WO2019205410A1 - Nps short message investigation method and system, computer device and storage medium - Google Patents

Abstract

The present application relates to the technical field of service assessment, and in particular, to an NPS short message investigation method and system, a computer device and a storage medium. The NPS short message investigation method comprises: screening and extracting data of customers to be investigated from a service system to generate an investigation list, performing short message investigation on the customers one by one according to the investigation list and collecting customer feedback information; extracting tag data and feature data in the customer feedback information and then merging the two to generate a feedback information list having a uniform data format; and analyzing the feedback information list by means of an algorithm and then obtaining two types of customer groups by screening, taking a high-score customer group as an object to which follow-up products or services are recommended, and regularly maintaining and managing a low-score customer group. Compared with the conventional NPS investigation technology, the present application has high real-time performance for customer demand processing, and achieves accurate data extraction, accurate positioning of customer demands, and clear classification.

Description

[Correction according to Rule 26 28.09.2018] NPS SMS research method, system, computer equipment and storage medium

This application claims priority to Chinese Patent Application No. 201810363670.8, entitled "NPS Short Message Research Method, System, Computer Equipment and Storage Media", filed on April 22, 2018, the entire contents of which are hereby incorporated by reference. Combined in this application.

Technical field

The present application relates to the field of service evaluation technologies, and in particular, to an NPS short message research method, system, computer device, and storage medium.

Background technique

NPS (Net Promoter Score), also known as net recommendation value, is one of the key indicators for measuring customer loyalty in a business by measuring the possibility that a customer will recommend a product or service to others. NPS was proposed by Fred Reichhold of Bain Consulting in the United States in 2003 for the benign benefits and real growth of the company. Information recommendation based on NPS can better meet the customer's actual situation, improve the recommendation success rate, and enhance the user experience. As one of the most popular customer loyalty analysis indicators in recent years, NPS has been widely used in the world.

Compared with customer satisfaction, NPS is only a holistic indicator. It is difficult to intuitively decompose it internally, and it is difficult to clearly define and expand services and work. The traditional service system lacks the satisfaction survey of customers, and further analyzes the customer feedback content. It can't get the real needs of customers in time, and can't cooperate with the marketing department to accurately locate the potential customer groups. Targeted marketing activities, the shortcomings are shown in the following aspects:

1) The huge amount of data in the service system lacks the function of extracting feature data by specific rules according to different service items, and it is impossible to analyze the satisfaction of the customer group based on the feature data;

2) The function of sending a short message to the customer corresponding to the feature data and collecting the feedback data of the customer;

3) Lack of feature analysis and deep mining of customer feedback data;

4) There is a lack of countermeasures for returning customers to customers corresponding to low-scoring data, and it is impossible to clarify the reason why customers are evaluated as low scores.

Summary of the invention

The purpose of the present application is to propose an NPS short message research method, system, computer device and storage medium to solve the deficiencies in the above background art. The traditional business system has huge data and many service project categories. It lacks the means of extracting and refining the feature data. It can not timely review the feature data, and carries out the progress of the feature analysis and deep mining of customer feedback information. The marketing needs cannot be met, and the analysis and return visits of the customers corresponding to the low score data cannot be carried out smoothly, which affects the overall operational efficiency.

To achieve the above objective, the present application provides the following technical solutions:

An NPS short message research method, including:

Filtering and extracting data of the customer to be researched in the business system, generating a research list, conducting short message research and collecting customer feedback information according to the research list, and the customer feedback information includes tag data;

Extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format;

After analyzing the feedback information list through K-clustering algorithm, two types of customer groups are selected, which are high-scoring customer group and low-scoring customer group respectively. The high-scoring customer group is the target of recommended follow-up products or services. The score customer group generates a return visit list, which is regularly maintained and sorted through a preset return visit method.

An NPS short message research system, including:

The research unit is configured to filter and extract data of the customer to be researched in the business system, generate a research list, perform short message research and collect customer feedback information according to the research list, and the customer feedback information includes tag data;

Extracting the merging unit, setting to extract the tag data and the feature data in the customer feedback information and merging, and generating a feedback information list having a unified data format;

The analysis countermeasure unit is set to analyze the feedback information list through the K-clustering algorithm and filter out two types of customer groups, namely a high score customer group and a low score customer group, and the high score customer group as a recommended follow-up product or The object of the service generates a return visit list for the low-scoring customer group, and performs regular maintenance and grooming through the preset return visit method.

A computer device comprising a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor such that the processor performs the following steps:

Filtering and extracting data of the customer to be researched in the business system, generating a research list, conducting short message research and collecting customer feedback information according to the research list, and the customer feedback information includes tag data;

Extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format;

After analyzing the feedback information list through K-clustering algorithm, two types of customer groups are selected, which are high-scoring customer group and low-scoring customer group respectively. The high-scoring customer group is the target of recommended follow-up products or services. The score customer group generates a return visit list, which is regularly maintained and sorted through a preset return visit method.

A storage medium storing computer readable instructions that, when executed by one or more processors, cause one or more processors to perform the steps of:

Filtering and extracting data of the customer to be researched in the business system, generating a research list, conducting short message research and collecting customer feedback information according to the research list, and the customer feedback information includes tag data;

Extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format;

After analyzing the feedback information list through K-clustering algorithm, two types of customer groups are selected, which are high-scoring customer group and low-scoring customer group respectively. The high-scoring customer group is the target of recommended follow-up products or services. The score customer group generates a return visit list, which is regularly maintained and sorted through a preset return visit method.

The above NPS short message research method, device, computer equipment and storage medium, for the huge user data and different business service items in the service system, according to the survey data obtained by the customer after the short message research, and the original characteristic data of the customer are pressed according to K The clustering algorithm analyzes and counts the high-value customer group and the low-value customer group in the research customer, so as to provide a corresponding basis for subsequent operations and sales, so that the marketing department can specifically recommend the appropriate business category for the customer. Or provide a targeted reference for customers with follow-up maintenance. By categorizing low-value customers, they can conduct targeted return visits to obtain the reasons for the low customer evaluation, so as to purposefully improve the corresponding services or Products to adapt to the needs of such customers, compared with the traditional NPS research technology, this application has high real-time performance, accurate data refinement, precise positioning of customer needs, and clear classification advantages when dealing with customer needs.

DRAWINGS

Various other advantages and benefits will become apparent to those skilled in the art from a The drawings are only for the purpose of illustrating the preferred embodiments and are not intended to be limiting.

1 is a flowchart of a method for investigating NPS short messages in an embodiment of the present application;

Figure 2 is a flow chart of step S1 of Figure 1;

Figure 3 is a flow chart of step S2 of Figure 1;

Figure 4 is a flow chart of step S3 of Figure 1;

5 is a structural diagram of an NPS short message research system in an embodiment of the present application;

6 is a block diagram of the research unit of FIG. 5;

7 is a block diagram of the extraction and merging unit in FIG. 5;

FIG. 8 is a block diagram of the analysis countermeasure unit of FIG. 5.

detailed description

In order to make the objects, technical solutions, and advantages of the present application more comprehensible, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the application and are not intended to be limiting.

The singular forms "a", "an", "the" It is to be understood that the phrase "comprise" or "an" Integers, steps, operations, components, components, and/or groups thereof.

FIG. 1 is a flowchart of a method for investigating NPS short messages in an embodiment of the present application. As shown in FIG. 1 , an NPS short message research method includes the following steps:

Step S1: screening and extracting data of the customer to be researched in the business system, generating a research list, performing short message research and collecting customer feedback information according to the research list, wherein the customer feedback information includes tag data; wherein the business system may be financial Or the real estate business, or the business business, the customer data with the customer contact method in the business as the basis for the extraction, select the customer sequence of one of the business categories as the basis for the extraction, and extract some of the customer data to be investigated, The customer conducts research one by one according to the pre-defined SMS information, and then collects the feedback information, and records the customer information that is not feedback, so as to aggregate and generate the research results.

Step S2, extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format; extracting the tag data in the customer feedback information, and extracting some fields in the complete basic information of the customer as features The data is set to analyze the characteristics of the customer, and the tag data and the feature data are combined and recorded into corresponding fields in the same data record, thereby generating a feedback information list having a unified data format; for example, when the service type is selected as the network sales policy service, When the survey content is a survey of customer satisfaction, the insured person's name, gender, age, occupation, place of residence and other personal characteristics information are inquired according to the insured number, and the tag information is extracted from the SMS content of the policyholder's reply, with 0-10 The points are used as label information. The data structure is as follows: (insurance number, gender, age, occupation, place of residence, scoring); the values of the corresponding data structure are: ('9001200011609315', 'female', '29', 'civil servant', 'Shanghai', 7); The following is presented as a list of feature data and tag data:

Feature data tag

(‘0’,’29’, 'stable occupation’, 'first-tier city’) 2

(‘1’,’30’, 'high-risk occupation’, 'second-tier city’) 7

(‘0’,’31’, 'general occupation’, 'third-tier city’) 5

(‘1’,’32’,’ stable career’, 'four-tier city’) 3

(‘0’,’33’, 'high-risk occupation’, 'rural and remote mountainous areas') 8

Among them, the female is marked with 0 and the male is marked with 1.

The field corresponding to the tag data records the score of the customer's satisfaction score, and the customer's feature data may select any one or more of the customer's complete basic information, and the corresponding complete basic information may also include: the policy number, the name of the policyholder , the insured number, the insured's home address, the insured's landline number, the insured's mobile phone number, the insured's e-mail address, the insured's employment in the industry, the insured's insurance time, and the insured's number of insurance applications. In step S3, the feedback information list is analyzed by the K-clustering algorithm, and two types of customer groups are selected, which are a high-scoring customer group and a low-scoring customer group, and the high-scoring customer group is the object of recommending a follow-up product or service. Generate a return visit list for low-value customer groups, perform regular maintenance and grooming through preset return visit methods; analyze and confirm high-value customer groups based on customer feedback information received, so as to accurately market potential customer groups, and Regularly maintain and sort out low-value customer groups, confirm the reasons for low scores of customers, and formulate corresponding countermeasures.

In an embodiment, as shown in FIG. 2, step S1 may include the following specific steps: Step S101, selecting a certain service category in the service system, obtaining a corresponding customer form, and extracting customer data having complete basic information therein. Record, when the business type is selected as the online sales policy business, the corresponding complete basic information includes: policy number, insured name, insured number, insured person's home address, insured person's landline number, insured's mobile phone number, insured electronic The email address, the insured person engaged in the industry, the insured person's insurance time, and the insured person's insurance number; in step S102, it is determined whether the contact method in the customer record extracted in the above step is complete and legal, and when at least one contact method is complete and legal, Add the customer data record to a pending research form. When there is no complete contact information, store the customer data record in a pending contact form and transfer it to the after-sales department for information supplement. At least one contact method to be supplemented. Appending the customer data record to the pending research form after complete and legal; in step S103, The customers in the summary of the research form are classified according to the first classification rule to generate different customer sequences to be investigated. The first classification rules include: classification according to the year of insurance, classification according to the number of insurances, classification according to the location of the contractor. Or generating a to-be-searched client sequence by using the first classification rule according to the classification performed by the contractor; the step S104, pre-storing a to-be-sent message list, where the to-be-sent message list includes at least one to-be-sent message record, Any one of the to-be-sent messages is recorded as a research text compiled in a specific template format, and is set to obtain the research result after researching the customer, and send the customer contact information in the form to be investigated to the customer to obtain the research result;

In one embodiment, any of the above-mentioned customer sequences to be researched prepares at least one format text of the short message to be sent, wherein the research information surveyed to the customer includes: research on the satisfaction of the business product, satisfaction with the service Investigate and recommend the recommendation of the business products to relatives and friends. The research information is prepared by using the scoring rules of obtaining customer ratings. The research information is sent by SMS, email, and mobile APP. It can also be carried out by using a mobile phone, an email address, a mobile APP client to push a link address of a webpage questionnaire including survey information content; wherein the scoring rule can take the following form: 0 to 10 points as a score basis. The scoring rule, the scoring rules based on A, B, and C as the scores. When the survey content is selected as the recommended strength for recommending business products to relatives and friends, the corresponding sample of the message records to be sent may be in the following format: How much are you likely to recommend this product when you purchase this product? Your relatives and friends? Please use 0-10 points to rate the evaluation, 0 point means very not recommended, 10 points means highly recommended, please SMS to reply to the number. Thank you for your participation!"; Step S105, select any of the above steps After a message is recorded, it is pushed to each customer in the corresponding customer sequence to be researched, and a corresponding information feedback list is established, which is set to record whether the customer has a reply and reply content.

In an embodiment, as shown in FIG. 3, step S2 may include the following specific steps: Step S201, formulating a statistical period table, and classifying the customers according to whether there is a reply according to a predetermined statistical period within the period. It is divided into two categories: replies and non-responses; in step S202, a list of unreacted customers is generated to generate a to-be-confirmed form, and is set to be sent to the after-sales department to confirm that the customer has not responded, so as to confirm whether such customers have the value of continuing maintenance; In step S203, the customer list that has been replied is summarized, and part of the field data in the complete basic information of the customer is extracted and recorded together with the tag data in the customer feedback information to a feedback information list.

In an embodiment, as shown in FIG. 4, step S3 may include: step S301, performing feature analysis on the feedback information list by using a K-clustering algorithm, and selecting a boundary value of the data distribution as a threshold according to the analysis result, to A customer corresponding to the threshold (on the left side of the threshold distribution) is a low-scoring customer, and a customer corresponding to the threshold (on the right side of the threshold distribution) is used as a high-scoring customer, thereby excavating a demand suitable for each type of characteristic customer; The K-Calcing algorithm is the K-MEANS algorithm, which accepts the input quantity k and the database containing n data objects, then divides the n data objects into k clusters, and outputs the minimum k standard clusters satisfying the variance. In order to make the obtained clusters satisfy: the similarity of objects in the same cluster is higher, and the similarity of objects in different clusters is smaller. In this step, the K-clustering algorithm is analyzed on the customer included in the feedback information list, and the demarcation value of the cluster (ie, the mean or the central object) is obtained as a threshold. The customer smaller than the threshold is a low-scoring customer, which is greater than the threshold. Customers as high score customers. Step S302: The client corresponding to the data that is greater than the threshold is selected as the target of the recommended follow-up product or service by the analysis result; in step S303, the customer corresponding to the data less than the threshold is filtered to generate a return visit list, and the customer appeal is obtained after the return visit according to the list one by one. Extracting the cause feature and the potential demand feature according to the return visit result; step S304, analyzing the cause feature, generating an improved project form, and transferring the improved project form to an improvement of the service project involved by the relevant department of product development; step S305, After analyzing the potential demand characteristics, a new product development plan form is generated, and the new product development plan form is transferred to the relevant department of product development to set up a new product for different customer groups.

In one embodiment, the return visit method when the return visit is performed one by one according to the list includes: a return visit mode by means of a voice call, an email or an application back feedback channel, or a return visit mode in the form of a web questionnaire.

In an embodiment, an NPS short message research system is proposed, as shown in FIG. 5, which includes: a research unit, configured to filter and extract data of a customer to be researched in a business system, generate a research list, and select one by one according to a research list. Performing short message research and collecting customer feedback information, the customer feedback information includes tag data; extracting and merging unit, setting to extract tag data and feature data in customer feedback information and merging, generating a feedback information list with unified data format; analyzing The countermeasure unit is set to analyze the feedback information list through the K-clustering algorithm and filter out two types of customer groups, namely a high score customer group and a low score customer group, and the high score customer group as a recommended follow-up product or service. The object is to generate a return visit list for the low-scoring customer group, perform regular maintenance and sorting through the preset return visit method, and formulate the way to recover the customer.

In an embodiment, as shown in FIG. 6, the research unit includes: a customer data acquisition module, configured to acquire a corresponding customer form after selecting a service category in the business system, and extract a customer data record having complete basic information therein; The customer data record judging module is configured to judge whether the contact information in the extracted customer data record is complete and legal, and when at least one contact method is complete and legal, the customer data record is added to a pending research form, when no In the complete contact mode, the customer data record is stored in a confirmation contact form and transferred to the after-sales department for information supplementation. At least one contact method to be supplemented is complete and legal, and the customer data record is added to the research to be investigated. In the form, the customer classification module is configured to generate a different customer sequence to be investigated according to the first classification rule by the customer who summarizes the form to be researched, wherein the first classification rule includes: classification according to the year of insured, Classification by number of insured times, classification by contractor location or by contractor The industry involved is classified, and the customer classification sequence is generated by the first classification rule; the research message prefabrication module is configured to pre-store the to-be-sent message list, and the to-be-sent message list includes at least one to-be-sent message record, wherein Any pending message is recorded as a research text compiled in a specific template format, and the research result is obtained after the customer is researched, and the customer contact information in the form to be researched is sent to the customer to obtain the research result, wherein any one is to be The research client sequence is prepared with at least one format text to be sent by the message; the message pushing module is configured to select any record to be sent and push it to each customer in the corresponding customer sequence to be investigated, and establish corresponding information feedback. A list that records whether the customer has a response and a response.

In an embodiment, as shown in FIG. 7, the extracting and merging unit comprises: a reply sorting module, configured to formulate a statistical periodic table, and according to a pre-defined statistical period therein, each period sorts the customer according to whether there is a reply, and points There are two types of replies and non-reply; the information to be confirmed module is set to generate a pending confirmation list for the unreviewed customer list, and is set to go to the after-sales department to confirm the customer's failure to reply, in order to confirm whether such customers have continued. The value of the maintenance; the reply information summary module is set to summarize the customer list that has been replied, and part of the field data in the complete basic information of such customer is extracted and recorded together with the tag data in the customer feedback information to a feedback information list. .

In an embodiment, as shown in FIG. 8, the analysis countermeasure unit includes: a feature analysis module configured to perform feature analysis on the feedback information list by using a K-clustering algorithm, and select a boundary value of the data distribution according to the analysis result as The threshold value is a low-value customer with a customer corresponding to the threshold (on the left side of the threshold distribution), and a customer corresponding to the threshold (on the right side of the threshold distribution) as a high-scoring customer, thereby excavating a customer suitable for each type of feature. The potential customer selection module is configured to select a customer corresponding to the data larger than the threshold by the analysis result as the object of the recommended follow-up product or service; the return list generation module is set to filter the customer corresponding to the data less than the threshold to generate a return visit The list is obtained after the return visit according to the list, and the customer's appeal is obtained. According to the return visit result, the reason feature and the potential demand feature are extracted. The reason feature countermeasure module is set to analyze the cause feature and generate an improved project form, and the improved project form is transferred to the product development. Relevant departments to improve the service items involved Measure of potential demand module, is set to feature after analysis of the potential demand for new product development programs generate the form, the form of new product development plans go to the relevant departments of product development to develop new products for different customer groups.

In an embodiment, the return visit mode in the return visit list generating module includes: a return visit mode by means of a voice call, an email, or a return visit form of an App-side feedback channel, or a return visit mode in the form of a webpage questionnaire.

In one embodiment, a computer apparatus is provided comprising a memory and a processor having stored therein computer readable instructions that, when executed by a processor, cause the processor to execute the computer program to implement the various embodiments described above The steps in the NPS SMS survey method.

In one embodiment, a storage medium storing computer readable instructions is provided that, when executed by one or more processors, cause one or more processors to perform NPS short message research in the various embodiments described above The steps in the method. The storage medium may be a non-volatile storage medium.

A person skilled in the art may understand that all or part of the various steps of the foregoing embodiments may be performed by a program to instruct related hardware. The program may be stored in a computer readable storage medium, and the storage medium may include: Read Only Memory (ROM), Random Access Memory (RAM), disk or optical disk.

The technical features of the above-described embodiments may be arbitrarily combined. For the sake of brevity of description, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction between the combinations of these technical features, All should be considered as the scope of this manual.

The above-described embodiments are only illustrative of some exemplary embodiments of the present application, and the description thereof is more specific and detailed, and is not to be construed as limiting the scope of the claims. It should be noted that a number of variations and modifications may be made by those skilled in the art without departing from the spirit and scope of the present application. Therefore, the scope of the invention should be determined by the appended claims.

Claims (20)

An NPS short message research method, including: Filtering and extracting data of the customer to be researched in the business system, generating a research list, conducting short message research and collecting customer feedback information according to the research list, and the customer feedback information includes tag data; Extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format; After analyzing the feedback information list through K-clustering algorithm, two types of customer groups are selected, which are high-scoring customer group and low-scoring customer group respectively. The high-scoring customer group is the target of recommended follow-up products or services. The score customer group generates a return visit list, which is regularly maintained and sorted through a preset return visit method. The NPS short message research method according to claim 1, wherein the filtering and extracting data of the customer to be researched in the business system, generating a research list, and performing the short message research and collecting the customer feedback information according to the research list include: Select a certain business category in the business system, obtain the corresponding customer form, and extract the customer data record with complete basic information; Determining whether the contact information in the extracted customer data record is complete and legal. When at least one contact method is complete and legal, the customer data record is added to a pending research form, and when there is no complete contact method, the customer is After the data record is stored in the confirmation contact form, the information is transferred to the after-sales department for information supplementation, and at least one contact method to be supplemented is complete and legal, and the customer data record is added to the to-be-invested form; The customer that summarizes the form to be researched is classified according to the first classification rule to generate different customer sequence to be investigated, wherein the first classification rule includes: classification according to the year of insurance, classification according to the number of insurances, and signing according to the contract. Classification of the location of the person or classification by the industry in which the contractor engages, and generating a customer sequence to be investigated through the first classification rule; Pre-storing a list of to-be-sent messages, the to-be-sent message list includes at least one to-be-sent message record, and any one of the to-be-sent message records is a research text compiled in a specific template format, and is set to obtain a survey after researching the customer. As a result, the customer contact information in the form to be researched is sent to the customer to obtain the research result, wherein any sample of the customer to be researched prepares at least one format text of the short message to be sent; After selecting any record to be sent, push it to each customer in the corresponding customer sequence to be investigated, and establish a corresponding information feedback list, which is set to record whether the customer has a reply and reply content. The NPS short message research method according to claim 2, wherein the extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format includes: Develop a statistical periodic table, according to the pre-established statistical cycle, each cycle will classify the customer according to whether there is a reply, divided into two categories: replies and non-responses; Aggregate the unreacted customer list to generate a pending confirmation form, set to transfer to the after-sales department to confirm the customer's failure to reply, in order to confirm whether such customers have the value of continuing maintenance; The customer list that has been replied is summarized, and part of the field data in the complete basic information of such customer is extracted and recorded together with the tag data in the customer feedback information to a feedback information list. The NPS short message research method according to claim 3, wherein the feedback information list is analyzed by a K-clustering algorithm, and two types of customer groups are selected, which are a high score customer group and a low score customer group, respectively. The high-scoring customer group, as the object of recommending the follow-up product or service, generates a return visit list for the low-scoring customer group, and performs regular maintenance and combing through the preset return visit method, including: The K-clustering algorithm is used to analyze the feature information list, and the demarcation value of the data distribution is selected as the threshold according to the analysis result, and the customer corresponding to the threshold is used as the low-scoring customer, and the customer corresponding to the threshold is used as the customer corresponding to the threshold. High-value customers, so as to dig out the needs of each type of customer; The customer corresponding to the data larger than the threshold is selected as the object of recommending the follow-up product or service through the analysis result; After filtering the customer corresponding to the data smaller than the threshold, a return visit list is generated, and the customer appeal is obtained after the return visit according to the list, and the reason feature and the potential demand feature are extracted according to the return visit result; After analyzing the cause characteristics, generating an improved project form, and transferring the improved project form to the relevant department of the product development to improve the service items involved therein; After analyzing the potential demand characteristics, a new product development plan form is generated, and the new product development plan form is transferred to the relevant department of product development to set up a new product for different customer groups. The NPS short message research method according to claim 4, wherein the returning mode when the returning one by one according to the list is performed one by one includes: A return visit method in the form of a return call by voice call, email or App-side feedback channel, or a return visit method in the form of a web questionnaire. An NPS short message research system, including: The research unit is configured to filter and extract data of the customer to be researched in the business system, generate a research list, perform short message research and collect customer feedback information according to the research list, and the customer feedback information includes tag data; Extracting the merging unit, setting to extract the tag data and the feature data in the customer feedback information and merging, and generating a feedback information list having a unified data format; The analysis countermeasure unit is set to analyze the feedback information list through the K-clustering algorithm and filter out two types of customer groups, namely a high score customer group and a low score customer group, and the high score customer group as a recommended follow-up product or The object of the service generates a return visit list for the low-scoring customer group, and performs regular maintenance and grooming through the preset return visit method. The NPS short message research system of claim 6, wherein the research unit further comprises: The customer data obtaining module is configured to obtain a corresponding customer form after selecting a business category in the selected business system, and extract a customer data record having complete basic information therein; The customer data record judging module is configured to judge whether the contact information in the extracted customer data record is complete and legal, and when at least one contact method is complete and legal, the customer data record is added to a pending research form, when no In the complete contact mode, the customer data record is stored in a confirmation contact form and transferred to the after-sales department for information supplementation. At least one contact method to be supplemented is complete and legal, and the customer data record is added to the research to be investigated. In the form; The customer classification module is configured to generate a different customer sequence to be investigated according to the first classification rule, and the first classification rule includes: classification according to the insured year, according to the number of insurance applications. The classification to be carried out, the classification according to the location of the contractor or the classification by the industry in which the contractor engages, and the customer sequence to be investigated is generated by the first classification rule; The research message prefabrication module is configured to pre-store a list of to-be-sent messages, where the to-be-sent message list includes at least one to-be-sent message record, and any one of the to-be-sent message records is a research text compiled in a specific template format, and is sent to the client. After the investigation, the research result is obtained, and the customer contact information in the form to be researched is sent to the customer to obtain the research result, wherein any sample of the customer to be researched prepares at least one format text of the short message to be sent; The message pushing module is configured to select any record to be sent and push it to each customer in the corresponding customer sequence to be investigated, and establish a corresponding information feedback list to record whether the customer has a reply and a reply content. The NPS short message research system of claim 7, wherein the extracting and merging unit comprises: The reply classification module is set to formulate a statistical periodic table. According to the pre-established statistical period, each cycle classifies the customer according to whether there is a reply, and is divided into two categories: replies and non-responses; The to-be-confirmed information generating module is configured to generate a to-be-confirmed list by summarizing the un-reacted customer lists, and set the reason to the after-sales department to confirm that the customer has not responded, so as to confirm whether such customers have the value of continuing maintenance; The information summary module has been replied to set up the customer list that has been replied, and part of the field data in the complete basic information of such customer is extracted and recorded together with the tag data in the customer feedback information to a feedback information list. The NPS short message research system according to claim 8, wherein the analysis countermeasure unit comprises: The feature analysis module is configured to perform feature analysis on the feedback information list by using a K-clustering algorithm, and select a boundary value of the data distribution according to the analysis result as a threshold value, and the customer corresponding to the threshold value is regarded as a low-scoring customer, which is greater than The customer corresponding to the threshold is used as a high-scoring customer to dig out the needs of each type of characteristic customer; The potential customer selection module is configured to select a customer corresponding to the data larger than the threshold by using the analysis result as an object of recommending a follow-up product or service; The returning list generating module is configured to generate a returning visit list after filtering the customer corresponding to the data smaller than the threshold, obtain a customer appeal according to the list one by one, and extract the reason feature and the potential demand feature according to the returning result; The reason feature countermeasure module is configured to generate an improvement project form after analyzing the cause feature, and transfer the improved project form to the relevant department of the product development to improve the service item involved therein; The potential demand feature countermeasure module is configured to generate a new product development plan form after analyzing the potential demand feature, and transfer the new product development plan form to the relevant department of product development to set up a new product for different customer groups. The NPS short message research system according to claim 9, wherein the return visit mode in the return visit list generating module comprises: a return visit mode by means of a voice call, an email, or a return visit form of an App end feedback channel, or a webpage questionnaire form Way of returning to visit. A computer device comprising a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor such that the processor performs the following steps: Filtering and extracting data of the customer to be researched in the business system, generating a research list, conducting short message research and collecting customer feedback information according to the research list, and the customer feedback information includes tag data; Extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format; After analyzing the feedback information list through K-clustering algorithm, two types of customer groups are selected, which are high-scoring customer group and low-scoring customer group respectively. The high-scoring customer group is the target of recommended follow-up products or services. The score customer group generates a return visit list, which is regularly maintained and sorted through a preset return visit method. The computer device according to claim 11, wherein the filtering and extracting data of the customer to be researched in the business system, generating a research list, performing short message research and collecting customer feedback information according to the research list, so that the processor executes The following steps: Select a certain business category in the business system, obtain the corresponding customer form, and extract the customer data record with complete basic information; Determining whether the contact information in the extracted customer data record is complete and legal. When at least one contact method is complete and legal, the customer data record is added to a pending research form, and when there is no complete contact method, the customer is After the data record is stored in the confirmation contact form, the information is transferred to the after-sales department for information supplementation, and at least one contact method to be supplemented is complete and legal, and the customer data record is added to the to-be-invested form; The customer that summarizes the form to be researched is classified according to the first classification rule to generate different customer sequence to be investigated, wherein the first classification rule includes: classification according to the year of insurance, classification according to the number of insurances, and signing according to the contract. Classification of the location of the person or classification by the industry in which the contractor engages, and generating a customer sequence to be investigated through the first classification rule; Pre-storing a list of to-be-sent messages, the to-be-sent message list includes at least one to-be-sent message record, and any one of the to-be-sent message records is a research text compiled in a specific template format, and is set to obtain a survey after researching the customer. As a result, the customer contact information in the form to be researched is sent to the customer to obtain the research result, wherein any sample of the customer to be researched prepares at least one format text of the short message to be sent; After selecting any record to be sent, push it to each customer in the corresponding customer sequence to be investigated, and establish a corresponding information feedback list, which is set to record whether the customer has a reply and reply content. The computer device according to claim 12, wherein said extracting tag data and feature data in the customer feedback information are combined and generating a feedback information list having a unified data format, such that said processor performs the following steps: Develop a statistical periodic table, according to the pre-established statistical cycle, each cycle will classify the customer according to whether there is a reply, divided into two categories: replies and non-responses; Aggregate the unreacted customer list to generate a pending confirmation form, set to transfer to the after-sales department to confirm the customer's failure to reply, in order to confirm whether such customers have the value of continuing maintenance; The customer list that has been replied is summarized, and part of the field data in the complete basic information of such customer is extracted and recorded together with the tag data in the customer feedback information to a feedback information list. The computer device according to claim 13, wherein the pair of feedback information is analyzed by a K-clustering algorithm, and two types of customer groups are selected, which are a high score customer group and a low score customer group, and the high score is As a target for recommending a follow-up product or service, the value customer group generates a return visit list for the low-scoring customer group, and performs regular maintenance and combing through a preset return visit mode, so that the processor performs the following steps: The K-clustering algorithm is used to analyze the feature information list, and the demarcation value of the data distribution is selected as the threshold according to the analysis result, and the customer corresponding to the threshold is used as the low-scoring customer, and the customer corresponding to the threshold is used as the customer corresponding to the threshold. High-value customers, so as to dig out the needs of each type of customer; The customer corresponding to the data larger than the threshold is selected as the object of recommending the follow-up product or service through the analysis result; After filtering the customer corresponding to the data smaller than the threshold, a return visit list is generated, and the customer appeal is obtained after the return visit according to the list, and the reason feature and the potential demand feature are extracted according to the return visit result; After analyzing the cause characteristics, generating an improved project form, and transferring the improved project form to the relevant department of the product development to improve the service items involved therein; After analyzing the potential demand characteristics, a new product development plan form is generated, and the new product development plan form is transferred to the relevant department of product development to set up a new product for different customer groups. The computer device according to claim 14, wherein said said processor performs the following steps when said returning to the list one by one is performed: A return visit method in the form of a return call by voice call, email or App-side feedback channel, or a return visit method in the form of a web questionnaire. A storage medium storing computer readable instructions that, when executed by one or more processors, cause one or more processors to perform the steps of: Filtering and extracting data of the customer to be researched in the business system, generating a research list, conducting short message research and collecting customer feedback information according to the research list, and the customer feedback information includes tag data; Extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format; After analyzing the feedback information list through K-clustering algorithm, two types of customer groups are selected, which are high-scoring customer group and low-scoring customer group respectively. The high-scoring customer group is the target of recommended follow-up products or services. The score customer group generates a return visit list, which is regularly maintained and sorted through a preset return visit method. The storage medium according to claim 16, wherein the filtering and extracting data of the customer to be researched in the business system, generating a research list, conducting short message research and collecting customer feedback information according to the research list, and causing one or more processing Perform the following steps: Select a certain business category in the business system, obtain the corresponding customer form, and extract the customer data record with complete basic information; Determining whether the contact information in the extracted customer data record is complete and legal. When at least one contact method is complete and legal, the customer data record is added to a pending research form, and when there is no complete contact method, the customer is After the data record is stored in the confirmation contact form, the information is transferred to the after-sales department for information supplementation, and at least one contact method to be supplemented is complete and legal, and the customer data record is added to the to-be-invested form; The customer that summarizes the form to be researched is classified according to the first classification rule to generate different customer sequence to be investigated, wherein the first classification rule includes: classification according to the year of insurance, classification according to the number of insurances, and signing according to the contract. Classification of the location of the person or classification by the industry in which the contractor engages, and generating a customer sequence to be investigated through the first classification rule; Pre-storing a list of to-be-sent messages, the to-be-sent message list includes at least one to-be-sent message record, and any one of the to-be-sent message records is a research text compiled in a specific template format, and is set to obtain a survey after researching the customer. As a result, the customer contact information in the form to be researched is sent to the customer to obtain the research result, wherein any sample of the customer to be researched prepares at least one format text of the short message to be sent; After selecting any record to be sent, push it to each customer in the corresponding customer sequence to be investigated, and establish a corresponding information feedback list, which is set to record whether the customer has a reply and reply content. The storage medium according to claim 17, wherein the extracting the tag data and the feature data in the customer feedback information and combining them to generate a feedback information list having a unified data format, such that the one or more processors perform the following steps: Develop a statistical periodic table, according to the pre-established statistical cycle, each cycle will classify the customer according to whether there is a reply, divided into two categories: replies and non-responses; Aggregate the unreacted customer list to generate a pending confirmation form, set to transfer to the after-sales department to confirm the customer's failure to reply, in order to confirm whether such customers have the value of continuing maintenance; The customer list that has been replied is summarized, and part of the field data in the complete basic information of such customer is extracted and recorded together with the tag data in the customer feedback information to a feedback information list. The storage medium according to claim 18, wherein the pair of feedback information is analyzed by a K-clustering algorithm, and two types of customer groups are selected, which are a high score customer group and a low score customer group, and the high score is As a target for recommending a follow-up product or service, the value customer group generates a return visit list for the low-scoring customer group, and performs regular maintenance and combing through a preset return visit mode, so that one or more processors perform the following steps: The K-clustering algorithm is used to analyze the feature information list, and the demarcation value of the data distribution is selected as the threshold according to the analysis result, and the customer corresponding to the threshold is used as the low-scoring customer, and the customer corresponding to the threshold is used as the customer corresponding to the threshold. High-value customers, so as to dig out the needs of each type of customer; The customer corresponding to the data larger than the threshold is selected as the object of recommending the follow-up product or service through the analysis result; After filtering the customer corresponding to the data smaller than the threshold, a return visit list is generated, and the customer appeal is obtained after the return visit according to the list, and the reason feature and the potential demand feature are extracted according to the return visit result; After analyzing the cause characteristics, generating an improved project form, and transferring the improved project form to the relevant department of the product development to improve the service items involved therein; After analyzing the potential demand characteristics, a new product development plan form is generated, and the new product development plan form is transferred to the relevant department of product development to set up a new product for different customer groups. The storage medium according to claim 19, wherein said one or more processors perform the following steps when said returning one by one according to a list: A return visit method in the form of a return call by voice call, email or App-side feedback channel, or a return visit method in the form of a web questionnaire.

Images