operations team can easily incorporate these products into the daily workflow for handling without initiating a special, centralized manual follow-up task.

[0045] The total number of demanded goods (13) is not greater than T_total_demand (15), and the total number of problematic goods has not exceeded the threshold. According to the rules, the process ends here, and it is determined that the goods do not need to be manually revisited.

[0046] S232 obtains the quantity of a type of demanded goods, and determines whether the quantity of the type of demanded goods is greater than a preset demanded goods quantity threshold. If yes, it determines the goods that need to be manually revisited; otherwise, it proceeds to the next step. Suppose a certain type of product has a high demand (triggers S232): If the number of goods in category 1 demand (N1) = 7, the number of goods in category 2 demand (N2) = 8, and the total number of goods demanded (N_total) = 15.

[0047] Execution process: S231: N_total (15)>T_total_demand (15)? Yes (usually "greater than" includes equal to), proceed to S232, S232: N1 (7)>T_priority1 (5), confirm the products that need to be manually revisited.

[0048] Based on the quantity of the first-class demand goods and the quantity of the second-class demand goods, and combined with the preset weight values ​​of the first-class demand goods and the second-class demand goods, S233 determines the sum of the preset weight values ​​of the first-class demand goods and the second-class demand goods, and determines whether the sum of the preset weight values ​​of the first-class demand goods and the second-class demand goods is greater than the preset weight threshold. If it is, it determines the goods that need to be manually revisited; if not, it determines the goods that do not need to be manually revisited.

[0049] Scenario 2: Weighted problem is prominent (triggering S233) Assume that the number of goods in the first category of demand (N1) = 4, the number of goods in the second category of demand (N2) = 20, and the total number of goods in demand (N_total) = 24; Specifically, the weighted sum = (N1 * W1) + (N2 * W2) = (4 * 1.0) + (20 * 0.5) = 4 + 10 = 14. The judgment is: the weighted sum (14) > T_weighted_sum (10). The weight of the problem is too high, and it is determined that the product needs to be manually revisited.

[0050] Specifically, such as Figure 4 As shown, the method for determining the manually followed-up products among the listed products is as follows: S31, based on the return visit optimization demand type of the listed products, determines the first-class demand products and the second-class demand products among the listed products; S32 determines, based on the customer association between the listed products and other listed products, the listed products that are associated with the listed products and are designated as associated listed products. Step Explanation: Locate the problem pool in the global product list and build an internal network of relationships within the problem pool. Focus management resources on the 30 problematic products to avoid wasting analysis resources on the 970 normal products.

[0051] This example data (30 products to be optimized): Category 1 demand products (8 products): SKU01, SKU02, SKU03, SKU04, SKU05, SKU06, SKU07, SKU08; Category 2 demand products (22 products): SKU09 to SKU30. Related listing products (example): Related listing products of SKU10: SKU01, SKU02, SKU15 (SKU15 is a second category), Related listing products of SKU11: SKU31, Related listing products of SKU12: (None), Related listing products of SKU13: SKU01, SKU04, SKU05, SKU31 (Very extensive relationships).

[0052] S33 determines whether the listed product is a product requiring manual follow-up based on the type of follow-up optimization requirement of the listed product and the associated listed products.

[0053] It is understandable that, based on the type of follow-up optimization requirement for the listed product and the associated listed products, determining whether the listed product is a product requiring manual follow-up includes: S331 When the listed product is a type of demand product, then the listed product is determined to be a product for manual follow-up visits; S331: A category of demand products are automatically selected, SKU01 to SKU08, a total of 8 products, and are automatically identified as products for manual follow-up.

[0054] S332 When the listed product does not belong to a category of demand products, obtain the quantity ratio of the category of demand products in the listed products, and determine whether the quantity ratio of the category of demand products in the listed products is greater than a preset percentage threshold. If yes, then the listed products other than the category of demand products are not manually revisited products. If no, proceed to the next step. S332: Determine the proportion of a certain type of demand product. The proportion of a certain type of demand product in the listed products is 8 / 1000 = 0.08. 0.08 is not greater than T_ratio_priority1 (5%). At this time, the number of a certain type of demand product is relatively small, so proceed to S333.

[0055] S333 determines the associated listed products of the second-category demand products based on the association between the second-category demand products and the listed products, and determines whether the number of associated listed products of the second-category demand products is greater than the preset threshold for the number of associated products. If yes, it is determined that the second-category demand products belong to the manually revisited products; otherwise, it proceeds to the next step. Further, the breadth of association of the second-class demand products is judged, and the "hub" products in the second-class products are found. A hub product is revisited. Since there is overlap between the customers of the hub product and other products, the processing efficiency is high. In this example, the judgment is made (taking SKU13 as an example): the number of related products of SKU13 = 4 (SKU01, SKU04, SKU05, SKU20). The judgment is: 4>T_related_sku (2), and SKU13 is determined to be a product that is manually revisited.

[0056] S334 Based on the association between the listed products, determine the first type of demand products and the associated listed products of the first type of demand products, and determine whether the quantity of the second type of demand products excluding the associated listed products of the first type of demand products is less than a preset quantity threshold. If yes, determine that the remaining second type of demand products are not products for manual follow-up. If no, proceed to the next step. In the above steps, the associated listed products of Category II demand products, excluding those that are associated with both Category II and Category I demand products, are classified as Category II associated products. If the number of Category II associated products is small, for example, less than 5, then the remaining Category II demand products are determined not to be products subject to manual follow-up. S335 Based on the related listed products of the two types of demand products, determine whether there are related listed products for the two types of demand products. If not, determine that the remaining two types of demand products are not products for manual follow-up. If so, proceed to the next step. In the above steps, isolated products with a small impact were eliminated, and related issues were addressed first because they have a greater impact on the overall customer experience and sales ecosystem. Among the remaining 20 products, SKU12 had no related listed products, so it was determined that SKU12 was not a product requiring manual follow-up.

[0057] S336 determines the association factor between the two types of demand products and different associated listed products based on the number of customers who need to be revisited for the overlapping of the two types of demand products and the associated listed products, and the proportion of such customers in the total number of customers who need to be revisited for the associated listed products. Based on the association factor between the two types of demand products and different associated listed products, S336 determines whether the two types of demand products are products for manual revisit.

[0058] Step explanation: For the last remaining related second-category products, calculate the customer overlap between them and the selected products that have been manually revisited.

[0059] Correlation Factor: The number of customers awaiting follow-up visits who overlap with the second-category demand products and related listed products / the total number of customers awaiting follow-up visits for related listed products. This identifies second-category products with a high degree of overlap with customers awaiting follow-up visits who are not part of the demand products. When following up on core products, these deeply tied products can be diagnosed simultaneously, making the process more efficient.

[0060] This example demonstrates the calculation and judgment process (using SKU11 as an example): The associated product of SKU11 is SKU31. Assuming that SKU03 has 150 customers to be contacted, of which 80 overlap with SKU31, the association factor between SKU11 and SKU31 is 80 / 150 ≈ 0.53. We determine that the sum of association factors (0.53) > T_total_factor (0.3). Therefore, SKU11 is a product that requires manual contact.

[0061] Final list of products requiring manual follow-up: S331 (8 items): SKU01, SKU02, SKU03, SKU04, SKU05, SKU06, SKU07, SKU08; S333 (2 items): SKU10, SKU13; S336 (1 item): SKU11; Total: 11 items It is understandable that, based on the association factors between the two types of demand products and different related listed products, determining whether the two types of demand products are products requiring manual follow-up visits specifically includes: If the sum of the association factors between the two types of demand products and different related listed products is greater than a preset association factor threshold, then the two types of demand products are determined to be products requiring manual follow-up. Otherwise, the two types of demand products are determined not to be products requiring manual follow-up.

[0062] It should be noted that the associated listed products are those with an association factor greater than a preset factor threshold.

[0063] Specifically, when the listed products are products requiring manual follow-up, all customers of the products requiring follow-up must be contacted manually.

[0064] It should be noted that the changes in the customer feedback data are determined based on the changes in the number of customers with different demand types in the customer feedback data of the listed products.

[0065] The core value of this method lies in its ability to act as a sophisticated "sieve" to filter products: First, it filters out 30 problematic products from 1,000 products; second, it filters out 11 products from the 30 problematic products that are most worth investing human resources in.

[0066] This ensures that, even in a complex environment with a massive number of products, the operations team can always concentrate its resources to solve key challenges, thereby maximizing operational efficiency and quality. The remaining 989 products are effectively managed through intelligent, tiered automation strategies.

[0067] S3 determines the changes in customer feedback data based on the results of the manual follow-up of the products. Based on the changes in customer feedback data of the manually followed products that are related to the listed products, S3 determines the follow-up management method for the listed products other than the manually followed products.

[0068] Let's assume we're continuing with the "Manual Follow-up for Product Identification" example. The store has 1000 products, of which 11 are identified as products requiring manual follow-up.

[0069] Analysis objective: To determine the follow-up management method for the remaining 989 products (19 products to be optimized + 970 normal products).

[0070] Specifically, the method for determining the follow-up management method for listed products, excluding those requiring manual follow-up, is as follows: S41 will identify manually visited products that are associated with the listed products as associated visit products; S42 determines the changes in the number of customers for the listed products across different demand types based on the changes in customer feedback data in the associated follow-up products; S43 determines the follow-up management method for the listed products based on the changes in the number of customers in different demand types and the associated follow-up product data of the listed products.

[0071] Furthermore, the types of demands include quality issues, user experience issues, and logistics issues.

[0072] Furthermore, the variation in the number of customers for the listed products across different demand types is determined based on the analysis results of manual follow-up data of customers for the listed products' associated follow-up products.

[0073] Furthermore, based on the changes in the number of customers across different demand types and the associated follow-up product data of the listed products, a follow-up management method for the listed products is determined, specifically including: S431 If the listed product does not have an associated follow-up product, then the follow-up management method of the listed product is determined to be the preset follow-up method; In the above steps, SKU12 is a second-category demand product, and it was determined in step S335 that there are no related listed products. Since there are no related follow-up products for SKU12, the follow-up management method for SKU12 is determined to be the [preset follow-up method], that is, 70% of customers are followed up by AI and 30% of customers are followed up by manual review.

[0074] SKU12 itself has issues (a type of requirement), but it cannot obtain "intelligence sharing" from any manual follow-up tasks, leaving it in an "information silo" state. In order not to miss its own potential unique problems, a high manual review ratio (30%) needs to be configured for risk control and in-depth diagnosis.

[0075] S432 If the listed product has associated return visit products, obtain the number of associated return visit products of the listed product, and determine whether the number of associated return visit products of the listed product is greater than the preset return visit product number threshold. If yes, determine that the return visit management method of the listed product is the second preset return visit method. If no, proceed to the next step. In the above steps, SKU100 (normal product) is a normally selling product with stable sales. However, it is related to SKU01 (a high-end headphone) which was manually followed up (for example, both are audio devices and have a high degree of overlap in customer groups).

[0076] S431: Determine if there are related return visit products. Judgment: SKU100 has related return visit product SKU01. Decision: No. Proceed to S432. S432: Determine the number of related return visit products. Judgment: The number of related return visit products for SKU100 = 1. Judgment: 1>T_related_manual (2)? Decision: No. Proceed to S433. S433: Based on the judgment of feedback change trend, check the distribution change of customer feedback type of SKU100 in manual processing of its related return visit product SKU01.

[0077] Assuming the percentage of customer feedback issues before and after a manual follow-up call for SKU01:

[0078] Determine if there are any abnormal change types of requirements: Change in quality issues (|-30%|) > T_change (10%), Change in logistics issues (|+25%|) > T_change (10%). Conclusion: There are abnormal change types of requirements.

[0079] Final decision: Therefore, the follow-up management method for SKU100 is determined as [Preset Follow-up Method], that is, 70% AI follow-up and 30% manual review.

[0080] Logic: There has been a drastic change in the feedback structure of the "information source" SKU01 of SKU100. Among the customers where SKU001 coincides with SKU100, the logistics problems exposed by its associated products may be systematic risks in the supply chain (for example, the service of a certain cooperating express delivery company has deteriorated). Such risks are likely to exist in large quantities in SKU100. Therefore, it is necessary to increase the proportion of manual work, focus on monitoring and reviewing the logistics-related feedback of SKU100, and achieve proactive risk control.

[0081] S433 determines the change in the proportion of the number of customers in different demand types based on the change in the number of customers in different demand types, and determines the follow-up management method of the shelved products based on the change in the proportion of the number of customers in different demand types.

[0082] Case 3: SKU11 (associated products to be optimized) Background: SKU11 is a type-two demand product, and its associated follow-up product is SKU03. In S336, it was落选 from manual follow-up because the sum of its associated factors (0.53) did not exceed the threshold.

[0083] S431&S432: There is an associated follow-up product SKU03, and the quantity is 1, which is not greater than the threshold. Enter S433.

[0084] S433: Judgment based on the feedback change trend; Analysis: Check the feedback changes of its associated follow-up product SKU03 before and after manual processing.

[0085] Assume the data of SKU03 before and after manual follow-up is as follows:

[0086] Judge whether there is an abnormal change in the demand type: The absolute value of the change amount of all types is <T_change (10%). Conclusion: There is no demand type with abnormal change. Therefore, the follow-up management method for SKU11 is determined as [Second Preset Follow-up Method], that is, 90% of the customers are followed up by AI, and only 10% of the customers are reviewed manually.

[0087] Specifically, the change in the proportion of the number of customers is determined according to the change amount of the proportion of the number of customers of the shelved products in the demand type at the current moment and the proportion of the number of customers before the manual follow-up of the manually followed-up products.

[0088] Furthermore, the follow-up management method for the listed products is determined based on the changes in the proportion of customers in different demand types, specifically including: Based on the changes in the proportion of the number of customers, determine whether there are any abnormal demand types in the changes in the proportion of the number of customers. If so, determine that the follow-up management method for the listed products is the preset follow-up method. If not, determine that the follow-up management method for the listed products is the second preset follow-up method.

[0089] It is understood that the abnormal change in the proportion of the number of customers is a demand type in which the difference between the proportion of the number of customers of the listed products in the demand type at the current moment and the proportion of the number of customers before the manual follow-up visit to the products is greater than the preset change threshold.

[0090] Furthermore, the preset follow-up method is to determine the proportion of customers who use AI and manual follow-up among the listed products using a preset ratio, that is, the proportion of customers who use AI for follow-up is the preset ratio. The second preset follow-up method is to determine the proportion of customers who use AI and manual follow-up among the listed products using a second preset ratio, that is, the proportion of customers who use AI for follow-up is the second preset ratio.

[0091] Through the above mechanisms, enterprises have established a dynamic and intelligent customer follow-up resource allocation system for a massive number of products:

[0092] In summary, the core of this approach lies in leveraging "insights" gained from high-value human follow-ups to dynamically guide automated strategies for a massive volume of low-to-medium priority products. It ensures that: Precision of resources: Human resources are concentrated on core issues (human follow-up on products) and high-risk scenarios (preset follow-up methods). Sensitivity of strategy: It can quickly capture and respond to signals of systemic risks through the network of relationships. Efficiency of operation: In most scenarios (second preset follow-up method), the best balance between cost and effect is achieved.

[0093] Example 2 Secondly, the present invention provides a computer system comprising: a memory and a processor connected in communication, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the above-described customer follow-up management method for an e-commerce platform when running the computer program.

[0094] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0095] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0096] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A customer follow-up management method for e-commerce platforms, characterized in that, Specifically, it includes: Based on the listing data of products on the e-commerce platform, identify the products listed on the e-commerce platform, and determine the types of follow-up optimization needs for the products based on the changes in the sales data of the listed products. Based on the composition data of listed products with different mutual visit optimization needs, when determining which products need to be manually revisited, the manual revisit products are determined from the revisit optimization needs of the listed products and the customer relationship between the listed products and other listed products. Based on the results of the follow-up visits to the products, the changes in customer feedback data are determined. Based on the changes in customer feedback data for the products that are related to the listed products, the follow-up management method for the listed products other than the products that are manually followed up is determined.

2. The customer follow-up management method for e-commerce platforms as described in claim 1, characterized in that, The products listed on the e-commerce platform are those that are listed on the e-commerce platform.

3. The customer follow-up management method for e-commerce platforms as described in claim 1, characterized in that, The changes in the sales data of the listed products are determined based on the changes in the sales volume of the listed products between different time periods.

4. The customer follow-up management method for e-commerce platforms as described in claim 1, characterized in that, The method for determining the type of return visit optimization requirement for the listed products is as follows: Based on the unit time period, the time is divided into multiple time intervals. Based on the sales data of the products listed in different time intervals, a benchmark value for the sales volume in the time interval is determined. Based on the sales data of products listed in different time intervals, the time intervals in which sales are lower than the benchmark value are determined and regarded as poor sales intervals. Based on the data of the poor sales period, determine the type of follow-up optimization requirement for the listed products.

5. The customer follow-up management method for e-commerce platforms as described in claim 4, characterized in that, The unit duration is divided based on half a month to one month.

6. The customer follow-up management method for e-commerce platforms as described in claim 4, characterized in that, The benchmark value for sales volume is a preset multiple of the average sales volume of the listed products in different time intervals.

7. The customer follow-up management method for e-commerce platforms as described in claim 1, characterized in that, The determination of which products require manual follow-up visits includes: Based on the composition data of listed products with different mutual visit optimization demand types, we determine the listed products of type 1 demand and type 2 demand, and treat them as type 1 demand products and type 2 demand products. Based on the composition data of the first-class and second-class demand products, determine the quantity of the first-class demand products and the quantity of the second-class demand products in the listed products; Based on the quantity of Category I and Category II demand products among the listed products, it is determined whether manual follow-up visits are required.

8. The customer follow-up management method for e-commerce platforms as described in claim 7, characterized in that, Based on the quantity of Category 1 and Category 2 demand products among the listed products, a determination is made as to whether manual follow-up is required for these products. This includes: Based on the quantity of Category 1 and Category 2 demand products among the listed products, the total quantity of demand products among the listed products is determined. If the total quantity of demand products among the listed products is not greater than a preset quantity threshold, then it is determined that no manual follow-up is required to identify the products.

9. The customer follow-up management method for e-commerce platforms as described in claim 1, characterized in that, The method for determining the follow-up management method for listed products, excluding those requiring manual follow-up, is as follows: Products that are associated with the listed products and are manually followed up are considered as associated follow-up products. Based on the changes in customer feedback data in the associated follow-up products, determine the changes in the number of customers for the listed products across different demand types; Based on the changes in the number of customers under different demand types and the associated return visit data of the listed products, a return visit management method for the listed products is determined.

10. A computer system, comprising: A memory and processor connected by communication, and a computer program stored in the memory and capable of running on the processor, characterized in that, when the processor runs the computer program, it executes a customer follow-up management method for an e-commerce platform as described in any one of claims 1-9.

Images