Sales clue score design method suitable for user value analysis and medium

Abstract

The invention discloses a sales clue score design method suitable for user value analysis and a medium, and the method comprises the steps: S1) selecting sales clues which should participate in scoring from a database, S2) calculating sales clue basic portrait scores based on multiple dimensions; s3) calculating a sales clue activity score; s4) calculating an effective follow-up score of the sales clue; s5) calculating a preliminary score of the sales clue; s6) calculating a sales clue dimension external weight 1; s7) judging whether the sales clue is the first sales clue of the customer, and calculating a weight 2 outside the dimension of the sales clue; and S8) calculating a final sales clue score. According to the method, the collected sales clues are subjected to multi-dimensional data processing and analysis, and calculated evaluation grading is performed, so that sales personnel can more accurately identify demands and preferences of potential customers, optimize resource configuration and sales strategies, better serve users, and finally realize business growth and sustainable development.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing, and in particular relates to a sales lead score design method and medium suitable for user value analysis. Background Art

[0002] With the rapid development of educational informatization, many people are constantly improving themselves. This has given rise to platforms that provide students with personalized learning resources. Taking the postgraduate entrance examination as an example, in recent years, the number of applicants has begun to stabilize after experiencing explosive growth, and the user's preparation path is gradually changing. More and more users have started preparing for the postgraduate entrance examination in the same year and are entering the preparation stage in earlier years. Therefore, the number of sales leads continues to increase.

[0003] However, in actual operation, due to the huge amount of sales leads and the large amount of information in the sales leads themselves, sales cannot quickly find high-quality sales leads among a large number of leads when blindly receiving leads or selectively following up on leads in the personal database. Passively following up sales leads when the quality of sales leads is unknown may reduce the efficiency of sales follow-up and seriously affect service performance. Therefore, it is urgent to build a sales lead scoring method to improve the efficiency of sales lead conversion, enhance the scientific nature of sales decision-making, and promote business growth. Summary of the invention

[0004] Purpose of the invention: In order to overcome the above shortcomings, the purpose of the present invention is to provide a sales lead scoring design method suitable for user value analysis. By subjecting the collected sales leads to multi-dimensional data processing and analysis, the calculated evaluation grades can help sales personnel to more accurately identify the needs and preferences of potential customers, optimize resource allocation and sales strategies, better serve users, and ultimately achieve business growth and sustainable development.

[0005] Technical solution: In order to achieve the above purpose, the present invention provides a sales lead score design method suitable for user value analysis, including the following steps:

[0006] S1): Circle the sales leads that should be included in the scoring from the database. The sales leads that should be included in the scoring must meet the requirements of not being closed, not being invalid, and not being in the transaction pool;

[0007] S2): Calculate the basic portrait score of sales leads based on multiple dimensions; the multiple dimensions are the user's postgraduate entrance examination year, sales stage, and reachable method;

[0008] S3): Calculate the sales lead activity score. Calculate the weight value of the sales lead record generation method based on the TF-IDF algorithm, calculate the interval between the sales lead record reported by the sales lead, calculate the time weight using the Gaussian decay function, and take the sum of the product of the weight value of the sales lead record generation method and the time weight as the sales lead activity score;

[0009] S4): Calculate the effective follow-up score of sales leads. Select three indicators based on the effective follow-up of sales leads by sales. Standardize the indicator values ​​according to the year of the sales lead's postgraduate entrance examination. Use the sigmoid function to calculate the score of each indicator. Use the entropy weight method to calculate the indicator data weight. The business weight of the indicator is determined by the business personnel based on the business rules. The effective follow-up score of the sales lead is the sum of the product of the sales lead effective follow-up indicator score, the indicator data weight and the indicator business weight.

[0010] S5): Calculate the preliminary score of the sales lead. The business personnel will determine the weight of each dimension based on the importance of each dimension and the reliability of the information source. The sum of the product of the sales lead score under each dimension and the dimension weight is used as the preliminary score of the sales lead;

[0011] S6): Calculate the weight 1 outside the dimension of sales leads, based on the fact that the shorter the interval between the creation time and the present, the less information the sales leads themselves carry; in addition to the dimension score, calculate the interval between the creation time of the sales leads and the present, and use the Gaussian decay function to calculate the weight 1 outside the dimension of sales leads. The weight value remains 1 within offset2 days, and the weight begins to decay after offset2 days. After scale2 days, the weight decays to decay2;

[0012] S7): Determine whether the sales lead is the first sales lead of the customer. Based on the particularity of the postgraduate entrance examination industry, the idea that the higher the sales lead's unit price, the lower the possibility of continued reporting is introduced. The weight of the sales lead dimension is calculated as 2. The weight of the first sales lead is kept at 1, and the weight of the non-first sales lead is decayed to 0.5;

[0013] S8) calculates the final sales lead score, that is, the product of the preliminary score of the sales lead calculated in S5), the out-of-dimensional weight 1 in S6) and the out-of-dimensional weight 2 in S7) is used as the final sales lead score.

[0014] The sales lead score design method and application suitable for user value analysis described in the present invention, the specific process of calculating the sales lead basic portrait score based on the predicted application year, sales stage and reachable method of the circled sales lead in S2) is as follows:

[0015]

[0016] Among them, i is the i-th clue, j is the j-th dimension, and d j Indicates the weight of the j-th dimension, d_score ij is the score of the i-th sales lead under the j-th dimension;

[0017] Among them, j|∈{1, 2, 3} respectively corresponds to the three dimensions of the customer's postgraduate entrance examination year, the lead sales stage and the customer's reachable method; the sales lead basic portrait score is composed of the three dimensions of the user's postgraduate entrance examination year, the lead sales stage and the customer's reachable method.

[0018] The sales lead scoring design method suitable for user value analysis described in the present invention is to calculate the scores of each dimension of the user's postgraduate entrance examination year, lead sales stage and customer reachability as follows:

[0019] When the dimension is the year in which the user took the postgraduate entrance examination, the score calculation is as follows:

[0020] The user's postgraduate entrance examination year is a label for the user's postgraduate entrance examination year. Since the customer's course purchase intention is related to the purchase time and the customer's postgraduate entrance examination year, customers in different time periods have different course purchase needs in different preparation cycles. Therefore, the user's postgraduate entrance examination year score is divided into different months for calculation;

[0021] If the sales lead does not have a postgraduate entrance examination year label, it means that the user has not left any intention information; if the sales lead has a postgraduate entrance examination year label, it means that the customer currently has no intention to purchase courses again;

[0022] If the sales lead has the label of this year's postgraduate entrance examination, the customer is more likely to buy high-priced courses from January to June of the year of the postgraduate entrance examination. During the summer vacation from July to August, there is still a tendency to buy the second half of the preparation courses. From September to December, as the time of the postgraduate entrance examination approaches, the tendency to buy this year's postgraduate entrance examination courses decreases and tends to 0;

[0023] If the sales lead has the label of "Next Year's Postgraduate Entrance Examination" or "Future Postgraduate Entrance Examination (later than next year), the customer's preparation path is relatively long, and the user's purchase tendency is gradually increasing. However, the closer the postgraduate entrance examination year is to the current year, the greater the purchase tendency is.

[0024] Therefore, for sales leads without a postgraduate entrance examination year label or with a postgraduate entrance examination year label that has passed, the postgraduate entrance examination year of the user is calculated as follows: Postgraduate entrance examination year score i =0, where i is the i-th clue;

[0025] If a sales lead with the tag "this year's postgraduate entrance examination" is extracted, the score of the postgraduate entrance examination year is calculated as follows:

[0026]

[0027] Where i is the i-th clue, t is the month corresponding to the statistical date, and is an integer in the range of [1, 12];

[0028] λ1 is the slope of the curve of this year's postgraduate entrance examination label when 7≤t≤8, that is, the decay rate of the curve in July and August;

[0029] λ2 is the additional attenuation coefficient of the curve of this year's postgraduate entrance examination label when 9≤t≤12, which is used to accelerate the attenuation speed of the curve from September to December;

[0030] If a sales lead with the tag "Next Year's Postgraduate Entrance Examination" is extracted, the score of the postgraduate entrance examination year is calculated as follows:

[0031]

[0032] If a sales lead with the tag "future postgraduate entrance examination year" is extracted, the score of the postgraduate entrance examination year is calculated as follows:

[0033]

[0034] α1 is the slope of the curve for next year's postgraduate entrance examination label, and α2 is the slope of the curve for future postgraduate entrance examination labels (later than next year), that is, the growth rate and satisfies α2<α1. Larger α1 and α2 values ​​will make the postgraduate entrance examination year score grow faster near the center point, and smaller α1 and α2 values ​​will make the postgraduate entrance examination year score grow more slowly;

[0035] β1 is the center point of the curve for next year's postgraduate entrance examination label, and β2 is the center point of the curve for future postgraduate entrance examination labels (later than next year), which is the point where the score increases fastest in the postgraduate entrance examination year;

[0036] When the dimension is lead sales stage, the calculation is as follows:

[0037] The lead sales stage is recorded in the CRM system, and the sales staff will mark the sales leads after following up;

[0038] A sales lead can be in any of the above sales stages before winning a deal, so the sales stage score is based on the historical sales lead data of winning deals. The sales stage score is the proportion of the number of sales leads in a certain sales stage to the number of all sales leads that win deals, that is,

[0039]

[0040] Among them, i is the i-th clue;

[0041] When the dimension is customer reachability, the score is calculated as follows:

[0042] The customer's contactable methods are recorded in the CRM system, which is composed of whether the customer has a mobile phone number or system account, whether the customer has a valid friend relationship, and whether the customer has a valid group chat relationship;

[0043] If the customer has a mobile phone number or system account, the sales staff can reach the customer through outbound calls, SMS, and APP private messages;

[0044] If the customer has a valid friend relationship, the salesperson can communicate with the customer directly through enterprise WeChat or other software;

[0045] If the customer has a valid group chat relationship, the salesperson can reach the customer by sending group chat messages.

[0046] Customers can be reached in multiple ways at the same time. The more ways a customer can be reached, the more ways sales staff can reach the customer, and the higher the possibility of the customer being reached. Therefore, the reachability score is based on historical winning sales lead data, that is,

[0047]

[0048] Among them, i is the i-th clue, and j is the j-th accessible method.

[0049] The sales lead score design method and application suitable for user value analysis described in the present invention, the specific algorithm for calculating the weight value of the sales lead record generation method according to the TF-IDF algorithm in S3) is as follows:

[0050] The weight of the sales lead record generation method is calculated based on the TF-IDF algorithm. TF-IDF is TF*IDF, which is word frequency*inverse text frequency index. In this context, word frequency is analogous to the proportion of sales lead records corresponding to a sales lead record generation method in all sales lead records for a certain winning sales lead.

[0051] The inverse text frequency index calculation formula is:

[0052] It can be compared to the ratio of all the winning sales leads to the sales leads with a certain sales lead generation method. To avoid the value being 0 and increase the discrimination, the denominator is increased by 1 and the logarithmic transformation is performed.

[0053] Therefore,

[0054] The time weight calculation process of the sales lead activity score is as follows:

[0055] The sales lead activity score is a comprehensive score calculated based on the sales lead records reported by the sales lead, the weight of the sales lead record generation method is calculated through historical winning data, and the time factor is taken into account;

[0056] Based on the idea that the longer the interval between the sales lead record reporting time and today is, the lower the importance should be, considering the time weight can emphasize the importance of the latest sales lead record, but at the same time does not ignore the contribution of the sales lead records reported earlier; the time weight refers to the Gaussian decay function, and the time weight formula for the interval between the sales lead record reporting time and today is set as follows:

[0057]

[0058] Among them, i is the i-th clue, x i Indicates the number of days between the reporting time of the i-th sales lead record and now;

[0059] Offset1 is the offset of the sales lead record time weight. The weight of the sales lead records reported within [0, offset1] days is set to 1 and remains unchanged. The weight starts to decay after offset1 day.

[0060] scale1 is the decay rate of the sales lead record time weight, that is, the speed at which the sales lead record time weight changes;

[0061] decay1 is the sales lead record time weight that decays from 0 days to the sales lead record time weight value corresponding to scale1;

[0062] For each sales lead record, this function can be used to obtain the time weight value of the interval between the sales lead record reporting time and the present time;

[0063] Reference parameter values: offset1=14, scale1=90, decay1=0.5;

[0064] Thus,

[0065] The sales lead activity score is the sum of the scores of all sales lead records reported by the sales lead.

[0066] The sales lead score design method and application suitable for user value analysis of the present invention, the specific calculation process of the sales lead effective follow-up score in S4) is as follows:

[0067]

[0068] Where i represents the i-th sales lead, j represents the j-th indicator, and biz_weigh jIndicates the business weight of the jth indicator, data-weigh j represents the data weight of the jth indicator, x ij It represents the effective follow-up score of the i-th sales lead under the j-th indicator.

[0069] The sales lead score design method and application suitable for user value analysis described in the present invention is characterized in that: the effective follow-up indicator score x of the i-th sales lead under the j-th indicator ij The calculation method is as follows:

[0070] In the same cycle, sales have different follow-up strategies for customers with different postgraduate entrance examination years. Therefore, when calculating the follow-up index score, the index values ​​are standardized based on the different postgraduate entrance examination years of the customers, that is,

[0071]

[0072] Among them, mean is the mean of the index values ​​grouped by postgraduate entrance examination year, and std is the standard deviation of the index values ​​distributed by postgraduate entrance examination year, so that the index values ​​of different years all satisfy the normal distribution with a mean of 0 and a standard deviation of 1.

[0073] The sales lead scoring design method and application suitable for user value analysis described in the present invention, the three major indicators of the sales lead effective follow-up score are the number of effective follow-up times, the effective follow-up rate and the interval from the last effective follow-up to the present.

[0074] Among them, the number of effective follow-up times and the effective follow-up rate are positive indicators, and the larger the indicator value, the higher the score should be; the interval from the last effective follow-up to the present is a negative indicator, and the smaller the indicator value, the higher the score should be:

[0075] The results are standardized by indicators, and the sigmoid function is used to calculate the scores of each indicator of sales leads. The calculation formula is as follows:

[0076] Positive indicators are:

[0077] Negative indicators are:

[0078] Among them, S ij It is the standardized result of the indicator value of the i-th sales lead under the j-th indicator.

[0079] The sales lead score design method and application suitable for user value analysis described in the present invention, the calculation process of the final sales lead score in S8) is as follows:

[0080]

[0081] Where i represents the i-th sales lead, j represents the j-th dimension, and weigh1 i Indicates the out-of-dimension weight of the i-th sales lead is 1, that is, the weight of the interval from the creation of the sales lead to the present;

[0082] weigh2 i represents the weight 2 of the ith sales lead, that is, whether it is the first sales lead of the customer, D_weigh j Indicates the weight of the j-th dimension, D_score ij Represents the score of the i-th sales lead under the j-th dimension.

[0083] The weigh1 i The calculation method is as follows:

[0084] The formula for setting the time weight of the sales lead creation interval is as follows:

[0085]

[0086] Where i represents the i-th sales lead, x 2_i Indicates the number of days since the creation of the i-th sales lead.

[0087] weigh1 i The out-of-dimension weight of the i-th sales lead is 1, that is, the time weight of the sales lead created since now;

[0088] Offset2 is the offset of weight 1 outside the sales lead dimension. The weight of sales leads created within [0, offset2] days is set to 1 and remains unchanged. The weight starts to decay after offset2 days.

[0089] scale2 is the decay rate of weight 1 outside the sales lead dimension, that is, the speed at which weight 1 outside the sales lead dimension changes;

[0090] decay2 is the value of the weight 1 outside the sales lead dimension that decays from 0 days to the weight 1 outside the sales lead dimension corresponding to scale2;

[0091] For each sales lead, this function can be used to obtain the value of weight 1 outside the sales lead dimension.

[0092] The sales lead scoring design method and application suitable for user value analysis described in the present invention, the weigh2 i is 1 or 0.5.

[0093] The present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the sales lead score design method suitable for user value analysis as described above is implemented.

[0094] It can be seen from the above technical solution that the present invention has the following beneficial effects:

[0095] 1. The sales lead scoring design method suitable for user value analysis described in the present invention performs multi-dimensional data processing and analysis on the collected sales leads to obtain a basic portrait score of the sales leads, and then calculates the preliminary score of the sales leads according to the proportion of sales lead record generation methods, the sales lead activity score, and the sales lead effective follow-up score. The final sales lead score is obtained by the preliminary score of the sales lead, the out-of-dimensional weight 1, and the out-of-dimensional weight 2, so that evaluation and grading can be performed according to the final sales lead score, which helps sales personnel to more accurately identify the needs and preferences of potential customers, optimize resource allocation and sales strategies, better serve users, and ultimately achieve business growth and sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 A schematic diagram of the structure of the sales lead score design method and application suitable for user value analysis according to the present invention;

[0097] Figure 2 A flowchart of the sales lead activity score calculation in the present invention;

[0098] Figure 3 is the weight of the generation method in the present invention;

[0099] Figure 4 A flow chart for calculating the effective follow-up score of sales leads in the present invention;

[0100] Figure 5 This is the definition of the sales stage in the basic portrait score in the present invention. DETAILED DESCRIPTION

[0101] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments.

[0102] Example

[0103] like Figure 1 A sales lead scoring design method suitable for user value analysis is shown, comprising the following steps:

[0104] S1): Circle the sales leads that should be included in the scoring from the database. The sales leads that should be included in the scoring must meet the following conditions: the sales leads are not closed, the sales leads are not invalid sales leads, and the sales leads are not sales leads in the transaction pool;

[0105] S2): Calculate the basic portrait score of sales leads based on multiple dimensions; the multiple dimensions are the user's postgraduate entrance examination year, sales stage, and reachable method;

[0106] S3): Calculate the sales lead activity score. Calculate the weight value of the sales lead record generation method based on the TF-IDF algorithm, calculate the interval between the sales lead record reported by the sales lead, calculate the time weight using the Gaussian decay function, and take the sum of the product of the weight value of the sales lead record generation method and the time weight as the sales lead activity score;

[0107] S4): Calculate the effective follow-up score of sales leads. Select three indicators based on the effective follow-up of sales leads by sales. Standardize the indicator values ​​according to the predicted application year of the sales leads. Use the sigmoid function to calculate the score of each indicator. Use the entropy weight method to calculate the data weight of each indicator. The business weight of the indicator is determined by the business personnel based on the business rules. The effective follow-up score of the sales lead is the sum of the product of the effective follow-up indicator score of the sales lead, the data weight of the indicator and the business weight of the indicator.

[0108] S5): Calculate the preliminary score of the sales lead. The business personnel will determine the weight of each dimension based on the importance of each dimension and the reliability of the information source. The sum of the product of the sales lead score under each dimension and the dimension weight is used as the preliminary score of the sales lead;

[0109] S6): Calculate the weight 1 outside the dimension of sales leads. The shorter the interval between the creation time and the present, the less information the sales leads themselves carry. Based on this, in addition to the scores of each dimension, calculate the interval between the creation time of the sales leads and the weight 1 outside the dimension of sales leads using the Gaussian decay function. The weight value remains 1 within offset2 days, and the weight begins to decay after offset2 days. After scale2 days, the weight decays to decay2.

[0110] S7): Determine whether the sales lead is the first sales lead of the customer. Based on the particularity of the postgraduate entrance examination industry, the idea that the higher the sales lead's unit price, the lower the possibility of continued reporting is introduced. The weight of the sales lead dimension is calculated as 2. The weight of the first sales lead is kept at 1, and the weight of the non-first sales lead is decayed to 0.5;

[0111] S8): Calculate the final sales lead score, that is, take the product of the preliminary score of the sales lead calculated in S5), the out-of-dimensional weight 1 in S6), and the out-of-dimensional weight 2 in S7) as the final sales lead score.

[0112] It should be noted that the steps from step S2) to step S4) are not in any particular order. In particular, there may be multiple sales lead records for one sales lead in S3).

[0113] Business personnel determine the weight of each dimension based on the importance of each dimension and the reliability of the information source. The weight assignment rules for each dimension are as follows:

[0114] Since active sales leads are behavioral records generated by users themselves and are not affected by sales staff, they are highly important.

[0115] Sales lead follow-up is a process where sales personnel reach out to users through outbound calls, text messages, WeChat sales, etc. The effectiveness of follow-up is influenced by many factors, such as the sales personnel's business capabilities, the user's determination to take the postgraduate entrance examination, and the perfection of follow-up effectiveness rules. It is therefore of high importance.

[0116] The basic sales lead profile is the basic information of the sales lead. It has low correlation with whether the sales lead is active and whether the sales staff is involved in follow-up, and is of low importance.

[0117] The importance of the intentional dimensions is: active sales leads > effective follow-up of sales leads > basic portrait of sales leads. The initial weights of each weight are: 50% weight for active sales leads score, 30% weight for effective follow-up of sales leads score, and 20% weight for basic portrait of sales leads score.

[0118] The sales lead scoring design method and application suitable for user value analysis described in this embodiment are as follows: Figure 5 As shown, the specific process of calculating the basic profile score of the sales lead in S2) based on the predicted application year, sales stage and reachable method of the circled sales lead is as follows:

[0119]

[0120] Among them, i is the i-th clue, j is the j-th dimension, and d j Represents the weight of the j-th dimension, d-score ij is the score of the i-th sales lead under the j-th dimension;

[0121] Among them, j|∈{1, 2, 3} corresponds to the three dimensions of the customer's postgraduate entrance examination year, the lead sales stage and the customer's reachable method; the sales lead basic portrait score is composed of the three dimensions of the user's postgraduate entrance examination year, the lead sales stage and the customer's reachable method.

[0122] In the sales lead scoring design method for user value analysis described in this embodiment, the scoring calculation method for each dimension of the user's postgraduate entrance examination year, lead sales stage, and customer reachability method is as follows:

[0123] When the dimension is the year in which the user took the postgraduate entrance examination, the score calculation is as follows:

[0124] The user's postgraduate entrance examination year is a label for the user's postgraduate entrance examination year. Since the customer's course purchase intention is related to the purchase time and the customer's postgraduate entrance examination year, customers in different time periods have different course purchase needs in different preparation cycles, so the user's postgraduate entrance examination year score is divided into different months for calculation;

[0125] If the sales lead does not have a postgraduate entrance examination year label, it means that the user has not left any intention information; if the sales lead has a postgraduate entrance examination year label, it means that the customer currently has no intention to purchase courses again;

[0126] If the sales lead has the label of this year's postgraduate entrance examination, the customer is more likely to buy high-priced courses from January to June of the year of the postgraduate entrance examination. During the summer vacation from July to August, there is still a tendency to buy the second half of the preparation courses. From September to December, as the time of the postgraduate entrance examination approaches, the tendency to buy this year's postgraduate entrance examination courses decreases and tends to 0;

[0127] If the sales lead has the label of "Next Year's Postgraduate Entrance Examination" or "Future Postgraduate Entrance Examination (later than next year), the customer's preparation path is relatively long, and the user's purchase tendency is gradually increasing. However, the closer the postgraduate entrance examination year is to the current year, the greater the purchase tendency is.

[0128] Therefore, for sales leads without a postgraduate entrance examination year label or with a postgraduate entrance examination year label that has passed, the postgraduate entrance examination year of the user is calculated as follows: Postgraduate entrance examination year score i =0, where i is the i-th clue;

[0129] If a sales lead with the tag "this year's postgraduate entrance examination" is extracted, the score of the postgraduate entrance examination year is calculated as follows:

[0130]

[0131] Where i is the i-th clue, t is the month corresponding to the statistical date, and is an integer in the range of [1,12];

[0132] λ1 is the slope of the curve of this year's postgraduate entrance examination label when 7≤t≤8, that is, the decay rate of the curve in July and August;

[0133] λ2 is the additional attenuation coefficient of the curve of this year's postgraduate entrance examination label when 9≤t≤12, which is used to accelerate the attenuation speed of the curve from September to December;

[0134] If a sales lead with the tag "Next Year's Postgraduate Entrance Examination" is extracted, the score of the postgraduate entrance examination year is calculated as follows:

[0135]

[0136] If a sales lead with the tag "future postgraduate entrance examination year" is extracted, the score of the postgraduate entrance examination year is calculated as follows:

[0137]

[0138] α1 is the slope of the curve for next year's postgraduate entrance examination label, and α2 is the slope of the curve for future postgraduate entrance examination labels (later than next year), that is, the growth rate and satisfies α2<α1. Larger α1 and α2 values ​​will make the postgraduate entrance examination year score grow faster near the center point, and smaller α1 and α2 values ​​will make the postgraduate entrance examination year score grow more slowly;

[0139] β1 is the center point of the curve for next year's postgraduate entrance examination label, and β2 is the center point of the curve for future postgraduate entrance examination labels (later than next year), which is the point where the score increases fastest in the postgraduate entrance examination year;

[0140] When the dimension is lead sales stage, the calculation is as follows:

[0141] The lead sales stage is recorded in the CRM system, and the sales staff will mark the sales leads after following up;

[0142] A sales lead can be in any of the above sales stages before winning a deal, so the sales stage score is based on the historical sales lead data of winning deals. The sales stage score is the proportion of the number of sales leads in a certain sales stage to the number of all sales leads that win deals, that is,

[0143]

[0144] Among them, i is the i-th clue;

[0145] When the dimension is customer reachability, the score is calculated as follows:

[0146] The customer's contactable methods are recorded in the CRM system, which is composed of whether the customer has a mobile phone number or system account, whether the customer has a valid friend relationship, and whether the customer has a valid group chat relationship;

[0147] If the customer has a mobile phone number or system account, the sales staff can reach the customer through outbound calls, SMS, and APP private messages;

[0148] If the customer has a valid friend relationship, the salesperson can communicate with the customer directly through enterprise WeChat or other software;

[0149] If the customer has a valid group chat relationship, the salesperson can reach the customer by sending group chat messages.

[0150] Customers can be reached in multiple ways at the same time. The more ways a customer can be reached, the more ways sales staff can reach the customer, and the higher the possibility of the customer being reached. Therefore, the reachability score is based on historical winning sales lead data, that is,

[0151]

[0152] Among them, i is the i-th clue, and j is the j-th accessible method.

[0153] The sales lead score design method and application suitable for user value analysis described in this embodiment, the specific algorithm for calculating the weight value of the sales lead record generation method according to the TF-IDF algorithm in S3) is as follows:

[0154] The weight of the sales lead record generation method shown is calculated based on the TF-IDF algorithm. TF-IDF is TF*IDF, which is the term frequency*inverse text frequency index. In this context, the term frequency is analogous to the proportion of the number of sales lead records corresponding to a sales lead record generation method in all sales lead records for a certain winning sales lead.

[0155] The inverse text frequency index calculation formula is:

[0156] It can be compared to the ratio of all the winning sales leads to the sales leads with a certain sales lead generation method. To avoid the value being 0 and increase the discrimination, the denominator is increased by 1 and the logarithmic transformation is performed.

[0157] Therefore,

[0158] like Figure 2 , based on the sales lead records reported by the winning sales leads in the past 6 months, calculate the TF-IDF weight value of each sales lead record generation method.

[0159] like Figure 3 As shown, the user generates behavioral data on the terminal, and the corresponding sales lead reports the sales lead records of the corresponding generation method to obtain the sales lead records of the winning sales leads in the past 6 months, and calculates the number of sales lead records based on the sales lead record generation method, the number of sales lead records based on the sales lead, and the number of sales lead records of each generation method based on the sales lead;

[0160] Then calculate the importance of each sales lead generation method:

[0161] Importance of sales lead a generation method = (number of records of sales lead a generation method / number of sales lead records) * lg (number of winning sales leads / (number of winning sales leads with sales lead records of generation method a + 1))

[0162] The weight of the production method = the average importance of the production method.

[0163] The sales lead activity score is a comprehensive score calculated based on the sales lead records reported by the sales lead, the weight of the sales lead record generation method is calculated through historical winning data, and the time factor is taken into account;

[0164] The time weight calculation process of the sales lead activity score is as follows:

[0165] The sales lead activity score is a comprehensive score calculated based on the sales lead records reported by the sales lead, the weight of the sales lead record generation method is calculated through historical winning data, and the time factor is taken into account;

[0166] Based on the idea that the longer the interval between the sales lead record reporting time and today is, the lower the importance should be, considering the time weight can emphasize the importance of the latest sales lead record, but at the same time does not ignore the contribution of the sales lead records reported earlier; the time weight refers to the Gaussian decay function, and the time weight formula for the interval between the sales lead record reporting time and today is set as follows:

[0167]

[0168] Among them, l is the lth clue, x 1_i Indicates the number of days between the reporting time of the i-th sales lead record and now;

[0169] Offset1 is the offset of the sales lead record time weight. The weight of the sales lead records reported within [0, offset1] days is set to 1 and remains unchanged. The weight starts to decay after offset1 day.

[0170] scale1 is the decay rate of the sales lead record time weight, that is, the speed at which the sales lead record time weight changes;

[0171] decay1 is the sales lead record time weight that decays from 0 days to the sales lead record time weight value corresponding to scale1;

[0172] For each sales lead record, this function can be used to obtain the time weight value of the interval between the sales lead record reporting time and the present time;

[0173] Reference parameter values: offset1=14, scale1=90, decay1=0.5;

[0174] Thus,

[0175] The sales lead activity score is the sum of the scores of all sales lead records reported by the sales lead.

[0176] It should be noted that business personnel determine the parameter value as the interval between the reporting time of the clue record based on business rules. The weight value remains at 1 within 14 days, and the weight begins to decay after 14 days. After 90 days, the weight decays to 0.5; the parameters are 14, 90, and 0.5 in the formula.

[0177] The sales lead scoring design method and application suitable for user value analysis described in this embodiment, the sales lead effective follow-up score in S4)

[0178] The specific calculation process is as follows:

[0179]

[0180] Where i represents the i-th sales lead, j represents the j-th indicator, and biz-weigh j Indicates the business weight of the jth indicator, data_weigh j represents the data weight of the jth indicator, x ij It represents the effective follow-up score of the i-th sales lead under the j-th indicator.

[0181] The business weights of the three major indicators in the sales lead effective follow-up score are as follows:

[0182] The number of effective follow-up statistics is the sum of the number of effective follow-up times of all sales personnel since the lead was created. Although the number of effective follow-up is not the only determining factor, the lack of sufficient follow-up times often leads to the loss of sales opportunities, so it is of high importance.

[0183] The effective follow-up rate is calculated as the proportion of effective follow-up times among all follow-up times. A high effective follow-up rate indicates that the salesperson has high execution and conversion rate, high customer positive feedback rate, and high importance.

[0184] The interval from the last effective follow-up reflects the salesperson's continued attention to and response speed to the sales leads. A shorter interval indicates that the salesperson is able to follow up with the customer in a timely manner and maintain communication. However, if the sales lead itself has lost its vitality or the customer has made a purchase decision, even a shorter interval may not change the final result, so it is less important.

[0185] The sales lead score design method and application for user value analysis described in this embodiment is as follows: the effective follow-up indicator score x of the i-th sales lead under the j-th indicator ij The calculation method is as follows:

[0186] In the same cycle, sales have different follow-up strategies for customers with different postgraduate entrance examination years. Therefore, when calculating the follow-up index score, the index values ​​are standardized based on the different postgraduate entrance examination years of the customers, that is,

[0187]

[0188] Among them, mean is the mean of the index values ​​grouped by postgraduate entrance examination year, and std is the standard deviation of the index values ​​distributed by postgraduate entrance examination year, so that the index values ​​of different years all satisfy the normal distribution with a mean of 0 and a standard deviation of 1.

[0189] The sales lead scoring design method and application suitable for user value analysis described in this embodiment, the three major indicators of the sales lead effective follow-up score are the number of effective follow-up times, the effective follow-up rate and the interval from the last effective follow-up to the present.

[0190] Among them, the number of effective follow-up times and the effective follow-up rate are positive indicators, and the larger the indicator value, the higher the score should be; the interval from the last effective follow-up to the present is a negative indicator, and the smaller the indicator value, the higher the score should be:

[0191] The results are standardized by indicators, and the sigmoid function is used to calculate the scores of each indicator of sales leads. The calculation formula is as follows:

[0192] Positive indicators are:

[0193] Negative indicators are:

[0194] Among them, s ij It is the standardized result of the indicator value of the i-th sales lead under the j-th indicator.

[0195] The sales lead score design method and application suitable for user value analysis described in this embodiment, the calculation process of the final sales lead score in S8) is as follows:

[0196]

[0197] Where i represents the i-th sales lead, j represents the j-th dimension, and weigh1 i Indicates the out-of-dimension weight of the i-th sales lead is 1, that is, the weight of the interval from the creation of the sales lead to the present;

[0198] weigh2 i represents the weight 2 of the ith sales lead, that is, whether it is the first sales lead of the customer, D_weigh j Indicates the weight of the j-th dimension, D_score ij represents the score of the i-th sales lead under the j-th dimension, and the weigh2 i is 1 or 0.5. As can be seen from the formula, the out-of-dimension weight refers to the secondary weighting after the dimension score is calculated.

[0199] The weigh1 i The calculation method is as follows:

[0200] The formula for setting the time weight of the sales lead creation interval is as follows:

[0201]

[0202] Where i represents the i-th sales lead, x 2_i Indicates the number of days since the creation of the i-th sales lead.

[0203] weigh1 i The out-of-dimension weight of the i-th sales lead is 1, that is, the time weight of the sales lead created since now;

[0204] Offset2 is the offset of weight 1 outside the sales lead dimension. The weight of sales leads created within [0, offset2] days is set to 1 and remains unchanged. The weight starts to decay after offset2 days.

[0205] scale2 is the decay rate of weight 1 outside the sales lead dimension, that is, the speed at which weight 1 outside the sales lead dimension changes;

[0206] decay2 is the value of the weight 1 outside the sales lead dimension that decays from 0 days to the weight 1 outside the sales lead dimension corresponding to scale2;

[0207] For each sales lead, this function can be used to obtain the value of weight 1 outside the sales lead dimension;

[0208] Reference parameter values: offset2=14, scale2=180, decay2=0.5.

[0209] Example 2

[0210] The present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the sales lead score design method suitable for user value analysis as described above is implemented.

[0211] The sales lead score design method suitable for user value analysis described in this embodiment is the same as that in Embodiment 1.

[0212] The above description is only a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements without departing from the principle of the present invention, and these improvements should also be regarded as within the protection scope of the present invention.

Claims

1. A sales lead scoring design method suitable for user value analysis, characterized by: Includes the following steps: S1): Circle the sales leads that should be included in the scoring from the database. The sales leads that should be included in the scoring must meet the requirements of not being closed, not being invalid sales leads, and not being sales leads in the transaction pool; S2): Calculate the basic portrait score of sales leads based on multiple dimensions; the multiple dimensions are the user's postgraduate entrance examination year, sales stage, and reachable method; S3): Calculate the sales lead activity score. Calculate the weight value of the sales lead record generation method based on the TF-IDF algorithm, calculate the interval between the sales lead record reported by the sales lead, calculate the time weight using the Gaussian decay function, and take the sum of the product of the weight value of the sales lead record generation method and the time weight as the sales lead activity score; S4): Calculate the effective follow-up score of sales leads. Select three indicators based on the effective follow-up of sales leads by sales. Standardize the indicator values ​​according to the application year of the sales leads. Use the sigmoid function to calculate the score of each indicator. Use the entropy weight method to calculate the indicator data weight. The business weight of the indicator is determined by the business personnel based on the business rules. The effective follow-up score of the sales lead is the sum of the product of the effective follow-up indicator score of the sales lead, the indicator data weight and the indicator business weight. S5): Calculate the preliminary score of the sales lead. The business personnel will determine the weight of each dimension based on the importance of each dimension and the reliability of the information source. The sum of the product of the sales lead score under each dimension and the dimension weight is used as the preliminary score of the sales lead; S6): Calculate the weight 1 outside the sales lead dimension, based on the fact that the shorter the interval between the creation time and the present, the less information the sales lead itself carries; in addition to the sales lead dimension score, calculate the interval between the creation time of the sales lead and the present, and use the Gaussian decay function to calculate the weight 1 outside the sales lead dimension. The weight value remains 1 within offset2 days, and the weight begins to decay after offset2 days. After scale2 days, the weight decays to decay2; S7): Determine whether the sales lead is the customer's first sales lead. Based on the particularity of the education industry, the idea that the higher the sales lead's unit price, the lower the possibility of re-subscription is introduced. The weight of the sales lead dimension is calculated as 2. The weight of the first sales lead is kept at 1, and the weight of the non-first sales lead is decayed to 0.5; S8): Calculate the final sales lead score, that is, take the product of the preliminary score of the sales lead calculated in S5), the out-of-dimensional weight 1 in S6), and the out-of-dimensional weight 2 in S7) as the final sales lead score.

2. The sales lead scoring design method suitable for user value analysis according to claim 1, characterized in that: The specific process of calculating the basic profile score of the sales lead in S2) based on the predicted application year, sales stage and reachable method of the circled sales lead is as follows: Among them, i is the i-th clue, j is the j-th dimension, and d j Indicates the weight of the j-th dimension, d_score ij is the score of the i-th sales lead under the j-th dimension; Among them, j∈{1, 2, 3}, respectively corresponds to the three dimensions of the customer's postgraduate entrance examination year, the lead sales stage and the customer's reachable method; the sales lead basic portrait score is composed of the three dimensions of the user's postgraduate entrance examination year, the lead sales stage and the customer's reachable method.

3. The sales lead score design method suitable for user value analysis according to claim 2, characterized in that: The scores for the user's postgraduate entrance examination year, lead sales stage, and customer reachability are calculated as follows: When the dimension is the year in which the user took the postgraduate entrance examination, the score calculation is as follows: The user's postgraduate entrance examination year is a label for the user's postgraduate entrance examination year. Since the customer's course purchase intention is related to the purchase time and the customer's postgraduate entrance examination year, customers in different time periods have different course purchase needs in different preparation cycles. Therefore, the user's postgraduate entrance examination year score is divided into different months for calculation; If the sales lead does not have a postgraduate entrance examination year label, it means that the user has not left any intention information; if the sales lead has a postgraduate entrance examination year label, it means that the customer currently has no intention to purchase courses again; If the sales lead has the label of this year's postgraduate entrance examination, the customer is more likely to buy high-priced courses from January to June of the year of the postgraduate entrance examination. During the summer vacation from July to August, there is still a tendency to buy the second half of the preparation courses. From September to December, as the time of the postgraduate entrance examination approaches, the tendency to buy this year's postgraduate entrance examination courses decreases and tends to 0; If the sales lead has the label of "Next Year's Postgraduate Entrance Examination" or "Future Postgraduate Entrance Examination (later than next year), the customer's preparation path is relatively long, and the user's purchase tendency is gradually increasing. However, the closer the postgraduate entrance examination year is to the current year, the greater the purchase tendency is. Therefore, for sales leads without a postgraduate entrance examination year label or with a postgraduate entrance examination year label that has passed, the user's postgraduate entrance examination year is calculated as follows: Score by year of entrance examination i =0, where i is the i-th clue; If a sales lead with the tag "this year's postgraduate entrance examination" is extracted, the score of the postgraduate entrance examination year is calculated as follows: Where i is the i-th clue, t is the month corresponding to the statistical date, and is an integer in the range of [1,12]; λ1 is the slope of the curve of this year's postgraduate entrance examination label when 7≤t≤8, that is, the decay rate of the curve in July and August; λ2 is the additional attenuation coefficient of the curve of this year's postgraduate entrance examination label when 9≤t≤12, which is used to accelerate the attenuation speed of the curve from September to December; If a sales lead with the tag "Next Year's Postgraduate Entrance Examination" is extracted, the score of the postgraduate entrance examination year is calculated as follows: If a sales lead with the tag "future postgraduate entrance examination year" is extracted, the score of the postgraduate entrance examination year is calculated as follows: α1 is the slope of the curve for next year's postgraduate entrance examination label, and α2 is the slope of the curve for future postgraduate entrance examination labels (later than next year), that is, the growth rate and satisfies α2<α1. Larger α1 and α2 values ​​will make the postgraduate entrance examination year score grow faster near the center point, and smaller α1 and α2 values ​​will make the postgraduate entrance examination year score grow more slowly; β1 is the center point of the curve for next year's postgraduate entrance examination label, and β2 is the center point of the curve for future postgraduate entrance examination labels (later than next year), which is the point where the score increases fastest in the postgraduate entrance examination year; When the dimension is lead sales stage, the calculation is as follows: The lead sales stage is recorded in the CRM system, and the sales staff will mark the sales leads after following up; A sales lead can be in any of the above sales stages before winning a deal, so the sales stage score is based on the historical sales lead data of winning deals. The sales stage score is the proportion of the number of sales leads in a certain sales stage to the number of all sales leads that win deals, that is, Among them, i is the i-th clue; When the dimension is customer reachability, the score is calculated as follows: The customer's contact methods are recorded in the CRM system, depending on whether the customer has a mobile phone number or system account. Whether the customer has a valid friend relationship, whether the customer has a valid group chat relationship; If the customer has a mobile phone number or system account, the sales staff can reach the customer through outbound calls, SMS, and APP private messages; If the customer has a valid friend relationship, the salesperson can communicate with the customer directly through enterprise WeChat or other software; If the customer has a valid group chat relationship, the salesperson can reach the customer by sending group chat messages. Customers can be reached in multiple ways at the same time. The more ways a customer can be reached, the more ways sales staff can reach the customer, and the higher the possibility of the customer being reached. Therefore, the reachability score is based on historical winning sales lead data, that is, Among them, i is the i-th clue, and j is the j-th accessible method.

4. The sales lead scoring design method for user value analysis according to claim 1, characterized in that: The specific algorithm for calculating the weight value of the sales lead record generation method according to the TF-IDF algorithm in S3) is as follows: The weight of the sales lead record generation method is calculated based on the TF-IDF algorithm. TF-IDF is TF*IDF, which is word frequency*inverse text frequency index. In this context, word frequency is analogous to the proportion of sales lead records corresponding to a sales lead record generation method in all sales lead records for a certain winning sales lead. The inverse text frequency index calculation formula is: Among them, D is the number of all winning sales leads, and d is the number of sales leads with sales lead records of a certain generation method, which can be compared to the ratio of all winning sales leads to the number of sales leads with sales lead records of a certain generation method. In order to avoid the value being 0 and increase the discrimination, the denominator is added by 1 and logarithmic transformation is performed respectively; Therefore, the The time weight calculation process of the sales lead activity score is as follows: The sales lead activity score is a comprehensive score calculated based on the sales lead records reported by the sales lead, the weight of the sales lead record generation method is calculated through historical winning data, and the time factor is taken into account; Based on the idea that the longer the interval between the sales lead record reporting time and today is, the lower the importance should be, considering the time weight can emphasize the importance of the latest sales lead record, but at the same time does not ignore the contribution of the sales lead records reported earlier; the time weight refers to the Gaussian decay function, and the time weight formula for the interval between the sales lead record reporting time and today is set as follows: Among them, i is the i-th clue, x 1_i Indicates the number of days between the reporting time of the i-th sales lead record and now; Offset1 is the offset of the sales lead record time weight. The weight of the sales lead records reported within [0, offset1] days is set to 1 and remains unchanged. The weight starts to decay after offset1 day. scale1 is the decay rate of the sales lead record time weight, that is, the speed at which the sales lead record time weight changes; decay1 is the sales lead record time weight that decays from 0 days to the sales lead record time weight value corresponding to scale1; For each sales lead record, this function can be used to obtain the time weight value of the interval between the sales lead record reporting time and the present time; Thus, The sales lead activity score is the sum of the scores of all sales lead records reported by the sales lead.

5. The sales lead scoring design method suitable for user value analysis according to claim 1, characterized in that: The specific calculation process of the sales lead effective follow-up score in S4) is as follows: Where i represents the i-th sales lead, j represents the j-th indicator, and biz_weigh j Indicates the business weight of the jth indicator, data_weigh j represents the data weight of the jth indicator, x ij It represents the effective follow-up score of the i-th sales lead under the j-th indicator.

6. The sales lead scoring design method for user value analysis according to claim 5 is characterized in that: The effective follow-up indicator x of the i-th sales lead under the j-th indicator ij The calculation method is as follows: In the same cycle, sales have different follow-up strategies for customers with different postgraduate entrance examination years. Therefore, when calculating the follow-up index score, the index values ​​are standardized based on the different postgraduate entrance examination years of the customers, that is, Among them, mean is the mean of the index values ​​grouped by postgraduate entrance examination year, and std is the standard deviation of the index values ​​distributed by postgraduate entrance examination year, so that the index values ​​of different years all satisfy the normal distribution with a mean of 0 and a standard deviation of 1.

7. The sales lead score design method suitable for user value analysis according to claim 5, characterized in that: The three major indicators of the sales lead effective follow-up score are the number of effective follow-up times, the effective follow-up rate and the interval from the last effective follow-up. Among them, the number of effective follow-up and the effective follow-up rate are positive indicators, and the larger the indicator value, the higher the score should be; The interval from the last effective follow-up is a negative indicator. The smaller the indicator value, the higher the score should be: The results are standardized by indicators, and the sigmoid function is used to calculate the scores of each indicator of sales leads. The calculation formula is as follows: Positive indicators are: Negative indicators are: Among them, s ij It is the standardized result of the effective follow-up indicator value of the i-th sales lead under the j-th indicator.

8. The sales lead scoring design method for user value analysis according to claim 1, characterized in that: The calculation process of the final sales lead score in S8) is as follows: Where i represents the i-th sales lead, j represents the j-th dimension, and weigh1 i Indicates the out-of-dimension weight of the i-th sales lead is 1, that is, the weight of the interval from the creation of the sales lead to the present; weigh2 i represents the weight 2 of the ith sales lead, that is, whether it is the first sales lead of the customer, D_weigh j Indicates the weight of the j-th dimension, D_score ij represents the score of the i-th sales lead under the j-th dimension; The weigh1 i The calculation method is as follows: The formula for setting the time weight of the sales lead creation interval is as follows: Where i represents the i-th sales lead, x 2_i Indicates the number of days since the creation of the i-th sales lead. weigh1 i The out-of-dimension weight of the i-th sales lead is 1, that is, the time weight of the sales lead created since now; Offset2 is the offset of weight 1 outside the sales lead dimension. The weight of sales leads created within [0, offset2] days is set to 1 and remains unchanged. The weight starts to decay after offset2 days. scale2 is the decay rate of weight 1 outside the sales lead dimension, that is, the speed at which weight 1 outside the sales lead dimension changes; decay2 is the value of the weight 1 outside the sales lead dimension that decays from 0 days to the weight 1 outside the sales lead dimension corresponding to scale2; For each sales lead, this function can be used to obtain the value of weight 1 outside the sales lead dimension.

9. The sales lead score design method suitable for user value analysis according to claim 8, characterized in that: The weigh2 i is 1 or 0.

5.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the sales lead score design method suitable for user value analysis as described in any one of claims 1 to 9 are implemented.

Images