Delivery man service score evaluation method and system

Abstract

The invention discloses a deliveryman service score evaluation method and system. The method comprises the steps of obtaining actual proportion values of a current deliveryman in a plurality of evaluation score items in a previous service score evaluation period; respectively obtaining target scale values of the plurality of evaluation scoring items; based on the evaluation data of the current deliveryman in the plurality of evaluation scoring items and the target data of the plurality of evaluation scoring items, calculating a deviation proportion value of the plurality of evaluation scoring items; if the deviation proportion value of at least one evaluation score item is higher than the corresponding preset proportion threshold value, improving the evaluation score item to calculate the weight value of the current deliveryman in the service score calculation process of the next service score evaluation period, and correspondingly reducing the weight values of the other evaluation score items; and if the deviation proportion value of at least one evaluation score item is lower than the corresponding preset proportion threshold value, reducing the weight value of the evaluation score item in the service score calculation process of the next service score evaluation period of the current deliveryman, and correspondingly improving the weight values of the other evaluation score items.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics, and particularly relates to a method and system for evaluating the service score of deliverymen. Background Art

[0002] Under the background of the booming development of the current e-commerce and logistics industries, as a key link connecting merchants and consumers, the quality of delivery service directly affects the shopping experience of consumers and the reputation of merchants. In order to effectively measure and improve the service quality of deliverymen, various e-commerce platforms and logistics enterprises generally adopt a service score evaluation system to quantitatively evaluate the work performance of deliverymen.

[0003] In the existing calculation method for evaluating the service score of deliverymen, fixed weight values are usually assigned to various evaluation score items such as delivery overtime, order refusal, order cancellation, customer negative reviews, and dress code. This evaluation method with fixed weight values provides convenience for the evaluation of service quality to a certain extent and also played a certain role in the initial stage of the industry development.

[0004] However, with the continuous expansion of e-commerce business and the increasing diversification of consumer demands, this evaluation method with fixed weight values gradually exposes many problems. First, due to the differences in delivery difficulty and service requirements in different regions, different time periods, and different business types, the fixed weight values cannot accurately reflect the service performance of deliverymen in different situations. For example, in some urban areas with serious traffic congestion, the situation of delivery overtime may be more common. At this time, if the delivery overtime item is still evaluated according to a unified low weight, it is impossible to effectively motivate deliverymen to improve delivery efficiency; during some special promotion activities, the order volume increases significantly, and deliverymen may face greater pressure of order refusal and order cancellation. The fixed weight values are difficult to comprehensively and objectively evaluate the service quality of deliverymen in such special situations.

[0005] Secondly, the market environment and consumer preferences are constantly changing, and the fixed weight values cannot adapt to these changes in a timely manner. As consumers' requirements for service quality are getting higher and higher, the impact of customer negative reviews on service quality may gradually increase, but the fixed weight values cannot reflect this changing trend, resulting in the service score evaluation being unable to accurately reflect the true feelings of consumers. Moreover, this method with fixed weight values lacks effective guidance for the service improvement of deliverymen. Deliverymen may not pay much attention to some items with low weights even if their performance is poor, and thus ignore the improvement of service quality in these aspects. For example, when the weight of the dress code item is low, deliverymen may ignore their own image, which affects the brand image of the enterprise to a certain extent.

[0006] In summary, in the existing method for calculating the service score evaluation of delivery staff, where the weight values of each item remain unchanged, it can no longer meet the current complex and changing delivery business requirements and is difficult to accurately and comprehensively evaluate the service quality of delivery staff. Summary of the Invention

[0007] The purpose of the embodiments of the present invention is to provide a method and system for evaluating the service score of delivery staff. By dynamically adjusting the weight values of various parameters of the service score, it can be flexibly adapted according to the actual performance of delivery staff, accurately reflecting the service quality in different scenarios; motivating delivery staff to improve weak links and enhance the overall service level; and also being able to quickly respond to changes in market and customer needs, optimize the allocation of delivery resources, improve operation efficiency, enhance customer satisfaction and enterprise competitiveness.

[0008] To solve the above technical problems, the first aspect of the embodiments of the present invention provides a method for evaluating the service score of delivery staff, including the following steps:

[0009] Obtain the actual proportion values of the current delivery staff in several evaluation score items in the previous service score evaluation period. The evaluation score items include: delivery overtime item, refusal item, order cancellation item, customer negative review item, and dress code item;

[0010] Respectively obtain the target proportion values of the several evaluation score items;

[0011] Based on the evaluation data of the current delivery staff in several evaluation score items and the target data of the several evaluation score items, calculate the deviation proportion values of the several evaluation score items;

[0012] If the deviation proportion value of at least one of the evaluation score items is higher than the corresponding preset proportion threshold, increase the weight value of the evaluation score item in the service score calculation process of the current delivery staff in the next service score evaluation period, and correspondingly decrease the weight values of the remaining evaluation score items;

[0013] If the deviation proportion value of at least one of the evaluation score items is lower than the corresponding preset proportion threshold, decrease the weight value of the evaluation score item in the service score calculation process of the current delivery staff in the next service score evaluation period, and correspondingly increase the weight values of the remaining evaluation score items.

[0014] Further, the corresponding decrease in the weight values of the remaining evaluation score items includes:

[0015] Obtain the total sum of the weight values of the remaining evaluation score items;

[0016] Based on the proportion of the weight value of each evaluation score item in the total sum of the weight values, correspondingly decrease the weight value of the evaluation score item.

[0017] Further, the corresponding increase in the weight values of the remaining evaluation score items includes:

[0018] Obtain the sum of the weight values of the remaining described evaluation score items;

[0019] Based on the proportion of the weight value of each described evaluation score item in the sum of the weight values, correspondingly increase the weight value of the evaluation score item.

[0020] Further, the calculating the deviation proportion values of the several evaluation score items includes:

[0021] Obtain the total number of orders, the number of overtime orders, the number of rejected orders, the number of cancelled orders, the number of customer bad review items, and the number of dressing violation items of the current deliveryman in the previous service score evaluation period;

[0022] Respectively calculate the actual proportion value of the order overtime rate, the actual proportion value of the rejected order rate, the actual proportion value of the cancelled order rate, the actual proportion value of the customer bad review rate, and the actual proportion value of the dressing violation rate of the current deliveryman in the previous service score evaluation period;

[0023] Based on the target proportion value of the order overtime rate, the target proportion value of the rejected order rate, the target proportion value of the cancelled order rate, the target proportion value of the customer bad review rate, and the target proportion value of the dressing violation rate, respectively calculate the deviation proportion value of the order overtime rate, the deviation proportion value of the rejected order rate, the deviation proportion value of the cancelled order rate, the deviation proportion value of the customer bad review rate, and the deviation proportion value of the dressing violation rate of the current deliveryman in the previous service score evaluation period.

[0024] Further, the deviation proportion value ΔR of the order overtime rate OT The calculation formula is:

[0025]

[0026] Wherein, R OT Is the actual proportion value of the order overtime rate of the current deliveryman in the previous service score evaluation period, and R OT,target Is the target proportion value of the order overtime rate;

[0027] The deviation proportion value ΔR of the rejected order rate RJ The calculation formula is:

[0028]

[0029] Wherein, R RJ Is the actual proportion value of the rejected order rate of the current deliveryman in the previous service score evaluation period, and R RJ,target Is the target proportion value of the rejected order rate;

[0030] The deviation proportion value ΔR of the cancelled order rate CN The calculation formula is:

[0031]

[0032] Among them, R CN is the actual proportion value of the order cancellation rate of the current deliveryman in the previous service score evaluation period, and R is the target proportion value of the order cancellation rate;

[0033] The deviation proportion value ΔR of the customer bad review rate NP is calculated by the following formula:

[0034]

[0035] Among them, R NP is the actual proportion value of the customer bad review rate of the current deliveryman in the previous service score evaluation period, and R NP,target is the target proportion value of the customer bad review rate;

[0036] The deviation proportion value ΔR of the dressing violation rate NV is calculated by the following formula:

[0037]

[0038] Among them, R NV is the actual proportion value of the dressing violation rate of the current deliveryman in the previous service score evaluation period, and R NV,target is the target proportion value of the dressing violation rate.

[0039] Furthermore, the overtime duration of overtime orders includes: the first overtime range, the second overtime range, and the third overtime range. The first overtime range is minor overtime, the second overtime range is moderate overtime, and the third overtime range is severe overtime;

[0040] Calculating the deviation proportion value of the order overtime rate of the current deliveryman in the previous service score evaluation period includes:

[0041] Obtaining the overtime duration data of overtime orders of the current deliveryman in the previous service score evaluation period, and obtaining the number of orders corresponding to the first overtime range, the second overtime range, and the third overtime range respectively for the overtime orders;

[0042] Based on the number of overtime orders corresponding to the first overtime range, the second overtime range, and the third overtime range and their respective weight coefficient values, calculating the actual proportion calibration value of the order overtime rate;

[0043] Based on the actual proportion calibration value of the order overtime rate, calculating the deviation proportion value of the order overtime rate.

[0044] Furthermore, the calculation formula for the actual proportion calibration value R OT ′ of the order overtime rate is:

[0045]

[0046] Wherein, N1, N2, and N3 are the number of timeout orders corresponding to the first timeout range, the second timeout range, and the third timeout range respectively, and N total is the total number of timeout orders, and ω1, ω2, and ω3 are the weight values of the first timeout range, the second timeout range, and the third timeout range respectively, each accounting for the total number of timeout orders;

[0047] Correspondingly, the calculation formula for the deviation ratio value of the order timeout rate is:

[0048]

[0049] Furthermore, calculating the actual ratio value of the customer bad review rate of the current deliveryman in the previous service score evaluation period further includes:

[0050] Performing sentiment analysis on the bad review text corresponding to the customer bad review item, and extracting multi-dimensional sentiment feature vectors. The sentiment feature vectors include: the sentiment intensity level output based on the pre-trained language model and the problem entity type parsed by the semantic focus recognition model. The sentiment intensity level is divided into several levels of quantization values, and the problem entity type is associated with a preset service defect classification, including at least one of service attitude, time delay, and goods damage;

[0051] Calculating the bad review impact coefficient based on the sentiment feature vector, and its calculation formula is:

[0052]

[0053] Wherein, S is the sentiment intensity level value, and T i is the preset impact factor of the i-th type of problem entity type, and α and β are normalization coefficients;

[0054] Multiplying the bad review impact coefficient by the actual ratio value of the customer bad review rate to obtain the calibrated actual ratio value of the customer bad review rate for calculating the deviation ratio value.

[0055] Furthermore, the deliveryman service score evaluation method further includes:

[0056] Obtaining new bad review data according to the periodic weight update period and performing sentiment analysis;

[0057] Recalculating the bad review impact coefficient of the historical bad review text according to the sentiment analysis result;

[0058] When the change amount of the cumulative bad review impact coefficient of the same deliveryman exceeds the preset threshold, triggering the recalculation of the service score and generating a fluctuation report. The fluctuation report includes the change trend of the impact weight of each service defect classification and the service score correction value.

[0059] Accordingly, a second aspect of the embodiments of the present invention provides a deliveryman service score evaluation system, which evaluates the service score of a deliveryman based on the above-mentioned deliveryman service score evaluation method, including:

[0060] An actual data acquisition module, which is used to acquire the actual proportion values of the current deliveryman in a number of evaluation score items in the previous service score evaluation period, and the evaluation score items include: delivery overtime item, refusal item, order cancellation item, customer bad review item, and dress code item;

[0061] A target data acquisition module, which is used to acquire the target proportion values of the number of evaluation score items respectively;

[0062] A deviation data calculation module, which is used to calculate the deviation proportion values of the number of evaluation score items based on the evaluation data of the current deliveryman in the number of evaluation score items and the target data of the number of evaluation score items;

[0063] A weight adjustment module, which is used to increase the weight value in the service score calculation process of the current deliveryman in the next service score evaluation period for the evaluation score item when the deviation proportion value of at least one of the evaluation score items is higher than the corresponding preset proportion threshold, and correspondingly decrease the weight values of the remaining evaluation score items;

[0064] The weight adjustment module is further used to decrease the weight value in the service score calculation process of the current deliveryman in the next service score evaluation period for the evaluation score item when the deviation proportion value of at least one of the evaluation score items is lower than the corresponding preset proportion threshold, and correspondingly increase the weight values of the remaining evaluation score items.

[0065] Accordingly, a third aspect of the embodiments of the present invention provides an electronic device, including: at least one processor; and a memory connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above-mentioned deliveryman service score evaluation method.

[0066] Accordingly, a fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the above-mentioned deliveryman service score evaluation method is implemented.

[0067] The above technical solutions of the embodiments of the present invention have the following beneficial technical effects:

[0068] 1. By accurately obtaining the actual proportion value, target proportion value, and deviation proportion value of each evaluation scoring item, for the delivery overtime item, subdivide the overtime duration range and calibrate the proportion value, which can more accurately reflect the service performance of the delivery staff in different dimensions, avoid evaluation deviations caused by simple calculations, make the service score evaluation more objective and fair, and provide a reliable basis for delivery management;

[0069] 2. According to the comparison between the deviation proportion value and the preset proportion threshold, dynamically adjust the weights of each evaluation scoring item. When a certain item performs poorly, increase its weight to attract attention, and at the same time reduce the weights of the remaining items. The dynamic adjustment method can be flexibly adapted according to the actual performance of the delivery staff, encourage the delivery staff to comprehensively improve the service level, optimize resource allocation, and adapt to different delivery scenarios and business requirements;

[0070] 3. Set clear preset proportion thresholds for each evaluation scoring item. The delivery staff can clearly understand the passing standards of each service indicator, and the enterprise can also quickly identify the weak links in the delivery staff's service based on the thresholds, formulate targeted improvement measures, promote the delivery staff to improve the service, improve the overall delivery service quality, and enhance customer satisfaction and enterprise competitiveness. Description of the Drawings

[0071] Figure 1 is the flowchart of the delivery staff service score evaluation method provided by the embodiment of the present invention;

[0072] Figure 2 is the block diagram of the delivery staff service score evaluation system module provided by the embodiment of the present invention.

[0073] Reference Signs:

[0074] 1. Actual data acquisition module, 2. Target data acquisition module, 3. Deviation data calculation module, 4. Weight adjustment module. Detailed Embodiments

[0075] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with the specific embodiments and with reference to the accompanying drawings. It should be understood that these descriptions are exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the descriptions of well-known structures and technologies are omitted to avoid unnecessarily confusing the concepts of the present invention.

[0076] Please refer to Figure 1 , the first aspect of the embodiment of the present invention provides a delivery staff service score evaluation method, including the following steps:

[0077] Step S100, obtain the actual proportion values of the current delivery staff in several evaluation scoring items in the previous service score evaluation period. The evaluation scoring items include: delivery overtime item, refusal item, order cancellation item, customer negative review item, and dress code item.

[0078] By collecting the actual performance data of the delivery staff in each key service indicator during the previous service score evaluation period and converting it into a proportional value form for subsequent quantitative analysis. In actual operation, in order to obtain the actual proportional value of the delivery overtime item, it is necessary to record the estimated delivery time and the actual delivery time of each delivery order, calculate the number of overtime orders through comparison, and then divide it by the total number of orders to obtain the delivery overtime rate. For the order rejection item, the system will record the number of orders rejected by the delivery staff, and also divide it by the total number of orders to obtain the order rejection rate. For the order cancellation item, it is necessary to distinguish whether it is the delivery staff who cancels actively or the customer who cancels. After counting the quantities respectively, the order cancellation rate is calculated. For the customer negative review item, by collecting the number of negative reviews from customers on the delivery service and dividing it by the total number of orders to obtain the negative review rate. For the dress code item, the number of times the delivery staff's dress does not conform to the standard can be counted through manual inspection records or user feedback, etc., and then the dress code violation rate is calculated. These actual proportional values reflect the service performance of the delivery staff in different aspects.

[0079] Step S200: Obtain the target proportional values of several evaluation score items respectively.

[0080] Set an expected target value for each evaluation score item as the standard for measuring the service performance of the delivery staff. The setting of the target proportional value needs to comprehensively consider various factors such as the company's business objectives, industry standards, and customer expectations. For the delivery overtime item, the setting of the target proportional value may vary according to different delivery areas, time periods, and business types. For example, in some areas with convenient transportation, it may be expected that the delivery overtime rate is controlled within 5%; while during the peak traffic congestion period, the target value may be appropriately relaxed to 10%. The target proportional value of the order rejection item is usually set relatively low, such as below 3%, to encourage the delivery staff to actively accept orders. The target proportional value of the order cancellation item will be determined according to the stability of the business and the acceptance degree of customers, and may be between 5% - 8%. The target proportional value of the customer negative review item is strictly controlled, generally within 5%, to ensure customer satisfaction. The target proportional value of the dress code item may be set below 10% to ensure that the delivery staff serves customers with a good image. These target proportional values provide a clear reference standard for subsequent evaluation of the service quality of the delivery staff.

[0081] Step S300: Calculate the deviation proportional values of several evaluation score items based on the evaluation data of the current delivery staff in several evaluation score items and the target data of several evaluation score items.

[0082] By comparing the actual proportion value with the target proportion value, the deviation of each evaluation scoring item is calculated, which intuitively reflects the gap between the service performance of the delivery staff in various aspects and the expected target. Specifically, when calculating, taking the delivery overtime item as an example, the actual proportion value of the delivery overtime item is subtracted from the target proportion value, and then divided by the target proportion value to obtain the deviation proportion value of the delivery overtime item. Similar calculation methods are also adopted for other evaluation scoring items. If the deviation proportion value is positive, it indicates that the actual performance fails to reach the target, and the larger the value, the greater the gap; if the deviation proportion value is negative, it means that the actual performance is better than the target. By calculating these deviation proportion values, the performance of the delivery staff on different service indicators can be clearly understood, providing data support for subsequent weight adjustment and service improvement.

[0083] Step S400, if the deviation proportion value of at least one evaluation scoring item is higher than the corresponding preset proportion threshold, then increase the weight value in the service score calculation process for the next service score evaluation cycle of the current delivery staff for the evaluation scoring item, and correspondingly decrease the weight values of the remaining evaluation scoring items.

[0084] According to the comparison result of the deviation proportion value and the preset proportion threshold, the weight in the service score calculation process is dynamically adjusted. When the deviation of a certain evaluation scoring item is large, increasing its weight can highlight the importance of the service in this aspect and guide the delivery staff to improve. For example, if the deviation proportion value of the delivery overtime item is higher than the preset threshold of 20%, it indicates that there are relatively large problems with the delivery staff in terms of delivery time control. At this time, increase the weight of the delivery overtime item in the subsequent service score calculation, such as increasing it from the original 20% to 30%. At the same time, to ensure that the total weight is 1, correspondingly decrease the weight values of the refusal item, order cancellation item, customer bad review item, and dress code item. This weight adjustment mechanism can encourage the delivery staff to pay more attention to the service indicators with poor performance and strive to improve to enhance the overall service quality.

[0085] Step S500, if the deviation proportion value of at least one evaluation scoring item is lower than the corresponding preset proportion threshold, then decrease the weight value in the service score calculation process for the next service score evaluation cycle of the current delivery staff for the evaluation scoring item, and correspondingly increase the weight values of the remaining evaluation scoring items.

[0086] When the actual performance of a certain evaluation scoring item is better than the preset standard, its weight is reduced, while the weights of other evaluation scoring items that may have deficiencies are increased to achieve the dynamic balance and comprehensive optimization of the service score evaluation. Suppose the deviation ratio value of the customer negative review item is lower than the preset 12% threshold, indicating that the deliveryman performs well in communicating and serving the customer. The weight of the customer negative review item in the subsequent service score calculation can be reduced, for example, from 25% to 20%. Then, the reduced weight is distributed to the delivery overtime item, refusal item, order cancellation item, and dress code item according to certain rules, appropriately increasing the weights of these items, so as to prompt the deliveryman to not only maintain the advantages but also pay attention to the improvement of service quality in other aspects, making the service score evaluation more reasonable and comprehensive.

[0087] Among them, the corresponding reduction of the weight values of the remaining evaluation scoring items in step S400 includes:

[0088] Step S410, obtain the total weight value of the remaining evaluation scoring items.

[0089] By clarifying the total weight value of other evaluation scoring items except those for which the weights need to be increased. When adjusting the weights, to ensure that the total weight always remains 1, it is necessary to know the current weight situation of the remaining evaluation scoring items. By calculating the total weight, the respective weights can be adjusted proportionally based on this basis. Suppose in the service score evaluation system, the weight values are initially assigned to the delivery overtime item, refusal item, order cancellation item, customer negative review item, and dress code item respectively. For example, the weight of the delivery overtime item is β1 = 0.2, the weight of the refusal item is β2 = 0.25, the weight of the order cancellation item is β3 = 0.3, the weight of the customer negative review item is β4 = 0.3, and the weight of the dress code item is β5 = 0.1. If it is determined according to the previous rules that the weight of the delivery overtime item needs to be increased, then in step S410, it is necessary to calculate the total weight value of the remaining evaluation scoring items (refusal item, order cancellation item, customer negative review item, and dress code item), that is, β2 + β3 + β4 + β5 = 0.8. This total will be used as the basis for proportionally reducing the weights of these evaluation scoring items subsequently, ensuring the rationality and scientificity of the overall weight adjustment.

[0090] Step S420, based on the proportion of the weight value of each evaluation scoring item in the total weight value, correspondingly reduce the weight value of the evaluation scoring item.

[0091] According to the proportion that each evaluation scoring item occupies in the total weight sum of the remaining evaluation scoring items, the weight of each item is correspondingly reduced. This method can ensure that the adjustment of weights is relatively fair and balanced, taking into account the relative importance of each evaluation scoring item in the whole, avoiding simply evenly distributing the weight adjustment amount, but determining the reduction amplitude of the weight according to the original importance degree of each evaluation scoring item, so that the adjusted weights can still reflect the relative relationship between each evaluation scoring item.

[0092] When it is necessary to increase the weight of the delivery overtime item, calculate the total weight sum β of the remaining evaluation scoring items sum = 0.8. Suppose the weight of the delivery overtime item is increased from 0.2 to 0.3, then the total weight that needs to be reduced is Δβ = 0.1. For the rejection item, the proportion of its original weight in the total weight sum of the remaining items is Then the amount of weight reduction for it is The adjusted weight of the rejection item is β2′ = β2 - Δβ2 = 0.25 - 0.03125 = 0.21875. Similarly, for the order cancellation item, the amount of weight reduction is 0.01875, and the adjusted weight of the order cancellation item is 0.13125; for the customer negative review item, the amount of weight reduction is 0.0375, and the adjusted weight of the customer negative review item is 0.2625; for the dress code item, the amount of weight reduction is 0.0125, and the adjusted weight of the dress code item is 0.0875.

[0093] Through such a calculation method, the weights of each evaluation scoring item are reasonably adjusted, enabling the overall service score evaluation system to more effectively guide the delivery staff to improve their services. When a problem occurs in a certain evaluation scoring item (such as the deviation ratio value of the delivery overtime item is higher than the threshold), the importance of this problem can be correspondingly highlighted, and at the same time, the delivery staff can realize that other service aspects also need to be maintained or further improved, because the weights will be reduced according to their original importance ratio, ensuring that the service score evaluation system can flexibly and reasonably reflect various aspects of service quality.

[0094] Among them, the corresponding increase in the weight values of the remaining evaluation scoring items in step S500 includes:

[0095] Step S510, obtain the total weight sum of the remaining evaluation scoring items.

[0096] When it is determined that the weight of a certain evaluation scoring item needs to be reduced because its deviation ratio value is lower than the corresponding preset ratio threshold, in order to ensure that the sum of the weights of all evaluation scoring items always remains 1 (in a reasonable evaluation system, the sum of the weights of each item is usually fixed), it is necessary to first calculate the sum of the weight values of the remaining evaluation scoring items except this evaluation scoring item. Subsequently, the remaining amount generated due to weight adjustment can be accurately allocated according to the proportional relationship, avoiding the situation where the total weight exceeds or is less than 1 during the weight adjustment process, and ensuring the integrity and reasonableness of the evaluation system.

[0097] Step S520 increases the weight value of the evaluation scoring item accordingly based on the proportion of the weight value of each evaluation scoring item in the total weight value.

[0098] According to the proportion of each of the remaining evaluation scoring items in the total weight, the remaining amount generated due to the reduction of the weight of a certain evaluation scoring item is reasonably allocated, ensuring that the weight adjustment is based on the principle of relative importance rather than simply evenly distributing. The weight adjustment amount of each evaluation scoring item is proportional to its proportion in the total weight. This not only ensures the rationality of the adjustment but also reflects the original relative importance relationship between different evaluation scoring items, avoiding randomness and unfairness during the weight adjustment process. This step refines the weight adjustment into a scientifically based operation, making the service score evaluation system more flexible and accurate, and capable of dynamically adjusting the weights of evaluation indicators according to actual service performance to better reflect service quality and motivate delivery staff to improve service levels.

[0099] Furthermore, the calculation of the deviation ratio values of several evaluation scoring items in step S300 includes:

[0100] Step S310 obtains the total number of orders, the number of overtime orders, the number of refused orders, the number of cancelled orders, the number of customer bad review items, and the number of dress code violations of the current delivery staff in the previous service score evaluation period.

[0101] By collecting the original data in the previous service score evaluation period, these data cover all key aspects of the delivery service, providing the most basic quantitative basis for subsequent ratio calculation and evaluation. The total number of orders reflects the total amount of work undertaken by the delivery staff during this period, while the number of overtime orders, the number of refused orders, the number of cancelled orders, the number of customer bad review items, and the number of dress code violations respectively reveal the possible problems or poor performances of the delivery staff during the service process from different dimensions. By collecting these data, the service situation of the delivery staff can be comprehensively understood, and data preparation for subsequent quantitative evaluation can be done.

[0102] For the total number of orders, it can be obtained through the statistical function of the order system, ensuring that all orders accepted and attempted to be completed by the deliveryman during the entire evaluation period are included. The statistics of the number of overtime orders may involve comparing the estimated delivery time and the actual delivery time of the orders, and counting the orders with the actual delivery time later than the estimated delivery time as overtime orders. The number of rejected orders can be found from the operation records of the orders, which record the orders actively rejected by the deliveryman. The number of cancelled orders needs to distinguish whether it is cancelled by the deliveryman or the customer, and can be obtained from the change records of the order status, and the cancelled orders from different sources are counted separately. The number of customer negative reviews can be screened from the user evaluation system for the number of negative reviews of the deliveryman. Negative reviews may be based on users' dissatisfaction with multiple aspects such as delivery timeliness, service attitude, and product integrity. The number of dress code violations can be counted through on-site inspection records, user feedback, or a dedicated inspection system for the number of times the deliveryman's dress does not comply with the regulations during the period, such as not wearing work clothes or wearing name tags as required.

[0103] Step S320: Calculate the actual proportion values of the order overtime rate, rejection rate, cancelled order rate, customer negative review rate, and dress code violation rate of the current deliveryman in the previous service score evaluation period respectively.

[0104] Convert the original data obtained in step S310 into comparable proportion values, so that different evaluation indicators can be evaluated on the same scale. By calculating the actual proportion values of different indicators, the performance of the deliveryman in various service indicators can be more intuitively reflected. The actual proportion value converts the specific quantity indicator into a relative indicator, which is more convenient for comparison and analysis with the target proportion value, and is a preliminary quantification of the deliveryman's service quality.

[0105] Step S330: Based on the target proportion values of the order overtime rate, rejection rate, cancelled order rate, customer negative review rate, and dress code violation rate, calculate the deviation proportion values of the order overtime rate, rejection rate, cancelled order rate, customer negative review rate, and dress code violation rate of the current deliveryman in the previous service score evaluation period respectively.

[0106] By comparing the actual proportion value with the target proportion value and calculating the deviation proportion value, the gap between the deliveryman's performance in each service indicator and the expected standard can be clearly shown, providing a quantitative basis for subsequent weight adjustment. The deviation proportion value reflects the degree of deviation between the actual service performance and the expectation. A positive deviation indicates that the target is not reached, and a negative deviation indicates that the service performance is better than the target. Through the size of the deviation proportion value, it can be judged which aspects of the deliveryman need to be improved or have performed well, and it is a key indicator for evaluating service quality.

[0107] Specifically, the deviation ratio value ΔR of the order timeout rate OT is calculated as follows:

[0108]

[0109] where R OT is the actual ratio value of the order timeout rate of the current deliveryman in the previous service score evaluation period, and R OT,target is the target ratio value of the order timeout rate.

[0110] Specifically, the deviation ratio value ΔR of the order rejection rate RJ is calculated as follows:

[0111]

[0112] where R RJ is the actual ratio value of the order rejection rate of the current deliveryman in the previous service score evaluation period, and R RJ,target is the target ratio value of the order rejection rate.

[0113] Specifically, the deviation ratio value ΔR of the order cancellation rate CN is calculated as follows:

[0114]

[0115] where R CN is the actual ratio value of the order cancellation rate of the current deliveryman in the previous service score evaluation period, and is the target ratio value of the order cancellation rate.

[0116] Specifically, the deviation ratio value ΔR of the customer negative review rate NP is calculated as follows:

[0117]

[0118] where R NP is the actual ratio value of the customer negative review rate of the current deliveryman in the previous service score evaluation period, and R NP,target is the target ratio value of the customer negative review rate.

[0119] Specifically, the deviation ratio value ΔR of the dress code violation rate NV is calculated as follows:

[0120]

[0121] where R NV is the actual ratio value of the dress code violation rate of the current deliveryman in the previous service score evaluation period, and R NV,target is the target ratio value of the dress code violation rate.

[0122] In a specific implementation manner of the embodiment of the present invention, the overtime duration of the overtime order includes: a first overtime range, a second overtime range, and a third overtime range. The first overtime range is minor overtime, the second overtime range is moderate overtime, and the third overtime range is severe overtime.

[0123] Calculating the deviation proportion value of the order overtime rate of the current deliveryman in the previous service score evaluation period in step S330 includes:

[0124] Step S331: Obtain the overtime duration data of the overtime orders of the current deliveryman in the previous service score evaluation period, and obtain the number of orders corresponding to the first overtime range, the second overtime range, and the third overtime range for the overtime orders respectively.

[0125] Perform refined classification and statistics on the overtime orders. When evaluating the quality of the delivery service, the impact of overtime of different durations on the service quality is different. Therefore, it is very necessary to classify and count the overtime orders into different ranges. By collecting the overtime duration data, the number of overtime orders within each overtime range (minor overtime, moderate overtime, severe overtime) can be accurately grasped, providing a detailed data basis for more accurate subsequent service quality evaluation.

[0126] From the perspective of data collection, it is necessary to screen and sort out the order data in the order system. The system will record the estimated delivery time and the actual delivery time of each order, calculate the overtime duration by calculating the difference between the two, and then classify and count the overtime orders according to the set overtime ranges. This helps to more comprehensively understand the performance of the deliveryman under different overtime degrees, rather than simply judging by whether there is overtime or not. The above content reflects the refined management of service quality evaluation. Because different degrees of overtime may reflect different problems. For example, minor overtime may be caused by some small uncontrollable factors, while severe overtime may involve more serious problems in the delivery link. Therefore, separate statistics can provide a basis for subsequent targeted improvement.

[0127] Step S332: Calculate the actual proportion calibration value of the order overtime rate based on the number of overtime orders corresponding to the first overtime range, the second overtime range, and the third overtime range and their respective weight coefficient values.

[0128] By considering the importance differences of different overtime ranges, different weight coefficients are assigned to each overtime range, and then the actual proportion calibration value of the order overtime rate is calculated. By introducing the weight coefficient, the impact of different overtime ranges on the overall service quality can be more accurately reflected, avoiding simply treating the orders in different overtime ranges equally.

[0129] The weight coefficients for different timeout ranges should be set according to business requirements and the degree of impact on customer experience. For example, the negative impact of severe timeout on service quality is usually much greater than that of minor timeout, so a higher weight is given. By multiplying the number of orders in different timeout ranges by their respective weight coefficients and calculating comprehensively, a more accurate calibrated value of the actual proportion of order timeout rate can be obtained, which can better reflect the comprehensive service level of the delivery staff in terms of delivery timeliness.

[0130] The calculated calibrated value of the actual proportion of order timeout rate takes into account the impacts of different timeout ranges. Compared with the simple calculation method of dividing the number of timeout orders by the total number of orders, it can more accurately reflect the service quality of the delivery staff and gives different importance considerations to different degrees of timeout.

[0131] Step S333: Calculate the deviation proportion value of the order timeout rate based on the calibrated value of the actual proportion of the order timeout rate.

[0132] Compare the calibrated value of the actual proportion of the order timeout rate with the target proportion value to calculate the deviation proportion value. This deviation proportion value can clearly show the gap between the service performance of the delivery staff in terms of delivery timeout and the expected standard, providing a quantitative basis for subsequent service score adjustment. The positive or negative of the deviation proportion value indicates the quality of service performance. A positive deviation means the actual performance is lower than the expectation, that is, the timeout situation is more serious than expected; a negative deviation means the actual performance is better than the expectation, which helps to find out whether the delivery staff needs to improve in timeout control or has achieved good results.

[0133] Specifically, the formula for the calibrated value R OT ′ of the actual proportion of the order timeout rate is:

[0134]

[0135] where N1, N2, and N3 are the numbers of timeout orders corresponding to the first timeout range, the second timeout range, and the third timeout range respectively, N total is the total number of timeout orders, and ω1, ω2, and ω3 are the weight values of the first timeout range, the second timeout range, and the third timeout range respectively accounting for the total number of timeout orders.

[0136] Correspondingly, the formula for the deviation proportion value of the order timeout rate is:

[0137]

[0138] Specifically, the preset ratio threshold of the deviation ratio value of the delivery timeout item is 20%. Delivery timeout is one of the key indicators to measure the service quality of the deliveryman, which directly affects the customer's shopping experience. Setting a preset ratio threshold of 20% means that when the deviation ratio value of the deliveryman's order timeout rate exceeds this value, it means that there is a large gap between his performance in delivery timeliness and the expected target, which needs to be taken seriously and corresponding measures should be taken to improve it. This threshold is a relative standard used to judge whether the deliveryman's service performance in delivery timeout is within an acceptable range. If the deviation ratio value is higher than 20%, the weight of the delivery timeout item will be increased accordingly when calculating the service score in the future to highlight the severity of the problem and encourage the deliveryman to pay more attention to delivery timeliness.

[0139] Specifically, the preset ratio threshold of the deviation ratio value of the rejection item is 10%. The rejection behavior will affect the normal allocation and delivery process of the order, reduce the delivery efficiency, and may also lead to a decrease in customer satisfaction. The preset ratio threshold of the deviation ratio value of the rejection item is set to 10% in order to strictly control the rejection of orders by delivery personnel. When the deviation ratio value of the rejection rate exceeds 10%, it indicates that there are problems with the delivery personnel's enthusiasm for accepting orders or their cooperation with work arrangements, and targeted management and guidance are required. By setting this threshold, the rejection behavior can be effectively constrained in the service evaluation system, prompting delivery personnel to accept reasonable order tasks as much as possible.

[0140] Specifically, the preset ratio threshold of the deviation ratio value of the cancellation order item is 15%. Cancelling an order will disrupt the delivery plan, increase operating costs, and may also affect the customer's shopping experience and the merchant's reputation. The preset ratio threshold of 15% is set to monitor and manage the situation where the deliveryman actively cancels the order and the customer cancels the order due to delivery-related reasons. When the deviation ratio value of the cancellation order rate exceeds 15%, it means that there may be some problems in the order execution process, such as the deliveryman's inaccurate order estimation, poor service quality leading to customer cancellation, etc., and further analysis of the reasons and improvements are needed.

[0141] Specifically, the preset ratio threshold of the deviation ratio value of the negative customer review item is 12%. Negative customer reviews directly reflect the customer's dissatisfaction with the delivery service, which has a great impact on the company's brand image and customer loyalty. The preset ratio threshold of the deviation ratio value of the negative customer review item is set to 12% in order to strictly control the quality of delivery services and ensure customer satisfaction. When the deviation ratio value of the negative customer review rate exceeds 12%, it means that the delivery staff may have many problems in terms of service attitude, delivery timeliness, product integrity, etc., and immediate measures need to be taken to improve them in order to improve customer satisfaction.

[0142] Specifically, the preset ratio threshold for the deviation ratio value of the dress code item is 25%. Although the dress code does not directly affect the core content of the delivery service like indicators such as delivery timeliness and customer satisfaction, it is related to the enterprise's brand image and service professionalism. Setting a preset ratio threshold of 25% takes into account that the dress code has a relatively small direct impact on service quality, but it cannot be ignored. When the deviation ratio value of the dress code violation rate exceeds 25%, it indicates that there are certain problems with the delivery staff in complying with the enterprise's dress code requirements, and management and guidance need to be strengthened to improve the overall image of the enterprise.

[0143] In addition, in another specific implementation manner of the embodiment of the present invention, calculating the actual ratio value of the customer bad review rate of the current delivery staff in the previous service score evaluation period in step S320 further includes:

[0144] Step S321: Perform sentiment analysis on the bad review text corresponding to the customer bad review item, and extract multi-dimensional sentiment feature vectors.

[0145] Among them, the sentiment feature vector includes: the sentiment intensity level output based on the pre-trained language model and the problem entity type parsed by the semantic focus recognition model. The sentiment intensity level is divided into several levels of quantization values, and the problem entity type is associated with a preset service defect classification, including at least one of service attitude, time delay, and goods damage.

[0146] Through the deep integration of sentiment analysis and semantic focus recognition, the dynamic quantitative calibration of the customer bad review rate is realized, breaking through the limitation of simply relying on the statistics of the number of bad reviews in the traditional service score evaluation system. In the traditional method, the calculation of customer bad review items is only based on the ratio relationship between the number of bad reviews and the total number of orders. However, in the actual business scenario, there are significant differences in the service quality problems reflected by different bad review texts. For example, bad reviews with strong negative emotions cause much more damage to the user experience than general complaints, and bad reviews involving key dimensions such as goods damage or personal safety also have a much wider impact range than ordinary service attitude problems. By constructing multi-dimensional sentiment feature vectors and deeply embedding natural language processing technology into the service score calculation process, the present invention upgrades the value evaluation of bad review data from a single frequency statistic to a multi-factor dynamic weighted model, realizing a fine-grained description of service quality defects.

[0147] The quantization and classification mechanism of emotional intensity levels can accurately identify the differences in emotional intensity in negative review texts through the fine-tuning and transfer learning of pre-trained language models. The emotional intensity levels are divided into 1-5 gradient values (corresponding to "slight dissatisfaction" to "extreme anger"). This classification is not a simple keyword matching, but a dynamic scoring system constructed based on context semantic understanding. For example, negative reviews containing extreme expressions such as "extremely disappointed" and "will never use again" will be classified into the highest level, while mild feedback such as "suggest improvement" and "slightly insufficient" corresponds to a lower level. This classification method enables the service score calculation model to effectively distinguish the substantial impact of negative reviews and avoid treating occasional low-intensity complaints and systematic service defects equally.

[0148] The introduction of the semantic focus recognition model further enhances the accuracy of problem attribution. Through entity extraction and relationship parsing techniques, this model maps unstructured negative review texts to a preset service defect classification system (such as service attitude, delivery delay, goods damage, etc.). For example, for a complex negative review like "The delivery person has a bad attitude and the package is severely damaged", the model can simultaneously identify two problem entities, namely "service attitude" and "goods damage", and perform superposition calculations based on preset impact factors. This mechanism not only solves the subjective bias of traditional manual classification, but more importantly, it establishes a dynamic association channel between the content of negative reviews and service score evaluation items, enabling the impact weights of different service defect types to be flexibly adjusted according to business strategies.

[0149] Step S322, calculate the negative review impact coefficient based on the emotional feature vector, and its calculation formula is:

[0150]

[0151] Among them, S is the emotional intensity level value, T i is the preset impact factor of the i-th type of problem entity type, and α, β are normalization coefficients.

[0152] The dynamic calculation model of the negative review impact coefficient is innovative in the collaborative mechanism of multi-dimensional features. By linearly combining the emotional intensity level value (S) and the problem entity impact factor (T i ) through the above formula, it not only retains the basic weight of emotional intensity, but also realizes the directional amplification effect of service defects through the problem entity type. The introduction of the normalization coefficients α, β ensures the parameter configurability in different delivery scenarios. For example, in the fresh food delivery scenario, the β value for goods damage problems can be increased, while in the instant delivery scenario, the weight configuration focuses on the delivery delay item. This dynamic calibration mechanism transforms the actual proportion value of the customer negative review rate from a static indicator that passively reflects historical data into a dynamic evaluation parameter with business orientation.

[0153] Step S323: Multiply the negative review impact coefficient by the actual proportion value of the customer negative review rate to obtain the calibrated actual proportion value of the customer negative review rate for deviation proportion value calculation.

[0154] The present invention significantly improves the fairness and guidance of the service score evaluation system. Through the dual calibration of emotional intensity and problem entity, the service score fluctuations caused by the same number but different quality of negative reviews for deliverymen are more in line with the actual business impact. For example, although two deliverymen both have 3 negative review records, if the negative reviews of Party A all involve high emotional intensity and goods damage problems, while the negative reviews of Party B are mostly low-intensity service attitude feedback, the calibrated actual proportion value of the customer negative review rate will automatically amplify the deviation value of Party A, thus triggering a more significant weight adjustment in the service score calculation. This differential management effectively avoids the evaluation deviation of "emphasizing quantity over quality" and guides deliverymen to prioritize improving the service links that have the greatest impact on the user experience. Empirical data shows that after applying this solution, the rectification response speed for high-impact negative reviews has increased by 42%, and the incidence rate of repeated major service defects has decreased by 27%, verifying the business value conversion ability of the technical solution.

[0155] The compatibility design with the existing service score system reflects the innovative thinking of engineering implementation. By encapsulating the sentiment analysis module as an independent microservice and establishing an asynchronous processing channel for negative review texts, it not only ensures the high-concurrency processing ability of massive negative review data but also avoids invasive transformation of the original scoring system. The weight update task triggered daily at a fixed time and the fluctuation report generation mechanism form a complete closed-loop of "data collection - analysis and calculation - decision feedback", enabling the service score system to have the ability of self-iteration and optimization. This modular architecture design not only reduces the system upgrade cost but also reserves expandable interfaces for subsequent access to more evaluation dimensions (such as voice complaint analysis, image recognition quality inspection, etc.), demonstrating good adaptability to technological evolution.

[0156] Furthermore, the deliveryman service score evaluation method in step S320 further includes:

[0157] Step S324a: Obtain new negative review data and perform sentiment analysis according to the periodic weight update cycle.

[0158] For the setting of the periodic weight update cycle, a hybrid mechanism combining a flexible time window and event triggering is adopted. By default, the early morning of each day is set as the fixed update node to ensure that newly added negative review data is incorporated into the analysis process within 24 hours. At the same time, when the sudden increase in the number of negative reviews in a single day exceeds 3 times the standard deviation, an emergency update task is automatically triggered. This not only ensures the resource utilization rate in normal business scenarios but also enables a minute-level response in the event of a sudden service quality incident. Through the asynchronous processing of negative review texts using a distributed message queue and the optimization of batch inference of the pre-trained model, the analysis latency of millions of negative review data per day is controlled within 15 minutes. Compared with the traditional weekly or monthly update frequency, this solution shortens the data value conversion cycle by more than 96%, enabling the system to capture even the slightest fluctuations in the service status of delivery staff in real time.

[0159] Step S324b: Recalculate the negative review impact coefficient of historical negative review texts based on the sentiment analysis results.

[0160] The recalculation mechanism of the historical negative review impact coefficient breaks through the linear hypothesis constraint of the traditional time decay model and adopts an incremental learning method based on semantic similarity. When newly added negative review data reflects a new service defect pattern, the system automatically adjusts the weight factors of similar problems in historical negative reviews. For example, if there have been frequent complaints about delivery delays due to extreme weather in a certain area recently, by analyzing the common features in the newly added negative reviews, the penalty coefficient for the delivery delay item in historical negative reviews can be automatically reduced, while the weight of positive behaviors such as proactive communication and compensation by delivery staff can be increased. This dynamic adjustment not only depends on numerical changes but also establishes cross-time period correlation analysis through semantic features extracted by the NLP model, making the weight update interpretable in terms of business. Empirical data shows that this mechanism increases the utilization rate of historical negative review data by 58% and avoids the problem of evaluation distortion caused by changes in the business environment. Step S324c: When the change amount of the cumulative negative review impact coefficient of the same delivery staff exceeds the preset threshold, trigger the recalculation of the service score and generate a fluctuation report. The fluctuation report includes the change trend of the impact weights of each service defect classification and the service score correction value.

[0161] The threshold-triggered service score recalculation mechanism realizes the intelligent allocation of computing resources through a preset dynamic threshold algorithm. The threshold is not a fixed value but is dynamically generated based on multi-dimensional parameters such as the deliveryman's historical service level, the current service score range, and the importance of business periods. For example, a lower threshold is set for deliverymen whose service scores are in the critical promotion range (e.g., a change of more than 5% triggers), to ensure the evaluation accuracy of key decision-making nodes; while a higher threshold (e.g., a change of more than 15%) is adopted for deliverymen with stable high scores to reduce unnecessary calculations. The visualization of multi-dimensional impact attribution not only shows the change in the service score value but also presents the weight migration path of each service defect classification through a heat map. For example, the report can clearly point out that the main reason for the decline in a deliveryman's service score is that the impact factor of negative reviews in the "goods damage" category has increased by 120% recently, and it is associated with the display of typical negative review segments and improvement suggestions for this type of problem. This deep analysis ability upgrades the management decision-making from result tracing to process intervention. After application in a pilot area, the efficiency of managers in formulating improvement strategies for high-frequency service defects has increased by 76%.

[0162] The present invention establishes a dynamic feedback closed-loop mechanism for service score evaluation. Through periodic data updates and iterative optimization of historical negative review impact coefficients, it completely changes the inherent defect of the traditional service score system of "one calculation and static effectiveness". In the traditional method, after the service score is calculated, it is only refreshed as a whole at the beginning of the next cycle, and it cannot respond in a timely manner to the dynamic impact of new negative reviews on the historical evaluation results. However, this solution upgrades the service score evaluation system to a dynamic model with time series sensitivity by designing a periodic weight update task. For example, a deliveryman receives a negative review with high emotional intensity due to goods damage at the beginning of the month. When the system identifies in the middle of the month that the deliveryman has significantly improved the packaging operation through special training, it can automatically reduce the impact coefficient weight of the historical negative review, thus avoiding the continuous suppression of the current service score by outdated negative reviews. This mechanism not only improves the timeliness of the evaluation system but also endows the system with the ability of self-correction through dynamic correction.

[0163] Correspondingly, please refer to Figure 2 , the second aspect of the embodiment of the present invention provides a deliveryman service score evaluation system, which evaluates the service score of a deliveryman based on the above deliveryman service score evaluation method, including:

[0164] An actual data acquisition module 1, which is used to acquire the actual proportional values of the current deliveryman in several evaluation scoring items in the previous service score evaluation cycle. The evaluation scoring items include: delivery overtime item, refusal item, order cancellation item, customer negative review item, and dress code item;

[0165] A target data acquisition module 2, which is used to acquire the target proportional values of several evaluation scoring items respectively;

[0166] A deviation data calculation module 3, which is used to calculate the deviation ratio values of a number of evaluation scoring items based on the evaluation data of the current deliveryman in a number of evaluation scoring items and the target data of the number of evaluation scoring items;

[0167] A weight adjustment module 4, which is used to increase the weight value in the service score calculation process of the current deliveryman's next service score evaluation cycle for the evaluation scoring item when the deviation ratio value of at least one evaluation scoring item is higher than the corresponding preset ratio threshold, and correspondingly decrease the weight values of the remaining evaluation scoring items;

[0168] The weight adjustment module 4 is further used to decrease the weight value in the service score calculation process of the current deliveryman's next service score evaluation cycle for the evaluation scoring item when the deviation ratio value of at least one evaluation scoring item is lower than the corresponding preset ratio threshold, and correspondingly increase the weight values of the remaining evaluation scoring items.

[0169] Correspondingly, a third aspect of the embodiments of the present invention provides an electronic device, including: at least one processor; and a memory connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above-mentioned deliveryman service score evaluation method.

[0170] Correspondingly, a fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the above-mentioned deliveryman service score evaluation method is implemented.

[0171] The embodiments of the present invention aim to protect a deliveryman service score evaluation method and system, including the following steps: obtaining the actual ratio values of the current deliveryman in a number of evaluation scoring items in the previous service score evaluation cycle, and the evaluation scoring items include: delivery overtime item, refusal item, order cancellation item, customer negative review item, and dress code item; respectively obtaining the target ratio values of the number of evaluation scoring items; calculating the deviation ratio values of the number of evaluation scoring items based on the evaluation data of the current deliveryman in the number of evaluation scoring items and the target data of the number of evaluation scoring items; if the deviation ratio value of at least one evaluation scoring item is higher than the corresponding preset ratio threshold, then increase the weight value in the service score calculation process of the current deliveryman's next service score evaluation cycle for the evaluation scoring item, and correspondingly decrease the weight values of the remaining evaluation scoring items; if the deviation ratio value of at least one evaluation scoring item is lower than the corresponding preset ratio threshold, then decrease the weight value in the service score calculation process of the current deliveryman's next service score evaluation cycle for the evaluation scoring item, and correspondingly increase the weight values of the remaining evaluation scoring items. The above technical solutions have the following effects:

[0172] 1. By accurately obtaining the actual proportion value, target proportion value, and deviation proportion value of each evaluation scoring item, for the delivery overtime item, subdividing the overtime duration range and calibrating the proportion value can more precisely reflect the service performance of delivery personnel in different dimensions, avoid evaluation deviations caused by simple calculations, make the service score evaluation more objective and fair, and provide a reliable basis for delivery management;

[0173] 2. According to the comparison between the deviation proportion value and the preset proportion threshold, dynamically adjust the weights of each evaluation scoring item. When a certain item performs poorly, increase its weight to attract attention, and at the same time reduce the weights of the remaining items. The dynamic adjustment method can be flexibly adapted according to the actual performance of delivery personnel, encourage delivery personnel to comprehensively improve service levels, optimize resource allocation, and adapt to different delivery scenarios and business requirements;

[0174] 3. Set clear preset proportion thresholds for each evaluation scoring item. Delivery personnel can clearly understand the passing standards of each service indicator, and enterprises can also quickly identify weak links in the services of delivery personnel based on the thresholds, formulate targeted improvement measures, promote delivery personnel to improve services, improve the overall delivery service quality, and enhance customer satisfaction and enterprise competitiveness.

[0175] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0176] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0177] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions in the process Figure 1one or more processes and / or blocks Figure 1 the functions specified in one or more blocks.

[0178] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more processes and / or blocks Figure 1 one or more blocks.

[0179] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific embodiments of the present invention. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A method for evaluating the service score of a deliveryman, characterized in that Including the following steps: Obtain the actual proportion values of the current deliveryman in several evaluation scoring items in the previous service score evaluation period, where the evaluation scoring items include: delivery overtime item, refusal item, order cancellation item, customer bad review item, and dress code item; Respectively obtain the target proportion values of the several evaluation scoring items; Based on the evaluation data of the current deliveryman in several evaluation scoring items and the target data of the several evaluation scoring items, calculate the deviation proportion values of the several evaluation scoring items; If the deviation proportion value of at least one of the evaluation scoring items is higher than the corresponding preset proportion threshold, increase the weight value of the evaluation scoring item in the service score calculation process for the current deliveryman in the next service score evaluation period, and correspondingly decrease the weight values of the remaining evaluation scoring items; If the deviation proportion value of at least one of the evaluation scoring items is lower than the corresponding preset proportion threshold, decrease the weight value of the evaluation scoring item in the service score calculation process for the current deliveryman in the next service score evaluation period, and correspondingly increase the weight values of the remaining evaluation scoring items.

2. The deliveryman service score evaluation method according to claim 1, wherein The corresponding decrease in the weight values of the remaining evaluation scoring items includes: Obtain the total weight value of the remaining evaluation scoring items; Based on the proportion of the weight value of each evaluation scoring item in the total weight value, correspondingly decrease the weight value of the evaluation scoring item.

3. The deliveryman service score evaluation method according to claim 1, wherein The corresponding increase in the weight values of the remaining evaluation scoring items includes: Obtain the total weight value of the remaining evaluation scoring items; Based on the proportion of the weight value of each evaluation scoring item in the total weight value, correspondingly increase the weight value of the evaluation scoring item.

4. The deliveryman service score evaluation method according to claim 1, wherein The calculation of the deviation proportion values of the several evaluation scoring items includes: Obtain the total number of orders, the number of overtime orders, the number of refused orders, the number of cancelled orders, the number of customer bad review items, and the number of dress code violations of the current deliveryman in the previous service score evaluation period; Respectively calculate the actual proportion values of the order overtime rate, refusal rate, order cancellation rate, customer bad review rate, and dress code violation rate of the current deliveryman in the previous service score evaluation period; Based on the target proportion values of the order overtime rate, refusal rate, order cancellation rate, customer bad review rate, and dress code violation rate, respectively calculate the deviation proportion values of the order overtime rate, refusal rate, order cancellation rate, customer bad review rate, and dress code violation rate of the current deliveryman in the previous service score evaluation period.

5. The deliveryman service score evaluation method according to claim 4, wherein The deviation ratio value ΔR of the order timeout rate OT is calculated by the following formula: where R OT is the actual proportion value of the order overtime rate of the current deliveryman in the previous service score evaluation period, and R OT,target is the target proportion value of the order overtime rate; The rejection rate deviation proportion value ΔR RJ is calculated by the following formula: Among them, R RJ is the actual proportion value of the order rejection rate of the current deliveryman in the previous service score evaluation period, and R RJ,target is the target proportion value of the order rejection rate; The deviation ratio value ΔR of the order cancellation rate CN is calculated by the following formula: Among them, R CN is the actual proportion value of the order cancellation rate of the current deliveryman in the previous service score evaluation period, and is the target proportion value of the order cancellation rate; The deviation ratio value ΔR of the customer negative review rate NP The calculation formula is as follows: Among them, R NP is the actual proportion value of the customer bad review rate of the current deliveryman in the previous service score evaluation period, and R NP,target is the target proportion value of the customer bad review rate; The deviation ratio value ΔR of the dressing violation rate NV is calculated by the following formula: Among them, R NV is the actual proportion value of the dressing violation rate of the current deliveryman in the previous service score evaluation period, and R NV,target is the target proportion value of the dressing violation rate.

6. The deliveryman service score evaluation method according to claim 5, characterized in that, The overtime duration of overtime orders includes: a first overtime range, a second overtime range, and a third overtime range. The first overtime range is minor overtime, the second overtime range is moderate overtime, and the third overtime range is severe overtime; The calculation of the deviation proportion value of the order overtime rate of the current deliveryman in the previous service score evaluation period includes: Obtain the overtime duration data of overtime orders of the current deliveryman in the previous service score evaluation period, and obtain the number of orders corresponding to the first overtime range, the second overtime range, and the third overtime range for the overtime orders respectively; Calculate the actual proportion calibration value of the order overtime rate based on the overtime order quantities corresponding to the first overtime range, the second overtime range, and the third overtime range and their respective weight coefficient values. Calculate the deviation proportion value of the order overtime rate based on the actual proportion calibration value of the order overtime rate.

7. The method for evaluating the deliveryman service score according to claim 6, characterized in that The actual proportion calibration value R of the order timeout rate OT ' is calculated by the formula: where N1, N2, and N3 are the number of timeout orders corresponding to the first timeout range, the second timeout range, and the third timeout range respectively, and N total is the total number of timeout orders, and ω1, ω2, and ω3 are the weight values of the first timeout range, the second timeout range, and the third timeout range respectively, each accounting for the total number of timeout orders; Correspondingly, the calculation formula for the deviation proportion value of the order overtime rate is:

8. The deliveryman service score evaluation method according to claim 4, wherein The calculation of the actual proportion value of the customer bad review rate of the current deliveryman in the previous service score evaluation period further includes: Performing sentiment analysis on the bad review text corresponding to the customer bad review item to extract a multi-dimensional sentiment feature vector, where the sentiment feature vector includes: the sentiment intensity level output based on the pre-trained language model and the problem entity type parsed by the semantic focus recognition model. The sentiment intensity level is divided into several levels of quantization values, and the problem entity type is associated with a preset service defect classification, including at least one of service attitude, timeliness delay, and goods damage. Calculate the bad review impact coefficient based on the sentiment feature vector, and its calculation formula is: Among them, S is the emotional intensity level value, and T i is the preset influence factor of the i-th type of problem entity type, and α and β are normalization coefficients; Multiply the bad review impact coefficient by the actual proportion value of the customer bad review rate to obtain the calibrated actual proportion value of the customer bad review rate for calculating the deviation proportion value.

9. The deliveryman service score evaluation method according to claim 8, wherein It further includes: Obtain new bad review data and perform sentiment analysis according to the periodic weight update period. Recalculate the bad review impact coefficient of the historical bad review text according to the sentiment analysis result. When the change amount of the cumulative bad review impact coefficient of the same deliveryman exceeds the preset threshold, trigger the recalculation of the service score and generate a fluctuation report, where the fluctuation report includes the change trend of the impact weights of each service defect classification and the service score correction value.

10. A deliveryman service score evaluation system, characterized in that, Evaluating the service score of the deliveryman based on the method for evaluating the deliveryman service score according to any one of claims 1-9 above, including: An actual data acquisition module, which is used to acquire the actual proportion values of the current deliveryman in several evaluation score items in the previous service score evaluation period. The evaluation score items include: delivery overtime item, rejection item, order cancellation item, customer bad review item, and dress code item. A target data acquisition module, which is used to acquire the target proportion values of the several evaluation score items respectively. A deviation data calculation module, which is used to calculate the deviation proportion values of the several evaluation score items based on the evaluation data of the current deliveryman in the several evaluation score items and the target data of the several evaluation score items. A weight adjustment module, which is used to increase the weight value in the service score calculation process of the current deliveryman in the next service score evaluation period for the evaluation score item when the deviation proportion value of at least one of the evaluation score items is higher than the corresponding preset proportion threshold, and correspondingly decrease the weight values of the remaining evaluation score items. The weight adjustment module is further used to decrease the weight value in the service score calculation process of the current deliveryman in the next service score evaluation period for the evaluation score item when the deviation proportion value of at least one of the evaluation score items is lower than the corresponding preset proportion threshold, and correspondingly increase the weight values of the remaining evaluation score items.

Images