CN116822700A

CN116822700A - Logistics customer order loss prevention method, device, equipment and storage medium

Info

Publication number: CN116822700A
Application number: CN202310437780.5A
Authority: CN
Inventors: 刘超
Original assignee: Dongpu Software Co Ltd
Current assignee: Dongpu Software Co Ltd
Priority date: 2023-04-21
Filing date: 2023-04-21
Publication date: 2023-09-29

Abstract

The invention relates to the technical field of data processing, in particular to a method, a device, equipment and a storage medium for preventing logistics customer order loss, which are characterized in that a customer loss prediction model is established according to key factors and customer characteristics, future orders are predicted through the customer loss prediction model to obtain customer loss prediction results, customers which are likely to be lost can be accurately predicted, root causes which are likely to be lost and customer demands corresponding to the root causes are found, corresponding prevention schemes are extracted from a preset processing scheme library according to the customer loss prediction results, the prevention schemes are sent to corresponding network points, the network points can refer to the prevention schemes to carry out rapid maintenance processing, different maintenance strategies are adopted aiming at the customers which are likely to be lost with different demands, the maintenance effect is improved, and further the experience and satisfaction of the customers are improved.

Description

Logistics customer order loss prevention method, device, equipment and storage medium

Technical Field

The present invention relates to the field of data processing technologies, and in particular, to a method, an apparatus, a device, and a storage medium for preventing loss of a physical distribution customer order.

Background

The slip of the consignment of the enterprise clients may be an operation problem, or may be caused by that other logistics service providers are changed for some reasons, if the later clients can be found early, corresponding measures can be taken for recovery, when a certain client which is placed for many times has not been placed for a long time, namely, the client is characterized as a client which is likely to be lost, the loss caused by the loss of the client is difficult to estimate, and the prior art is difficult to predict in advance what reason the client is likely to be lost, so that the client cannot be maintained in time, and the corresponding measures cannot be adopted for maintaining the reason of the loss of the client.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention aims to provide a physical distribution customer order loss prevention method, a physical distribution customer order loss prevention device, physical distribution customer order loss prevention equipment and a physical distribution customer order loss prevention storage medium, wherein the physical distribution customer order loss prevention method, the physical distribution customer order loss prevention equipment and the physical distribution customer order loss prevention equipment can accurately predict customers which are likely to be lost and find the root cause of the customer loss and the customer demand corresponding to the root cause, and different maintenance strategies are adopted for the customers which are likely to be lost according to different demands, so that the maintenance effect is improved, and the experience and the satisfaction of the customers are improved.

The first aspect of the present invention provides a method for preventing loss of a physical distribution customer order, comprising: acquiring historical logistics order data, and cleaning the historical logistics order data to obtain historical logistics order effective data; acquiring historical abnormal condition data, and determining key factors of customer loss according to the historical abnormal condition data and the historical logistics order effective data; establishing a customer loss prediction model according to the key factors and the customer characteristics, and predicting a future order through the customer loss prediction model to obtain a customer loss prediction result; and extracting a corresponding prevention scheme from a preset processing scheme library according to the customer loss prediction result, and sending the prevention scheme to a corresponding website.

Optionally, in a first implementation manner of the first aspect of the present invention, the acquiring historical logistics order data, and cleaning the historical logistics order data to obtain historical logistics order valid data includes: acquiring historical logistics order data, and performing outlier rejection, missing data filling and denoising filtering processing on the historical logistics order data to obtain first historical logistics order processing data; identifying an order generation time field of the first historical logistics order processing data; sequencing the first historical logistics order processing data according to the order generation time field to obtain second historical logistics order processing data; identifying a customer name field of the second historical logistics order processing data; and grouping the second historical logistics order processing data according to the customer name field to obtain the historical logistics order effective data.

Optionally, in a second implementation manner of the first aspect of the present invention, the obtaining historical abnormal situation data and determining a key factor of customer loss according to the historical abnormal situation data and the historical logistics order valid data includes: screening the effective data of the historical logistics orders to obtain client group data of which the order number is in a preset first threshold value; identifying the generation time of the last order of each client in the client group data, and acquiring real-time; judging whether the time period between the generation time and the real-time of the last order is satisfied with a preset second threshold value or not; if yes, marking the client as an inactive state client, and marking the generation time of the last order as a loss time point; acquiring historical abnormal condition data, and identifying an abnormal record event corresponding to the loss time point in the historical abnormal condition data; and determining key factors of customer churn according to the abnormal record event, and acquiring the customer characteristics of the inactive state customer.

Optionally, in a third implementation manner of the first aspect of the present invention, the acquiring historical abnormal situation data, identifying an abnormal record event corresponding to the churn time point in the historical abnormal situation data includes: acquiring historical abnormal condition data, and identifying the network point of the inactive state client; extracting corresponding abnormal condition data of the network point in the historical abnormal condition data, wherein the abnormal condition data of the network point comprises customer complaint records, dispatch delay records, cargo loss records, network point order error records, data loss records, abnormal climate records, public control records and three-party outage records; identifying a record generation time field in the network point abnormal situation data; and matching the record generation time field with the loss time point to obtain a corresponding abnormal record event.

Optionally, in a fourth implementation manner of the first aspect of the present invention, the determining a key factor of customer churn according to the anomaly record event, and obtaining a customer feature of the inactive state customer includes: determining key factors of customer churn according to the abnormal record events, wherein each abnormal record event corresponds to one key factor, and the key factors comprise service attitude problems, delivery time problems, delivery quality problems, ordering problems, information security problems and unreliability problems; extracting data corresponding to the inactive state clients from the client group data to obtain inactive state client data; and acquiring the client characteristics of the inactive state client from the inactive state client data, wherein the client characteristics comprise the type of the article, the type of the express delivery, the delivery address, the order frequency, the order quantity and the order amount.

Optionally, in a fifth implementation manner of the first aspect of the present invention, the establishing a customer loss prediction model according to the key factors and the customer characteristics, predicting a future order by using the customer loss prediction model, to obtain a customer loss prediction result, includes: pairing key factors corresponding to each inactive state client with client features corresponding to the inactive state clients to obtain a pairing relation; integrating all the pairing relations to obtain a pairing relation library; establishing a customer loss prediction model according to the pairing relation library; and inputting the customer characteristic data of the new order into the customer loss prediction model as an input variable to obtain a customer loss prediction result.

Optionally, in a sixth implementation manner of the first aspect of the present invention, the extracting, according to a client churn prediction result, a corresponding prevention scheme from a preset processing scheme library, and sending the prevention scheme to a corresponding website includes: invoking a pre-trained processing scheme library, and inputting the customer loss prediction result into the processing scheme library for processing policy matching to obtain a matching result; extracting a corresponding reference processing scheme from the processing scheme library according to the matching result, and generating the reference processing scheme into a prevention scheme in a text form; and sending the prevention scheme to the corresponding network point.

The second aspect of the present invention provides a physical distribution customer order loss prevention apparatus, comprising: the cleaning module is used for acquiring historical logistics order data, cleaning the historical logistics order data and obtaining historical logistics order effective data; the acquisition determining module is used for acquiring historical abnormal condition data and determining key factors of customer loss according to the historical abnormal condition data and the historical logistics order effective data; the establishment prediction module is used for establishing a customer loss prediction model according to the key factors and the customer characteristics, and predicting future orders through the customer loss prediction model to obtain a customer loss prediction result; and the extraction and transmission module is used for extracting a corresponding prevention scheme from a preset processing scheme library according to the customer loss prediction result and transmitting the prevention scheme to a corresponding website.

Optionally, in a first implementation manner of the second aspect of the present invention, the acquiring and cleaning module includes: the acquisition processing unit is used for acquiring historical logistics order data, and performing outlier rejection, missing data filling and denoising filtering processing on the historical logistics order data to obtain first historical logistics order processing data; a first identifying unit, configured to identify an order generation time field of the first historical logistics order processing data; the ordering unit is used for ordering the first historical logistics order processing data according to the order generation time field to obtain second historical logistics order processing data; a second identifying unit, configured to identify a customer name field of the second historical logistics order processing data; and the grouping unit is used for grouping the second historical logistics order processing data according to the customer name field to obtain the historical logistics order effective data.

Optionally, in a second implementation manner of the second aspect of the present invention, the acquisition determining module includes: the screening unit is used for screening the effective data of the historical logistics orders to obtain client group data of which the order number is satisfied with a preset first threshold value; the identification acquisition unit is used for identifying the generation time of the last order of each client in the client group data and acquiring real-time; the judging unit is used for judging whether the time period between the generation time and the real-time of the last order is satisfied with a preset second threshold value; the marking unit is used for marking the client as an inactive state client if yes, and marking the generation time of the last order as a loss time point; the acquisition and identification unit is used for acquiring historical abnormal condition data and identifying an abnormal record event corresponding to the loss time point in the historical abnormal condition data; and the determining and acquiring unit is used for determining key factors of customer churn according to the abnormal record event and acquiring the customer characteristics of the inactive state customer.

Optionally, in a third implementation manner of the second aspect of the present invention, the acquiring and identifying unit is specifically configured to acquire historical abnormal situation data, and identify a website where the inactive state client is located; extracting corresponding abnormal condition data of the network point in the historical abnormal condition data, wherein the abnormal condition data of the network point comprises customer complaint records, dispatch delay records, cargo loss records, network point order error records, data loss records, abnormal climate records, public control records and three-party outage records; identifying a record generation time field in the network point abnormal situation data; and matching the record generation time field with the loss time point to obtain a corresponding abnormal record event.

Optionally, in a fourth implementation manner of the second aspect of the present invention, the determining and obtaining unit is specifically configured to determine a key factor of customer loss according to the abnormal recording events, where each abnormal recording event corresponds to one of the key factors, and the key factors include a service attitude problem, a delivery time problem, a delivery quality problem, an ordering problem, an information security problem, and an unreliability problem; extracting data corresponding to the inactive state clients from the client group data to obtain inactive state client data; and acquiring the client characteristics of the inactive state client from the inactive state client data, wherein the client characteristics comprise the type of the article, the type of the express delivery, the delivery address, the order frequency, the order quantity and the order amount.

Optionally, in a fifth implementation manner of the second aspect of the present invention, the establishing a prediction module includes: the pairing unit is used for pairing the key factors corresponding to each inactive state client with the client characteristics corresponding to the inactive state clients to obtain a pairing relationship; the integration unit is used for integrating all the pairing relations to obtain a pairing relation library; the establishing unit is used for establishing a customer loss prediction model according to the pairing relation library; and the input unit is used for inputting the customer characteristic data of the new order into the customer loss prediction model as an input variable to obtain a customer loss prediction result.

Optionally, in a sixth implementation manner of the second aspect of the present invention, the extracting and sending module includes: the call matching unit is used for calling a pre-trained processing scheme library, inputting the client loss prediction result into the processing scheme library for processing policy matching, and obtaining a matching result; the extraction generating unit is used for extracting a corresponding reference processing scheme from the processing scheme library according to the matching result and generating the reference processing scheme into a prevention scheme in a text form; and the transmitting unit is used for transmitting the prevention scheme to the corresponding network point.

A third aspect of the present invention provides a physical distribution customer order loss prevention apparatus comprising: a memory and at least one processor, the memory having instructions stored therein; at least one of the processors invokes the instructions in the memory to cause the customer order loss prevention device to perform the steps of the customer order loss prevention method of any of the above.

A fourth aspect of the present invention provides a computer readable storage medium having instructions stored thereon which when executed by a processor perform the steps of the method of preventing loss of a physical distribution customer order of any one of the above.

According to the technical scheme, a customer loss prediction model is established according to key factors and customer characteristics, future orders are predicted through the customer loss prediction model to obtain customer loss prediction results, customers which are likely to be lost can be accurately predicted, root causes of the customer loss and customer demands corresponding to the root causes are found, corresponding prevention schemes are extracted from a preset processing scheme library according to the customer loss prediction results, the prevention schemes are sent to corresponding network points, the network points can refer to the prevention schemes to carry out rapid maintenance processing, different maintenance strategies are adopted for the customers which are likely to be lost with different demands, the maintenance effect is improved, and the experience and satisfaction of the customers are further improved.

Drawings

FIG. 1 is a first flowchart of a method for preventing loss of a customer order according to an embodiment of the present invention;

FIG. 2 is a second flowchart of a method for preventing loss of a customer order according to an embodiment of the present invention;

FIG. 3 is a third flow chart of a method for preventing loss of a customer order according to an embodiment of the present invention;

FIG. 4 is a fourth flowchart of a method for preventing loss of a customer order according to an embodiment of the present invention;

Fig. 5 is a schematic structural diagram of a device for preventing loss of a physical distribution customer order according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of another embodiment of a customer order loss prevention apparatus according to the present invention;

fig. 7 is a schematic structural diagram of a customer order loss prevention apparatus according to an embodiment of the present invention.

Detailed Description

The invention provides a method, a device, equipment and a storage medium for preventing logistics customer order loss, which can accurately predict customers which are likely to be lost, find out root causes of customer loss and customer demands corresponding to the root causes, adopt different maintenance strategies for the customers which are likely to be lost with different demands, improve maintenance effects, and further improve experience and satisfaction of the customers.

The terms "first," "second," "third," "fourth" and the like in the description and in the claims and in the above drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments described herein may be implemented in other sequences than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.

For ease of understanding, a specific flow of an embodiment of the present invention is described below with reference to fig. 1, where an embodiment of a method for preventing loss of a customer order in a physical distribution system according to the embodiment of the present invention includes:

101. acquiring historical logistics order data, and cleaning the historical logistics order data to obtain historical logistics order effective data;

in this embodiment, the logistics order data refers to data including key information such as product information, receiving and sending information, order number, cargo weight, order amount, and logistics path generated in the logistics field, but because the sources of the historical order data are different and the formats are various, the data needs to be cleaned to obtain effective data, and the data cleaning refers to a process of making the original data into usable data which meets the requirements and has no errors through a series of operations, and mainly includes the following three cases: 1. abnormal value elimination, namely deleting or replacing abnormal data and outliers from a data set to ensure that the obtained effective data are real and reliable, wherein in actual enterprise operation, data abnormal values caused by abnormal behaviors of various logistics links, such as damage or recording errors of goods, and the like, possibly exist, and the abnormal values influence data analysis and processing, so that the abnormal values need to be eliminated or corrected, and the specific practice can be realized through professional software or algorithm identification, manual judgment and other modes; 2. filling missing data, if the data is missing, performing data filling by adopting methods such as interpolation algorithm, regression analysis and the like to enable the data set to be more complete, and in a specific original data record, sometimes some data is empty or missing, if the data with empty value or missing is directly used, tilting or error results are caused by data analysis, so that filling strategies are needed to be adopted to enable the data set to be more complete, and performing data filling by adopting methods such as interpolation algorithm, regression analysis and the like; 3. the denoising and filtering processing is carried out on the data by using a digital signal processing technology, noise interference is removed, data quality is improved, the historical logistics order data of enterprises often have noise and singal to a certain extent, the meaning of the data is often influenced when the data is intuitively expressed, the data can be subjected to the smoothing processing based on the digital signal processing technology, the noise interference is removed, and the data quality is optimized.

102. Acquiring historical abnormal condition data, and determining key factors of customer loss according to the historical abnormal condition data and the historical logistics order effective data;

in this embodiment, historical abnormal condition data is obtained, the historical abnormal condition data is compared with the effective data of the historical logistics order, the historical abnormal record is matched with the historical order record, abnormal events related to the order abnormality, such as complaints, delay in delivery time, damage to goods, loss of goods, network point delivery errors, data loss, abnormal climate, public control, three-party outage and the like, are found, and are one of key factors causing customer loss, and the abnormal event corresponding to the time point where the historical order record is located can be determined to be the key factor causing customer loss.

103. Establishing a customer loss prediction model according to the key factors and the customer characteristics, and predicting a future order through the customer loss prediction model to obtain a customer loss prediction result;

In this embodiment, key factors corresponding to each inactive state customer are paired with customer features corresponding to the inactive state customer to obtain pairing relationships, all pairing relationships are integrated to obtain a pairing relationship library, a statistical learning method and a machine learning technology are applied, a customer loss prediction model is built by combining relevant customer information retained in a pairing database, when a new order arrives, customer feature data in the new order is input into the built customer loss prediction model as input variables to obtain a customer loss prediction result, namely, customers of the new order can be lost for a certain reason, and further, advanced maintenance can be performed for the reason.

104. Extracting a corresponding prevention scheme from a preset processing scheme library according to a customer loss prediction result, and sending the prevention scheme to a corresponding website;

in this embodiment, according to the predicted risk level and type, an optimal prevention scheme is found in a preset processing scheme library, and is extracted, and sent to a corresponding logistics website or responsible person, notifying them that the risk of customer loss may occur and discussing the implementation of the prevention scheme together with the risk, the website or responsible person starts a corresponding process, and takes feasible prevention measures, including making service packages, enhancing complaints and compensation mechanisms, optimizing distribution timeliness to retrieve and retain potential loss customers, helping to take necessary actions at the beginning of predicting customer loss, reducing the risk of potential customer loss as much as possible, in addition, among key factors causing customer loss, factors such as public control, three-way outage, etc. are also non-negligible, and when such an irresistible emergency occurs, emergency measures can be adopted in advance to coordinate resources of each party, so as to ensure the customer logistics service requirements, and reduce the influence and loss to customers as much as possible, and further maintain customers.

In the embodiment of the invention, a customer loss prediction model is established according to key factors and customer characteristics, future orders are predicted through the customer loss prediction model to obtain customer loss prediction results, customers which are likely to be lost can be accurately predicted, root causes of the customer loss and customer demands corresponding to the root causes are found, corresponding prevention schemes are extracted from a preset processing scheme library according to the customer loss prediction results, the prevention schemes are sent to corresponding network points, the network points can refer to the prevention schemes to carry out rapid maintenance processing, different maintenance strategies are adopted for the customers which are likely to be lost with different demands, the maintenance effect is improved, and the experience and satisfaction of the customers are further improved.

Referring to fig. 2, a second embodiment of a method for preventing loss of a customer order according to an embodiment of the present invention includes:

201. acquiring historical logistics order data, and performing outlier rejection, missing data filling and denoising filtering processing on the historical logistics order data to obtain first historical logistics order processing data;

in this embodiment, the logistics order data refers to data including key information such as product information, receiving and sending information, order number, cargo weight, order amount, and logistics path generated in the logistics field, but because the sources of the historical order data are different and the formats are various, the data needs to be cleaned to obtain effective data, and the data cleaning refers to a process of making the original data into usable data which meets the requirements and has no errors through a series of operations, including the following three cases: 1. abnormal value elimination, namely deleting or replacing abnormal data and outliers from a data set to ensure that the obtained effective data are real and reliable, wherein in actual enterprise operation, data abnormal values caused by abnormal behaviors of various logistics links, such as damage or recording errors of goods, and the like, possibly exist, and the abnormal values influence data analysis and processing, so that the abnormal values need to be eliminated or corrected, and the specific practice can be realized through professional software or algorithm identification, manual judgment and other modes; 2. filling missing data, if the data is missing, performing data filling by adopting methods such as interpolation algorithm, regression analysis and the like to enable the data set to be more complete, and in a specific original data record, sometimes some data is empty or missing, if the data with empty value or missing is directly used, tilting or error results are caused by data analysis, so that filling strategies are needed to be adopted to enable the data set to be more complete, and performing data filling by adopting methods such as interpolation algorithm, regression analysis and the like; 3. the denoising and filtering processing is carried out on the data by using a digital signal processing technology, noise interference is removed, data quality is improved, the historical logistics order data of enterprises often have noise and singal to a certain extent, the meaning of the data is often influenced when the data is intuitively expressed, the data can be subjected to the smoothing processing based on the digital signal processing technology, the noise interference is removed, and the data quality is optimized.

202. Identifying an order generation time field of the first historical logistics order processing data;

in this embodiment, a plurality of fields of the first historical logistics order processing data are identified, and words associated with an order generation time, such as "order time", are identified in a field name or interpretation, typically the order generation time will employ date and time type data, such as: 2023/01/01.

203. Ordering the first historical logistics order processing data according to the order generation time field to obtain second historical logistics order processing data;

in this embodiment, in the first historical logistics order processing data, a field related to the order generation time is found, the meaning and format specification represented by the field are confirmed, the order generation time field is designated as a sorting field, so as to ensure that the data are arranged in time sequence, the order generation time field is selected to be sorted in ascending (from early to late) or descending (from late to early) according to service requirements, and after sorting, a data set which is already arranged in time sequence can be obtained, wherein the data set is the second historical logistics order processing data.

204. Identifying a customer name field of the second historical logistics order processing data;

In the present embodiment, a field in which a client name exists is identified using a keyword search function or a filtering function.

205. Grouping the second historical logistics order processing data according to the customer name field to obtain the historical logistics order effective data;

in this embodiment, in the second historical logistics order processing data, a field related to a customer name is found, and the meaning and format specification represented by the field are confirmed, and the data processor or the grouping function of the programming language is used to divide the order data into different customer groups by using the customer name as a classification index, so as to obtain the historical logistics order effective data.

In the embodiment of the invention, the first historical logistics order processing data is sequenced according to the order generation time field to obtain the second historical logistics order processing data, the second historical logistics order processing data is grouped according to the customer name field, the order information of the same customer can be grouped together, for each customer group, a plurality of key statistical variables such as the order quantity, the freight volume, the transportation time and the like of the customer can be counted to obtain the logistics order data condition of the customer, the data of each customer service condition can be obtained, and the support can be provided for customer feature modeling.

Referring to fig. 3, a third embodiment of a method for preventing loss of a customer order according to an embodiment of the present invention includes:

301. screening the effective data of the historical logistics orders to obtain client group data of which the order number meets a preset first threshold value;

in this embodiment, a threshold value of the number of orders to be screened is determined, where the threshold value may be determined according to a specific service requirement, for example, a client group with the number of orders greater than or equal to a certain value is selected, client group data meeting the condition is selected from the historical logistics order effective data table, and the number of orders of each client group in the historical logistics order effective data table is compared with a preset threshold value, so that the client group data with the number of orders meeting the first threshold value requirement is screened.

302. Identifying the generation time of the last order of each client in the client group data, and acquiring real-time;

in this embodiment, the order data in each client group is ordered according to the order generation time field, the latest order is arranged at the forefront, the generation time of the order is the generation time of the last order of the client, the first order data of each client is extracted from the ordered client group data, that is, the last order record of the client, the order generation time of the record is obtained, and the UTC time or the local time of the current system is obtained by using a programming language, an operating system and other modes and is used as the real-time.

303. Judging whether the time period between the generation time and the real-time of the last order is satisfied with a preset second threshold value or not;

in this embodiment, a time period threshold to be determined is determined, where the threshold may be determined according to a service requirement, for example, a group of clients whose time interval between the generation of the last order does not exceed a certain time period is selected, and the time of the generation of the last order is subtracted from the real-time to obtain the amount of time interval between the generation of the last order and the current time, and if the time interval exceeds a preset second threshold, it is indicated that the client is inactive for a relatively long time or not in contact with the enterprise (for example, the client is not given for consumption by placing a list).

304. If yes, marking the client as an inactive state client, and marking the generation time of the last order as a loss time point;

in this embodiment, if it is determined that the period of time between the last order generation time and the real-time exceeds the preset second threshold, the customer may be considered to be in an inactive state. At this point, the customer may be marked as a "churn customer" using churn customer marking techniques, and the time of its last order generation may be marked as the churn point in time.

305. Acquiring historical abnormal condition data, and identifying the network point where the inactive state client is located;

in this embodiment, historical abnormal condition data is obtained, and the site information of the inactive state customer is found through information such as freight bill number, logistics tracking information and order address, and in general, the relevant connection line display between the customer location information and store/warehouse location information can be realized through tools such as websis.

306. Extracting corresponding abnormal condition data of the network point in the historical abnormal condition data, wherein the abnormal condition data of the network point comprise customer complaint records, dispatch delay records, cargo loss records, network point order error records, data loss records, abnormal climate records, public control records and three-party outage records;

in this embodiment, the network node that needs to extract abnormal situation data is determined, and abnormal situation data records related to the network node are screened out from the historical abnormal situation data, so as to obtain the network node abnormal situation data including customer complaint records, dispatch delay records, cargo loss records, network node order error records, data loss records, abnormal climate records, public control records and three-party shutdown records.

307. Identifying a record generation time field in the network point abnormal condition data;

in this embodiment, a plurality of fields in the site abnormality data are identified, and words related to the order generation time, such as "order time", are identified in the field names or interpretations, and typically, the order generation time will employ date and time type data, such as: 2023/01/01.

308. Matching the record generation time field with the loss time point to obtain a corresponding abnormal record event;

in this embodiment, an abnormal recording event corresponding to the churn time point is found by a screening method, specifically, data that does not match the churn time point may be filtered out by using the abnormal time generation time as a standard, and then the remaining abnormal recording event is determined to be an abnormal situation related to the churn event.

309. Determining key factors of customer loss according to abnormal record events, wherein each abnormal record event corresponds to one key factor, and the key factors comprise service attitude problems, delivery time problems, delivery quality problems, ordering problems, information security problems and unreliability problems;

in this embodiment, the website anomaly data includes customer complaint records, sending delay records, cargo loss records, website delivery error records, data loss records, abnormal climate records, public control records and three-way shutdown records, each record is an abnormal record event, each abnormal record event corresponds to a key factor, the key factors include service attitude problem, sending time problem, sending quality problem, order issue, information security problem and non-resistance problem, for example, the customer complaint records correspond to the service attitude problem, the sending delay records correspond to the sending time problem, the cargo loss records correspond to the sending quality problem, the website delivery error records correspond to the order issue, the data loss records correspond to the information security problem, and the abnormal climate records, the public control records and the three-way shutdown records all correspond to the non-resistance problem.

310. Extracting data corresponding to the inactive state clients from the client group data to obtain the inactive state client data;

in this embodiment, data corresponding to the inactive state clients in the client group data is extracted to obtain inactive state client data, and then data of all inactive state clients are screened out.

311. Acquiring client characteristics of the inactive state client from the inactive state client data, wherein the client characteristics comprise an article type, an express delivery type, a delivery address, order frequency, order quantity and order amount;

in this embodiment, the customer characteristics of the inactive state customer are obtained from the inactive state customer data, where the customer characteristics include an item type, an express type, a delivery address, an order frequency, an order number, and an order amount, so as to facilitate analysis of the item type of the inactive state customer ordered in the past, analysis of the main delivery address of the inactive state customer, analysis of the historical order amount statistics of the inactive state customer, and analysis of information such as the frequency and time period of the inactive state customer ordered in the past.

The method and the device can be used for evaluating the liveness of different clients, accurately predicting the clients which are likely to run off, finding out the root cause of the client running off and the client requirements corresponding to the root cause, and facilitating the subsequent establishment of a more refined and targeted service scheme.

Referring to fig. 4, a fourth embodiment of a method for preventing loss of a customer order according to an embodiment of the present invention includes:

401. pairing the key factors corresponding to each inactive state client and the client characteristics corresponding to the inactive state clients to obtain a pairing relation;

in this embodiment, the key factors corresponding to each inactive state client are paired with the client features corresponding to the inactive state client to obtain the pairing relationship, for example, the key factors corresponding to the inactive state client are service attitude problems, and it can be considered that clients having the same or similar client features as the inactive state client are also likely to be lost due to the service attitude problems.

402. Integrating all the pairing relations to obtain a pairing relation library;

in this embodiment, all the pairing relations are integrated to obtain a pairing relation library.

403. Establishing a customer loss prediction model according to the pairing relation library;

in this embodiment, the customer churn prediction model is trained and built on the training set according to the pairing relation library using algorithms such as a classifier or a regressor.

404. Inputting the customer characteristic data of the new order into a customer loss prediction model as an input variable to obtain a customer loss prediction result;

In this embodiment, the customer characteristic data of the new order is input as an input variable into the customer loss prediction model, so as to obtain a customer loss prediction result, specifically, the characteristic value of the customer of the new order needs to be brought into the trained customer loss prediction model, and the model predicts whether the customer will lose.

405. Invoking a pre-trained processing scheme library, and inputting a customer loss prediction result into the processing scheme library for processing policy matching to obtain a matching result;

in this embodiment, a pre-trained processing scheme library is invoked, and a customer loss prediction result is input into the processing scheme library for processing policy matching, so that a corresponding matching result can be obtained, specifically, the processing scheme library generally includes a plurality of loss prediction models and processing policies for different situations, and appropriate models and policies are selected for processing according to the customer loss prediction result, for example, if the prediction result is that the customer is likely to be lost due to service attitude problems, a processing policy for improving service attitude is selected.

406. Extracting a corresponding reference processing scheme from the processing scheme library according to the matching result, and generating the reference processing scheme into a prevention scheme in a text form;

In this embodiment, a specific solution to the problem of customer loss can be obtained by extracting a corresponding reference processing scheme from the processing scheme library according to the matching result, and these reference processing schemes are generated into a text-form prevention scheme, which can be used as an alternative scheme for enterprise reference selection and optimization on the one hand, and can also be used for staff education and knowledge sharing on the other hand, so as to help all relevant staff better understand and cope with the problem of loss, for example, the prediction result shows that the customer is easy to lose due to the problem of service attitude, so that the processing strategy for improving the service attitude can be selected pertinently, the feedback awareness can be enhanced in the whole customer experience process, the demands of the customer can be focused actively, and the service experience of the customer can be improved.

407. Transmitting the preventive scheme to the corresponding website;

in this embodiment, the formulated prevention scheme is sent to the corresponding website, so that the network branch office can be helped to know and master the specific measures and schemes of customer loss prevention, and further execute and implement the prevention scheme according to specific situations in actual work, and the prevention scheme can be sent in a mail mode.

According to the embodiment of the invention, according to the prediction result, enterprises can timely take corresponding measures to recover customers about to lose, such as providing better service, giving preferential promotion, developing targeted marketing activities and the like, and meanwhile, the enterprises can pay attention to the loss reasons of different customer groups through analysis and summarization of the prediction result, further improve the service level and the product quality, enhance the market competitiveness and the profitability, formulate a more refined and targeted service scheme to recover the customers, improve the maintenance effect and further improve the experience and satisfaction of the customers.

The method for preventing loss of a physical distribution customer order in the embodiment of the present invention is described above, and the apparatus for preventing loss of a physical distribution customer order in the embodiment of the present invention is described below, referring to fig. 5, where an embodiment of the apparatus for preventing loss of a physical distribution customer order in the embodiment of the present invention includes:

the acquiring and cleaning module 501 is configured to acquire historical logistics order data, clean the historical logistics order data, and obtain historical logistics order valid data;

the acquiring and determining module 502 is configured to acquire historical abnormal condition data, and determine key factors of customer loss according to the historical abnormal condition data and the historical logistics order valid data;

The establishing prediction module 503 is configured to establish a customer loss prediction model according to the key factors and the customer characteristics, and predict a future order through the customer loss prediction model to obtain a customer loss prediction result;

and the extraction and transmission module 504 is configured to extract a corresponding prevention scheme from a preset processing scheme library according to the client churn prediction result, and transmit the prevention scheme to a corresponding website.

In this embodiment, a customer loss prediction model is established according to key factors and customer characteristics, future orders are predicted through the customer loss prediction model to obtain customer loss prediction results, customers which may be lost can be accurately predicted, root causes of customer loss and customer demands corresponding to the root causes are found, corresponding prevention schemes are extracted from a preset processing scheme library according to the customer loss prediction results, the prevention schemes are sent to corresponding network points, the network points can refer to the prevention schemes to perform quick maintenance processing, different maintenance strategies are adopted for the customers which may be lost according to different demands, the maintenance effects are improved, and further experience and satisfaction of the customers are improved.

Referring to fig. 6, another embodiment of the apparatus for preventing loss of a physical distribution customer order according to the present invention includes:

the extracting and sending module 504 is configured to extract a corresponding prevention scheme from a preset processing scheme library according to a customer churn prediction result, and send the prevention scheme to a corresponding website;

in this embodiment, the acquiring the cleaning module 501 includes: the acquiring and processing unit 5011 is configured to acquire historical logistics order data, perform outlier rejection, missing data filling and denoising filtering processing on the historical logistics order data, and obtain first historical logistics order processing data; a first identifying unit 5012 for identifying an order generation time field of the first historical logistics order processing data; the sorting unit 5013 is configured to sort the first historical logistics order processing data according to the order generation time field to obtain second historical logistics order processing data; a second identifying unit 5014 for identifying a customer name field of the second historical logistics order processing data; and the grouping unit 5015 is configured to group the second historical logistics order processing data according to the customer name field to obtain the historical logistics order valid data.

In this embodiment, the acquisition determination module 502 includes: the screening unit 5021 is configured to screen the effective data of the historical logistics orders to obtain client group data with the order number meeting a preset first threshold; an identification acquisition unit 5022 for identifying the generation time of the last order of each customer in the customer group data and acquiring real-time; a determining unit 5023, configured to determine whether a time period between the generation time and the real-time of the last order is satisfied with a preset second threshold; the marking unit 5024 is configured to mark the client as an inactive client if yes, and mark the generation time of the last order as a loss time point; the acquiring and identifying unit 5025 is used for acquiring historical abnormal condition data and identifying an abnormal record event corresponding to the loss time point in the historical abnormal condition data; the determining and acquiring unit 5026 is configured to determine key factors of the customer churn according to the abnormal record event, and acquire the customer characteristics of the inactive state customer.

In this embodiment, the acquiring and identifying unit 5025 is specifically configured to acquire historical abnormal situation data, and identify a website where the inactive state client is located; extracting corresponding abnormal condition data of the network point in the historical abnormal condition data, wherein the abnormal condition data of the network point comprise customer complaint records, dispatch delay records, cargo loss records, network point order error records, data loss records, abnormal climate records, public control records and three-party outage records; identifying a record generation time field in the network point abnormal condition data; and matching the record generation time field with the loss time point to obtain a corresponding abnormal record event.

In this embodiment, the determining and acquiring unit 5026 is specifically configured to determine a key factor of customer churn according to an abnormal recording event, where each abnormal recording event corresponds to one key factor, and the key factors include a service attitude problem, a delivery time problem, a delivery quality problem, an ordering problem, an information security problem, and an unreliability problem; extracting data corresponding to the inactive state clients from the client group data to obtain the inactive state client data; customer characteristics of the inactive state customer are obtained from the inactive state customer data, the customer characteristics including item type, express type, delivery address, order frequency, order quantity, and order amount.

In this embodiment, the establishing prediction module 503 includes: pairing unit 5031, configured to pair key factors corresponding to each inactive state client with client features corresponding to inactive state clients to obtain a pairing relationship; an integrating unit 5032, configured to integrate all the pairing relationships to obtain a pairing relationship library; the establishing unit 5033 is configured to establish a customer churn prediction model according to the pairing relation library; the input unit 5034 is configured to input the customer characteristic data of the new order as an input variable into the customer loss prediction model, so as to obtain a customer loss prediction result.

In this embodiment, the extraction and transmission module 504 includes: the calling matching unit 5041 is used for calling a pre-trained processing scheme library, inputting a customer loss prediction result into the processing scheme library for processing policy matching, and obtaining a matching result; an extraction generating unit 5042, configured to extract a corresponding reference processing scheme from the processing scheme library according to the matching result, and generate the reference processing scheme as a prevention scheme in a text form; a transmitting unit 5043 for transmitting the preventive scheme to a corresponding mesh point.

The above-mentioned fig. 5 and fig. 6 describe the physical distribution customer order loss prevention apparatus in the embodiment of the present invention in detail from the shooting direction of the modularized functional entity, and the following describes the physical distribution customer order loss prevention device in the embodiment of the present invention in detail from the shooting direction of the hardware processing.

Fig. 7 is a schematic diagram of a physical distribution customer order loss prevention device 600 according to an embodiment of the present invention, where the physical distribution customer order loss prevention device 600 may have a relatively large difference according to a configuration or performance, and may include one or more processors (central processing units, CPU) 610 (e.g., one or more processors) and a memory 620, and one or more storage media 630 (e.g., one or more mass storage devices) storing application programs 633 or data 632. Wherein the memory 620 and the storage medium 630 may be transitory or persistent storage. The program stored on the storage medium 630 may include one or more modules (not shown), each of which may include a series of instruction operations for the logistics customer order loss prevention apparatus 600. Still further, the processor 610 may be configured to communicate with the storage medium 630 and execute a series of instruction operations in the storage medium 630 on the physical distribution customer order loss prevention device 600 to implement the steps of the physical distribution customer order loss prevention method provided by the above-described method embodiments.

The logistics customer order loss prevention apparatus 600 may also include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input/output interfaces 660, and/or one or more operating systems 631, such as Windows Serve, mac OS X, unix, linux, freeBSD, and the like. It will be appreciated by those skilled in the art that the configuration of the customer order loss prevention apparatus shown in fig. 7 is not limiting on the customer order loss prevention apparatus, and may include more or fewer components than shown, or may be combined with certain components, or may be arranged with different components.

The present invention also provides a computer readable storage medium, which may be a non-volatile computer readable storage medium, and which may also be a volatile computer readable storage medium, having stored therein instructions that, when executed on a computer, cause the computer to perform the steps of a method for preventing loss of a physical distribution customer order.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process of the system or apparatus and unit described above may refer to the corresponding process in the foregoing method embodiment, which is not repeated herein.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a storage medium, including instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a read-only memory (ROM), a random access memory (random access memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

Finally, it should be noted that: the foregoing is merely a preferred example of the present invention, and the present invention is not limited thereto, but it is to be understood that modifications and equivalents of some of the technical features described in the foregoing embodiments may be made by those skilled in the art, although the present invention has been described in detail with reference to the foregoing embodiments. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for preventing loss of a physical distribution customer order, comprising:

acquiring historical logistics order data, and cleaning the historical logistics order data to obtain historical logistics order effective data;

acquiring historical abnormal condition data, and determining key factors of customer loss according to the historical abnormal condition data and the historical logistics order effective data;

establishing a customer loss prediction model according to the key factors and the customer characteristics, and predicting a future order through the customer loss prediction model to obtain a customer loss prediction result;

and extracting a corresponding prevention scheme from a preset processing scheme library according to the customer loss prediction result, and sending the prevention scheme to a corresponding website.

2. The method for preventing loss of a physical distribution customer order according to claim 1, wherein the step of obtaining historical physical distribution order data, and cleaning the historical physical distribution order data to obtain historical physical distribution order valid data, comprises:

acquiring historical logistics order data, and performing outlier rejection, missing data filling and denoising filtering processing on the historical logistics order data to obtain first historical logistics order processing data;

Identifying an order generation time field of the first historical logistics order processing data;

sequencing the first historical logistics order processing data according to the order generation time field to obtain second historical logistics order processing data;

identifying a customer name field of the second historical logistics order processing data;

and grouping the second historical logistics order processing data according to the customer name field to obtain the historical logistics order effective data.

3. The method for preventing customer loss according to claim 1, wherein said obtaining historical abnormal situation data and determining key factors for customer loss based on said historical abnormal situation data and said historical physical distribution order validity data comprises:

screening the effective data of the historical logistics orders to obtain client group data of which the order number is in a preset first threshold value;

identifying the generation time of the last order of each client in the client group data, and acquiring real-time;

judging whether the time period between the generation time and the real-time of the last order is satisfied with a preset second threshold value or not;

if yes, marking the client as an inactive state client, and marking the generation time of the last order as a loss time point;

Acquiring historical abnormal condition data, and identifying an abnormal record event corresponding to the loss time point in the historical abnormal condition data;

and determining key factors of customer churn according to the abnormal record event, and acquiring the customer characteristics of the inactive state customer.

4. The method for preventing loss of a physical distribution customer order according to claim 3, wherein said obtaining historical abnormal situation data, identifying an abnormal recorded event corresponding to the loss time point in the historical abnormal situation data, comprises:

acquiring historical abnormal condition data, and identifying the network point of the inactive state client;

extracting corresponding abnormal condition data of the network point in the historical abnormal condition data, wherein the abnormal condition data of the network point comprises customer complaint records, dispatch delay records, cargo loss records, network point order error records, data loss records, abnormal climate records, public control records and three-party outage records;

identifying a record generation time field in the network point abnormal situation data;

and matching the record generation time field with the loss time point to obtain a corresponding abnormal record event.

5. The method for preventing loss of a physical distribution customer order according to claim 3, wherein determining key factors of customer loss according to the abnormal recording event and obtaining customer characteristics of the inactive state customer comprises:

determining key factors of customer churn according to the abnormal record events, wherein each abnormal record event corresponds to one key factor, and the key factors comprise service attitude problems, delivery time problems, delivery quality problems, ordering problems, information security problems and unreliability problems;

extracting data corresponding to the inactive state clients from the client group data to obtain inactive state client data;

and acquiring the client characteristics of the inactive state client from the inactive state client data, wherein the client characteristics comprise the type of the article, the type of the express delivery, the delivery address, the order frequency, the order quantity and the order amount.

6. The method for preventing loss of a physical distribution customer order according to claim 1, wherein the step of establishing a customer loss prediction model according to the key factors and the customer characteristics, predicting a future order by using the customer loss prediction model, and obtaining a customer loss prediction result includes:

Pairing key factors corresponding to each inactive state client with client features corresponding to the inactive state clients to obtain a pairing relation;

integrating all the pairing relations to obtain a pairing relation library;

establishing a customer loss prediction model according to the pairing relation library;

and inputting the customer characteristic data of the new order into the customer loss prediction model as an input variable to obtain a customer loss prediction result.

7. The method for preventing loss of a physical distribution customer order according to claim 1, wherein the step of extracting a corresponding prevention scheme from a preset processing scheme library according to a customer loss prediction result, and transmitting the prevention scheme to a corresponding website comprises:

invoking a pre-trained processing scheme library, and inputting the customer loss prediction result into the processing scheme library for processing policy matching to obtain a matching result;

extracting a corresponding reference processing scheme from the processing scheme library according to the matching result, and generating the reference processing scheme into a prevention scheme in a text form;

and sending the prevention scheme to the corresponding network point.

8. A physical distribution customer order loss prevention apparatus, comprising:

The cleaning module is used for acquiring historical logistics order data, cleaning the historical logistics order data and obtaining historical logistics order effective data;

the acquisition determining module is used for acquiring historical abnormal condition data and determining key factors of customer loss according to the historical abnormal condition data and the historical logistics order effective data;

the establishment prediction module is used for establishing a customer loss prediction model according to the key factors and the customer characteristics, and predicting future orders through the customer loss prediction model to obtain a customer loss prediction result;

and the extraction and transmission module is used for extracting a corresponding prevention scheme from a preset processing scheme library according to the customer loss prediction result and transmitting the prevention scheme to a corresponding website.

9. A physical distribution customer order loss prevention apparatus, characterized in that the physical distribution customer order loss prevention apparatus comprises: a memory and at least one processor, the memory having instructions stored therein;

at least one of the processors invokes the instructions in the memory to cause the physical distribution customer order loss prevention device to perform the steps of the physical distribution customer order loss prevention method of any of claims 1-7.

10. A computer readable storage medium having instructions stored thereon, which when executed by a processor, perform the steps of the method of preventing loss of a physical distribution customer order as claimed in any one of claims 1 to 7.