WO2022105107A1

WO2022105107A1 - Method for optimizing telephone sales working performance by using smart watch

Info

Publication number: WO2022105107A1
Application number: PCT/CN2021/089730
Authority: WO
Inventors: 姜平
Original assignee: 焦点科技股份有限公司
Priority date: 2020-11-17
Filing date: 2021-04-25
Publication date: 2022-05-27
Also published as: CN112396313B; CN112396313A

Abstract

Disclosed in the present invention is a method for optimizing telephone sales working performance by using a smart watch, characterized by: selecting a batch of sampled telephone sales personnel, performing associative classification analysis by means of data of the heart rates and environment volumes detected in nearly real time by smart watches worn by the personnel and in combination with historical working data summaries of the telephone sales personnel in a sales system at the same time period, and establishing a data classification model. That is, relevant data collected by the smart watches according to in a certain time period is inputted into a data classification model to obtain corresponding working performance categories, materials of clients to be contacted of different categories and quantities are pushed according to the performance categories, so that a sales system can automatically allocate appropriate sales workload and work objects according to the current working performance and appropriate policy scheduling, and thus the effect of improving the sales working performance is achieved.

Description

A way to optimize telemarketing job performance with a smartwatch

technical field

The invention relates to the field of combining smart wearable devices and telemarketing systems, in particular to a method for optimizing telemarketing work performance by using a smart watch.

Background technique

The core profit of a sales company comes from the performance of its sales team. Because the sales team has a large number of personnel, high liquidity and high management costs, its performance improvement method can generally only use monthly or quarterly performance appraisal to conduct performance appraisal on its actual sales orders and motivate them.

Due to the poor timeliness, traditional methods cannot monitor and promote the work performance of sales in real time during the sales process. For the company's precious pending marketing customer leads distribution, it can only use some pre-agreed rules, which cannot be dynamically modified in real time, so that the lead resources cannot be reasonably utilized to the greatest extent, and resources are easily wasted. Due to the continuous and rapid popularization of smart wearable devices, the use of smart watches, which can be carried along with the sound, and can collect data information of each sample in real time, can use these data to adjust the sales resource allocation strategy of the company's sales system in real time, so as to maximize the Develop sales potential customers to a high degree and improve sales performance.

Patent CN202010155127 - Method for obtaining characteristic value of physiological data, method for establishing classifier, classification method, classifier and classification system, including extracting physiological data, training the classifier and classifying the physiological data, but its application scenario is the collection of digestive tract data, wherein The selection and processing of physiological data and the use of classifiers have strong pertinence and limitations, and cannot be directly or indirectly applied to the telemarketing scenario of this case.

Patent CN201910375153-A fatigue evaluation system and evaluation method based on dual-modal bioelectrical signals and physiological data, including monitoring and collection of physiological information, using the analysis of physiological data to intervene in work, but the specific monitoring of physiological information. The content is too much, the relationship with the work scene is very complex, and it has strong pertinence and limitations. It cannot be directly transferred to the telemarketing scene of this case, and there is no obvious inspiration, so it cannot be directly or indirectly applied to telemarketing work. middle.

Patent CN201911222034 - Workload evaluation method weakly coupled with job task scene, including the collection of physiological data, the workload index is given through index evaluation rules, and the allocation of work content can be optimized according to it, but it is evaluated through stimulation sequences. It will inevitably cause interference and impact on the normal work, the timeliness is poor, and the setting of the stimulus sequence is too complicated, and the implementation is difficult. After the workload is evaluated, no specific follow-up methods are given, which cannot be applied to the actual work scene of this case. .

Therefore, there is a need for a method of optimizing telemarketing job performance utilizing a more efficient method.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to use various sensors of the smart watch to collect various information of the wearing person in real time, so as to provide a more reasonable distribution strategy of sales leads based on a specific strategy, thereby improving the utilization efficiency of sales leads , to enhance the work of telemarketing.

In order to solve the above technical problems, a method for optimizing telemarketing work performance by using a smart watch is characterized in that a group of telemarketing personnel is selected by sampling for a period of data sampling, classification training and association. In practical applications, the information returned by the telesales staff's smart watch can be collected in near real time to classify the work performance, and then according to the specific sales lead allocation strategy, the pending processing seen by each telesales staff can be adjusted in near real time. The content and quantity of the sales lead list, including the following steps:

Step 1: Monitoring and collection of information, using the smart watch hardware to monitor the target object, collecting the information of the target object, the information includes heartbeat data, ambient volume, geographic location, and reporting the information through the mobile communication device connected to the smart watch hardware , the reporting frequency is set to be 10 minutes each time interval;

Step 2: collect the work performance data of the target object during the monitoring period through the data analysis system, the work performance data includes the corresponding data in the corresponding sales management system, and the time span is set to be a complete evaluation cycle of sales;

Step 3: establish a data classification model in the data analysis system for combining the information reported in step 1 and the work performance data collected in step 2 to determine the work performance classification of the target object;

Step 4: classify the target object by using the data classification model, and give a judgment on the classification of the work performance of the target object;

Step 5: After receiving the judgment result of the work performance classification corresponding to the target object, the sales management system allocates different types of marketing customer lead lists to be processed according to the results of different work performance classifications, and the marketing customer lead list includes marketing customer leads. The marketing customer leads are classified into three categories: high-quality, general, and unknown according to their quality from high to low;

Step 6: Retrain and optimize the classification model according to the information monitored by the smart watch hardware and the work performance data of the target object.

In the step 1, the specific steps include:

Step 1-1: The smart watch hardware is connected to the mobile communication device through Bluetooth;

Step 1-2: The smart watch hardware sends the current wearer's time information T _i i∈[1,n] and the current average heart rate HR _i i∈[1, n], the current ambient sound decibel DB _{i i} ∈ [1, n], the current geographic location information GS _{i i} ∈ [1, n], the accuracy of the geographic location information requires the absolute value of the error to be within 1.2 meters, where n is the target The number of objects, the n satisfies n ≥ 100, and the subscript i represents different target objects participating in the sample collection;

Step 1-3: After the mobile communication device receives a set of [T, HR, DB, GS] message information, it sends it to the data analysis system to monitor the network status. It will be sent together when waiting for the next sending, and the interval between each sending from the mobile communication device to the data analysis system is 10 minutes.

In the step 2, the specific steps are:

Step 2-1: After receiving the [T, HR, DB, GS] message information in Step 1-3, continue to collect a complete assessment cycle, according to T _ij i∈[1, n], j∈[ 1, m] Set the sample observation points for extracting work performance data from the sales management system, where n represents the number of target objects, m represents the total number of messages returned by each target object, T represents the sample observation points, and i represents the participating samples Different target objects collected, j represents the serial number of the sample observation point in the assessment period;

Step 2-2: According to the target object corresponding to each T _j , take out the corresponding work performance data of one complete assessment cycle from the sales management system, the work performance data includes the total number of phone calls C _j i∈[1,m] , telephone call duration L _j i ∈ [1, m], number of marketing customer leads identified as high-quality S _j i ∈ [1, m], number of characters filled in contact records R _j i ∈ [1, m], sales management system Information query times Q _j i∈[1, m];

Step 2-3: Transfer the work performance data obtained in Step 2-2 into the data analysis system to construct analysis samples [C, L, S, R, Q].

In the step 3, the specific steps are:

Step 3-1: Perform Softmax function classification model training on the [T, HR, DB, GS] message information in step 1, and obtain an information classification WL _{i i} ∈ [1, k] from the smart watch hardware, where , k represents the number of categories of information collected by the smart watch hardware;

Step 3-2: Perform Softmax function classification model training on [C, L, S, R, Q] in step 2, and obtain a classification PL _i i∈[1, k′ from the work performance data of the sales management system ], where k′ represents the number of job performance categories;

Step 3-3: Perform Sigmod function correlation model training on the information classification of step 3-1 and the classification of work performance data in step 3-2, according to the time factor T, get each WL _{i i} ∈ [1, k] corresponding to The conversion rate of PL _i i∈[1, k′], the correlation degree TR between any WL and any PL is [0, 1];

Step 3-4: Take the value with the largest correlation degree TR in WL and PL, any WL _{i i} ∈ [1, k] only retains one corresponding PL _{i i} ∈ [1, k′], and store it in a two-dimensional table on TV.

In the step 4, the specific steps are:

Step 4-1: According to the working time range of the target object, the smart watch hardware collects [T, HR, DB, GS] message information every 2 minutes and sends it to the corresponding mobile communication device;

Step 4-2: The mobile communication device sends a message to the data analysis system every 10 minutes;

Step 4-3: Calculate the average [T, HR, DB, GS] of the target object in this time period according to the valid messages received within 10 minutes.

Step 4-4: The average [T, HR, DB, GS] is passed into the Softmax function classification model in step 3-1, and the obtained WL value is the corresponding information classification result.

In the step 5, the specific steps are:

Step 5-1: Classify all pending marketing customer leads in the sales relationship system, and mark the classification type as CL;

Step 5-2: Preset marketing customer lead distribution rules, specifically: according to PL _i i∈[1, k′], where k′ represents the number of job performance classifications, take all [C, L, S for calculating PL , R, Q] sample data, set the mean

m is the number of all total samples, and calculate the number of each PL _i classification in turn

k is the number of all samples classified as PL _i in the total samples, and the maximum AvgPL _i is set as MaxAvgPL, then if the corresponding PL _i

Then the marketing customer leads corresponding to the two-dimensional table TV in the configuration step 3-4 are high-quality, and the number to be processed is _WN =25.

Then the marketing customer leads corresponding to the configuration two-dimensional table TV are general, and the number to be processed is WN=50. If the corresponding PL _i

Then the marketing customer leads corresponding to the configuration two-dimensional table TV are unknown, and the number to be processed is WN=10;

Step 5-3: The sales management system receives the [T, HR, DB, GS] sent in step 1-3, calls the model trained in step 3-1 to obtain the corresponding WL _i , and finds the corresponding CL and WN, then modify the current classification type in the list of pending marketing customer leads for the target object in the list of pending leads of the sales management system to the corresponding classification type CL in the two-dimensional table, and the current quantity to the corresponding one in the two-dimensional table. quantity WN;

In the step 6, the specific steps are:

Step 6-1: Collect [C, L, S, R, Q] and [T, HR, DB, GS] of the target object every day;

Step 6-2: According to the data collected in Step 6-1, the data analysis system recalculates the model training in Step 3-1 and Step 3-2 at the beginning of each assessment cycle, and calculates the positive and negative deviation EV;

Step 6-3: Set the positive and negative deviation threshold Y, if when |EV|>Y, then re-update the data [WL, PL, CL, WN] recorded in the TV table, and the update includes the category of raising or lowering CL , increase or decrease WN.

The beneficial effects achieved by the present invention:

(1) It can change the traditional sales performance appraisal method based on the final performance result, and effectively intervene by collecting data in the daily sales process.

(2) In the present invention, considering the practical constraints of wearable smart devices in terms of battery life and portability, smart watches are selected as the hardware carrier for collecting and judging the work performance of sales personnel, and the function and cost are balanced.

(3) The method for optimizing the performance of telesales by using a smart watch provided by the present invention can not only use the example in the invention to allocate different types and quantities of customer leads to be contacted, but also conduct other management interventions based on the work performance status at the time of sales .

(4) The present invention provides a set of differentiated automatic model update mechanism according to actual operation result data and model preset, thereby continuously improving the accuracy of model classification and having automatic maintenance characteristics.

Description of drawings

FIG. 1 is a simplified method flow diagram of an exemplary embodiment of the present invention.

Detailed ways

Below in conjunction with accompanying drawing, the present invention is further described:

As shown in FIG. 1, an exemplary embodiment of a method for optimizing telemarketing work performance using a smart watch selects a sample of telemarketers, and uses the near-real-time detection data of the heartbeat and ambient volume of the smart watch worn by the telemarketing personnel. Combined with the summary of historical work data of telemarketers in the sales system at the same time period, correlation classification analysis is carried out, and a data classification model is established. According to the relevant data collected by the smart watch in a certain period of time, enter the data classification model to obtain the corresponding work performance classification, and then push different types and quantities of customer leads to be contacted according to the performance classification, so that it can be classified according to the current situation. Work performance and appropriate strategy scheduling, the sales system automatically matches the appropriate sales workload and work objects, so as to achieve the effect of improving sales performance. Specifically include the following steps:

Step 1: Establish an information monitoring and collection system for the monitoring and collection of target object information. Customized smart watch hardware, use smart watch hardware to monitor the target object, regularly collect information during working hours, including heartbeat data, environmental volume, and geographic location information, and the frequency of reporting data is set at 10-minute intervals;

Step 2: Collect the work performance data of the target object (that is, the corresponding electricity sales participating in the sampling) during wearing the smart watch hardware through step 1. The work performance data mainly includes its corresponding data in the corresponding sales management system, and the time span is set to one. The complete evaluation cycle of sales, such as monthly evaluation, is set to one month;

Step 3: Use big data modeling technology to model and analyze the information and data in

steps

1 and 2, and establish a data classification model in the data analysis system, which can combine the information collected by the smart watch hardware and the sales management system. The work performance data in , determine the current work performance classification of electricity sales;

Step 4: All power pins are required to wear smart watch devices during work, and report heartbeat, ambient volume, and geographic location data to the monitoring data analysis system every 10 minutes, and the system will call the data classification model obtained in step 3 for classification. , giving a judgment on the classification of the work performance of the current electricity sales;

Step 5: After the sales management system receives the judgment result of the work performance classification corresponding to the telemarketing, it will push different types of pending marketing customer lead lists according to the results of different performance classification judgments; the marketing customer lead list includes marketing customer leads, marketing Customer leads are divided into three categories: high-quality, general, and unknown according to the quality from high to low;

Step 6: The monitoring data collection system will regularly collect the performance data of all electricity sales and the monitoring information returned by the smart watch hardware, and the data analysis system will retrain and optimize the data classification model accordingly.

In the step 1, a monitoring data collection system is established, and the heartbeat data, ambient volume, and geographic location information are regularly collected during the working hours through the smart watch hardware worn during the working hours. The specific steps include:

Step 1-1: Develop a customized smart phone App application so that the smart watch hardware can connect to the mobile phone device through Bluetooth;

Step 1-2: The smart watch hardware sends the current wearer's time information T _i i∈[1,n], the current average heart rate HR _i i∈[1,n], The current ambient sound decibel DB _i i∈[1,n], the current geographic location information GS _i i∈[1,n], the accuracy requirement is the civil accuracy of the current Beidou system version 2.0 is 1.2 meters, and n is the number of electricity sales participating in the sampling , according to the total size of the telemarketing team, the sampling satisfies n ≥ 100; the subscript i represents the different salespersons participating in the sample collection;

Step 1-3: After the mobile terminal receives a set of messages from [T, HR, DB, GS], it sends it to the corresponding monitoring data analysis system (MS for short) to monitor the network status. It is cached on the local mobile phone and sent together when waiting for the next sending window. The interval between each sending from the mobile phone to the MS system is 10 minutes.

In the step 2, the work performance data of the corresponding electric pins participating in the sampling during the wearing of the smart watch are collected, mainly including their corresponding data in the corresponding sales management system, and the specific steps are as follows:

Step 2-1: After receiving the relevant [T, HR, DB, GS] message information in Step 1-3, continue to collect a complete sales assessment cycle. According to T _ij i∈[1,n],j∈[1,m], set the sample observation point for extracting work performance data from the sales management system, n represents the corresponding sampled number of salespeople, m represents each salesperson The total number of returned messages, i represents the different target objects participating in the sample collection, j represents the serial number of the sample observation point in the assessment period;

Step 2-2: According to the telemarketing staff corresponding to each T _j , extract the telemarketing performance data within the time range from the sales management system, including the total number of telephone calls C _j i∈[1, m] and the duration of telephone calls L _j i∈[1,m], the number of high-quality customer leads actively identified by telemarketers S _j i∈[1,m], the number of characters filled in contact records R _j i∈[1,m], the number of sales management system information queries Q _j i∈[1,m];

Step 2-3: Pass all the above work performance data into the MS system to construct analysis samples [C, L, S, R, Q].

In the step 3, according to the step 1 and the step 2, the modeling analysis is performed to establish a model for the classification of the current work performance of the sales, and the specific steps are as follows:

Step 3-1: Perform Softmax function classification model training on [T, HR, DB, GS] in Step 1, and obtain an information classification WL _i i∈[1, k] from the smart watch hardware, where k represents the smart watch The number of categories of information collected by the hardware;

Step 3-2: Perform softmax function classification model training on [C, L, S, R, Q] in step 2, and obtain a classification PL _i i∈[1, k from the performance data of the sales management system '], k' represents the number of job performance classifications;

Step 3-3: Perform Sigmod function correlation model training on the classification of work performance data in steps 3-1 and 3-2, and obtain the PL _i corresponding to each WL _{i i} ∈ [1, k] according to the time factor T The conversion rate of i∈[1, k′], that is, the relationship degree TR between any WL and any PL valued in the [0,1] interval, the closer to 1, the higher the correlation between the two;

Step 3-4: Take the largest value of the correlation between WL and PL, that is, any WL _i i∈[1, k] only retains a corresponding PL _i i∈[1, k′], and stores it in a two-dimensional table (TV table).

In the step 4, it is required that all electric pins need to wear smart watch devices during the working period, and report the heartbeat, ambient volume, and geographic location data to the monitoring data analysis system every 10 minutes, and the system will call the model obtained in step 3 to carry out. Classification, to give the classification judgment of the current work performance of the electricity sales, the specific steps are:

Step 4-1: According to the working time range of the electric pin, the smart watch worn by the electric pin collects [T, HR, DB, GS] every 2 minutes, and then sends it to the App side of the corresponding mobile phone system;

Step 4-2: The App terminal corresponding to the mobile phone system sends a message to the monitoring data analysis system MS every 10 minutes;

Step 4-3: According to the valid messages received within 10 minutes, find the average [T, HR, DB, GS] of the sales employee corresponding to this time period.

Step 4-4: Pass the average [T, HR, DB, GS] into the model of Step 3-1, and get a WL value, which is the information classification result of the current electricity sales.

In the step 5, after receiving the classification judgment of the work performance of the corresponding telemarketing, the sales management system will push different categories of pending marketing customer lead lists according to different performance classification judgments. The specific steps are:

Step 5-1: Classify all pending marketing customer leads in the sales relationship system, mark the classification type as CL, and divide them into three categories: high-quality, general, and unknown according to the quality of marketing customer leads from high to low;

Step 5-3: The sales management system receives the [T, HR, DB, GS] that the current salesperson's smart watch hardware passes in through the App, and calls the model trained in step 3-1 to obtain the corresponding WL _i , which is in the TV table If the corresponding CL and WN are found, the current classification type in the pending marketing customer leads list of the target object in the list of pending leads of the sales management system is modified to the corresponding classification type CL in the two-dimensional table, and the current quantity is modified is the corresponding quantity WN in the two-dimensional table;

In the step 6, the monitoring data collection system will periodically collect the work performance data of all electricity pins and the monitoring information returned by the smart watch hardware, and the data analysis system will retrain and optimize the data classification model accordingly. The specific steps are:

Step 6-1: During the official launch of the system, the monitoring data collection system will regularly collect the actual daily [C, L, S, R, Q] and daily actual [T, HR, DB, GS] of the actual sales staff;

Step 6-2: According to the data collected in Step 6-1, the MS system will recalculate the model training in Step 3-1 and Step 3-2 at the beginning of each assessment cycle, and calculate the positive and negative deviation EV;

Step 6-3: The system can set the positive and negative deviation threshold Y, if |EV|>Y, then re-update the data [WL, PL, CL, WN] recorded in the TV table, and the update principle includes increasing or decreasing CL category, increase or decrease WN.

Compared with the existing technology, smart wearable devices such as smart watches are generally used for motion detection and health monitoring. However, in the daily work of actual phone sales, the heart rate curve, ambient sound and noise decibels (such as the The four-dimensional data of telephone volume of communication), its moving frequency and space example can establish a common classification feature of telemarketers at work, and this classification feature can be compared with the sales performance indicators, so that real-time performance can be established. Automatically adjust the work content assigned by the sales system to telesales, so as to guide sales to improve work performance, solve the randomness and hysteresis of customer allocation policies in traditional telesales systems, and effectively improve the use efficiency of customer resources. Specifically:

(1) This method needs to collect sampling data in advance for preliminary data analysis and modeling.

(2) The present invention considers the basic functions of smart watch hardware currently on the market, takes into account the cost and functional requirements, and has practical value in large-scale promotion.

(3) The main job content of telemarketing is to make calls to the customer leads resources allocated by the system, so the impact of different types of customer leads on their work performance at that time is obvious. The sales management department can completely adjust the sales policy flexibly according to the method in the present invention in combination with the actual sales situation of the department, and has wide adaptability.

(4) This method takes into account the variability of the model in the early and late stages. After the system continuously self-collects data to correct the deviation, the classification model can make the classification and judgment of telemarketers more accurate, so that the system has the function of self-improvement.

The present invention can push interventions in real time, and because different strategies can be configured after systemization, it can push forward incentives based on superiors, and reverse challenge strategies based on superiors. General telemarketing is often unable to intervene in real time, and can only be adjusted after accumulating data over a long period of time, while telemarketing leads are very time-sensitive. potential loss of performance.

The above embodiments do not limit the present invention in any way, and all other improvements and applications made in the form of equivalent transformations to the above embodiments belong to the protection scope of the present invention.

Claims

A method for optimizing telemarketing work performance by using a smart watch, characterized in that it specifically includes the following steps:

Step 1: Monitoring and collection of information, using the smart watch hardware to monitor the target object, collecting the information of the target object, the information includes heartbeat data, ambient volume, geographic location, and reporting the information through the mobile communication device connected to the smart watch hardware , the reporting frequency is set to be 10 minutes each time interval;

Step 2: collect the work performance data of the target object during the monitoring period through the data analysis system, the work performance data includes the corresponding data in the corresponding sales management system, and the time span is set to be a complete evaluation cycle of sales;

Step 3: establish a data classification model in the data analysis system for combining the information reported in step 1 and the work performance data collected in step 2 to determine the work performance classification of the target object;

Step 4: classify the target object by using the data classification model, and give a judgment on the classification of the work performance of the target object;

Step 5: After receiving the judgment result of the work performance classification corresponding to the target object, the sales management system allocates different types of marketing customer lead lists to be processed according to the results of different work performance classifications, and the marketing customer lead list includes marketing customer leads. The marketing customer leads are classified into three categories: high-quality, general, and unknown according to their quality from high to low;

Step 6: Retrain and optimize the classification model according to the information monitored by the smart watch hardware and the work performance data of the target object.
A method for optimizing telemarketing work performance using a smart watch as claimed in claim 1, wherein in the step 1, the specific steps include:

Step 1-1: The smart watch hardware is connected to the mobile communication device through Bluetooth;

Step 1-2: The smart watch hardware sends the current wearer's time information T i i∈[1,n] and the current average heart rate HR i i∈[1, n], the current ambient sound decibel DB i i∈[1,n], the current geographic location information GS i i∈[1,n], the accuracy of the geographic location information requires the absolute value of the error to be within 1.2 meters, where n is the target The number of objects, the n satisfies n ≥ 100, and the subscript i represents different target objects participating in the sample collection;

Step 1-3: After the mobile communication device receives a set of [T, HR, DB, GS] message information, it sends it to the data analysis system to monitor the network status. It will be sent together when waiting for the next sending, and the interval between each sending from the mobile communication device to the data analysis system is 10 minutes.
A method for optimizing telemarketing work performance using a smart watch as claimed in claim 2, wherein in the step 2, the specific steps are:

Step 2-1: After receiving the [T, HR, DB, GS] message information in Step 1-3, continue to collect a complete assessment cycle, according to T ij i∈[1, n], j∈[ 1, m] Set the sample observation points for extracting work performance data from the sales management system, where n represents the number of target objects, m represents the total number of messages returned by each target object, T represents the sample observation points, and i represents the participating samples Different target objects collected, j represents the serial number of the sample observation point in the assessment period;

Step 2-2: According to the target object corresponding to each T j , take out the corresponding work performance data of one complete assessment cycle from the sales management system, the work performance data includes the total number of phone calls C j i∈[1,m] , telephone call duration L j i ∈ [1, m], number of marketing customer leads identified as high-quality S j i ∈ [1, m], number of characters filled in contact records R j i ∈ [1, m], sales management system Information query times Q j i∈[1, m];

Step 2-3: Transfer the work performance data obtained in Step 2-2 into the data analysis system to construct analysis samples [C, L, S, R, Q].
A method for optimizing telemarketing work performance using a smart watch as claimed in claim 3, wherein in the step 3, the specific steps are:

Step 3-1: Perform Softmax function classification model training on the [T, HR, DB, GS] message information in step 1, and obtain an information classification WL i i ∈ [1, k] from the smart watch hardware, where , k represents the number of categories of information collected by the smart watch hardware;

Step 3-2: Perform Softmax function classification model training on [C, L, S, R, Q] in step 2, and obtain a classification PL i i∈[1, k′ from the work performance data of the sales management system ], where k′ represents the number of job performance categories;

Step 3-3: Perform Sigmod function correlation model training on the information classification of step 3-1 and the classification of work performance data in step 3-2, according to the time factor T, get each WL i i ∈ [1, k] corresponding to The conversion rate of PL i i∈[1, k′], the correlation degree TR between any WL and any PL is [0, 1];

Step 3-4: Take the value with the largest correlation degree TR in WL and PL, any WL i i ∈ [1, k] only retains one corresponding PL i i ∈ [1, k′], and store it in a two-dimensional table on TV.
A method for optimizing telemarketing work performance using a smart watch as claimed in claim 4, wherein in the step 4, the specific steps are:

Step 4-1: According to the working time range of the target object, the smart watch hardware collects [T, HR, DB, GS] message information every 2 minutes and sends it to the corresponding mobile communication device;

Step 4-2: The mobile communication device sends a message to the data analysis system every 10 minutes;

Step 4-3: Calculate the average [T, HR, DB, GS] of the target object within this time period according to the valid messages received within 10 minutes;

Step 4-4: The average [T, HR, DB, GS] is passed into the Softmax function classification model in step 3-1, and the obtained WL value is the corresponding information classification result.
A method for optimizing telemarketing work performance using a smart watch as claimed in claim 5, wherein in the step 5, the specific steps are:

Step 5-1: Classify all pending marketing customer leads in the sales relationship system, and mark the classification type as CL;

Step 5-2: Preset marketing customer lead distribution rules, specifically: according to PL i i∈[1, k′], where k′ represents the number of job performance classifications, take all [C, L, S for calculating PL , R, Q] sample data, set the mean

m is the number of all total samples, and calculate the number of each PL i classification in turn

k is the number of all samples classified as PL i in the total samples, and the maximum AvgPL i is set as MaxAvgPL, then if the corresponding PL i
Then the marketing customer leads corresponding to the two-dimensional table TV in the configuration step 3-4 are high-quality, and the number to be processed is WN =25.
Then the marketing customer leads corresponding to the configuration two-dimensional table TV are general, and the number to be processed is WN=50. If the corresponding PL i

Then the marketing customer leads corresponding to the configuration two-dimensional table TV are unknown, and the number to be processed is WN=10;

Step 5-3: The sales management system receives the [T, HR, DB, GS] sent in step 1-3, calls the model trained in step 3-1 to obtain the corresponding WL i , and finds the corresponding CL and WN, then modify the current classification type in the list of pending marketing customer leads for the target object in the list of pending leads of the sales management system to the corresponding classification type CL in the two-dimensional table, and the current quantity to the corresponding one in the two-dimensional table. Quantity WN.
A method for optimizing telemarketing work performance using a smart watch as claimed in claim 6, wherein in the step 6, the specific steps are:

Step 6-1: Collect [C, L, S, R, Q] and [T, HR, DB, GS] of the target object every day;

Step 6-2: According to the data collected in Step 6-1, the data analysis system recalculates the model training in Step 3-1 and Step 3-2 at the beginning of each assessment cycle, and calculates the positive and negative deviation EV;

Step 6-3: Set the positive and negative deviation threshold Y, if |EV|>Y, then re-update the data [WL, PL, CL, WN] recorded in the TV table, and the update includes the category of increasing or decreasing CL , increase or decrease WN.